Monthly Archives: May 2024

FIG 2

The Application of Case-Based Learning in Endodontics

Leave a reply

DOI: 10.31038/JDMR.2024712

Abstract

Introduction: Theoretical teaching in endodontics is based on lectures delivered by qualified professionals. Recent advancements explore options such as case based learning (CBL), that allow students to apply their knowledge to real-world clinical scenarios.

Objective: To evaluate the effect of CBL on clinical problem solving in endodontics, in a cohort of dentists enrolled in an “endodontic case series” workshop.

Methodology: An Endodontics Case Series Activity (ECSA) was organized at the Aga Khan University Hospital, Karachi. The enrolled participants and attendees participated in a pre-activity assessment, through Google Form. The form consisted of 5 clinical scenario based multiple choice questions (MCQs), based on dental trauma, iatrogenic errors, regenerative endodontics and guided endodontics. The participants then attended the ECSA, where post-graduate trainees presented the management of complex endodontic cases, surrounding the same themes, which was followed by an interactive discussion. After the workshop, the same MCQs were re-attempted to assess any changes in managing the same five clinical scenarios after attending the ECSA. Additionally, nine questions regarding the perception of CBL were also included in the post-test questionnaire.

Results: Of the 28 participants, 64.3% were post-graduate trainees of Operative Dentistry and Endodontics, whereas the remaining participants were trainees from other dental specialties (10.7%) general dentists (17.9%), undergraduate dental students (7.1%). Fifty percent of the participants reported that CBL improved the implementation of key concepts, 51% responded that CBL allowed an improved treatment planning and problem-solving skills and 68.2% reported that CBL encouraged their interest in endodontics and self-learning.

Conclusion: CBL may improve the clinical problem-solving skills for students and trainees, however, large scale studies are required to further establish the true effectiveness of CBL in training and education.

Introduction

When discussing the different methods to teach endodontics, it goes without saying that there can be no single ‘best method’ [1]. Didactic lecture-based learning formats have been considered highly effective in disseminating a large quantity of information to a large number of students. However, it is a passive form of learning, which often leaves students uninterested or demotivated. This passivity may impede active engagement, critical thinking, and the application of theoretical knowledge to practical scenarios. Recognizing these disadvantages, there has been a shift towards more interactive student-centered learning approaches which include problem based learning (PBL) and Case based learning (CBL) [2-4].

CBL and PBL are both student-centered active learning methods that aim to engage students and foster deep understanding [5]. However, each method has its own distinct characteristics [6]. CBL, initially applied in medical education by the Anatomy Department of the Medical School in Newfoundland, Canada, is an interactive, instructor-led learning technique. Conversely, PBL is a student driven learning method in which students takes the lead in identifying problems, conducting research and finding solutions.In PBL no prior knowledge regarding subject is required whereas, CBL requires students to have some past knowledge that can benefit in problem solving [7]. Though both the methods connects theory to practice by applying knowledge to cases utilizing inquiry-based learning methods, but CBL stands out in its emphasis on a more structured learning environment with instructor guidance, contributing to the preparation of students for clinical practice by exposing them to real-life clinical cases [8].

In recent years, studies have proven CBL to be an effective teaching method currently used in various health disciplines such as medicine, allied health, child developments and some aspects of dentistry [9]. As depicted in literature, in a study by Bi M et al. conducted on postgraduate trainees of medical oncology, reported CBL is an efficient teaching method for improving problem-solving abilities when compared to traditional teaching method [10]. Another study by Shigli et al. conducted to evaluate the effectiveness of CBL in the field of prosthodontics, concluded CBL to be a useful method in enhancing the knowledge of dental interns [11]. Despite this positive outcome, there is a notable gap in the literature concerning the implementation of this innovative approach in the field of endodontics, particularly in our geographic region. Moreover, endodontics is inherently procedure-based, underscoring the significance of integrating clinical experience into training programs. Given this context, the aim of our study is to evaluate the perception and to compare the knowledge of participants related to endodontic clinical cases both pre and post CBL activity using a questionnaire, providing valuable insights into the potential effectiveness of CBL in the field of endodontics.

Materials and Methods

The participants of this study were post-graduate trainees of Operative Dentistry and Endodontics from several renowned institutes, along with their supervisors. However, attendees included undergraduate dental students, post-graduate dental students of all dental specialties and general dental practitioners. Ethical approval was not considered necessary for the activity. Figure 1 presents a diagrammatic representation of the process of data collection.

FIG 1

Figure 1: Diagrammatic representation of data collection illustrating CBL activity

Endodontic Case Series Activity

Post-graduate trainees of Operative Dentistry and Endodontics from various institutes were invited to present their clinical cases at the Aga Khan University Hospital, to participate in the “Endodontic Case Series Activity” (ECSA). The trainees were requested to share a pre-recorded presentation of their clinical case, with well-documented photographs and radiographs. Among the received cases, 5 cases were selected by two faculty members to include the following themes: Dental Trauma, Regenerative Endodontics, Guided Endodontics, Complex Endodontics and Iatrogenic Errors. The presenting candidates were requested to prepare a 5-minute pre-recorded video presentation of their case according to a provided template. The template included the relevant medical and dental history, presenting complaint, treatment planned, treatment provided and follow-up. After each presentation, the presenter was addressed regarding any questions and an interactive panel discussion took place, encouraging participation from the audience. The panel consisted of 2 international and 3 national specialists in the field of Operative Dentistry and Endodontics, with over ten years of clinical experience.

Questionnaire Development

To assess the responses of the participants and attendees regarding clinical problem solving in endodontics, a questionnaire was developed, with 5 multiple choice questions (MCQS) based on: Dental trauma, Endodontic treatment planning, iatrogenic errors, application of guided endodontics and regenerative endodontics. These MCQs were part of the Operative Dentistry and Endodontics MCQ bank, where each MCQ is reviewed by 7 post-graduate trainees and 3 endodontists, and an answer key is decided. However, since the questions were modified, Content Validation Index was employed (CVI) to evaluate the validity of the questions. A panel of 4 experts were tasked with reviewing the questionnaire items for relevance and clarity. These 4 experts included general dentist, consultant, biostatistician, and an epidemiologist. Each questionnaire item was assessed by the experts based on relevance and clarity and was rated on a scale of ‘1’ to ‘4’ with ‘1’ being not relevant/not clear to ‘4’ being highly relevant/very clear. A score of ‘1’ or ‘2’ rated by experts is designated as 0 while a score of ‘3’ or ‘4’ is designated as 1. An average of this score is calculated to determine the CVI. Typically, a CVI score of 0.80 or higher is considered indicative of satisfactory content validity. In our study, the combined evaluations of all four experts yielded an exceptionally high CVI score of 0.95, affirming the questionnaire’s outstanding precision and reliability in effectively capturing the required information.

Pre-Activity and Post Activity Assessment

The formulated questionnaire was distributed amongst the participants and attendees using online GoogleForm and the total scores were recorded. After the ECSA, the same questions were then distributed along with another questionnaire which assessed the perception of the ECSA in the attendees and participants, using a Likert’s scale.

Statistical Analysis

Responses from the study questionnaires were recorded using GoogleForm. The data was only shared with the three authors carrying out the study and was stored in a password protected file. Data was analyzed by using SPSS version 21. Descriptive statistics were reported, including the designation of the participants. The percentage and mean of correct responses was calculated according to each theme for both the pre-activity assessment and the post-activity assessment. To compare the mean pre-activity and post-activity scores, the paired sample’s t-test was applied. The level of significance was kept at<0.05.

Results

A total number of 28 participants were enrolled in the ECSA. Eighteen of these participants were post-graduate trainees from Operative Dentistry and Endodontics, three were residents from other specializations, five were general dentists and two were undergraduate dental students as depicted in Figure 2. The percentage of correct responses for the pre-activity assessment for dental trauma, iatrogenic errors, regenerative endodontics, surgical endodontics and guided endodontics were 20%, 86.7%, 60%, 86.7% and 60% respectively as evidenced in Table 1. The mean pre-activity score was 3.20 (1.01), whereas the post-activity score was 4.13 (0.83). A statistically significant improvement was noted in the post-activity score (p-value=0.014) as shown in Table 2. The participants feedback revealed a positive response, with a majority of the participants rating the activity as ‘4’ for improvement in treatment planning, encouraging interest, self-learning and enthusiasm, as evidenced in Figure 3.

FIG 2

Figure 2: Graphical representation of demographic data

Table 1: Participant Response on Pre-Test and Post-Test Assessment

Themes

 Correct responses (Total number of participants: 28)
Pre-ECSA (%)

Post-ECSA (%)
Dental Trauma

20%

33%
Iatrogenic Errors

86.7%

100%
Regenerative Endodontics

60%

86.7%
Surgical Endodontics

86.7%

93.3%
Guided Endodontics

60%

80%
ECSA: Endodontic Case Series Activity

Table 2: Comparison of pre & post activity score

Time of Assessment (No. of participants)

Mean scores (SD)

p-value
Pre-Activity (28 participants)

3.2 (1.01)

0.014*
Post-Activity (28 participants)

4.13 (0.83)
*Paired sample t-test, p-value < 0.05.

FIG 3

Figure 3: Post activity feedback assessing participants’ perception of CBL

Discussion

It’s intriguing how, despite global efforts to embrace more learner-centered teaching approaches in medical education, seminars and lectures continue to dominate in certain regions of the world [12]. The problem with traditional teaching is that it does not promote deep learning. It mainly emphasizes rote memorization and information transmission rather than promoting critical thinking, problem-solving, and a thorough comprehension of the subject matter. On the other hand, small group discussions using CBL model has number of benefits in teaching institutes as it utilizes collaborative learning, develops students’ intrinsic and extrinsic motivation to learn, supplements existing knowledge and supports the development of variety of clinical skills.

The present study uses strategic learning CBL model and investigated its effectiveness by comparing pre-test and post-test results of the participants enrolled in endodontic case series (ECS) activity. ECS activity was a single day workshop conducted in Aga Khan University Hospital, Karachi in which 28 candidates registered for the workshop. The participants enrolled had different levels of expertise ranging from undergraduates to general dentist to postgraduate trainees. In this cohort of variety of participants, majority of them were postgraduate trainees of endodontics (64%), followed by general dentist (18%), post graduate trainees of other specialty (11%) and a smaller proportion of undergraduates (7%). This diversity in expertise level is potentially advantageous as it allows for a comprehensive exploration of how individuals at different stages of their educational or professional journey engage with and benefit from the ECS activity using CBL approach.

The workshop session included a pre-test questionnaire followed by visual-audio presentation by participants on the assigned topics and team based interactive discussion after which a post-test questionnaire assessment was carried out. The questionnaire used in this present study consists of multiple-choice questions retrieved from MCQ bank of department of ‘Operative Dentistry and Endodontics’, AKUH. These questions underwent adaptations based on our study’s specific themes. Themes around which questions were formulated include dental trauma, iatrogenic errors, regenerative, surgical, and guided endodontics. These themes were chosen as they are normally encountered in our dental practice and are a subject of dental education which includes anatomy, microbiology, pathology, radiology and pharmacology.

Furthermore, the modified questionnaire underwent validation by 4 experts of different specialty and CVI was calculated to be 0.95, proving it to be accurate. This high CVI score indicates strong agreement among these experts regarding the relevance and clarity of the questionnaire items concerning the study’s specified themes.

The type of CBL activity employed in the present study is different from those employed in previous studies. For example, a of study by Chutinan et al. was conducted on second year dental students using lengthy survey-based approach to evaluate their perception regarding case-based activity. The authors carried out a survey at three different times to gain a comprehensive feedback at each stage. However, it is possible that the repetitive assessment may have inadvertently led to participant disengagement due to its prolonged nature, which defeats the purpose of active learning methods [13]. On the contrary, the current study adopted a more focused assessment, aiming to capture specific and immediate feedback following the CBL activity. This approach aimed to quickly collect accurate observations, enabling participants to express their responses while the experience was still fresh in their minds.

Interestingly, the mean scores significantly improved after the ECSA in all the five domains. These results are in agreement with those by Shigli et al. who conducted a study on dental interns assessing their knowledge related to hyperplastic tissues in complete denture patients. The authors reported a significant improvement in the post activity assessment (p<0.001). It is noteworthy that the results of our study found a drastic improvement after the ECSA in each theme, except dental trauma management. It appears that this discrepancy might stem from differences in participant knowledge derived from textbooks, IADT guidelines, or practical experiences. Comparing how different resources were used or emphasizing specific areas of their learning process may shed light on why this specific domain did not exhibit a substantial increase post-CBL activity.

Another area highlighted in this study is the perception of participants regarding CBL using Likert scale. When responses were analyzed, majority of them acknowledged that they enjoyed CBL and it also promoted self-learning, improved implementation of key concepts and encourages interest in the field of the subject taught. These results were in agreement with those Shigli et al. who reported that CBL stimulates their study interest, promotes self-learning and facilitates solving clinical problems. The participants also perceived that CBL improved their ability to develop diagnosis & treatment planning skills, expand related knowledge and improve their confidence in solving any clinical problems. The results of the present study are consistent with those of Zhang et al who concluded that CBL is an effective method for improving students’ clinical diagnosis, reasoning, and logical thinking [14]. Interestingly, when participants were asked if ‘CBL was less beneficial than lectures’ a variable response was evident. Majority of them disagreed that CBL was less beneficial than lectures (41.3%) followed by those who neither agreed nor disagreed (34.5%) and a small proportion who agreed with this statement (24.2%). This ambiguity could be due to the diverse learning preferences and experiences among individuals [15,16]. Understanding the reasons behind these disparities is critical to increasing the effectiveness and acceptability of CBL. Exploring the factors influencing participants’ perspectives, such as prior experience to teaching methods, comfort levels with various learning approaches, and perceived strengths and shortcomings of both CBL and lectures, should shed light on this ambiguity.

Despite its novelty, certain limitations were encountered while carrying out this study. Since the study was based on a single day event, it was not possible to provide a comparison of CBL with lecture-based learning. Moreover, since this was a preliminary study, the sample size was limited, and the results should be interpreted keeping these limitations in mind. Our recommendations are that more multicenter longitudinal and randomized clinical trials should be conducted with large sample size to evaluate long term results of CBL in Endodontics.

Conclusion

Participants perceived an improvement in diagnosis, treatment planning and clinical judgement after the ECS activity. Moreover, the CBL activity significantly improved the scores of the participants. However, since this was a preliminary assessment, further research is warranted to develop a better understanding of the role of CBL in teaching endodontics.

References

  1. Zhao B, Potter DD (2016) Comparison of Lecture Based Learning vs Discussion Based Learning in Undergraduate Medical Students. J Surg Educ 73: 250-7. [crossref]
  2. Miller CJ, Metz MJ (2014) A comparison of professional-level faculty and student perceptions of active learning: its current use, effectiveness, and barriers. Adv Physiol Educ 38: 246-52. [crossref]
  3. Tayem YI (2013) The Impact of Small Group Case-based Learning on Traditional Pharmacology Teaching. Sultan Qaboos Univ Med J 13: 115-20. [crossref]
  4. Zhao WH, Deng L, Zhu W, Su J, Zhang A (2020) The effectiveness of the combined problem-based learning (PBL) and case based learning (CBL) teaching method in the clinical practical teaching of thyroid disease. BMC Med Educ.20: 381-91. [crossref]
  5. IlgüyMI, Fişekçioğlu EO (2014) Comparison of case-based and lecture-based learning in dental education using the SOLO taxonomy. J Dent Educ 78: 1521-7. [crossref]
  6. Williams B (2005) Case based learninga review of the literature: is there scope for this educational paradigm in prehospital education? Emerg Med J.22: 577-81. [crossref]
  7. Nadershahi NA, Bender DJB, Lyon CB (2013) An overview of casebased and problembased learning methodologies for dental education. J Dent Educ 77: 1300-5. [crossref]
  8. Thistlethwaite JE, Davies DEea (2012) The effectiveness of casebased learning in health professional education. A BEME systematic review: BEME Guide No. 23. Med Teach 34: e421-44. [crossref]
  9. McLean SF (2016) Case-Based Learning and its Application in Medical and Health-Care Fields: A Review of Worldwide Literature. J Med Educ Curric Dev.3: 39-49.
  10. Bi, M, Zhao Z, Yang J, Wang Y (2019) Comparison of casebased learning and traditional method in teaching postgraduate students of medical oncology. Med Teach 41: 1124-8. [crossref]
  11. Shigli KA, Fulari DS, Huddar D, Vikneshan M (2017) Case-based learning: A study to ascertain the effectiveness in enhancing the knowledge among interns of an Indian dental institute. J Indian Prosthodont Soc 17: 29-34. [crossref]
  12. McManus I, Richards PW, BC Sproston, KA (1998) Clinical experience, performance in final examinations, and learning style in medical students: prospective study. BMJ 316: 345-50. [crossref]
  13. Chutinan SK,Chien et al (2021) Can an interactive casebased activity help bridge the theory-practice gap in operative dentistry? Eur J Dent Educ 25: 199-206. [crossref]
  14. Zhang SYZ, YangJW, Zhang C, Shen ZY, Zhang GF, et al (2012) Case-based learning in clinical courses in a Chinese college of stomatology. J Dent Educ 76: 1389-92. [crossref]
  15. Alfarsi W, Elaghoury AH, Kore SE (2023) Preferred Learning Styles and Teaching Methods Among Medical Students: A CrossSectional Study. Cureus 15: e46875. [crossref]
  16. İlçin NT, Yeşilyaprak SS (2018) The relationship between learning styles and academic performance in TURKISH physiotherapy students. BMC Med Educ 18: 291. [crossref]
FIG 2

Accelerating Critical Thinking to Industrial Pace and Scale Through AI: Addressing the Global Issue of Food Sustainability

Leave a reply

DOI: 10.31038/NRFSJ.2024714

Abstract

We present a new, systematized way to teach critical thinking, using AI (artificial intelligence) incorporated into a research tool created for a newly emerging science, Mind Genomics, that is concerned with how people respond to ideas concerning everyday experiences. Mind Genomics methodology requires the researcher to develop four questions which ‘tell a story,’ and for each question to provide four alternative answers. Previous studies showed that many users experienced difficulty creating the questions. To overcome this problem, Mind Genomics incorporates AI through the mechanism of the Idea Coach. This mechanism allows the researcher to describe the problem being addressed, and then generates 15 questions the researcher evaluates and chooses for returns with 15 questions during the course of setting up the study’s story. Idea Coach provides additional analyses on the questions returned to reveal deeper structure and stimulate critical thinking by the researcher. We demonstrate the capabilities of the process by comparing the results for ‘food sustainability’ for people who are defined to be poverty stricken, first in the United States, and then in Ghana, and finally in Egypt. The effort requires approximately 10 minutes in total and is scalable for purposes of education and practical use.

Introduction: The Importance of Critical Thinking to Solve Problems

In order to address issues facing humanity, such as sustainability, it is important to be able to think clearly about the nature of the problem, and from there proceed to solutions. The importance of critical thinking cannot be underestimated, most apparently in education [1,2], but also in other areas, such as dentistry [3], not to mentioned the very obvious importance of critical thinking in areas where there are opposing parties confronting each other with the weapons of knowledge and thinking, such as the law [4]. The very idea of dealing with the United Nations’ (UN) 24 defined Global Issues (United Nations, undated) calls into play the need to understand and then deal with the problem. Critical thinking, or its absence has been recognized as a key feature in the solution of these problems. From the UN’s perspective, their 24 issues need to be addressed continually over time, strongly suggesting that the need for critical thinking is not limited in time but needs to be engaged with through time.

In today’s world, critical thinking is recognized as important for society [5]. The key question is not the recognition of critical thinking, but rather how to encourage it in a way which itself is sustainable, in a way which is cost-effective, scalable, and productive in terms of what it generates. To the degree that one can accelerate critical thinking, and even more so to focus critical thinking on a problem, one will most likely be successful . Finally, if such critical thinking can be aided by technical aids, viz., TACT (Technical Aids to Creative Thought), there is a greater chance of success. The notion of the aforementioned approach TACT was first introduced to the senior author HRM by the late professor Anthony Oettinger of Harvard University in 1965, almost 60 years ago. This paper shows how today’s AI can become a significant contributor to TACT, and especially to critical thinking about UN based problems, this one being food sustainability [6].

The topic of food sustainability is just one of many different topics of the United Nations, but one seeing insufficient progress (UN undated). From the point of view of behavioral science, how does one communicate issues regarding food sustainability? And how does one move beyond the general topic to specific topics? It may well be that with years of experience in a topic the questions become easier, but what about the issue of individuals wanting to explore the topic but individuals without deep professional experience? Is it possible to create a system using AI which can teach in a manner best called Socratic, i.e., a system which teaches by laying out different questions that a person could ask about a topic?

The Contribution of Mind Genomics to Critical Thinking about a Problem

During the past 30 years, researchers have begun to explore the way people think about the world of the everyday. The approach has been embodied in an emerging science called Mind Genomics (REF). The foundation of Mind Genomics is the belief that we are best able to understand how people think about a topic by presenting them with combinations of ideas, and instructing these people to rate the combination of ideas on a particular rating scale, such scales as relevance to them, interest to them, perceived solvability, etc. The use of combinations of ideas is what is new, these combinations created in systematic manner by an underlying structure called an experimental design. The respondent who participates does not have to consciously think about what is important, but rather do something that is done every day, namely choose or better ‘rate’ the combinations on a scale. The analysis of the relation between what is presented and what is rated, usually through statistics (e.g., regression) ends up showing what is important.

The process has been used extensively to uncover the way people think about social problems [7], legal issues [8], etc.. The process is simple, quick and easy to do, prevents guessing, and ends up coming up with answers to problems.

The important thing here is that the researcher has to ask questions, provide answers, and then the computer program matches the answers together into small groups, vignettes, presents these to the respondent, who has to rate he group or the combination.

Of interest here is the front end of the process, namely, how to ask the right question. It is asking questions which has proved to be the stumbling block for Mind Genomics, since its founding in 1993 (REF). Again, and again researchers have request help to formulate the studies. It is no exaggeration to state that the creation of questions which tell a story has become one of the stumbling blocks to the adoption of Mind Genomics.

Early efforts to ameliorate the problem involved work sessions, where a group of experts would discuss the problem. Although one might surmise that a group of experts in a room certainly could come up with questions, the opposite was true. What emerged was irritation, frustration, and the observation that the experts attending either could not agree on a question, or in fact could even suggest one. More than a handful of opportunities to do a Mind Genomics project simply evaporated at this point, with a great deal of disappointment and anger covering what might have been professional embarrassment. All would not be lot, however, as many of the researchers who had had experienced continued to soldier on, finding the process relatively straightforward. Those who continued refused to let the perfect get in the way of the good. This experience parallels what has been previously reported, namely that people can ask good questions, but they need a ‘boost’ early on [9].

The Contribution of AI in 2023

The announcement of AI by Open AI in the early months of 2023 proved to provide the technology which would cut the Gordian knot of frustration. Rather than having people have to ‘think’ through the answer to the problem with all of the issues which would ensue, it appeared to be quite easy to write a query about a topic and have the Mind Genomics process come up with questions to address that query. It was, indeed, far more enjoyable to change the ingoing query, and watch the questions come pouring out. It would be this process, a ‘box for queries’ followed by a standardized report, which would make the development fun to do.

Figure 1 shows what confronted the researcher before the advent of AI, namely an introduction page which required the researcher to name the study, followed immediately by a dauntingly empty page, requesting the research to provide our questions which tell a story. The researcher has the option to invoke AI for help by pressing the Idea Coach button.

FIG 1

Figure 1: Panel A shows the first screen, requiring the respondent to name the study. Panel B shows the second screen, presenting the four questions to be provided by the respondent.

Results Emerging Immediately and After AI Summarization

The next set of tables shows the questions submitted through the query to Idea Coach, the immediate set of 15 questions returned within 5-15 seconds. Later on, we will see the results after AI summarization has been invoked on the different set of questions.

In the typical use of Mind Genomics, the researcher often ends up submitting the squib to Idea Coach from a minimum of one time, but more typically 3-5 times, occasionally modifying the squib, but often simply piling up different questions. These different questions, 15 per page, provide a valuable resource of understanding the topic through the question. Typically, about 2/3 of the questions are different from those obtained just before, but over repeated efforts many of the questions will repeat.

Table 1 shows the first set of 15 questions for each of three countries, as submitted to Idea Coach. Note that the squib presented to Idea Coach is only slightly different for each country, that difference being only the name of the country. The result, however, ends up being 15 quite different questions for each country, questions which appear to be appropriate for the country. It is important to emphasize here that the ‘task’ of AI is to ask questions, not to provide factual information. Thus, the issue of factual information is not relevant here The goal is to drive thinking.

Table 1: Query & Questions for United States, Ghana and Egypt. These 15 questions emerged 10-15 seconds after the query was submitted to Idea Coach.

TAB 1

It is important to note that Table 1 can be replicated as many time as the researcher wishes. The questions end up allowing the researcher to look at different aspects of the problem. The results come out immediately to the researcher, as well as being stored in a file for subsequent AI ‘summarization’ described below. At the practical level, one can imagine a student interested in a topic looking at the questions for a topic again and again, as the student changes some of the text of the query (viz, the squib shown in Figure 2, Panel B). It is worth emphasizing that the Idea Coach works in real time, so that each set of 15 questions can be re-run and presented in the span of 5-15 seconds when the AI system is ‘up and running.’ Thus, the reality ends up being a self-educating system, at least one which provides the ‘picture of the topic’ through a set of related questions, 15 questions at a time. The actual benefit of this self-pacing learning by reading questioning is yet to be quantified in empirical measures, however (Figure 3).

FIG 2

Figure 2: Panel A shows the information about Idea Coach. Panel B shows the ‘box’ where the researcher creates the query for Idea Coach, in terms of ‘shaping’ the structure and information of the question.

FIG 3

Figure 3: The first six questions out of the 15 returned by Idea Coach to answer the request shown in Figure 2, Panel B. The remaining nine questions are accessed by scrolling through the screen.

It is relevant to note that AI-generated questions are beginning to be recognized as an aid to critical thinking, so that the Idea Coach strategy can be considered as part of the forefront of what might be the 21st century TACT program, Technical Aid to Creative thought (Oettinger, 1965, personal communication). Papers such as the new thesis by Danry [10] of MIT reflect this new thinking. Half-way around the world the same approaches are being pioneering in the Muslim world [11].

Once the questions are presented, it is left to the researcher to move on to completing the set-up of the Mind Genomics study, or to further request additional sets of 15 questions. When the creation of questions is complete, the researcher is instructed to provide four answer for each question. A separate paper will deal with the nature of ‘answers’ to the questions. This paper deals only with the additional analysis of the questions generated by Idea Coach.

AI Summarization and Extensions of Sets of 15 Questions

The second part of Idea Coach occurs after the researcher has competed the selection of the four questions, as well as completing the generation or selection of the four answers for each question. This paper does not deal with the creation of answers, but the process is quite similar to the creation of questions. The researcher creates the set of four questions, perhaps even editing/polishing the questions to ensure proper understanding, and tone. Once the questions are published, the Idea Coach generates four answer to each question. The entire process of summarization, for all of the set of 15 questions, takes about 15-30 minutes. The Excel file containing the ‘Answer Book’ with all summarizations is generally available 20 minutes after the questions and answers have been selected. The Answer book is available for download at the website (www.BimiLeap.com) and is emailed to the researcher as well.

We now go into each part of the summarization. The actual summarization for each set of 15 questions is presented on one tab of the Idea Book. We have broken up the summarizations into each major section, and then present the summarization by AI for the USA, followed by for Ghana, and finally for Egypt In this way the reader can see how the initial squib, the prompt to Idea Coach, differing only in the country, ends up with radically different ideas.

Key Ideas

The output from the first prompt had produced full questions. The ‘Key Ideas’ prompt strips the question format away, to show the idea or issue underlying the question. In this way, the ‘Key Ideas prompt can be considered simply as a change in format, with no new ideas emerging. Table 2 shows these ideas. It is not clear which is better to use. To the authors, it seems to be more engaging to present the ideas in the form of a question. When presenting the same material as ideas seems to be more sterile, less engaging, and without grounding.

Table 2: Key Ideas underlying the 15 questions

TAB 2

The use both of questions and of the ideas on which these questions are based have been addressed as part of an overall study of the best ways to learn. In the authors’ own words ‘Likewise, being constructive is better than being active because being constructive means that a learner is creating new inferences and new connections that go beyond the information that is presented, whereas being active means only that old knowledge is retrieved and activated.’ [12]

Before moving on to the next section, one may rightfully ask whether a student really learns by being given questions which emerge from a topic, or whether it is simply better to let the student flounder around, come up with questions, and hopefully discover other questions, either by accident, or by listening to the other students answer the same question and gleaning from those other answers new points of view [13]. The point of view taken here is that these aids to creative thought do not provide answers to questions, but rather open up the vistas, so that the questioner, research or student, can think is new but related directions. The output are additional, newly focused questions, rather than answers which put the question to rest. Quite the opposite [14]. The question opens up to reveal many more dimensions perhaps unknown to the researcher of the student when the project was first begun. In other words, perhaps the newly surfaced questions provide more of an education than one might have imagined.

Themes

With themes Idea Coach moves toward deconstructing the ideas, to identify underlying commonalities of issues, and the specific language in the questions supporting those commonalities. With ‘Themes’ the AI begins the effort to teach in a holistic manner, moving away simply from questions to themes which weave through the questions. For the current version of Idea Coach, the effort to uncover themes is done separately for each set of 15 questions, in order to make the task manageable. In that way the researcher or the student can quickly compare the themes generated from questions invoking the United States versus questions invoking Ghana, or questions invoking Egypt. Table 3 gives a sense of how the pattern of themes differ [15]. It is also important that the organization shown in Table 3, is the one provided by the Idea Coach AI, and not suggested by the researcher. Note that for Egypt, as contrasted with the USA and Ghana, Idea Coach refrained from grouping ideas into themes, but treated each idea as its own theme.

Table 3: Themes emerging from the collection of 15 questions for each country

TAB 3

Perspectives, an Elaboration of Themes

Perspectives advances the section of themes, which had appeared in Table 4. Perspectives takes the themes, and puts judgment around these themes, in terms of positive aspects, negative aspects, and interesting aspects. Perspectives are thus elaborations of themes. In other words, perspectives ends up being an elaboration of themes, useful as a way to cement the themes into one’s understanding.

Table 4: Perspectives (an elaboration of Themes)

TAB 4(1)

TAB 4(2)

TAB 4(3)

What is Missing

As the analysis moves away from the clarification of the topic, it moves towards more creative thought. The first step is to find out what is missing, or as stated by Idea Coach, ‘Some missing aspects that can complete the understanding of the topic include: ’ It is at this point that AI moves from simple providing ideas to combining ideas, and suggesting ideas which may be missing.

It is at this stage, and as the stage of ‘innovation’ that AI reaches a new level. Rather than summarizing what has been asked, AI now searches for possible ‘holes’ and a path towards greater completeness in thinking. Perhaps it is at this level of suggesting missing ideas that the user begins to move into a more creative mode, although with AI suggesting what is missing one cannot be clear whether it is the person who is also thinking in these new directions, or whether the person is simply moving with the AI, taking in the information, and enhancing their thinking (Table 5).

Table 5: What is missing

TAB 5(1)

TAB 5(2)

Alternative Viewpoints

Alternative viewpoints involve arguing for the opposite of the question. We are not accustomed to thinking about counterarguments in the world of the everyday. Of course, we recognize counterarguments such as what occurs when people disagree. Usually, however, the disagreement is about something that people think to be very important, such as the origin of climate change or the nature of what climate change is likely to do. In such cases we routinely accept alternative viewpoints.

The Idea Coach takes alternative viewpoints and counterarguments to a deeper stage, doing so for the various issues which emerge from the questions. The embedded AI takes an issue apart and looks for the counterargument. The counterargument is not put forward as fact, but simply as a possible point of view that can be subject to empirical investigation for proof or disproof (Tables 6-9).

Table 6: Alternative viewpoint, showing negative arguments countering each point uncovered previously by Idea Coach using AI.

TAB 6(1)

TAB 6(2)

Table 7: Interested audiences.
The next AI analysis deals with the interested audiences for each topic. Rather than just listing the audience for each topic, the Idea Coach goes into the reasons why the audience would be interested, once again providing a deeper analysis into the topic, along with a sense of the stakeholders, their positions, their areas of agreement and disagreement.

TAB 7(1)

TAB 7(2)

Table 8: Opposing audiences.
Once again, in the effort to promote critical thinking, the Idea Coach provides a list of groups who would oppose the topic, and for each group explain the rationale for their opposition.

TAB 8

Table 9: Innovations.
The final table selected for the Idea Coach summarization is innovations, shown in Table 9. The table suggests new ideas emerging from the consideration of the questions and the previous summarizations. Once again the ideas are maintained with the constraints of the topic and reflect a disciplined approach to new ideas.

TAB 9(1)

TAB 9(2)

TAB 9(3)

Discussion and Conclusions

The goal of the paper has been to show what is currently available to students and researchers alike. The objective of the demonstration has been to take a simple problem, one that might be part of everyday discourse, and use that problem to create a ‘book of knowledge’ from the topic, using questions and AI elaboration of the questions.

We hear again and again about the importance of critical thinking, but we are not given specific tools to enhance critical thinking. As noted in the introduction, in the 1960’s, the late professor Anthony Gervin Oettinger of Harvard University began his work on creative thought. We might not think that programming a computer to go shopping is an example of creative thought, but in the 1960’s it was (Oettinger, xxx). Now, just about six decades later, we have the opportunity to employ a computer and AI create books that help us thinking critically about a problem. We are not talking here about giving factual answers, actual ‘stuff,’ but really coaching us how to think and how to think comprehensively about the ideas within a societal milieu, a milieu of competing ideas, of proponents and opposers who may eventually agree on solutions that address or resolve issues as thorny as food sustainability. If sixty years ago teaching a computer (the EDSAC) to go shopping was considered a TACT, a technical aid to creative thought, perhaps now co-creating a book of pointed inquiry about a topic might be considered a contribution of the same type, albeit one more attuned to today. The irony is that sixty years ago the focus was on a human programming a machine ‘to think,’ whereas today it is the case of a machine coaching a human how to think. And, of course, in keeping with the aim of this new to the world journal, the coach is relevant to thinking about any of the topics germane to the journal. This same paper could be created in an hour for any topic.

References

  1. Cojocariu VM, Butnaru CE (2014) Asking questions–Critical thinking tools. Procedia-Social and Behavioral Sciences 128: 22-28.
  2. Lai ER 2011. Critical thinking: A literature review. Pearson’s Research Reports 6: 40-41.
  3. Miller SA& Forrest JL (2001) Enhancing your practice through evidence-based decision making: PICO, learning how to ask good questions. Journal of Evidence Based Dental Practice 1: 136-141.
  4. Nicholar J, Hughes C & Cappa C (2010) Conceptualising, developing and accessing critical thinking in law. Teaching in Higher Education 15: 285-297.
  5. ŽivkoviĿ S 2016. A model of critical thinking as an important attribute for success in the 21st century. Procedia-Social and Behavioral Sciences 232: 102-108.
  6. Bossert WH, Oettinger AG 1973. The Integration of Course Content, Technology and Institutional Setting. A Three-Year Report, 31 May 1973. Project TACT, Technological Aids to Creative Thought.
  7. Moskowitz H, Kover A & Papajorgji P (eds), (2022) Applying Mind Genomics to Social Sciences. IGI Global.
  8. United Nations, Undated. “Global Issues,” accessed January 27, 2024.
  9. Rothe A, Lake BM & Gureckis TM 2018. Do people ask good questions? Computational Brain & Behavior 1: 69-89.
  10. Danry VM (2023) AI Enhanced Reasoning: Augmenting Human Critical Thinking with AI Systems (Doctoral dissertation, Massachusetts Institute of Technology).
  11. Fariqh N 2023, October. Developing Literacy and Critical Thinking with AI: What Students Say. In .Proceedings Annual International Conference on Islamic Education (AICIED) 1: 16-25.
  12. Chi MTH 2009. Active-Constructive-Interactive: A Conceptual Framework for Differentiating Learning Activities. Topics in Cognitive Science 1: 73-105. [crossref]
  13. Moskowitz HR, Wren J & Papajorgji P 2020. Mind Genomics and the Law. LAP LAMBERT Academic Publishing.
  14. Niklova, N (2021) The art of asking questions: Flipping perspective. In: EDULEARN21 Proceedings Publication 2816-2825.
  15. Oettinger AG. Machine translation at Harvard 2003. In: Early Years in Machine Translation, Memoirs and Biographies of Pioneers, (ed. W.J. Hutchins), John Benjamin’s Publishing Company, Amsterdam/Philadelphia, pp. 73-86.

Understanding the Mind and Inventing the Future: The Problem of Failure to Show Up for Follow-Up Appointments with One’s Health Provider

Leave a reply

DOI: 10.31038/ASMHS.2024813

Abstract

The paper introduces a system to deal with problems of society using SCAS, Socrates as a Service. SCAS is provided with a detailed description of a conventional problem faced by people, and in turn instructed to defined prospective mind-sets in the population who suffer with this problem. SCAS further provides information on the nature of these hypothesized mind-sets, what the mind-sets are thinking, and how the mind-sets would respond to topic-relevant slogans that would be generated to solve the problem. Finally, the paper finishes with the use of SCAS to summarize the issue, provide perspectives that people might have, and identify what next steps need to be taken, as well as innovations that should be introduced which deal with and even solve the problem. SCAS is a general approach. The paper here uses SCAS to investigate the ‘why’ patients fail to keep their doctor’s visits, and what innovations might solve the problem.

Introduction

This paper grew out of the recognition that all too often patients fail to follow the suggestions of their medical and health professionals. The topic of compliance is a large one. The focus of this paper is on the simple problem of patients not showing up at the prescribed time for their follow-up appointments. The damage which ensues can be enormous, impacting the health of the patient, the cost to the medical practice, and the disruption of a system which must accommodate the schedules of a variety of people who then must regroup and update the schedules [1].

When dealing with this problem, we are actually dealing with issues of communication interacting with motivation and habit. How does the medical establishment work with individuals to ensure that they come to scheduled appointments. The importance of this question can be easily understood when one realizes the number of reminder messages which appear on the smartphones of patients, telling them of the upcoming appointment, asking them to ‘e-check in’ and then giving them the chance to cancel and reschedule. This and other actions such as reminder phone call are the obvious effort to minimize the expensive ‘no-shows’. In recent years, the process has been automated, with AI-driven chatbots and voice interactions finding their place in the seemingly impossible to solve conundrum of getting patients to sow for their appointments [2].

The business literature recognizes the problems of ‘no-shows’. The issues underlying the no-shows are extensive, as are the suggestions for improvement. The case of medicine is particular serious for no-shows simply because one cannot necessarily move the appointment to some later time and ‘go from there.’ A person’s health is labile. Moving a scheduled appointment a month or two later, when a slot opens up, may be too late when the issue is the follow up from what can be a serious problem, and when not treated can evolve to a life-threatening one. One serious illness often comes to the fore, diabetes. The consequence of missing a follow up appoint with a doctor when the person has diabetes 2 can be severe [3-6].

The Contribution of Mind Genomics Enhanced by SCAS (Socrates as a Service)

The problem of no-shows was first brought into the world of Mind Genomics through collaboration with physicians in Chicago, IL, specifically anesthesiologist Dr. Glen Zemel. Author Moskowitz collaborated with Dr. Zemel on a variety of studies dealing with the mind of the patient in the hospital. As a practicing anesthesiologist, Zemel often recognized the issues involved in patients who fail to follow up, often even having to forego surgery on the particular scheduled date because either they ‘forgot’ (rare) or forgot to follow the requirements of avoid food for the previous 12 hours and so forth. It was these immediate issues which ended up costing the medical practice many thousands of dollars.

The problem became more acute when authors Braun and Mulvey, and later Cooper, became involved in the issue of patients who failed to follow up at specific times. These individuals suffered from a variety of metabolic disorders; the most common one being diagnosed as pre-diabetic. The failure to return at the scheduled time for a follow-up morphed from being a financial loss to a medical practice into the possibility that diabetes might develop because the pre-diabetic essentially disappeared, but presumably the condition remained with the individual.

The evolution of Mind Genomics into a much deeper use of AI opened up the questions about what SCAS might be able to contribute to an understanding of why people fail to go to follow-up appoints with their doctor after learning that they are suffering from a serious condition. Could AI provide insights, especially with the newly discovered ability to ‘prime’ AI with a detailed background of an issue, and then instruct AI to ‘flesh out’ what might be going on in the mind of a person? As we move through the topics in this paper we must keep in mind that everything presented here regarding ‘thinking’ is the result of instructing Socrates as a Service (SCAS), viz., a version of AI powered by Chat GPT 3.5 [7].

Demonstration: Priming AI to Simulate Poor Patients Living in Public Housing

The remainder of this paper presents the results of a simulation using SCAS (Socrates as a Service, a form of AI growing out of ChatGPT 3.5), and the secondary analysis, viz., AI summarization of the data generated by the SCAS simulation. The important thing to keep in mind is that there is almost no information of any substantive import presented by the user, other than the initial framing of the situation, and what the user wants to ‘discover’ by having AI simulate the answers in place of having a human being do so.

The process begins with the orientation provided to AI, shown in Table 1. The table divides into three sections.

Table 1: The input to Socrates as a Service (SCAS)

tab 1

Section 1 – Input Information to SCAS

Here, the user creates a general picture of the situation. The input positions the user as a person working in a clinic in a poor area in Brooklyn. One might this simulation with a variety of different so-called general pictures, such as stating that the area is inhabited by upper middle classes, that the person works in a concierge medical service, that the location is somewhere else. With that flexibility the user would be well on the way to parametrically exploring the different alternatives. The opportunities are endless.

Section 2 – Understanding the Mind-sets

Here the user presents SCAS with a minimum amount of information, sufficient however to allow SCAS to create mind-sets. The user does not define the concept of mind-set, nor does the user give any hint about what properties are possessed by the three mind-sets. Given only this minimal amount of information, really only one piece of information, that there are three mind-sets, the system requests AI to create names, and inner thoughts of these three mind-sets.

Section 3 – Request that SCAS Produce 12 Message and Estimate the Performance of Each Message among the Three Mind-sets

The final request generated the desired 12 messages to be evaluated by three mind-sets. It is important to emphasize that nowhere in the instructions is any information presented to SCAS program that could be considered to be a subject-relevant prompt. All of the information generated by SCAS comes from the way SCAS processes the request.

Table 2 present the first part of the output, viz., the three mind-sets, explicated in terms of what each mind-set thinks at the time of making the appointment, and then a week before the appointment. The remarkable thing emerging from Table 2 is the realistic nature of the mind-sets and their thoughts. Once could easily think that these are verbatim quotes emerging from a discussion with the patient about the issue of making and keeping medical appointments.

Table 2: The description of the three mind-sets emerging from SCAS. As noted in the text, SCAS was not given any specific material on mind-sets which to base what it returned to the user.

tab 2

Table 3 shows how each of the three mind-sets would estimate the likelihood of showing up for the follow up medical appointment if the mind-set were to be reminded through the slogan. The slogans were created by SCAS. SCAS ‘predicts’ that all 12 would be effective for Mind-Set 1 (proactive), effective for Mind-Set 3 (Anxious), but not particularly effective for Mind-Set (Carefree). Once again it should be noted that these results make sense. We expected a mind-set named Carefree not to care about any of the messages, and thus not pay attention to follow up messages with the slogans shown in Table 3.

Table 3: Estimated likelihood of showing up for the follow-up appointment, for each of 12 slogans by each of the three mind-groups. Everything was generated by SCAS, using only the input to SCAS shown in Figure 1.

tab 3

Inventing the Future Using Today’s Topics

The second part of this paper focuses on the use of SCAS to understand what to do in order to improve the compliance of patients regarding their requested follow up visit. The original use of SCAS was to allow the user to type a ‘squib’ or information about a topic and have SCAS return with a set of 15 questions. The same feature was available for SCAS to return 15 answer to a given question. These feature remain in SCAS, and led to an effort to compare the answers to the same questions when SCAS was told that the answers had to be appropriate for the 21st century (now), and then that the answers had to be appropriate for the 22nd century (75 years hence).

The same set of 15 questions was used to compare the answers for the two centuries. The SCAS was primed to provide four separate answers to each of the 15 questions, requiring the answers to be appropriate for the 21st century (Table 4, answers A-D), and then be appropriate for the 22nd century (Tabe 4, answers E-H, italicized). Table 4 suggest that the answers for the 22nd century seem reasonable, and to be extensions of current day technology.

Table 4: Fifteen SCAS-generated topic-related questions about office visits to the medical professional. Each question shows four SCAS-generated questions assuming a year in the 21st century, and then a year in the 22nd century.

tab 4(1)

tab 4(2)

tab 4(3)

Summarization of Information Proposed – Broad Overviews Produced by SCAS

When the Mind Genomics study has been closed, SCAS creates a set of summarizations for each iteration, doing the summarizations separately. These summarizations are returned to the user in an email, usually within a half hour after the close of the study. Thus, in the not-unusual case of the user doing 10-15 iterations with different squibs, e.g., exploring different time periods with the same instructions, the user will receive one page for each effort, all of the pages becomes tabs in the one Excel workbook.

Table 5 shows one set of summarization, aptly summarized ‘Ideas’. The three summarization are key ideas, themes, and then perspectives.

Table 5: Summarization of the output from SCAS in terms of key ideas, themes emerging from the key ideas, and then a discussion of the positives and negatives of each theme.

tab 5

Key ideas simply highlights what the term suggests, namely what are the ideas presented to the user. This study generates a great number of key ideas because input to the studies comprises the basic questions and the answers pertaining only to the 21st Century both shown in Table 4.

Themes further summarize the key ideas, this time using SCAS to group together the related group of key ideas, perspectives, in turn, take these themes and provide the basis for ongoing discussion and learning, showing two alternative points of view for each theme.

The ’Human Reaction’ to These Ideas, as Envisioned by SCAS

As part of the summarization, SCAS returns with three different analyses of the sets of ideas. The analyses look at populations of people, whether these populations be defined by who they are (for both interested and opposing audiences), or by the way they think (alternative viewpoints). Table 6 shows the various groups and their reactions to the ideas uncovered by SCAS. It is important to keep in mind that these reactions are to the general ideas, not to any specific idea.

Table 6: The ‘human’ reaction to these ideas as envision by SCAS

tab 6

The final analysis deals with SCAS as an inventor. Table 7 shows two sections. The first section lists questions about what may be missing. These are typically questions which ask: How do we… ? The second section lists possible innovations, based upon the information processed by SCAS. The list of possible innovations is organized by topic.

Table 7: Using SCAS to suggest new products and services

tab 7(1)

tab 7(2)

Discussion and Conclusions

This paper emerged from recurring discussions about the real problem of ‘no shows’ in the world of medicine. The problem is a vexing one, perhaps growing because of the increasing difficulties encountered in the practice of medicine. One problem is the growing lack of affordability of medical treatments, the cost perhaps acting as a mechanism to discourage visits because of the fear of incurring expenses that are unaffordable to the patient. A second problem is the reality that doctors no longer make house calls. The patient must go to the doctor, a trip which might be difficult to schedule in view of the competing demands on the patient’s time. The third is the loss of the personal relationship between patient and doctor as the small, perhaps long-time practices are incorporated into the large medical practices. What was a personal relationship between patient and doctor (or other medical professional) now becomes a short interaction, often with the doctor’s assistant taking the necessary measurements, and the doctor meeting the patient for a few minutes debrief [8].

The importance of this paper is not in the solution in provides, but rather in the way SCAS can help focus the problem, providing a source of ideas. The speed (minutes), the extensive results in terms of the ‘human element’, and the presentation of the results in an easy-to-understand format, all suggest that those in the medical profession might avail themselves of SCAS as they enter a new subject area, if only to understand some of the issues from the part of the patient, the doctor, and the system. Scattered publication suggested only the positive, the ‘up-side’, and not the down-side of using AI and such offshoots as SCAS to solve the problem of no-shows [9].

A second aspect of the approach presented here comes from the potential of instructing SCAS to ‘imagine’ what will happen in the years to come, or even to imagine what things were like a century ago or even longer. By simply asking SCAS to assume that all the topics are to be asked from the framework of the year 2200, almost 75 years into the future, it is possible to jump-start futuristic thinking. There is no reason to assume that the answers will be ‘correct.’ On the other hand, to SCAS there is no penalty for being ‘wrong’, so that SCAS dutifully produces its best guess, once it has been properly instructed. It is that potential to focus on the future in terms of concrete questions and suggestions which make the approach attractive, especially in light of the simplicity of executing just another ‘iteration,’ albeit this time priming SCAS to guess about the future or guess about the past [10,11].

References

  1. Parsons J, Bryce C, Atherton H (2021) Which patients miss appointments with general practice and the reasons why: a systematic review. British Journal of General Practice 71(707): e406-e412. [crossref]
  2. Nadarzynski T, Miles O, Cowie A, Ridge D (2019) Acceptability of artificial intelligence (AI)-led chatbot services in healthcare: A mixed-methods study. DIGITAL HEALTH.v.5. [crossref]
  3. Kreps GL, Neuhauser L (2013) Artificial intelligence and immediacy: designing health communication to personally engage consumers and providers. Patient Education And Counseling 92: 205-10. [crossref]
  4. Lacy NL, Paulman A, Reuter MD, Lovejoy B (2004) Why we don’t come: patient perceptions on no-shows. The Annals of Family Medicine 2: 541-545. [crossref]
  5. Miller AJ, Chae E, Peterson E, Ko AB (2015) Predictors of repeated “no-showing” to clinic appointments Am J Otolaryngol 36: 411-4. [crossref]
  6. Sun CA, Taylor K, Levin S, Renda SM, Han HR (2021) Factors associated with missed appointments by adults with Type 2 Diabetes Mellitus: a systematic review. BMJ Open Diabetes Research & Care 9(1).
  7. Wu T, He S, Liu J, Sun S, Liu, K,et al.( 2023) A brief overview of ChatGPT: The history, status quo and potential future development. IEEE/CAA Journal of Automatica Sinica 10: 1122-1136.
  8. Bjerring JC, Busch, J (2021) Artificial intelligence and patient-centered decision-making. Philosophy & Technology 34: 349-371.
  9. Yun JH, Lee EJ, Kim DH (2021) Behavioral and neural evidence on consumer responses to human doctors and medical artificial intelligence. Psychology & Marketing 38: 610-625.
  10. Fogel AL, Kvedar JC (2018) Artificial intelligence powers digital medicine. npj Digital Med 1: 5. [crossref]
  11. Kuper A, Whitehead C, Hodges BD (2013) Looking back to move forward: using history, discourse and text in medical education research: AMEE Guide No. 73. Medical Teacher 35: e849-60. [crossref]

Explorations in Time Using SCAS (Socrates as a Service): Reimagining the Doctor’s Waiting Room of 1850 versus 2150 and the Evolution of that Room from 1600

Leave a reply

DOI: 10.31038/PSYJ.2024632

Abstract

With the help of AI-based SCAS (Socrates as a Service), developed to support Mind Genomics, the study considered the nature of the doctor’s waiting room of the year 1850, followed by a paragraph about the doctor’s waiting room in 50-year intervals, from years 1600 to 2350. SCAS produced basic information about the doctor’s office as it changed over the centuries and was able to use that basic information to create even more information regarding ideas for innovation. Mind Genomics was also prompted to suggest responses of acceptors versus rejectors of the features of the 1850 doctor’s office. The paper demonstrates the simplicity, speed, and depth of information that can be obtained using AI, and the promise of the coupling of interesting reading with deeper information.

Introduction: The ‘Draw’ of the ‘What Was’ and ‘What Will Be’

A continuing theme in many aspects of life is the fascination of what was and what will be. The world of history gives people a chance to experience what happened before, and the world of ‘future studies’ for want of a better term gives people a chance to look at trends and peer into a future which might be. Indeed, the focus on the world over time, before, now, and in the future, has given the world wonderful works of history, literature, philosophy, just to name a few disciplines.The introduction of AI, artificial intelligence, has made it possible to move beyond what has been published in history and in ‘futurology.’ Through its own mechanisms of deep learning, it may be possible to get a sense of what the past may have been, not so much from reading books, but from asking AI to paint a picture of a specific issue. Even more interesting may be the attempt to do the same, not so much painting a picture of the past as a picture of the reasonably near future, a few decades from now, or perhaps a century or so. It was the development of two technologies which, when combined, opened up the focus on the past. The first was the emerging technology of Mind Genomics [1]. In simplest terms, Mind Genomics is the study of the everyday, the ordinary events, material things, and behaviors. The second is the new availability of user-friendly AI, artificial intelligence, embedded in Mind Genomics as SCAS, (Socrates as a Service), and based upon current AI systems [2]. Mind Genomics opened up the possibility of studying the everyday more deeply, looking into features, painting a picture of a situation and understanding what is important to people. The result was the realization that the ordinary events of everyday, the quotidian life, are worth studying. SCAS, embodying easy to use AI, allowed the investigation of the everyday life, not by doing experiments but rather by asking the embedded AI to assume a situation, and then report on its details.

It is important to note that this paper follows in a stream of previous work, much of it trying to digitize the historical narrative, to make history ‘come alive’ to students [3-5]. Furthermore, a great deal of interest in AI-based simulation comes from the desire to add reality and depth to history-games, which are very popular. These games try to create a realistic ‘set’ and realistic ‘behaviors.’

Exploring a Simple Topic: The Doctor’s Waiting Room Across Years

The ‘research’ presented here began with a request to SCAS to present a short description of what the doctor’s office was like:

The year is: [provided by user]. Everything that is talked about here happened in [provided by user].The doctor is a general practitioner in New York

What should the waiting room of the doctor be like, in terms of decor, in terms of people, in terms of the way people are greeted. Write you answer as five sentences in one long paragraph, simply written, in order to give the reader a complete description. Make the writing lively, and fun to read, and make the description realistic, as if the person reading the description were to be right there, at this time and this place

Table 1 shows the results for three years, 1900, 2000, and 2150., respectively. The appendix to this paper shows many more years, beginning with 1600 and going to 2300 in 50-year leaps. The first reaction to the ‘first fruits’ of this effort are summarized by the ‘astonishment.’ The paragraphs describing the mundane topic of the doctor’s waiting room seem real, as if someone were there. This led to doing the ‘experiment’ with 50-year intervals, starting in the year 1600, and proceeding to the year 2250. The Appendix those short descriptions.

Table 1: Descriptions of the doctor’s office, product for three years, 1900, 2000 and 2150

TAB 1

Exploring the Doctor’s Waiting Room in Detail – The Year is 1850

The remainder of this paper shows an AI-based exploration using SCAS. The year is 1850. The general instructions appeared above. SCAC produces the immediate output shown in Table 2. The material is similar to what appears in Table 1, as well as in the Appendix. Once again, it is important to emphasize that the paragraph is synthesized by SCAS without any information other than the year, and the directive to provide the answer as a story in five sentences.

Shortly after the completion of the session, after the Mind Genomics program finishes, SCAS produces a summarization of the results. Within the summarization appear a detailed expansion of ideas, all based upon the five sentences shown in Table 2.

Table 2: The SCAS-generated description of the doctor’s waiting room of 1850

TAB 2

Key Ideas, Themes, and Perspectives

The first set of subsequent analyses present the various ideas, this time expanded. Once again, SCAS returns with an easy-to-read analysis, all based on what SCAS had produced initially in answer to a simple question. Essentially, therefore, SCAS is producing ‘new knowledge’ based upon ‘knowledge’ it had developed simply knowing the topic and the year. Note that the perspectives are different points of view about the topics presented in the section on themes (Table 3).

Table 3: Expansion of knowledge through the key ideas, themes, and perspective regarding those themes

TAB 3

SCAS provides a sense of who would be interested in the materials, and who would be ‘opposed’ to the materials. These appear in Table 4. Once again, it is SCAS which is working on the information it first generated to provide additional information or real points of view.

Table 4: Points of view, interested versus opposed

TAB 4

Steps Towards Innovation (of Knowledge)

The final summarizations deal with questions and ideas for innovation. For this historical exploration using SCAS there is no ‘innovation’ per se. Rather, the ‘innovations’ comprise questions to answer. These are presented in the sections called ‘Alternative Viewpoint,’ and ‘What is Missing,’ both in Table 5. SCAS does return with ‘innovations,’ but this is the one section in SCAS which as yet cannot put itself into the mind of the 1850 doctor to look at the innovation of that time.

Table 5: Questions to answer, to create new knowledge about the doctor’s waiting room in 1850

TAB 5

Discussion and Conclusions

The objective of this paper is to explore how deeply one can ‘flesh out’ an otherwise modestly interesting topic, the doctor’s waiting room, although a topic which has received attention in the popular literature [6]. There is a relevant academic literature dealing with the history of doctor’s offices and their furnishings [7,8]. It is likely, however, that the material being published will interest the experts, whether these experts be those who study the history of interior design [9], or the history of medicine [10]. There is also a developing literature on the additional aspects of the doctor’s waiting room, such as design, content, etc., based upon the recognition that the waiting room is not only a place to store people, but also to make their visit pleasant [11,12], and a chance to teach them [13]. There is always a need for solid academic work the topic. It is hoped that the simulation efforts with SCAS shown here adds to the bank of knowledge and contributes to the study of the history and sociology of those in the health field and those in the field of interior design.

The real opportunity presented in this paper emerges in the world of education. The use of Mind Genomics, and especially its easy use AI embodied in SCAS can result in a great deal of relevant information being produced in minutes, with the student able to modify the requests to SCAS, and in turn get new information in virtually seconds. Afterwards, there is the major contribution to education products from the SCAS-based summarization of the information. Each iteration, the effort taking about 30 seconds per iteration, is returned with a full summarization, one Excel tab for each iteration. A student excited about the prospects, can work for 30 minutes, generating a great deal of information, with the nature of the requested information dynamically changing according to the instructions written into the squib by the user, in this case the student. One can only imagine the level of excitement as the student works with SCAS and Mind Genomics a coaches, teaching the study many things in dept, and actively interacting with the student who wants to explore the topic in different ways.

A question that can be posed is how does this AI image of a doctors waiting room across the eras, past and future, coalesce with reality? One thing that can be considered in the current and future eras is the post-COVID 19 world where telehealth and social distancing has become the norm, particularly in healthcare settings. We therefore must consider the potential future of waiting rooms with the emergence of telemedicine as less crowded [14].This is important when we consider the impact of COVID on the layout of waiting rooms, with aspects such as social distancing, spacing the time between appointments in order to prevent crowded waiting rooms, and so forth. The emergence of ancillary healthcare personnel, from the licensure of higher and higher rankings in nurse practice levels, as well as the introduction of physician assistants, have made the visit to the doctors’ office a place where there could potentially be more individuals working at the back end than patients waiting in the front.

A more casual flair is also being approached in medical offices, from patients to healthcare workers alike, with a “casualization of the workforce” occurring [15], keeping in line with recent trends in society as a whole. This casualization may likely show itself in the change of the patient waiting room, from a room psychologically separate from where the medical professionals work to simple part of a continuum of space, with far less psychological separation. This change will manifest the evolving change in power of dominance by the medical professional over the patient to one of cooperation and collaboration. One need only see the change from the formal living, dining and kitchen spaces of traditional homes to their blending in new homes, as designed by forward looking architects with their forward-looking clients.

References

  1. Moskowitz HR (2012) ‘Mind genomics’: The experimental, inductive science of the ordinary, and its application to aspects of food and feeding. Physiol Behav 107: 606-13. [crossref]
  2. Kalyan KS (2023) A survey of GPT-3 family large language models including ChatGPT and GPT-4. Natural Language Processing Journal, p. 100048.
  3. Danyun L, Jiun CY (2016) October. Historical cultural art heritage come alive: Interactive design in Taiwan palace museum as a case study. In 2016 22nd International Conference on Virtual System & Multimedia (VSMM) (pp. 1-8) IEEE.
  4. Staley DJ (2015) Computers, Visualization, and History: How New Technology Will Transform Our Understanding of the Past. Routledge.
  5. Taylor T (2003) Historical simulations and the future of the historical narrative. Ann Arbor, MI: MPublishing 6: 2.
  6. Tanner LE (2002) Bodies in waiting: representations of medical waiting rooms in contemporary American fiction. American Literary History 14: 115-130.
  7. Gainty C (2019) Why Wait? Modern American History 2: 249-255.
  8. Waltz M (2016) Patient Patients: An Ethnography of Medical Waiting Rooms. Case Western Reserve University
  9. Figueroa NI (2016) Culture, gender, and medical waiting rooms: A Kuwaiti case study. Journal of Interior Design 41: 33-46.
  10. Devlin AS (2022) Seating in doctors’ waiting rooms: Has COVID-19 changed our choices?. HERD: Health Environments Research & Design Journal 15: 41-62. [crossref]
  11. Berkhout C, Zgorska-Meynard-Moussa S, Willefert-Bouche A, Favre J. et al (2018) Audiovisual aids in primary healthcare settings’ waiting rooms. A systematic review. European Journal of General Practice 24: 202-210.
  12. Lai JCY, Amaladoss N (2022) Music in waiting rooms: A literature review. HERD: Health Environments Research & Design Journal 15: 347-354. [crossref]
  13. Maskell K, McDonald P, Paudyal P (2018) Effectiveness of health education materials in general practice waiting rooms: a cross-sectional study. British Journal of General Practice 68: e869-e876. [crossref]
  14. Swicki B (2021) “The future of the waiting room, and how telemedicine and mobile health could change it”, Healthcare IT News.
  15. Taub P (2001) The trend towards casual address and dress in the medical profession. Virtual Mentor 3: 169-171.
fig 3

High XPO1 Expression can Stratify Gastric Cancer Patients with Poor Clinical Outcome

Leave a reply

DOI: 10.31038/CST.2024921

Abstract

Objective: Local recurrence and abdominal metastasis are the main reasons for reducing the survival. It is of great clinical value to identify patients with more malignant biological features at high recurrence and metastasis risk. We want to evaluate the efficacy and sensitivity of XPO1 as a biomarker to stratify gastric cancer patients at high biological aggressive risk.

Method: We retrospectively analyzed the pathological records of 100 enrolled patients with gastric cancer who underwent gastric cancer resection in the department of surgery of our hospital from January 2017 to December 2022; all enrolled patients had complete pathological data and follow up for survival. In this study, we analyzed the immunohistochemical staining patterns of gastrectomy tissue specimens with patients with follow-up survival information and evaluated the efficacy of a novel biomarker XPO1/CRM1, also called Exportin 1.

Results: The positive IHC of XPO1 was correlated with the following factors: primary tumor volume (P value=0.05), regional lymph node invasion (P value=0.008) and TNM staging (P value=0.069). We noticed a sequential upregulation of XPO1 IHC intensity in benign lesions, borderline tumors, invasive carcinomas biological changes. Kaplan-Meier survival analysis indicated that XPO1 positivity was associated with poor survival.

Conclusions: Our results revealed XPO1 as a sensitive and useful biomarker to stratify gastric cancer patients at high biological aggressive risk. We recommend supplementing XPO1 IHC to routine pathology test to stratify individual patients for intensive therapy and stringent follow-up plans.

Highlights

  • High XPO1 can stratify tumors with more biology malignancy trend
  • High XPO1 predicts poor prognosis in gastric cancer
  • High XPO1 patients need stringent treatment and follow-up

Keywords

Biomarker, Gastric cancer, Immunohistochemistry, Prognosis, XPO1

Introduction

Gastric cancer has a high incidence and poor prognosis, particularly in China. Each year, most new cases of gastric cancer are diagnosed among Asians and Eastern Europeans [1]. In 2020, approximately 27,000 new cases will be diagnosed [2]. The survival rates for patients with gastric disease are 31% in the United States and 25% globally [3]. In addition to its high incidence, gastric cancer has a poor prognosis and survival rate. Local recurrence and abdominal metastasis substantially impaired long-term survival. Common causes of poor prognosis [3,4] include late-stage diagnosis with regional or distant metastases, intratumor heterogeneity, and chemotherapeutic resistance. The identification of novel and specific biomarkers with prognostic significance and novel targets in gastric cancer is urgently required. At present, gastric cancer remains a fatal disease with limited treatment options. In clinical practice, clinicians execute TNM staging for patients primarily based on imaging; we believe it would be more beneficial if biomarkers that can predict the intrinsic metabolic characteristics of tumor cells could be identified for clinical applications. In addition to TNM staging, for instance, more effective prognostic assessment methods for gastric cancer can be identified, and patients who are more likely to experience recurrence can be identified. Recent reports have linked elevated XPO1 expression to a poor prognosis in a variety of tumors. XPO1, also known as CRM1, is a nuclear pheherin that belongs to the importin-superfamily [5-7] and can export at least 221 NES containing proteins and several nuclear Rnas to the cytoplasm [8,9]. The presence of conserved hydrophobic NES on carrier molecules was identified by XPO [9-11]. XPO1 participates in the localization and passive transport of diverse regulatory proteins between the nucleus and cytoplasm. Presently, it is known that XPO1 regulates a number of tumor suppressor genes that play a significant role in the pathogenesis and progression of cancer. Among the cargo proteins detected to be transported by XPO1 are the tumor suppressor p53, CDK1, adenomatous colonic polyposis (APC), BRCA1 and BRCA2, survivin, etc. [12,13]. Therefore, targeting XPO1 has promising potential as a cancer treatment. Intriguingly, XPO1 inhibitors effectively discriminate between tumor and normal tissue. XPO1 inhibitors are more likely to selectively and preferentially target tumor cells. The mechanism may be that, compared to non-malignant tumors, tumor cells express more XPO1 and cancer cells have an increased rate of cell proliferation and metabolism, making them more susceptible to nuclear trafficking inhibition [14,15]. First, we selected 100 gastric cancer patients with comprehensive clinical data from the pathology center of our hospital; all of these patients underwent surgical resection of gastric cancer in our hospital. Immunohistochemical staining was used to determine the XPO1 protein expression level in paraffin-embedded specimens of gastric carcinoma. We analyzed XPO1 IHC results in various TNM stages, as well as the correlation between XPO1 positivity and patient clinical data. Second, we analyzed XPO1 positivity variations in benign lesions, ambiguous tumors, and invasive carcinomas. We observed a pathological upregulation of XPO1 in malignant transformation of tumors, indicating its role in tumorigenesis. We performed a Kaplan-Meier analysis of survival to determine the impact of XPO1 on the clinical prognosis and survival of patients with gastric cancer. High XPO1 was able to stratify high-risk patients and predict a poorer prognosis, according to the findings. We advise these patients to adhere to rigorous treatment regimens and frequent follow-up appointments. Finally, we extended our findings to additional cancer categories. By comparing pan-cancer XPO1 expression and conducting survival analyses, we identified XPO1 as a biomarker for a poor prognosis in a variety of cancer types.

Methods

Patients’ Enrollment

We selected 100 patients who underwent surgical resection for gastric cancer at Suqian Hospital Affiliated with Xuzhou Medical University between January 1, 2017 and December 31, 2022. All patients enrolled in this study were informed of the study’s purpose and procedures, and all provided written consent to participate. The included patients must have comprehensive basic and clinic pathological information. Their paraffin-embedded tissue specimens were retrieved from the pathology department archives. Bormann grade of gross morphology and WHO grade of histopathology were used as the pathological diagnostic criteria [16]. The TNM classification of the 5th edition of the International Union against Cancer (UICC) was utilized for cancer staging [17]. Patients’ clinical information was gathered, recorded, and analyzed in detail. Indicators analyzed included the patient’s gender, age, tumor size, gastric wall invasion depth, histopathological grade, regional lymph nodes, and distant metastasis. Each patient was individually contacted via telephone to inquire about their survival status and to obtain a death date from their family. None of the patients included in the study received radiotherapy, chemotherapy, or immunotherapy prior to surgery. The study protocol was approved by the Ethics Committee Board of Suqian Hospital Affiliated to Xuzhou Medical University, and all experiments were carried out in accordance with Xuzhou Medical University’s guidelines.

Immunohistochemistry Staining

Tissues embedded in paraffin were sliced into 5-mm-thick sections. The portions were deparaffinized with xylene three times for five minutes each and rehydrated with 90, 75, and 50 percent ethanol in each container for two minutes. To recover antigenicity, the sections were submerged in a 10 mmol/L citrate buffer solution (pH 6.0) and microwaved for 12 minutes. To inhibit the activation of endogenous peroxidase, the samples were treated for 12 minutes with 3% hydrogen peroxide–methanol and then rinsed with distilled water. Anti-XPO1 rabbit polyclonal antibody (sc5595; Santa Cruz Biotechnology, Santa Cruz, California; 1: 100 dilution) was applied and incubated for one hour. Following washing, sections were rinsed with TBS and incubated with horseradish peroxidase-conjugated anti-rabbit antibody (Dako Cytomation, Carpinteria, CA). Phosphate-buffered saline (PBS) was substituted for the primary antibody to create negative controls.

Interpretation and analysis of immunohistochemistry results: Two pathologists independently examined the radiographs without knowledge of the patient’s clinical history. Each slide was investigated individually using a light microscope. When the results of two pathologists’ reviews are incongruent, the conclusion of the review is reached through mutual consultation between the two pathologists. The following criteria were used to interpret the XPO1 staining results: The intensity and proportion of positive cells were used to evaluate the immunostaining for XPO1. The staining intensity scores were as follows: 0 (negative), 1 (mild positive), 2 (medium positive), and 3 (strong positive). The following four kinds of scores were calculated based on the proportion of XPO1-positive cells: 0% to 10% was 1, 11 to 50% was 2, 51 to 80% was 3, and 81 to 100% was 4. As indicated previously, the final XPO1 staining score was calculated by multiplying the intensity score by the percentage score [18]. Positive results were defined as > 10% of cells with dark brown nuclei staining, and negative results were defined as < 10% of cells with staining. We determined the cutoff point for XPO1 IHC scores using the X-tile software (Rimm Lab at Yale University, http: //www.tissuearray.org/rimmlab).

Kaplan-Meier Survival Analysis in TCGA

The Kaplan-Meier curves for overall survival (OS) have been calculated for the high/low XPO1 expression group dichotomized by the 75% quantile of XPO1 expression. The log-rank test was utilized to investigate the difference in survival between those with high and low XPO1 expression.

Statistical Analysis

We used the chi-square test to assess the relationship between XPO1 expression and various clinicopathological features of gastric cancer. Cox’s proportional hazards regression models were used to determine univariate and multivariate analyses in order to identify independent factors associated with disease-free survival and overall survival. The Kaplan-Meier method was utilized to assess the relationships between XPO1 expression and patient outcomes. *, P < 0.05, **, P < 0.01, ***, P < 0.005, and exact P values are stated in the source data for each figure panel.

Results

Gastric Cancer Exhibits Higher XPO1 with Immunohistochemistry

The clinical characteristics of the patients were summarized and exhibited (Table 1). The ages of the patients ranged from 30 to 85 years. High XPO1 expression was specifically correlated with TNM stage (p=0.003), tumor stage (p=0.05), and lymph node metastasis positivity (p=0.007). In contrast, no significant correlation was found between XPO1 expression and other clinical factors, including gender, tumor diameter, age, and status of distant metastasis. To determine if there are any differences in XPO1 expression between gastric patient samples and normal gastric tissues, we compared XPO1 expression between gastric cancer tumors and normal gastric epithelial tissues. Each clinic sample contains comprehensive information regarding the pathology cell type and tumor stage. Two pathologists independently evaluated the results of XPO1 immunohistochemistry staining, with no knowledge of the patient’s clinical history. When there was disagreement, a conclusion was reached via consensus. Evaluation of the immunostaining was based on the intensity and percentage of XPO1-positive cells. The stain’s intensity was measured as follows: 0 (negative), 1 (weakly positive), 2 (moderately positive), and 3 (strongly positive). In addition, the percentage of XPO1-positive cells was scored based on four categories: 1 for 0 to 10%, 2 for 11 to 50%, 3 for 51 to 80%, and 4 for 81 to 100%. Multiplying the intensity and percentage scores produced the final XPO1 staining score. The emblematic images of IHC were displayed. We observed a substantial difference in XPO1 expression between tumor and normal tissue samples. Strong XPO1 positivity was observed in gastric tumors, and XPO1 intensity increased with TNM stages II, III, and IV (Figure 1).

Table 1: Demographic characteristics of the 100 gastric cancer patients. High XPO1 expression was associated with TNM stage (p=0.003), tumor stage (p=0.05) and positive lymph node metastasis (p=0.007). No significant correlation was discovered between XPO1 expression and other clinical parameters, such as gender, age, tumor diameter, and distant metastasis status.

 

n

 Negative (%) (n = 55)   Positive (%) (n = 45)

P-value
Gender (M: F)

Age (years)

Longest diameter (cm)

T stage

T1

T2

T3

T4

Nodal stage

N0

N1

N2

N3

Distant metastasis

M0

M1

TNM stage

I

II

III

IV

  
 
 
 

33

24

38

5

48

34

14

4

97

3

44

17

33

6

  
 
 
38: 17(69%: 31%)

56.38

3.65

27(82%)

11(46%)

16(42%)

1(20%)

37(39%)

20(59%)

4(29%)

1(25%)

62(64%)

0(0%)

34(77%)

11(65%)

15(45%)

2(33%)

  
 
 
 
26: 19(58%: 42%)

57.65

4.76

6(18%)

13(54%)

22(58%)

4(80%)

11(61%)

14(51%)

10(71%)

3(75%)

35(36%)

3(100%)

10(23%)

6(35%)

18(55%)

4(67%)

  
 
 
 
 
 
 

0.531

0.723

0.125

0.05

0.007

0.072

0.03

 

fig 1

Figure 1: Gastric tumors in TNM II, III, IV stages exhibit increased XPO1 intensity with immunohistochemistry staining. Representative immunohistochemistry results for anti-human XPO1 staining were presented. Gastric tumors were strongly positive for XPO1, and XPO1 intensity increased with TNM II, III, IV stages. Compared with gastric cancer samples, the expression of XPO1 in normal tissues was limited or absent.

Compared with gastric cancer samples, the expression of XPO1 in normal tissues was limited or absent. There was a statistically significant difference between adjacent non-tumor tissues and tumor-infiltrated areas in XPO1 expression, P = 0.001. The rate of positivity in normal tissue was 6%, whereas the rate of positivity in tumor areas was significantly higher (45%). High XPO1 expression was detected in 45 of 100 (45%) gastric cancer tissues, while only 6 of 100 (6%) normal gastric tissues displayed XPO1 expression (Table 2). These results indicated that XPO1 signaling was strongly activated in gastric cancer.

Table 2: Overall XPO1 expression in tumor and surrounding normal tissues. IHC was employed to investigate the expression of XPO1 in gastric cancer. There was a statistical difference in the XPO1 expression between tissues adjacent non-tumor tissues and tumor-infiltrated areas (p=0.0001).

XPO1

Normal tissue

Cancer
negative

1–10%

11–50%

51–100%

P-value

94 (94%)

4(4%)

2(2%)

0(0%)

0.0001

25(55%)

13(23%)

18(18%)

44(4%)

XPO1/CRM1, also called Exportin 1, Cancer-related genes, FDA approved drug targets.

XPO1 Plays a Role in Tumor Initiation and Progression

Previous research indicates that XPO1 exports tumor suppressor genes from the nucleus and promotes tumorigenesis. We hypothesized that XPO1 facilitated tumor initiation, i.e., that XPO1 levels would increase during the carcinogenesis process. We compared the variance in XPO1 IHC intensity among benign lesions, ambiguous tumors, and invasive carcinoma groups. Although there was no XPO1 positivity in benign lesions, there was an increase in borderline tumors. Strong XPO1 positivity was observed in invasive carcinomas (Figure 2).

fig 2

Figure 2: Gastric tumors exhibit higher XPO1 expression, which predicts shorter disease-free survival and overall survival. (A) Presentative XPO1 IHC staining in benign lesions, borderline tumors, and invasive carcinoma groups. There was no XPO1 positive staining in benign lesions, however, invasive carcinoma showed a very strong XPO1 positive staining. (B) STRING analysis showed the top genes interacting with XPO1 in gastric cancer, of which the top correlated genes were TP53, CDKN1B, RANBP2, NUP98, NUP214. (C) Gastric cancer had higher XPO1 expression than the normal tissues, results calculated from TCGA gastric cancer cohort. (D) Higher XPO1 expression predicts shorter disease-free survival in gastric cancer, p=0.031. (E) Higher XPO1 expression predicts shorter overall survival in gastric cancer, p=0.039.

To gain a more detailed understanding of how XPO1 may interact with other genes. We analyzed the top genes in TCGA gastric cancer cohorts that correlate with XPO1. According to gene STRING analysis, the XPO1 gene is closely related to a number of genes that promote malignancy. STRING analysis showed the top interacting genes with XPO1 in gastric cancer, of which the top genes were TP53 (responds to diverse cellular stresses, induce cell cycle arrest, apoptosis, senescence, DNA repair, or changes in metabolism), CDKN1B (cyclin-dependent kinase inhibitor, which shares a limited similarity with CDK inhibitor CDKN1A/p21), RANBP2 (RAN binding protein 2, enables SUMO ligase activity), NUP98 (the 96 kDa nucleoporin is a scaffold component of the nuclear pore complexes), NUP214 (the protein encoded by this gene is localized to the cytoplasmic face of the nuclear pore complex). Gastric cancer had higher XPO1 expression than the normal tissues. Higher XPO1 expression was related with shorter disease-free survival, p=0.031 and overall survival, p=0.039 in gastric cancer. To gain a more detailed understanding of how XPO1 may interact with other genes. We analyzed the top genes in TCGA gastric cancer cohorts that correlate with XPO1. According to gene STRING analysis, the XPO1 gene is closely related to a number of genes that promote malignancy. Results revealed a considerable increase in XPO1 during the progression of gastric cancer from benign lesions to borderline tumors and then to the terminal invasive carcinoma (Table 3).

Table 3: Increased expression of exportin 1/XPO1 located both in nuclear and cytoplasm. The specific number of XPO1 IHC stain location in benign lesions, borderline tumors, and invasive carcinoma groups were summarized and shown.

Number. of patients (%)

XPO1
Expression

Invasive carcinomas, N=70 Borderline tumors, n=20 Benign lesions, n=10

P*
Nuclear and Cytoplasmic Negative 20 (28.5)

Positive 50 (71.4)

16(80)

4 (20)

  10 (100)

0 (0)

0.001
Nuclear Negative 41 (58.6)

Positive 29 (41.4)

18 (90)

2 (10)

 10 (100)

0 (0)

0.002
Cytoplasmic Negative 49 (70)

Positive 21 (30)

18 (90)

2 (10)

 10 (100)

0 (0)

0.000
XPO1/CRM1, also called Exportin 1, Cancer-related genes, FDA approved drug targets.

Nuc and Cyt, Nuclear and cytoplasmic.

*Chi-square test.

Since XPO1 is present in both the nucleus and cytoplasm, both patterns were evaluated separately. Immunohistochemistry for XPO1 was negative in benign lesions. In borderline tumors, XPO1 positivity was more prominent than in benign lesions. Nuclear (2 of 20) and cytoplasmic (2 of 20) expression was moderate in 4 of 20 borderline tumors. XPO1 nuclear expression was detected in 29 of 70 invasive carcinomas (41.4%), whereas XPO1 cytoplasmic expression was detected in 21 of 74 tumors (30%). In the majority of tumors, both expression patterns were found concurrently, albeit with differing intensities. XPO1 facilitates the transport of tumor suppressor genes outside of the nucleus and may facilitate and accelerate tumorigenesis, as suggested by these findings.

TP53 Mutant Gastric Cancer had Higher XPO1 Expression

To confirm the localization of XPO1, we examined the human protein atlas and cell atlas. We discovered that XPO1 was predominantly localized in the nucleus and cytoplasm of cancer cells, which is consistent with our results. The specific intracellular XPO1 localization were examined and analyzed (Figure 3).

fig 3

Figure 3: XPO1 has many subcellular locations, with cytosol and nucleus as the two most frequent sites. (A) Representative confocal images stained with anti-XPO1 (CAB010184) antibody. In addition to localized at the cytosol & vesicles, XPO1 mainly localize to the nucleoplasm & nuclear membrane. XPO1, also called exportin 1, was cancer-related genes, a transporter which localized to the nucleoplasm (enhanced), and nuclear membrane (enhanced). (B) The specific subcellular XPO1 location were examined and analyzed from COMPARTMENTS. (C) TP53 mutant gastric cancer exhibited higher XPO1 expression. (D) Subcellular locations of XPO1 from the Human Protein Atlas (HPA) COMPARTMENTS, cytosol (5), nuclear membrane (5), nucleoplasm (2), vesicles (2).

The XPO1 localization intensity was calculated. We observed that XPO1 can be found in numerous locations within the cell, with the cytosol and nucleus being the most common. XPO1 facilitates tumorigenesis and confers drug resistance by transporting the tumor suppressor TP53. Given that p53 was a cargo protein for XPO1, it was hypothesized that inhibiting XPO1 could activate TP53. We observed that TP53 mutant gastric cancer has increased XPO1 expression. The significance of TP53 mutational and functional status on XPO1 inhibitor sensitivity in gastric cancer cell lines and the functional role of apoptosis signaling mediated by TP53 were correlated with nuclear accumulation of TP53.

Pan-Cancer XPO1 Expression and Survival Analysis

Finally, we wanted to extend our discovery to other cancer types. We investigated The Cancer Genome Atlas (TCGA) database for pan-cancer XPO1 expression analysis in cancer and normal tissues. Box graphs were used to illustrate the differential gene expressions (Figure 4).

fig 4

Figure 4: Pan cancer XPO1 expression and Kaplan-Meier survival analysis. (A) We studied the differential expression between tumor and adjacent normal tissues for XPO1 in all TCGA tumors. Distributions of gene expression levels were displayed using box plots. The statistical significance computed by the Wilcoxon test was annotated by the number of stars (*, p-value < 0.05). (B) Kaplan-Meier survival analysis between XPO1 expression and clinical outcome in multiple cancer types. In the kidney renal papillary (KRP) carcinoma, bladder tumor, cervical squamous (CS), liver hepatocellular cancer (LHC), and esophageal adenocarcinoma (EA) cohort, high XPO1 predicts shorter overall survival. The p value was labeled on each graph. (C) In lung adenocarcinoma (LA), pancreatic ductal adenocarcinoma (PDA), pheochromocytoma and paraganglioma (PP), sarcoma, and uterine corpus endometrial carcinoma (UCE) cohort, high XPO1 predicts shorter overall survival. The p value was labeled on each graph.

The Wilcoxon test’s statistical significance was indicated by the number of stars (*: p-value 0.05). We were able to determine whether XPO1 was up- or down-regulated in tumors relative to their normal counterparts for each cancer type. The 33 malignancies analyzed by the TCGA/Pan Cancer Initiative were represented schematically according to their tissue of origin. XPO1 was expressed substantially more (red) in cancerous tissues than in normal tissues, with the exception of ovarian, prostate, thyroid, and uterine cancers (black). We then performed a Kaplan-Meier analysis of survival between XPO1 expression and clinical prognosis in multiple types of cancer. In the cohort of patients with kidney renal papillary (KRP) carcinoma, bladder tumor, cervical squamous (CS), liver hepatocellular cancer (LHC), and esophageal adenocarcinoma (EA), a high XPO1 level predicts a shortened overall survival. In lung adenocarcinoma (LA), pancreatic ductal adenocarcinoma (PDA), pheochromocytoma and paraganglioma (PP), sarcoma, and uterine corpus endometrial carcinoma (UCE) cohorts, XPO1 expression was associated with shorter overall survival at a 75% quantile threshold. Collectively, these findings demonstrated that XPO1 is a potential broad-spectrum biomarker for cancer prognosis and could be a therapeutic target for treatment.

Discussion

In this retrospective study, we evaluated the immunohistochemical staining of XPO1 in gastric tumor samples and investigated the correlation between XPO1 level and multiple clinicopathology factors in predicting its clinical significance in patients with gastric cancer. XPO1 levels were substantially elevated in cancer samples compared to normal counterparts. Statistically, the degree of XPO1 positivity did not correlate with tumor size; however, higher expressions were found in patients with higher T values, more regional lymph node invasion, and advanced TNM staging, which could predict a substantially lower survival rate. We examined the differential expression of XPO1 in various phases of gastric cancer and the correlation between XPO1 immunohistochemical staining and patient clinical characteristics. Our results demonstrated that XPO1 is a valuable biomarker for stratifying gastric cancer patients based on their biologically malignant nature. Chemotherapy and surgery for gastric cancer have improved over the past few decades [19,20]. Nonetheless, patients with gastric cancer continue to have a poor prognosis due to therapeutic failure and disease progression [21]. Identification of novel and validated prognostic biomarkers in practice has clinically significance for gastric cancer. In this study, we discovered that XPO1 was a useful marker in gastric cancer that had the potential to be used as a candidate for targeted therapy. The regulation of material transport across the nuclear membrane was essential for maintaining homeostasis, which required the correct nuclear-cytoplasm positioning of large molecules; nevertheless, this process was typically dysregulated in cancer cells [22]. XPO1, an export receptor responsible for the nuclear-cytoplasm transport of multiple proteins and RNA species, was frequently overexpressed or mutated in human malignancies and served as a potential oncogenic driver [23]. Unlike small molecules, which can passively diffuse through the nuclear pore complex (NPC), larger cargo molecules (>40 kDa) require active transport via transport receptors [24-26], which belong to the karyopherin beta family and are classified as importins (nuclear import), exportins (nuclear export), and transportins (for both import and export) [25]. Studies show that exportins are potential targets in tumorigenesis [27,28], of which XPO1 was the most important and well-studied target. XPO1 was initially identified as a chromosomal mutation in the yeast Schizosaccharomyces pombe [29]. XPO1, also known as CRM1, transported over 200 proteins, the majority of which were tumor suppressors and oncoproteins [29-31]. CRM1-mediated cargos include p27, p53, FOXOs, nucleophosmin, PI3K/AKT, Wnt/-catenin, BCR-ABL, p21, NF-kB, APC, and Rb; these cargos all play important roles in tumorigenesis [28,32]. For the first time, we investigated the clinical and prognosis value of XPO1 in gastric cancer. IHC analysis revealed a higher XPO1 concentration in gastric cancer tissues compared to normal gastric tissues. Consistent with previous research, our findings indicated that a variety of malignancies exhibited a higher level of XPO1 expression than their normal counterparts [33,34]. In addition, elevated XPO1 levels in gastric cancer were associated with certain clinical-pathologic factors, including AJCC stage, positive lymph node metastasis, and tumor grade. The Kaplan-Meier analysis demonstrated that the disease-free survival and overall survival of patients with increased XPO1 expression were shorter than those of patients with negative or decreased expression. A univariate analysis revealed that XPO1 expression, AJCC stage, and lymph node metastasis were correlated with gastric cancer patients’ survival (both disease-free survival and overall survival). High levels of XPO1 and advanced AJCC staging independently predicted unfavorable disease-free survival and overall survival outcomes for patients with gastric cancer, as determined by multivariate analysis. XPO1 overexpression was identified in solid tumors and hematologic malignancies and was reported as an indicator of poor prognosis and potential drug resistance in cancers [35]. One potential mechanism for XPO1 overexpression was associated with altered transport, which promoted cancer-promoting outcomes [36]. XPO1 facilitated the import of growth regulatory proteins, such as c-myc or BCR-ABL, into the cytoplasm and consequently activated downstream signaling, resulting in sustained cell proliferation. Similarly, tumor suppressor proteins (TSPs), such as p53, p21, Rb, and p27, were rendered inactive by exportin and lost their ability to inhibit uncontrolled cell proliferation. Collectively, these findings support the notion that XPO1 inhibition is an attractive therapeutic target for its ability to target a variety of hallmarks of oncogenesis signaling. In addition, the combination of SINE compounds with existing standard regimens in multiple cancer types was feasible and well tolerated in clinical trials. Common inhibitors of nuclear export (SINE) XPO1 antagonists included KPT-185, KPT-276, KPT-251, and KPT-330, which were reported to inhibit the proliferation of triple-negative breast cancer (TNBC) cell lines and also demonstrated efficacy in human breast cancer xenograft models. Mechanically, SINE compounds inhibit XPO1 and suppress STAT3 trans-activation, thereby inhibiting the oncogenic potential of TNBC and their clinical application [36]. Priming cancer cells with XPO1 inhibitors followed by doxorubicin, melphalan, bortezomib, or carfiltiamob may sensitize de novo and adaptive cancer cell lines to drug resistance [37]. Inhibiting the activation of the XPO1 pathway would accelerate the apoptosis of tumor cells and induce cell cycle arrest [38,39]. In summary, XPO1 expression or upregulation may replicate the natural process of gastric cancer bio-evolution, and XPO1 may therefore predict and stratify patients with a poor prognosis. In another sense, we may consider the XPO1 level as a molecular staging biomarker for oncologists employing intensive surgical intervention or chemotherapy. High XPO1 expression in gastric cancer was a reliable molecular biomarker for staging and prognostic prediction during both the diagnostic and treatment phases. High XPO1 expression in gastric cancer is indicative of an aggressive phenotype requiring intensive treatment and careful monitoring. Our findings supported XPO1 as a novel prognostic biomarker for patients with gastric cancer, and targeting XPO1 may provide a beneficial strategy for gastric cancer patients with positive XPO1 expression, which is typically accompanied by TP53 mutation. As mentioned previously, inhibiting XPO1 signaling with SINE may restore the functions of common tumor suppressors. Thus, targeting XPO1 in gastric cancer may provide new treatment options for gastric cancer patients, particularly those with advanced disease and a high recurrence risk. In addition, our pan-cancer analysis of the TCGA dataset revealed that XPO1 was commonly elevated in all cancer types. Consequently, our findings illuminated the potential universal application of XPO1 inhibitors in multiple types of cancer. Future clinical studies are required to evaluate the therapeutic effects of KPT-SINE compounds (small molecules for XPO1) alone and in combination with XPO1-targeted therapy. Our research had several limitations. We detected XPO1 positivity solely through immunohistochemistry, so there is a possibility for diagnostic error. Several other techniques, such as immune blotting and qRT-PCR for mRNA expression, have been considered in an effort to achieve more precise diagnostic results. Second, all clinical data, including recurrence and survival rates, were retrospectively collected. Thirdly, the relatively small number of patients enrolled in our study may result in a lack of statistical power; therefore, a larger prospective study is needed in the future.

Availability of Data and Materials

The data generated in the present study may be requested from the corresponding author through 1822991734@qq.com

Authors’ Contributions

Conception and design were performed by Rui Wang and Yanli Cheng. Data analysis and interpretation were performed by Ruimin Wang. Manuscript writing was performed by Rui Wang and revised by Yanli Cheng. Final approval of manuscript was performed by all authors who read and approved the final manuscript.

Grant Support

Rui Wang is founded by China Scholarship Council (202206920039). This research was supported by funds from Natural Science Foundation of Suqian Science and Technology Bureau (K201903, Z2018076, Z2018213 and Z2022065). Jiangsu Association for Science and Technology (JSTJ-2022-004).

Ethics Approval and Consent to Participate

The patient reported in this study was informed for the purpose and process of this study and had written informed consent according to the guidelines of the hospital’s human associated research.

Patient Consent for Publication

Not applicable.

Competing Interests

The authors indicated no potential conflicts of interest.

Declaration of Interest

The authors declare that there is no competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Financial Support

Rui Wang is founded by China Scholarship Council (202206920039). This research was supported by funds from Natural Science Foundation of Suqian Science and Technology Bureau (K201903, Z2018076, Z2018213 and Z2022065). Jiangsu Association for Science and Technology (JSTJ-2022-004).

Acknowledgements

The authors would like to thank Dr. Xiaohong Shi for comments and discussion on the manuscript. We also would like to thank Dr. Quanquan Guo for data analysis.

References

  1. Rawla P, A Barsouk (2019) Epidemiology of gastric cancer: global trends, risk factors and prevention. Prz Gastroenterol 14(1): 26-38. [crossref]
  2. Gu E (2020) . SCDb: an integrated database of stomach cancer. BMC Cancer 20(1): 490. [crossref]
  3. Rawla P, A Barsouk (2019) Epidemiology of gastric cancer: global trends, risk factors and prevention. Prz Gastroenterol 14(1): 26-38. [crossref]
  4. Stahl P (2015) Heterogeneity of amplification of HER2, EGFR, CCND1 and MYC in gastric cancer. BMC Gastroenterol 15(1): 70-78. [crossref]
  5. Fornerod M (1997) CRM1 is an export receptor for leucine-rich nuclear export signals. Cell 90(6): 1051-60. [crossref]
  6. Fukuda M (1997) CRM1 is responsible for intracellular transport mediated by the nuclear export signal. Nature 390(6657): 308-11. [crossref]
  7. Ossareh-Nazari B F. Bachelerie, C. Dargemont (1997) Evidence for a role of CRM1 in signal-mediated nuclear protein export. Science 278(5335): 141-4. [crossref]
  8. Fu SC (2013) ValidNESs: a database of validated leucine-rich nuclear export signals. Nucleic Acids Res 41(Database issue): D338-43. [crossref]
  9. Okamura M H. Inose and S. Masuda (2015) RNA Export through the NPC in Eukaryotes. Genes (Basel) 6(1): 124-49. [crossref]
  10. Monecke T A. Dickmanns, R. Ficner (2014) Allosteric control of the exportin CRM1 unraveled by crystal structure analysis. FEBS J 281(18): 4179-94. [crossref]
  11. Cautain B (2015) Components and regulation of nuclear transport processes. FEBS J 282(3): 445-62. [crossref]
  12. Wang A Y, H Liu (2019) The past, present, and future of CRM1/XPO1 inhibitors. Stem Cell Investig 6(3): 69-78. [crossref]
  13. Schmidt J (2013) Genome-wide studies in multiple myeloma identify XPO1/CRM1 as a critical target validated using the selective nuclear export inhibitor KPT-276. Leukemia 27(12): 2357-65. [crossref]
  14. Angus L, PJ. van der Watt, VD Leaner (2014) Inhibition of the nuclear transporter, Kpnbeta1, results in prolonged mitotic arrest and activation of the intrinsic apoptotic pathway in cervical cancer cells. Carcinogenesis 35(5): 1121-31. [crossref]
  15. Kuusisto HV, DA Jans (2015) Hyper-dependence of breast cancer cell types on the nuclear transporter Importin beta1. Biochim Biophys Acta 1853(8): 1870-8. [crossref]
  16. Schlemper RJ ( 2000) The Vienna classification of gastrointestinal epithelial neoplasia. Gut 47(2): 251-5. [crossref]
  17. Hermanek P (1987) [TNM classification of malignant tumors: the new 1987 edition]. Rontgenblatter. 40(6): 200. [crossref]
  18. Gong W (2005) Expression of activated signal transducer and activator of transcription 3 predicts expression of vascular endothelial growth factor in and angiogenic phenotype of human gastric cancer. Clin Cancer Res 11(4): 1386-93. [crossref]
  19. Barchi LC (2016) MINIMALLY INVASIVE SURGERY FOR GASTRIC CANCER: TIME TO CHANGE THE PARADIGM. Arq Bras Cir Dig 29(2): 117-20. [crossref]
  20. Kinoshita J (2021) Current status of conversion surgery for stage IV gastric cancer. Surg Today 11(4): 1386-93. [crossref]
  21. Digklia A, AD Wagner (2016) Advanced gastric cancer: Current treatment landscape and future perspectives. World J Gastroenterol 22(8): 2403-14. [crossref]
  22. Gravina GL (2014) Nucleo-cytoplasmic transport as a therapeutic target of cancer. J Hematol Oncol (7): 85-96. [crossref]
  23. Azizian NG, Y Li (2020) XPO1-dependent nuclear export as a target for cancer therapy. J Hematol Oncol 13(1): 61. [crossref]
  24. Schmidt HB, D Gorlich (2016) Transport Selectivity of Nuclear Pores, Phase Separation, and Membraneless Organelles. Trends Biochem Sci 41(1): 46-61. [crossref]
  25. Ullman KS, MA Powers, DJ Forbes (1997) Nuclear export receptors: from importin to exportin. Cell 90(6): 967-70.
  26. Jamali T (2011) Nuclear pore complex: biochemistry and biophysics of nucleocytoplasmic transport in health and disease. Int Rev Cell Mol Biol 287(3): 233-86. [crossref]
  27. Sun Q et al. (2016) Inhibiting cancer cell hallmark features through nuclear export inhibition. Signal Transduct Target Ther (1): 160-175. [crossref]
  28. Das A et al. (2015) Selective inhibitors of nuclear export (SINE) in hematological malignancies. Exp Hematol Oncol (4): 79-86. [crossref]
  29. Stade K et al. (1997) Exportin 1 (Crm1p) is an essential nuclear export factor. Cell 90(6): 1041-50. [crossref]
  30. Senapedis WT, E Baloglu, Y Landesman (2014) Clinical translation of nuclear export inhibitors in cancer. Semin Cancer Biol (27): 74-86. [crossref]
  31. Xu D, NV Grishin, YM.Chook (2012) NESdb: a database of NES-containing CRM1 cargoes. Mol Biol Cell 23(18): 3673-6. [crossref]
  32. Senapedis WT, E Baloglu, Y Landesman (2014) Clinical translation of nuclear export inhibitors in cancer. Semin Cancer Biol (27): 74-86. [crossref]
  33. Zhang S (2012) ROR1 is expressed in human breast cancer and associated with enhanced tumor-cell growth. PLoS One 7(3): e31127. [crossref]
  34. Zhang S (2012) The onco-embryonic antigen ROR1 is expressed by a variety of human cancers. Am J Pathol 181(6): 1903-10. [crossref]
  35. Sun Q et al. (2016) Inhibiting cancer cell hallmark features through nuclear export inhibition. Signal Transduct Target Ther (1): 160-179. [crossref]
  36. [36]. Cheng Y et al. (2014) XPO1 (CRM1) inhibition represses STAT3 activation to drive a survivin-dependent oncogenic switch in triple-negative breast cancer. Mol Cancer Ther 13(3): 675-86. [crossref]
  37. Turner JG (2014) Inhibition of CRM1-dependent nuclear export sensitizes malignant cells to cytotoxic and targeted agents. Semin Cancer Biol (27): 62-73. [crossref]
  38. Huang WY (2009) Prognostic value of CRM1 in pancreas cancer. Clin Invest Med 32(6): E315. [crossref]
  39. Kim J (2016) XPO1-dependent nuclear export is a druggable vulnerability in KRAS-mutant lung cancer. Nature 538(7623): 114-117. [crossref]