The review, published in Artificial Intelligence Review, focused specifically on post-pretraining interventions applied to existing foundation models. Its systematic, auditable workflow identified 719 eligible reports representing 705 independent studies. Of those reports, 473 provided controlled comparative evidence. The primary synthesis comprised 322 formally published studies, while 151 preprint-only studies formed a separate sensitivity layer. The authors classified adaptation into five non-mutually exclusive families: parameter-level, representation-level, objective-level, data-centric, and architectural or sequence-level adaptation. Model use and update regimes were analyzed separately.

Clinical-readiness evidence was notably limited. Only 173 of the 705 studies, or 24.5%, had confirmed independent external validation, while 10 studies, or 1.4%, reported calibration or uncertainty evidence. Among the 322 formally published comparative studies, 97, or 30.1%, included a qualifying externally validated arm. These results describe the coverage and composition of the available evidence; they do not show that adaptation is ineffective or that any particular clinical application is ready for use.

The central finding is that adaptation outcomes depend on the conditions in which they are assessed. A method that performs well in one comparison cannot therefore be treated as universally superior, and evidence from a narrowly defined task or evaluation setting does not establish the same performance under different conditions. The review is also bounded by its focus on interventions applied after pretraining and by the separation of formally published research from the preprint sensitivity layer. Overall, the available evidence portrays foundation-model adaptation for medical imaging as a design space whose performance and clinical readiness must be evaluated in context, with independent validation, calibration and uncertainty evidence remaining uncommon across the reviewed literature.

Subject of Research: Post-pretraining adaptation strategies and clinical-readiness evidence for foundation models in medical image analysis

Article Title: Adaptation of foundation models in medical image analysis: mechanisms, evidence, and deployment considerations

Article References: Phuntsho, K., Abdullah, Lee, K., Lee, I., & Ahn, E. (2026). Adaptation of foundation models in medical image analysis: mechanisms, evidence, and deployment considerations. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11733-0

Image Credits: AI Generated

DOI: 10.1007/s10462-026-11733-0

Keywords: medical imaging, foundation models, model adaptation, external validation, calibration, uncertainty, clinical readiness, systematic review, artificial intelligence, Adaptation, foundation, models

The study, led by researchers at Al-Quds University in Jerusalem, set out to measure knowledge, attitudes, readiness, and ethical concerns about precision medicine and pharmacogenomics in diabetes care among medical students across Palestinian universities. Diabetes was chosen as the clinical focus for good reason. Patients with type 2 diabetes can respond very differently to commonly prescribed antidiabetic medications, and genetic differences are increasingly recognized as one driver of that variability. As genomic testing becomes cheaper and more widely available, the question of whether newly trained physicians can actually use this information is becoming urgent in health systems everywhere, not only in the Middle East.

Between June and August 2026, the team conducted a multicenter cross-sectional study using a self-administered online questionnaire. The instrument covered sociodemographic characteristics, educational background, and several carefully structured assessment domains. Knowledge was tested with eighteen true-or-false questions probing the scientific foundations of pharmacogenomics and its application to diabetes treatment. Attitudes, readiness, and ethical concerns were measured with five-point Likert-scale items, a standard survey technique that asks respondents to rate their agreement with statements on a scale ranging from strong disagreement to strong agreement. The researchers then analyzed the responses using descriptive statistics, chi-square tests, Kruskal-Wallis tests, and multivariable logistic regression, a modeling approach that allows investigators to determine which factors independently predict an outcome while controlling for the influence of other variables.

The headline numbers tell a nuanced story. Among the 352 students who participated, 62.2 percent, or 219 individuals, demonstrated adequate knowledge of precision medicine and pharmacogenomics. That is a respectable figure, suggesting that the conceptual building blocks of genomic medicine are reaching students through lectures, textbooks, and self-directed learning. Yet when the researchers turned to attitudes, only 29.3 percent, or 103 students, expressed favorable attitudes toward these technologies. And readiness, the measure of whether students felt prepared to actually use pharmacogenomic information in clinical decision-making, was even lower: just 18.8 percent, or 66 students, met the criterion for readiness. In other words, knowing the science is not the same as being able to practice it.

What separated the well-prepared students from the rest? The regression analysis identified several independent predictors. Greater experience with diabetes pathogenesis and treatment was associated with both adequate knowledge, with an odds ratio of 1.78 and a 95 percent confidence interval of 1.36 to 2.34, and favorable attitudes, with an odds ratio of 1.32 and a confidence interval of 1.06 to 1.65. An odds ratio above 1 means the factor increases the likelihood of the outcome, and because these confidence intervals do not cross 1, the associations are statistically meaningful. Students who had dug deeper into how diabetes develops and how it is treated simply knew more and felt more positive about genomic approaches to care.

Readiness followed a different logic. It was independently associated with greater experience in biostatistics, with an odds ratio of 1.55 and a confidence interval of 1.17 to 2.06, and with previous formal study of genetics, with an odds ratio of 5.17 and a confidence interval of 1.21 to 22.15. That last figure is remarkable: students who had previously studied genetics were more than five times as likely to feel ready to apply pharmacogenomics in practice. The finding makes intuitive sense. Applying genomic medicine requires the confidence to interpret genetic data, weigh probabilities, and translate laboratory results into treatment decisions, and that confidence appears to be built through direct exposure to genetics coursework and quantitative training rather than through general clinical knowledge alone.

The relationships among the three main outcomes were also revealing. Knowledge was significantly associated with both attitudes and readiness, but the strongest link appeared between attitudes and readiness, measured with a Cramér’s V statistic of 0.39 and a p-value below 0.001. Cramér’s V is a measure of association that ranges from 0 to 1, where higher values indicate stronger relationships between categorical variables. A value of 0.39 suggests a moderate-to-strong connection, meaning that students who viewed pharmacogenomics favorably were much more likely to feel ready to use it. This pattern hints at a psychological pathway: positive attitudes may be a necessary bridge between theoretical knowledge and genuine clinical preparedness, which has direct implications for how educators should design curricula.

Perhaps the most consequential findings concern the curriculum itself. Only 23 percent of the surveyed students believed that pharmacogenomics was adequately covered in their medical education. Yet demand for the subject was overwhelming: 89.2 percent of students supported its inclusion in the curriculum, and 90.9 percent wanted additional training. That gap between what is taught and what students want to learn represents both a problem and an opportunity. The problem is a generation of future physicians entering practice without the tools to interpret an expanding stream of genomic information. The opportunity is an unusually receptive audience; few areas of medical education enjoy near-unanimous student demand, and the authors argue that their findings support integrating structured, clinically oriented, and experiential pharmacogenomics education into medical curricula in Palestine.

Ethical concerns, often cited as a barrier to genomic medicine, formed another dimension of the survey. While the study found that students held measurable concerns about issues such as privacy and the implications of genetic testing, the dominant pattern in the data was enthusiasm constrained by lack of preparation. This matters because the successful deployment of pharmacogenomics depends not only on laboratory technology but on clinicians who can explain genetic results to patients, obtain meaningful informed consent, and navigate questions about data security and equitable access. Education that ignores the ethical dimension risks producing physicians who are technically capable but unprepared for the conversations that genomic medicine inevitably triggers in the clinic.

The broader lesson extends well beyond Palestine. Health systems worldwide are grappling with how to move pharmacogenomics from research papers into routine prescribing, and workforce readiness has emerged as a universal bottleneck. This study offers a template for measuring that readiness and a clear diagnosis of where the pipeline breaks: conceptual knowledge is being transmitted, but the applied, statistical, and ethical competencies that turn knowledge into practice are not. The authors received ethical approval from the Al-Quds University Research Ethics Committee, conducted the work without external funding, and published their findings open access so that educators elsewhere can build on them. If the next generation of doctors is to prescribe drugs matched to a patient’s DNA rather than to a population average, medical schools will need to close the gap this study has so precisely measured, and they will need to do it before their students graduate into clinics where genomic data is already waiting in the chart.

Subject of Research: Medical students' knowledge and readiness for pharmacogenomics in diabetes care

Article Title: Knowledge and readiness toward precision medicine and pharmacogenomics in diabetes care among medical students: a cross-sectional study

Article References: Faroun, J., Ayasa, Y., Madaka, S., Alayasa, S., Bael, P., & Ereqat, S. (2026). Knowledge and readiness toward precision medicine and pharmacogenomics in diabetes care among medical students: a cross-sectional study. BMC Medical Education. https://doi.org/10.1186/s12909-026-10548-5

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10548-5

Keywords: pharmacogenomics, precision medicine, diabetes, medical education, medical students, genetics, curriculum, cross-sectional study, clinical readiness, Palestine, biostatistics, personalized medicine

The problem the team set out to address is well documented in health professions education. Traditional dental curricula tend to rely on didactic, teacher-centered instruction that produces passive learning environments and fragments knowledge into silos. Educational researchers distinguish between horizontal integration, which connects disciplines such as anatomy and pathology at the same stage of training, and vertical integration, which links foundational science to advanced clinical topics across time. Both are considered essential for clinical problem-solving, yet conventional teaching methods often deliver neither, leaving students to assemble the pieces on their own, usually under the pressure of real patient care.

Case-based learning has emerged as one of the most promising answers. By anchoring instruction in clinically relevant scenarios, CBL builds on students’ existing knowledge, encourages critical thinking, and has repeatedly outperformed lecture-based methods in knowledge retention and integration. But the Mahidol team identified a weakness in the standard format: traditional CBL typically drops students into undifferentiated groups without a dedicated phase for discipline-specific inquiry, which can produce a superficial grasp of specialized knowledge before students are asked to synthesize it with peers. Problem-based learning and team-based learning share similar limitations, prioritizing either open-ended inquiry or group accountability testing rather than a deliberate progression from individual expertise to multidisciplinary synthesis.

PM-CBL was designed to fix that. The framework unfolds in three sequenced phases, each grounded in a distinct learning theory. In Phase 1, grounded in constructivist theory, students work independently within an assigned discipline—microbiology and immunology, pharmacology, anatomy, physiology, or pathology—conducting literature reviews and presenting findings to a discipline instructor for structured feedback. Phase 2, informed by social constructivism, brings the five discipline representatives together to integrate their findings around a shared case and prepare a group presentation. Phase 3, drawing on Kolb’s experiential learning cycle, culminates in formal presentations to all instructors, cross-disciplinary questioning, and an individual written reflection. The activity ran twice during the preclinical curriculum, once in the third year focusing on systemic diseases and once in the fourth year on oral conditions with systemic involvement.

One representative fourth-year case illustrates the approach in action: a 26-year-old pregnant patient presenting with gum swelling, difficulty chewing, and a grey-white ulcerated tongue lesion. The pharmacology group examined drugs contraindicated in pregnancy, the pathology group described histopathological features, the anatomy group considered fetal safety in dental treatment, the physiology group explained the hormonal mechanisms behind gingival enlargement, and the microbiology group identified possible causative organisms. Only when all five perspectives converged did the full clinical picture emerge—a deliberate demonstration that no single discipline can manage a complex patient alone.

To evaluate the framework, the researchers conducted semi-structured interviews with fourteen participants—five fifth-year students, five sixth-year students, and four recent graduates—who had completed PM-CBL in their preclinical years and subsequently entered clinical practice. Interviewers had no teaching or grading role in the activity, and the analysis employed reflexive thematic analysis with a collaborative team that blended former learners, the course developer, and more distanced researchers. Four main themes emerged: learning activities and processes, concerns in learning design, educational outcomes, and clinical application.

The educational payoffs were striking. Ten of the fourteen participants reported that PM-CBL deepened their knowledge acquisition and reinforced preclinical material they had studied years earlier. Ten described genuine multidisciplinary integration, with one noting it was the first time they had linked their preclinical knowledge into a coherent whole rather than viewing each subject in isolation. Ten also reported gains in research and self-directed learning skills, learning to navigate databases, screen academic papers critically, and rely on evidence rather than blogs or simple websites. Most tellingly, all fourteen participants connected the experience to clinical application, describing how early exposure to diverse cases helped them diagnose lesions, assess patients with systemic conditions, and identify risks with greater confidence during their clinical years.

The study is equally candid about the framework’s shortcomings. Eleven participants questioned the realism of the cases, noting that some were overly rare or complex and that real clinical diagnosis requires investigating information rather than receiving it all at once. Nine complained that lengthy, information-dense presentations eroded engagement. Eight wanted the freedom to choose their assigned discipline rather than being randomly allocated, and several noted that rigid assignments meant they barely discussed disciplines outside their own. Participants also called for more frequent instructor interaction beyond the two fixed meetings, and seven argued that the activity should be repositioned later in the curriculum, closer to the start of clinical training, so that reinforced knowledge does not fade before it is needed.

From these findings the team constructed a conceptual model proposing how design concerns shape learning processes, which in turn drive educational outcomes and ultimately clinical readiness. The authors frame the model as hypothesis-generating rather than empirically validated, and they acknowledge the limitations inherent in a single-center qualitative study with a modest sample and potential recall bias among graduates reflecting on experiences from years past. Perceived readiness, they note, is not the same as measured competence; future work should incorporate objective metrics such as standardized assessments and direct clinical observation.

Nevertheless, the study offers a compelling and theoretically grounded template for reform. By sequencing constructivist knowledge-building, socially constructed integration, and experiential application, PM-CBL maps a deliberate cognitive journey from passive reception to interactive engagement—a progression the authors also interpret through the ICAP framework of cognitive engagement. Their recommendations are practical: use common, clinically realistic cases; allow flexible discipline selection; increase interactivity and instructor feedback; consider gamified elements such as roleplay; and place the activity close to the clinical years. For dental schools wrestling with the persistent chasm between preclinical theory and patient care, the message is clear—integration does not happen by accident. It must be engineered, phase by phase, into the very architecture of learning.

Subject of Research: Multidisciplinary case-based learning in preclinical dental education

Article Title: Exploring learning experiences and perceived benefits of multidisciplinary case-based learning for knowledge integration and clinical readiness among preclinical dental students

Article References: Sipiyaruk, K., Asavisanu, M., Chirachanchai, J., Ruangvanish, W., Wachirawuttikai, U., & Supa-amornkul, S. (2026). Exploring learning experiences and perceived benefits of multidisciplinary case-based learning for knowledge integration and clinical readiness among preclinical dental students. Heliyon, 12(15), Article e45526. https://doi.org/10.1016/j.heliyon.2026.e45526

Image Credits: AI Generated

DOI: 10.1016/j.heliyon.2026.e45526

Keywords: dental education, case-based learning, preclinical training, knowledge integration, clinical readiness, qualitative research, constructivism, experiential learning, multidisciplinary curriculum, thematic analysis, Mahidol University, health professions education