Artificial intelligence for detection of microsatellite instability in colorectal cancer-a multicentric analysis of a pre-screening tool for clinical application.

BACKGROUND: Microsatellite instability (MSI)/mismatch repair deficiency (dMMR) is a key genetic feature which should be tested in every patient with colorectal cancer (CRC) according to medical guidelines. Artificial intelligence (AI) methods can detect MSI/dMMR directly in routine pathology slides, but the test performance has not been systematically investigated with predefined test thresholds. METHOD: We trained and validated AI-based MSI/dMMR detectors and evaluated predefined performance metrics using nine patient cohorts of 8343 patients across different countries and ethnicities. RESULTS: Classifiers achieved clinical-grade performance, yielding an area under the receiver operating curve (AUROC) of up to 0.96 without using any manual annotations. Subsequently, we show that the AI system can be applied as a rule-out test: by using cohort-specific thresholds, on average 52.73% of tumors in each surgical cohort [total number of MSI/dMMR = 1020, microsatellite stable (MSS)/ proficient mismatch repair (pMMR) = 7323 patients] could be identified as MSS/pMMR with a fixed sensitivity at 95%. In an additional cohort of N = 1530 (MSI/dMMR = 211, MSS/pMMR = 1319) endoscopy biopsy samples, the system achieved an AUROC of 0.89, and the cohort-specific threshold ruled out 44.12% of tumors with a fixed sensitivity at 95%. As a more robust alternative to cohort-specific thresholds, we showed that with a fixed threshold of 0.25 for all the cohorts, we can rule-out 25.51% in surgical specimens and 6.10% in biopsies. INTERPRETATION: When applied in a clinical setting, this means that the AI system can rule out MSI/dMMR in a quarter (with global thresholds) or half of all CRC patients (with local fine-tuning), thereby reducing cost and turnaround time for molecular profiling.

Direct and indirect effects of alcohol expectancies through drinking motives on alcohol outcomes among students in Vietnam.

OBJECTIVE: To investigate whether the links between alcohol expectancies (tension reduction; global positive change; improved cognitive and motor abilities; and change in social behavior) and alcohol outcomes (drinking volume, 6+ drinks, alcohol problems, and symptoms of alcohol dependence) are mediated by drinking motives (social, enhancement, conformity, and coping). METHOD: A multi-stage sampling strategy was used in four Vietnamese provinces, resulting in a final sample of 4756 students (43.2% females) with mean age 20.6 (SD 1.8) years. Structural equation models, including indirect effects, were estimated for women and men separately. RESULTS: Overall, there were many cases of full mediation (indirect effects range from -0.006 to 0.083 and p-values from <0.05 to <0.001) and little indication of partial mediation (indirect effects range from -0.009 to 0.025 and p-values from <0.05 to <0.001). In both men and women, coping motives most frequently mediated the influence of expectancies on alcohol outcomes. Among men, enhancement motives and, to a lesser extent, social motives also played a role in mediating the effects of expectancies on alcohol outcomes. Among women, full mediation was found far less often and less consistently. CONCLUSION: By confirming that, in Vietnam, motives mediate the link between expectancies and drinking behavior, this study supports the cultural robustness of a key assumption of the motivational model (i.e. that drinking motives are more closely associated with alcohol use than expectancies). Enhancement, coping and social motives are most frequently found as mediators among male students whereas coping motive only is most frequently found as a mediator among female students. As most of the effects of expectancies were mediated by motives, drinking motives appear to be a promising factor for interventions.