Supplementary Figure S2 from Machine Learning Links T-cell Function and Spatial Localization to Neoadjuvant Immunotherapy and Clinical Outcome in Pancreatic Cancer

Supplementary Figure S2. A. SHAP plots showing the top 30 features driving each histopathologic model. Features are ordered on the y-axis such that those with a larger impact on model’s predictions appear at the top of the SHAP plots. SHAP values are shown on the x-axis, with a value of zero (center) indicating no impact on the model, and negative or positive SHAP values predicting treatment-naive or αCD40-treated tissues, respectively. Red or blue dots indicate presence or absence, respectively, of the corresponding feature in the tissue. B-E. Box plots showing feature values for each of the top 15 features for models derived from T, IA, TAS, or NAP sites, respectively, split by treatment cohort. Each dot represents the log10+1 normalized feature value for one tissue region, inputted into the classifier model. Boxes = quartile 1 (Q1) to quartile 3 (Q3); whiskers = smallest and largest datapoints within 1.5*interquartile range (IQR) +/- Q3/Q1; solid line = median. Mann–Whitney U-test used to determine statistical significance. P-values corrected using the Benjamini–Hochberg procedure. *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001. B. T site, n= 55 treatment-naive and n = 48 αCD40-treated regions per feature. C. IA site, n= 89 treatment-naive and n = 43 αCD40-treated regions per feature. D. TAS site, n = 25 treatment-naive and n = 27 αCD40-treated regions per feature. E. NAP site, n = 6 treatment-naive and n = 13 αCD40-treated regions per feature.