Teacher-Referenced Heatmap Coverage for Auditing Attribution Shifts in Updated CT Classifiers

Model updates may change the attribution patterns of computed tomography (CT) classifiers without producing corresponding changes in predictive performance. This study evaluated whether peak-aware heatmap measures could provide a more informative signal than thresholded overlap for auditing such changes. Gradient-weighted class activation mapping (Grad-CAM) maps from baseline and decision-distilled ResNet18 classifiers were compared with a frozen, model-derived reference. The primary Heatmap Coverage Metric (HCM) endpoints measure missed circular coverage and nearest-peak distance from the reference to the evaluated map; the reverse direction was analyzed separately. Three held-out CT cohorts were evaluated at the slice level with three training seeds. Fixed-threshold intersection over union was frequently near zero. On the designated seed-42 tests, decision distillation increased circular discrepancy on Priv-EGFR by 0.083 (95% confidence interval, 0.032-0.137), decreased Euclidean discrepancy on Chest CT-Scan by 0.026 (95% confidence interval, $-0.037$ to $-0.016$ ), and produced inconclusive changes for both RADGEN endpoints. Both primary Chest endpoints were numerically lower under decision distillation in all three seeds, whereas Priv-EGFR and RADGEN showed endpoint- or seed-dependent behavior. Direction reversal changed some small or uncertain effects. Direction is an essential part of the HCM specification. The corrected endpoints complement thresholded overlap by detecting and characterizing update-related attribution shifts under a fixed audit protocol; they measure internal consistency rather than clinical correctness.