One-Slide Calibration of Pathology Foundation Models

A new method called SlideRuler calibrates pathology foundation models from a single slide scan, reducing mean target-to-source embedding distance by 16.3-38.5% relative to raw embeddings across two encoders and five SCORPION scanners, according to the arXiv paper 2610.08944v1. SlideRuler learns a transfer map from paired rescans and uses regions within a slide as internal controls to correct acquisition-induced shifts while keeping the foundation model frozen. A source-anchored variant cuts source-feature displacement by 47.7-83.6% relative to learned transfer while retaining most of its alignment gain, with positive same-slide contribution across all four evaluation settings, including scanner holdout.

arXiv:2610.08944v1 Announce Type: new Abstract: Scanner variation changes how pathology foundation models represent the same tissue. We introduce SlideRuler, which uses regions within a slide as internal controls to estimate and correct acquisition-induced shifts in other regions. A transfer map learned from paired rescans enables calibration from a single scan at inference while keeping the foundation model fixed. Across two encoders and five SCORPION scanners, learned transfer reduces mean target-to-source embedding distance by 16.3-38.5% relative to raw embeddings. Comparisons with unrelated same-scanner controls reveal a positive same-slide contribution across all four evaluation settings, including scanner holdout. A source-anchored variant reduces source-feature displacement by 47.7-83.6% relative to learned transfer while retaining most of its alignment gain. By drawing calibration information from the slide itself, SlideRuler offers a path toward more consistent use of frozen pathology models across imaging systems.