RACR-MIL: Rank-aware contextual reasoning for weakly supervised grading of squamous cell carcinoma using whole slide images

Anirudh Choudhary, Mosbah Aouad, Krishnakant V. Saboo, Angelina Susie Hwang, Jacob Kechter, Blake Bordeaux, Puneet Bhullar, David DiCaudo, Steven Nelson, Nneka Comfere, Emma Johnson, Olayemi Sokumbi, Jason Sluzevich, Leah Swanson, Dennis Murphree, Aaron Mangold, Ravi Iyer

Abstract:

Squamous cell carcinoma (SCC) is the most common cancer subtype, with an increasing incidence and a signi cant
impact on cancer-related mortality. SCC grading using whole slide images is inherently challenging due to the lack
of a reliable protocol and substantial tissue heterogeneity. We propose RACR-MIL, the rst weakly-supervised SCC
grading approach achieving robust generalization across multiple anatomies (skin, head & neck, lung). RACR-MIL is
an attention-based multiple-instance learning framework that enhances grade-relevant contextual representation learn-
ing and addresses tumor heterogeneity through two key innovations: (1) a hybrid WSI graph that captures both local
tissue context and non-local phenotypical dependencies between tumor regions, and (2) a rank-ordering constraint in
the attention mechanism that consistently prioritizes higher-grade tumor regions, aligning with pathologists’ diagnos-
tic process. Our model achieves state-of-the-art performance across multiple SCC datasets, achieving 3-9% higher
grading accuracy, resilience to class imbalance, and up to 16% improved tumor localization. In a pilot study, patholo-
gists reported that RACR-MIL improved grading e ciency in 60% of cases, underscoring its potential as a clinically
viable cancer diagnosis and grading assistant.