RAA-MIL: A Novel Framework for Classification of Oral Cytology

TL;DR

This work addresses automated, patient-level classification of oral cytology WSIs under weak supervision. It introduces Region-Affinity Attention MIL (RAA-MIL), a four-stage framework that tokenizes patches with a frozen self-supervised ViT, refines tokens via region affinity, and pools them with gated MIL to produce patient-level predictions, validated with cross-validation and ensemble. It extends the Oral Cytology Dataset with expert-verified, four-class patient labels and establishes the first patient-level benchmark for oral cytology, achieving 72.7% accuracy and 0.697 weighted F1 on an unseen test set, with notable gains in OSCC and OPMD. The results demonstrate the viability of region-aware MIL for reliable AI-assisted digital cytology and point to future generalizability improvements via dataset expansion and additional modalities.

Abstract

Cytology is a valuable tool for early detection of oral squamous cell carcinoma (OSCC). However, manual examination of cytology whole slide images (WSIs) is slow, subjective, and depends heavily on expert pathologists. To address this, we introduce the first weakly supervised deep learning framework for patient-level diagnosis of oral cytology whole slide images, leveraging the newly released Oral Cytology Dataset [1], which provides annotated cytology WSIs from ten medical centres across India. Each patient case is represented as a bag of cytology patches and assigned a diagnosis label (Healthy, Benign, Oral Potentially Malignant Disorders (OPMD), OSCC) by an in-house expert pathologist. These patient-level weak labels form a new extension to the dataset. We evaluate a baseline multiple-instance learning (MIL) model and a proposed Region-Affinity Attention MIL (RAA-MIL) that models spatial relationships between regions within each slide. The RAA-MIL achieves an average accuracy of 72.7%, weighted F1-score of 0.69 on an unseen test set, outperforming the baseline. This study establishes the first patient-level weakly supervised benchmark for oral cytology and moves toward reliable AI-assisted digital pathology.

Figures (2)

• Figure 1: Overview of the proposed RAA–MIL pipeline. Frozen ViT tokens are refined using RAA and pooled with gated MIL to produce a patient-level embedding, which a lightweight MLP maps to the final four-class diagnosis.
• Figure 2: Qualitative visualization of gated-MIL attention maps for three representative cases from the test set: Healthy, OPMD, and OSCC.