Pan-cancer Histopathology WSI Pre-training with Position-aware Masked Autoencoder

TL;DR

This paper introduces PAMA, a novel self-supervised pre-training framework for whole-slide image (WSI) analysis in pan-cancer studies, focusing on WSI-level features with innovative modules to improve semantic understanding and robustness.

Contribution

The paper presents a new position-aware masked autoencoder (PAMA) with cross-attention, kernel reorientation, and anchor dropout for WSI pre-training, addressing the gap in WSI-level feature learning.

Findings

PAMA outperforms 8 state-of-the-art WSI analysis methods.

Demonstrates strong generalization across 7 large-scale datasets.

Effective in discriminative pan-cancer classification.

Abstract

Large-scale pre-training models have promoted the development of histopathology image analysis. However, existing self-supervised methods for histopathology images primarily focus on learning patch features, while there is a notable gap in the availability of pre-training models specifically designed for WSI-level feature learning. In this paper, we propose a novel self-supervised learning framework for pan-cancer WSI-level representation pre-training with the designed position-aware masked autoencoder (PAMA). Meanwhile, we propose the position-aware cross-attention (PACA) module with a kernel reorientation (KRO) strategy and an anchor dropout (AD) mechanism. The KRO strategy can capture the complete semantic structure and eliminate ambiguity in WSIs, and the AD contributes to enhancing the robustness and generalization of the model. We evaluated our method on 7 large-scale datasets from multiple organs for pan-cancer classification tasks. The results have demonstrated the effectiveness and generalization of PAMA in discriminative WSI representation learning and pan-cancer WSI pre-training. The proposed method was also compared with 8 WSI analysis methods. The experimental results have indicated that our proposed PAMA is superior to the state-of-the-art methods. The code and checkpoints are available at https://github.com/WkEEn/PAMA.