AdvDINO: Domain-Adversarial Self-Supervised Representation Learning for Spatial Proteomics

TL;DR

AdvDINO introduces a domain-adversarial self-supervised learning framework that enhances robustness to domain shifts in biomedical imaging, demonstrated on multiplex immunofluorescence data from lung and breast cancer cohorts.

Contribution

It integrates a gradient reversal layer into DINOv2 to promote domain-invariant features, improving biological relevance and prognostic performance in spatial proteomics.

Findings

AdvDINO outperforms non-adversarial baselines in learning robust representations.

It uncovers phenotype clusters with prognostic significance.

The approach improves survival prediction in lung and breast cancer datasets.

Abstract

Self-supervised learning (SSL) has emerged as a powerful approach for learning visual representations without manual annotations. However, the robustness of standard SSL methods to domain shift -- systematic differences across data sources -- remains uncertain, posing an especially critical challenge in biomedical imaging where batch effects can obscure true biological signals. We present AdvDINO, a domain-adversarial SSL framework that integrates a gradient reversal layer into the DINOv2 architecture to promote domain-invariant feature learning. Applied to a real-world cohort of six-channel multiplex immunofluorescence (mIF) whole slide images from lung cancer patients, AdvDINO mitigates slide-specific biases to learn more robust and biologically meaningful representations than non-adversarial baselines. Across more than 5.46 million mIF image tiles, the model uncovers phenotype clusters with differing proteomic profiles and prognostic significance, and enables strong survival prediction performance via attention-based multiple instance learning. The improved robustness also extends to a breast cancer cohort. While demonstrated on mIF data, AdvDINO is broadly applicable to other medical imaging domains, where domain shift is a common challenge.