Multimodal Learning for Scalable Representation of High-Dimensional Medical Data

TL;DR

This paper presents MarbliX, a self-supervised multimodal framework that embeds pathology images and genomic data into binary codes for scalable, interpretable, and high-resolution patient similarity analysis, improving diagnostic accuracy.

Contribution

Introduces MarbliX, a novel self-supervised method for joint embedding of multimodal medical data into binary codes, enabling scalable retrieval and cross-modal analysis.

Findings

MarbliX achieves 85-89% accuracy in lung cancer classification.

Binary monograms perform comparably to real-valued embeddings in kidney cancer.

The framework outperforms unimodal models in clinical case retrieval.

Abstract

Integrating artificial intelligence (AI) with healthcare data is rapidly transforming medical diagnostics and driving progress toward precision medicine. However, effectively leveraging multimodal data, particularly digital pathology whole slide images (WSIs) and genomic sequencing, remains a significant challenge due to the intrinsic heterogeneity of these modalities and the need for scalable and interpretable frameworks. Existing diagnostic models typically operate on unimodal data, overlooking critical cross-modal interactions that can yield richer clinical insights. We introduce MarbliX (Multimodal Association and Retrieval with Binary Latent Indexed matriX), a self-supervised framework that learns to embed WSIs and immunogenomic profiles into compact, scalable binary codes, termed ``monogram.'' By optimizing a triplet contrastive objective across modalities, MarbliX captures high-resolution patient similarity in a unified latent space, enabling efficient retrieval of clinically relevant cases and facilitating case-based reasoning. \textcolor{black}{In lung cancer, MarbliX achieves 85-89\% across all evaluation metrics, outperforming histopathology (69-71\%) and immunogenomics (73-76\%). In kidney cancer, real-valued monograms yield the strongest performance (F1: 80-83\%, Accuracy: 87-90\%), with binary monograms slightly lower (F1: 78-82\%).