ImmuVis: Hyperconvolutional Foundation Model for Imaging Mass Cytometry

TL;DR

ImmuVis is a novel foundation model for imaging mass cytometry that adapts to varying marker sets, enabling efficient large-scale tissue profiling with calibrated uncertainty estimation.

Contribution

Introduces marker-adaptive hyperconvolutions in a foundation model for IMC, trained on the largest dataset, outperforming baselines with lower computational cost.

Findings

Outperforms state-of-the-art in virtual staining and classification.

Operates efficiently on arbitrary marker subsets without retraining.

Provides calibrated uncertainty estimates.

Abstract

We present ImmuVis, a family of efficient foundation models for imaging mass cytometry (IMC), a high-throughput multiplex imaging technology that handles molecular marker measurements as image channels and enables large-scale spatial tissue profiling. Unlike natural images, multiplex imaging lacks a fixed channel space, as real-world marker sets vary across studies, violating a core assumption of standard vision backbones. To address this, ImmuVis introduces marker-adaptive hyperconvolutions that generate convolutional kernels from learned marker embeddings, enabling a single model to operate on arbitrary measured marker subsets without retraining. We pretrain ImmuVis on the largest dataset to date, IMC17M (28 cohorts, 24,405 images, 265 markers, over 17M patches), using self-supervised masked reconstruction. ImmuVis outperforms state-of-the-art baselines and ablations in virtual staining and downstream classification tasks at substantially lower compute cost than transformer-based alternatives, and is the sole model that provides calibrated uncertainty via a heteroscedastic likelihood objective. These results position ImmuVis as a practical framework for real-world IMC modeling.