Explainable Techniques for Analyzing Flow Cytometry Cell Transformers

TL;DR

This paper introduces explainability techniques tailored for Flow Cytometry data using a transformer model, enabling better understanding of model decisions in clinical cell classification tasks.

Contribution

It adapts and evaluates attention and gradient-based visualization methods specifically for FCM data, demonstrating their effectiveness in clinical cell analysis.

Findings

Gradient visualization highlights key cells influencing predictions.

Attention heads focus on biologically meaningful cell sub-populations.

Proposed methods improve interpretability of FCM cell classification models.

Abstract

Explainability for Deep Learning Models is especially important for clinical applications, where decisions of automated systems have far-reaching consequences. While various post-hoc explainable methods, such as attention visualization and saliency maps, already exist for common data modalities, including natural language and images, little work has been done to adapt them to the modality of Flow CytoMetry (FCM) data. In this work, we evaluate the usage of a transformer architecture called ReluFormer that ease attention visualization as well as we propose a gradient- and an attention-based visualization technique tailored for FCM. We qualitatively evaluate the visualization techniques for cell classification and polygon regression on pediatric Acute Lymphoblastic Leukemia (ALL) FCM samples. The results outline the model's decision process and demonstrate how to utilize the proposed techniques to inspect the trained model. The gradient-based visualization not only identifies cells that are most significant for a particular prediction but also indicates the directions in the FCM feature space in which changes have the most impact on the prediction. The attention visualization provides insights on the transformer's decision process when handling FCM data. We show that different attention heads specialize by attending to different biologically meaningful sub-populations in the data, even though the model retrieved solely supervised binary classification signals during training.