Tissue Concepts: supervised foundation models in computational pathology

TL;DR

This paper introduces a supervised training approach for foundation models in computational pathology, significantly reducing data and computational costs while maintaining high performance and generalizability across multiple cancer types.

Contribution

It presents a multi-task supervised training method that creates a versatile encoder, called Tissue Concepts, which rivals self-supervised models and outperforms ImageNet pre-trained models in pathology tasks.

Findings

Achieves comparable performance to self-supervised models with only 6% of training data.

Outperforms ImageNet pre-trained encoder on in-domain and out-of-domain data.

Reduces training expenses significantly in foundation model development.

Abstract

Due to the increasing workload of pathologists, the need for automation to support diagnostic tasks and quantitative biomarker evaluation is becoming more and more apparent. Foundation models have the potential to improve generalizability within and across centers and serve as starting points for data efficient development of specialized yet robust AI models. However, the training foundation models themselves is usually very expensive in terms of data, computation, and time. This paper proposes a supervised training method that drastically reduces these expenses. The proposed method is based on multi-task learning to train a joint encoder, by combining 16 different classification, segmentation, and detection tasks on a total of 912,000 patches. Since the encoder is capable of capturing the properties of the samples, we term it the Tissue Concepts encoder. To evaluate the performance and generalizability of the Tissue Concepts encoder across centers, classification of whole slide images from four of the most prevalent solid cancers - breast, colon, lung, and prostate - was used. The experiments show that the Tissue Concepts model achieve comparable performance to models trained with self-supervision, while requiring only 6% of the amount of training patches. Furthermore, the Tissue Concepts encoder outperforms an ImageNet pre-trained encoder on both in-domain and out-of-domain data.