Yoni Schirris

Yoni Schirris, Eric Marcus, Jonas Teuwen et al.

Explaining deep learning models is essential for clinical integration of medical image analysis systems. A good explanation highlights if a model depends on spurious features that undermines generalization and harms a subset of patients or, conversely, may present novel biological insights. Although techniques like GradCAM can identify influential features, they are measurement tools that do not themselves form an explanation. We propose a human-machine-VLM interaction system tailored to explaining classifiers in computational pathology, including multi-instance learning for whole-slide images. Our proof of concept comprises (1) an AI-integrated slide viewer to run sliding-window experiments to test claims of an explanation, and (2) quantification of an explanation's predictiveness using general-purpose vision-language models. The results demonstrate that this allows us to qualitatively test claims of explanations and can quantifiably distinguish competing explanations. This offers a practical path from explainable AI to explained AI in digital pathology and beyond. Code and prompts are available at https://github.com/nki-ai/x2x.

Yoni Schirris, Rosie Voorthuis, Mark Opdam et al.

The level of tumour-infiltrating lymphocytes (TILs) is a prognostic factor for patients with (triple-negative) breast cancer (BC). Computational TIL assessment (CTA) has the potential to assist pathologists in this labour-intensive task, but current CTA models rely heavily on many detailed annotations. We propose and validate a fundamentally simpler deep learning based CTA that can be trained in only ten minutes on hundredfold fewer pathologist annotations. We collected whole slide images (WSIs) with TILs scores and clinical data of 2,340 patients with BC from six cohorts including three randomised clinical trials. Morphological features were extracted from whole slide images (WSIs) using a pathology foundation model. Our label-efficient Computational stromal TIL assessment model (ECTIL) directly regresses the TILs score from these features. ECTIL trained on only a few hundred samples (ECTIL-TCGA) showed concordance with the pathologist over five heterogeneous external cohorts (r=0.54-0.74, AUROC=0.80-0.94). Training on all slides of five cohorts (ECTIL-combined) improved results on a held-out test set (r=0.69, AUROC=0.85). Multivariable Cox regression analyses indicated that every 10% increase of ECTIL scores was associated with improved overall survival independent of clinicopathological variables (HR 0.86, p<0.01), similar to the pathologist score (HR 0.87, p<0.001). We demonstrate that ECTIL is highly concordant with an expert pathologist and obtains a similar hazard ratio. ECTIL has a fundamentally simpler design than existing methods and can be trained on orders of magnitude fewer annotations. Such a CTA may be used to pre-screen patients for, e.g., immunotherapy clinical trial inclusion, or as a tool to assist clinicians in the diagnostic work-up of patients with BC. Our model is available under an open source licence (https://github.com/nki-ai/ectil).

Yoni Schirris, Mendel Engelaer, Andreas Panteli et al.

We present WeakSTIL, an interpretable two-stage weak label deep learning pipeline for scoring the percentage of stromal tumor infiltrating lymphocytes (sTIL%) in H&E-stained whole-slide images (WSIs) of breast cancer tissue. The sTIL% score is a prognostic and predictive biomarker for many solid tumor types. However, due to the high labeling efforts and high intra- and interobserver variability within and between expert annotators, this biomarker is currently not used in routine clinical decision making. WeakSTIL compresses tiles of a WSI using a feature extractor pre-trained with self-supervised learning on unlabeled histopathology data and learns to predict precise sTIL% scores for each tile in the tumor bed by using a multiple instance learning regressor that only requires a weak WSI-level label. By requiring only a weak label, we overcome the large annotation efforts required to train currently existing TIL detection methods. We show that WeakSTIL is at least as good as other TIL detection methods when predicting the WSI-level sTIL% score, reaching a coefficient of determination of $0.45\pm0.15$ when compared to scores generated by an expert pathologist, and an AUC of $0.89\pm0.05$ when treating it as the clinically interesting sTIL-high vs sTIL-low classification task. Additionally, we show that the intermediate tile-level predictions of WeakSTIL are highly interpretable, which suggests that WeakSTIL pays attention to latent features related to the number of TILs and the tissue type. In the future, WeakSTIL may be used to provide consistent and interpretable sTIL% predictions to stratify breast cancer patients into targeted therapy arms.

Yoni Schirris, Efstratios Gavves, Iris Nederlof et al.

We propose a Deep learning-based weak label learning method for analyzing whole slide images (WSIs) of Hematoxylin and Eosin (H&E) stained tumor tissue not requiring pixel-level or tile-level annotations using Self-supervised pre-training and heterogeneity-aware deep Multiple Instance LEarning (DeepSMILE). We apply DeepSMILE to the task of Homologous recombination deficiency (HRD) and microsatellite instability (MSI) prediction. We utilize contrastive self-supervised learning to pre-train a feature extractor on histopathology tiles of cancer tissue. Additionally, we use variability-aware deep multiple instance learning to learn the tile feature aggregation function while modeling tumor heterogeneity. For MSI prediction in a tumor-annotated and color normalized subset of TCGA-CRC (n=360 patients), contrastive self-supervised learning improves the tile supervision baseline from 0.77 to 0.87 AUROC, on par with our proposed DeepSMILE method. On TCGA-BC (n=1041 patients) without any manual annotations, DeepSMILE improves HRD classification performance from 0.77 to 0.81 AUROC compared to tile supervision with either a self-supervised or ImageNet pre-trained feature extractor. Our proposed methods reach the baseline performance using only 40% of the labeled data on both datasets. These improvements suggest we can use standard self-supervised learning techniques combined with multiple instance learning in the histopathology domain to improve genomic label classification performance with fewer labeled data.