Multiple instance learning (MIL) has enabled substantial progress in computational histopathology, where a large amount of patches from gigapixel whole slide images are aggregated into slide-level predictions. Heatmaps are widely used to validate MIL models and to discover tissue biomarkers. Yet, the validity of these heatmaps has barely been investigated. In this work, we introduce a general framework for evaluating the quality of MIL heatmaps without requiring additional labels. We conduct a large-scale benchmark experiment to assess six explanation methods across histopathology task types (classification, regression, survival), MIL model architectures (Attention-, Transformer-, Mamba-based), and patch encoder backbones (UNI2, Virchow2). Our results show that explanation quality mostly depends on MIL model architecture and task type, with perturbation ("Single"), layer-wise relevance propagation (LRP), and integrated gradients (IG) consistently outperforming attention-based and gradient-based saliency heatmaps, which often fail to reflect model decision mechanisms. We further demonstrate the advanced capabilities of the best-performing explanation methods: (i) We provide a proof-of-concept that MIL heatmaps of a bulk gene expression prediction model can be correlated with spatial transcriptomics for biological validation, and (ii) showcase the discovery of distinct model strategies for predicting human papillomavirus (HPV) infection from head and neck cancer slides. Our work highlights the importance of validating MIL heatmaps and establishes that improved explainability can enable more reliable model validation and yield biological insights, making a case for a broader adoption of explainable AI in digital pathology. Our code is provided in a public GitHub repository: https://github.com/bifold-pathomics/xMIL.

Deep learning models have made significant advances in histological prediction tasks in recent years. However, for adaptation in clinical practice, their lack of robustness to varying conditions such as staining, scanner, hospital, and demographics is still a limiting factor: if trained on overrepresented subpopulations, models regularly struggle with less frequent patterns, leading to shortcut learning and biased predictions. Large-scale foundation models have not fully eliminated this issue. Therefore, we propose a novel approach explicitly modeling such metadata into a Me tadata-guided generative Di ffusion model framework (MeDi). MeDi allows for a targeted augmentation of underrepresented subpopulations with synthetic data, which balances limited training data and mitigates biases in downstream models. We experimentally show that MeDi generates high-quality histopathology images for unseen subpopulations in TCGA, boosts the overall fidelity of the generated images, and enables improvements in performance for downstream classifiers on datasets with subpopulation shifts. Our work is a proof-of-concept towards better mitigating data biases with generative models.

Multiple instance learning (MIL) is an effective and widely used approach for weakly supervised machine learning. In histopathology, MIL models have achieved remarkable success in tasks like tumor detection, biomarker prediction, and outcome prognostication. However, MIL explanation methods are still lagging behind, as they are limited to small bag sizes or disregard instance interactions. We revisit MIL through the lens of explainable AI (XAI) and introduce xMIL, a refined framework with more general assumptions. We demonstrate how to obtain improved MIL explanations using layer-wise relevance propagation (LRP) and conduct extensive evaluation experiments on three toy settings and four real-world histopathology datasets. Our approach consistently outperforms previous explanation attempts with particularly improved faithfulness scores on challenging biomarker prediction tasks. Finally, we showcase how xMIL explanations enable pathologists to extract insights from MIL models, representing a significant advance for knowledge discovery and model debugging in digital histopathology.

We present DiffInfinite, a hierarchical diffusion model that generates arbitrarily large histological images while preserving long-range correlation structural information. Our approach first generates synthetic segmentation masks, subsequently used as conditions for the high-fidelity generative diffusion process. The proposed sampling method can be scaled up to any desired image size while only requiring small patches for fast training. Moreover, it can be parallelized more efficiently than previous large-content generation methods while avoiding tiling artifacts. The training leverages classifier-free guidance to augment a small, sparsely annotated dataset with unlabelled data. Our method alleviates unique challenges in histopathological imaging practice: large-scale information, costly manual annotation, and protective data handling. The biological plausibility of DiffInfinite data is evaluated in a survey by ten experienced pathologists as well as a downstream classification and segmentation task. Samples from the model score strongly on anti-copying metrics which is relevant for the protection of patient data.

OBJECTIVE: The objective was to develop a dataset definition, information model, and FHIR® specification for key data elements contained in a German molecular genomics (MolGen) report to facilitate genomic and phenotype integration in electronic health records. MATERIALS AND METHODS: A dedicated expert group participating in the German Medical Informatics Initiative reviewed information contained in MolGen reports, determined the key elements, and formulated a dataset definition. HL7's Genomics Reporting Implementation Guide (IG) was adopted as a basis for the FHIR® specification which was subjected to a public ballot. In addition, elements in the MolGen dataset were mapped to the fields defined in ISO/TS 20428:2017 standard to evaluate compliance. RESULTS: A core dataset of 76 data elements, clustered into 6 categories was created to represent all key information of German MolGen reports. Based on this, a FHIR specification with 16 profiles, 14 derived from HL7®'s Genomics Reporting IG and 2 additional profiles (of the FamilyMemberHistory and RiskAssessment resources), was developed. Five example resource bundles show how our adaptation of an international standard can be used to model MolGen report data that was requested following oncological or rare disease indications. Furthermore, the map of the MolGen report data elements to the fields defined by the ISO/TC 20428:2017 standard, confirmed the presence of the majority of required fields. CONCLUSIONS: Our report serves as a template for other research initiatives attempting to create a standard format for unstructured genomic report data. Use of standard formats facilitates integration of genomic data into electronic health records for clinical decision support.

MOTIVATION: Semiconductor sequencing directly translates chemically encoded information (A, C, G or T) into voltage signals that are detected by a semiconductor device. Changes of pH value and thereby of the electric potential in the reaction well are detected during strand synthesis from nucleotides provided in cyclic repeated flows for each type of nucleotide. To minimize time requirement and costs, it is necessary to know the number of flows that are required for complete coverage of the templates. RESULTS: We calculate the number of required flows in a random sequence model and present exact expressions for cumulative distribution function, expected value and variance. Additionally, we provide an algorithm to calculate the number of required flows for a concrete list of amplicons using a BED file of genomic positions as input. We apply the algorithm to calculate the number of flows that are required to cover six amplicon panels that are used for targeted sequencing in cancer research. The upper bounds obtained for the number of flows allow to enhance the instrument throughput from two chips to three chips per day for four of these panels. AVAILABILITY AND IMPLEMENTATION: The algorithm for calculation of the flows was implemented in R and is available as package ionflows from the CRAN repository. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.