Artificial intelligence (AI) has shown promise for diagnosing prostate cancer in biopsies. However, results have been limited to individual studies, lacking validation in multinational settings. Competitions have been shown to be accelerators for medical imaging innovations, but their impact is hindered by lack of reproducibility and independent validation. With this in mind, we organized the PANDA challenge—the largest histopathology competition to date, joined by 1,290 developers—to catalyze development of reproducible AI algorithms for Gleason grading using 10,616 digitized prostate biopsies. We validated that a diverse set of submitted algorithms reached pathologist-level performance on independent cross-continental cohorts, fully blinded to the algorithm developers. On United States and European external validation sets, the algorithms achieved agreements of 0.862 (quadratically weighted κ, 95% confidence interval (CI), 0.840–0.884) and 0.868 (95% CI, 0.835–0.900) with expert uropathologists. Successful generalization across different patient populations, laboratories and reference standards, achieved by a variety of algorithmic approaches, warrants evaluating AI-based Gleason grading in prospective clinical trials.

Deep learning algorithms have been successfully used in medical image classification. In the next stage, the technology of acquiring explainable knowledge from medical images is highly desired. Here we show that deep learning algorithm enables automated acquisition of explainable features from diagnostic annotation-free histopathology images. We compare the prediction accuracy of prostate cancer recurrence using our algorithm-generated features with that of diagnosis by expert pathologists using established criteria on 13,188 whole-mount pathology images consisting of over 86 billion image patches. Our method not only reveals findings established by humans but also features that have not been recognized, showing higher accuracy than human in prognostic prediction. Combining both our algorithm-generated features and human-established criteria predicts the recurrence more accurately than using either method alone. We confirm robustness of our method using external validation datasets including 2276 pathology images. This study opens up fields of machine learning analysis for discovering uncharted knowledge.

Diffuse large B-cell lymphoma (DLBCL) is an aggressive malignancy and is the most common type of malignant lymphoid neoplasm. While some DLBCLs exhibit strong cell-autonomous survival and proliferation activity, others depend on interactions with non-malignant cells for their survival and proliferation. Recent next-generation sequencing studies have linked these interactions with the molecular classification of DLBCL. For example, germinal center B-cell-like DLBCL tends to show strong associations with follicular T cells and epigenetic regulation of immune recognition molecules, whereas activated B-cell-like DLBCL shows frequent genetic aberrations affecting the class I major histocompatibility complex. Single-cell technologies have also provided detailed information about cell–cell interactions and the cell composition of the microenvironment of DLBCL. Aging-related immunological deterioration, i.e., immunosenescence, also plays an important role in DLBCL pathogenesis, especially in Epstein-Barr virus-positive DLBCL. Moreover, DLBCL in “immune-privileged sites”—where multiple immune-modulating mechanisms exist—shows unique biological features, including frequent down-regulation of immune recognition molecules and an immune-tolerogenic tumor microenvironment. These advances in understanding the immunology of DLBCL may contribute to the development of novel therapies targeting immune systems.

The fifth edition of the WHO Classification of Tumours of the Urinary and Male Genital Systems encompasses several updates to the classification and diagnosis of prostatic carcinoma as well as incorporating advancements in the assessment of its prognosis, including recent grading modifications. Some of the salient aspects include: (1) recognition that prostatic intraepithelial neoplasia (PIN)‐like carcinoma is not synonymous with a pattern of ductal carcinoma, but better classified as a subtype of acinar adenocarcinoma; (2) a specific section on treatment‐related neuroendocrine prostatic carcinoma in view of the tight correlation between androgen deprivation therapy and the development of prostatic carcinoma with neuroendocrine morphology, and the emerging data on lineage plasticity; (3) a terminology change of basal cell carcinoma to “adenoid cystic (basal cell) cell carcinoma” given the presence of an underlyingMYB::NFIBgene fusion in many cases; (4) discussion of the current issues in the grading of acinar adenocarcinoma and the prognostic significance of cribriform growth patterns; and (5) more detailed coverage of intraductal carcinoma of prostate (IDC‐P) reflecting our increased knowledge of this entity, while recommending the descriptive term atypical intraductal proliferation (AIP) for lesions falling short of IDC‐P but containing more atypia than typically seen in high‐grade prostatic intraepithelial neoplasia (HGPIN). Lesions previously regarded as cribriform patterns of HGPIN are now included in the AIP category. This review discusses these developments, summarising the existing literature, as well as the emerging morphological and molecular data that underpins the classification and prognostication of prostatic carcinoma.