Histopathologic reassessment of melanoma and other melanocytic skin lesions excised in 2009 and 2018-2019. / Histopatologisk revurdering av melanom og andre melanocytære hudlesjoner eksidert i 2009 og 2018­19.

BACKGROUND: Histopathological assessment of melanoma and other melanocytic skin lesions can be difficult and can vary between pathologists. MATERIAL AND METHOD: Histopathological slides of 196 melanocytic skin lesions from 2009 and 2018-2019 were obtained from the archive of the Department of Pathology at Oslo University Hospital and classified into six diagnostic categories: 1) benign nevus, 2) irregular/dysplastic nevus, i.e. dysplastic nevus with moderate atypia, 3) nevus with severe atypia, i.e. dysplastic nevus with severe atypia, 4) melanoma in situ, 5) superficial spreading or lentiginous melanoma and 6) nodular melanoma. The slides were then examined independently and blindly by three experienced pathologists and categorised in the same way. Interobserver agreement was assessed with Cohen's kappa, and agreement with the original diagnosis was assessed by the proportion of assessments in the same diagnostic category. RESULTS: The kappa values for the assessments from the three pathologists ranged from 0.45 to 0.50. The proportion of reassessments in agreement with the original diagnostic category was 85.7 % (95 % CI 75.7 to 92.1), 29.2 % (19.9 to 40.5), 27.8 % (20.9 to 36.0), 78.3 % (70.4 to 84.5), 81.2 % (73.7 to 86.9) and 93.3 % (82.1 to 97.7), respectively, i.e. highest for nodular melanoma. The proportion of reassessments in which the diagnosis was more serious or less serious than the original diagnosis was higher and lower, respectively, for slides from 2009 than for slides from 2018-2019. INTERPRETATION: The differences between the pathologists' assessments and deviations from the original diagnoses can be explained by poorly reproducible diagnostic criteria, diagnostic entities with overlapping morphology and increasing awareness of early signs of malignancy. Some evolution in diagnostic practice cannot be ruled out.

The immuneML ecosystem for machine learning analysis of adaptive immune receptor repertoires.

Adaptive immune receptor repertoires (AIRR) are key targets for biomedical research as they record past and ongoing adaptive immune responses. The capacity of machine learning (ML) to identify complex discriminative sequence patterns renders it an ideal approach for AIRR-based diagnostic and therapeutic discovery. To date, widespread adoption of AIRR ML has been inhibited by a lack of reproducibility, transparency, and interoperability. immuneML (immuneml.uio.no) addresses these concerns by implementing each step of the AIRR ML process in an extensible, open-source software ecosystem that is based on fully specified and shareable workflows. To facilitate widespread user adoption, immuneML is available as a command-line tool and through an intuitive Galaxy web interface, and extensive documentation of workflows is provided. We demonstrate the broad applicability of immuneML by (i) reproducing a large-scale study on immune state prediction, (ii) developing, integrating, and applying a novel deep learning method for antigen specificity prediction, and (iii) showcasing streamlined interpretability-focused benchmarking of AIRR ML.