LymphoML: An interpretable artificial intelligence-based method identifies morphologic features that correlate with lymphoma subtype
Shankar, V.; Yang, X.; Krishna, V.; Tan, B.; Silva, O.; Rojansky, R.; Ng, A.; Valvert, F.; Briercheck, E.; Weinstock, D.; Natkunam, Y.; Fernandez-Pol, S.; Rajpurkar, P.
Show abstract
Lymphomas vary in terms of clinical behavior, morphology, and response to therapies and thus accurate classification is essential for appropriate management of patients. In this study, using a set of 670 cases of lymphoma obtained from a center in Guatemala City, we propose an interpretable machine learning method, LymphoML, for lymphoma subtyping into eight diagnostic categories. LymphoML sequentially applies steps of (1) object segmentation to extract nuclei, cells, and cytoplasm from hematoxylin and eosin (H&E)-stained tissue microarray (TMA) cores, (2) feature extraction of morphological, textural, and architectural features, and (3) aggregation of per-object features to create patch-level feature vectors for lymphoma classification. LymphoML achieves a diagnostic accuracy of 64.3% (AUROC: 85.9%, specificity: 88.7%, sensitivity: 66.9%) among 8 lymphoma subtypes using only H&E-stained TMA core sections, at a level similar to experienced hematopathologists. We find that the best models set of nuclear and cytoplasmic morphological, textural, and architectural features are most discriminative for diffuse large B-cell lymphoma (F1: 78.7%), classic Hodgkin lymphoma (F1 score: 74.5%), and mantle cell lymphoma (F1: 71.0%). Nuclear shape features provide the highest diagnostic yield, with nuclear texture, cytoplasmic, and architectural features providing smaller gains in accuracy. Finally, combining information from the H&E-based model together with the results of a limited set of immunohistochemical (IHC) stains resulted in a similar diagnostic accuracy (accuracy: 85.3%, AUROC: 95.7%, sensitivity: 84.5%, specificity: 93.5%) as with a much larger set of IHC stains (accuracy: 86.1%, AUROC: 96.7%, specificity: 93.2%, sensitivity: 86.0%). Our work suggests a potential way to incorporate machine learning tools into clinical practice to reduce the number of expensive IHC stains while achieving a similar level of diagnostic accuracy.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Attention-based whole-slide image compression achieves pathologist-level pre-screening of multi-organ routine histopathology biopsies 92%
- Built to last? Reproducibility and Reusability of Deep Learning Algorithms in Computational Pathology 92%
- Clinical-Grade Validation of an Autofluorescence Virtual Staining System with Human Experts and a Deep Learning System for Prostate Cancer 92%
Similar papers in this journal
- Deep learning accurately quantifies plasma cell percentages on CD138-stained bone marrow samples 95%
- Cell projection plots: a novel visualization of bone marrow aspirate cytology 92%
- Independent assessment of a deep learning system for lymph node metastasis detection on the Augmented Reality Microscope 92%
Similar papers in this journal
- Classification of human white blood cells using machine learning for stain-free imaging flow cytometry 93%
- Hematologist-level classification of mature B-cell neoplasm using deep learning on multiparameter flow cytometry data 93%
- DeepIFC: virtual fluorescent labeling of blood cells in imaging flow cytometry data with deep learning 92%
Similar papers in this journal
- Breast invasive ductal carcinoma classification on whole slide images with weakly-supervised and transfer learning 92%
- Topological Structures in the Space of Treatment-Naive Patients With Chronic Lymphocytic Leukemia 92%
- piNET: An Automated Proliferation Index Calculator Framework for Ki67 Breast Cancer Images 91%
Similar papers in this journal
- CREBBP and STAT6 co-mutation and 16p13 and 1p36 loss define the t(14;18)-negative diffuse variant of follicular lymphoma 90%
- Transcriptome sequencing of Hodgkin lymphoma Hodgkin and Reed-Stenberg cells reveals escape from NK cell recognition and an unfolded protein response 89%
- Dual T-cell constant β chain (TRBC)1 and TRBC2 staining for the identification of T-cell neoplasms by flow cytometry 89%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.