Image-Based Deep Learning Detection of High-Grade B-Cell Lymphomas Directly from Hematoxylin and Eosin Images

Abstract

Deep learning applications are emerging as promising new tools that can support the diagnosis and classification of different cancer types. While such solutions hold great potential for hematological malignancies, there have been limited studies describing the use of such applications in this field. The rapid diagnosis of double/triple-hit lymphomas (DHLs/THLs) involving MYC, BCL2 and/or BCL6 rearrangements is obligatory for optimal patient care. Here, we present a novel deep learning tool for diagnosing DHLs/THLs directly from scanned images of biopsy slides. A total of 57 biopsies, including 32 in a training set (including five DH lymphoma cases) and 25 in a validation set (including 10 DH/TH cases), were included. The DHL-classifier demonstrated a sensitivity of 100%, a specificity of 87% and an AUC of 0.95, with only two false positive cases, compared to FISH. The DHL-classifier showed a 92% predictive value as a screening tool for performing conventional FISH analysis, over-performing currently used criteria. The work presented here provides the proof of concept for the potential use of an AI tool for the identification of DH/TH events. However, more extensive follow-up studies are required to assess the robustness of this tool and achieve high performances in a diverse population.

Keywords: BCL2 rearrangement; MYC rearrangement; artificial intelligence; deep learning; diffuse large B-cell lymphoma (DLBCL); double hit; high-grade B-cell lymphoma (HGBL).

Figure 1

Study schematic representation. A self-supervised step was performed on a pan-cancer cohort (including cases of solid cancers with no lymphoma biopsies), establishing a foundation model, followed by a fine-tuning step using the training set’s WSIs, generating the final DHL-classifier. For the DHL-classifier performance evaluation, the DH/TH status of 25 cases included in the validation set was evaluated, and the results were compared to the official results reported in the FISH analysis. WSI—whole slide image, MIL—multiple instance learning, DHL—double-hit lymphoma, DH—double-hit, TH—triple-hit.

Figure 2

Performance of the DHL-classifier. (A) DHL-classifier results and performance in the validation cohort. N—negative, P—positive, TN—true negative, TP—true positive, FN—false negative, FP—false positive, AUC—area under the curve. (B) Predictive values of conventional methods vs. the DHL-classifier as a screening tool for FISH analysis. Presented are the number of samples in the relevant bars and predictive values for each screening method used. The number on the bars represents the number of cases in the relevant group.