Artificial intelligence strategy integrating morphologic and architectural biomarkers provides robust diagnostic accuracy for disease progression in chronic lymphocytic leukemia

Abstract

Artificial intelligence-based tools designed to assist in the diagnosis of lymphoid neoplasms remain limited. The development of such tools can add value as a diagnostic aid in the evaluation of tissue samples involved by lymphoma. A common diagnostic question is the determination of chronic lymphocytic leukemia (CLL) progression to accelerated CLL (aCLL) or transformation to diffuse large B-cell lymphoma (Richter transformation; RT) in patients who develop progressive disease. The morphologic assessment of CLL, aCLL, and RT can be diagnostically challenging. Using established diagnostic criteria of CLL progression/transformation, we designed four artificial intelligenceconstructed biomarkers based on cytologic (nuclear size and nuclear intensity) and architectural (cellular density and cell to nearest-neighbor distance) features. We analyzed the predictive value of implementing these biomarkers individually and then in an iterative sequential manner to distinguish tissue samples with CLL, aCLL, and RT. Our model, based on these four morphologic biomarker attributes, achieved a robust analytic accuracy. This study suggests that biomarkers identified using artificial intelligence-based tools can be used to assist in the diagnostic evaluation of tissue samples from patients with CLL who develop aggressive disease features.

Biology and Treatment of Richter Transformation

“…Artificial intelligence tools can assist the diagnostic process for patients with a suspected RT. Four biomarkers have been recently identified to have consistent value for an RT-diagnosis model, according to cytologic (nuclear size and nuclear intensity) and architectural (cellular density and cell to nearest-neighbor distance) characteristics ( 114 ). This model was used to distinguish CLL from aCLL and RT cases with a good performance, and could be of support for further studies.…”

Quantitative features to assist in the diagnostic assessment of chronic lymphocytic leukemia progression†

“…In a recent paper in The Journal of Pathology , El Hussein, Chen et al [2] proposed an algorithmic procedure to assist pathologists in the diagnostic assessment of CLL progression using whole‐slide image analysis of tissue samples. They proposed four quantitative descriptors (biomarkers) obtained from cell nuclei segmentation using a convolutional neural network (CNN).…”

“…After learning, CNNs are able to extract abstract quantitative features that represent the images and are relevant to tasks such as segmentation or classification. As briefly reported in ref 2, CNNs have been used to develop automatic classification tools for digital pathology images. The authors highlighted that features obtained from CNN models are impossible or difficult to interpret, which is a motivation to bring into play other characteristics that can be directly introduced by pathologists, such as those proposed in ref 2.…”

“…In addition to individual nuclear descriptors, El Hussein, Chen et al [2] proposed two descriptors for the entire slide: cellular density and the distance between neighboring cells. Density is intuitive and is easily calculated by counting pixels.…”

“…The purpose of this commentary is to give a conceptual and functional overview of the strategy developed in ref 2, and to discuss the most relevant aspects. The strategy was performed following the steps illustrated in Figure 1: (1) image analysis and selection of regions of interest (ROIs); (2) nuclear segmentation; (3) calculation of morphological and architectural descriptors; and (4) classification of stage progression.…”