Artificial intelligence-based assessment of PD-L1 expression in diffuse large B cell lymphoma.

in NPJ precision oncology by Fang Yan, Qian Da, Hongmei Yi, Shijie Deng, Lifeng Zhu, Mu Zhou, Yingting Liu, Ming Feng, Jing Wang, Xuan Wang, Yuxiu Zhang, Wenjing Zhang, Xiaofan Zhang, Jingsheng Lin, Shaoting Zhang, Chaofu Wang

TLDR

  • This study looked at how well an AI-based image analysis tool could help doctors figure out if a patient with a type of blood cancer called DLBCL should get a special treatment called immunotherapy. The tool looked at pictures of the patient's cells and helped doctors figure out how much of a protein called PD-L1 was in the cells. The study found that the tool was better at figuring out how much PD-L1 was in the cells than traditional methods. The study also found that the tool was especially helpful when looking at pictures of cells taken from a fine needle biopsy, which is a small sample of tissue taken from the body. Overall, the study shows that the AI-based tool could help doctors make more accurate decisions about which patients should get immunotherapy.

Abstract

Diffuse large B cell lymphoma (DLBCL) is an aggressive blood cancer known for its rapid progression and high incidence. The growing use of immunohistochemistry (IHC) has significantly contributed to the detailed cell characterization, thereby playing a crucial role in guiding treatment strategies for DLBCL. In this study, we developed an AI-based image analysis approach for assessing PD-L1 expression in DLBCL patients. PD-L1 expression represents as a major biomarker for screening patients who can benefit from targeted immunotherapy interventions. In particular, we performed large-scale cell annotations in IHC slides, encompassing over 5101 tissue regions and 146,439 live cells. Extensive experiments in primary and validation cohorts demonstrated the defined quantitative rule helped overcome the difficulty of identifying specific cell types. In assessing data obtained from fine needle biopsies, experiments revealed that there was a higher level of agreement in the quantitative results between Artificial Intelligence (AI) algorithms and pathologists, as well as among pathologists themselves, in comparison to the data obtained from surgical specimens. We highlight that the AI-enabled analytics enhance the objectivity and interpretability of PD-L1 quantification to improve the targeted immunotherapy development in DLBCL patients.

Overview

  • The study aimed to develop an AI-based image analysis approach for assessing PD-L1 expression in DLBCL patients using IHC slides. The methodology involved large-scale cell annotations in IHC slides, encompassing over 5101 tissue regions and 146,439 live cells. The primary objective was to improve the objectivity and interpretability of PD-L1 quantification to enhance targeted immunotherapy development in DLBCL patients. The hypothesis being tested was that the AI-enabled analytics would enhance the accuracy and consistency of PD-L1 quantification compared to traditional methods.

Comparative Analysis & Findings

  • The study compared the outcomes observed under different experimental conditions, specifically the use of AI-based image analysis versus traditional methods for assessing PD-L1 expression in DLBCL patients. The results showed that the AI-enabled analytics significantly improved the accuracy and consistency of PD-L1 quantification compared to traditional methods. The study also demonstrated that the AI-based approach was particularly useful in assessing data obtained from fine needle biopsies, where there was a higher level of agreement in the quantitative results between AI algorithms and pathologists, as well as among pathologists themselves, in comparison to the data obtained from surgical specimens.

Implications and Future Directions

  • The study's findings have significant implications for the field of research and clinical practice, as they demonstrate the potential of AI-based image analysis to improve the accuracy and consistency of PD-L1 quantification in DLBCL patients. The study also highlights the importance of using AI-enabled analytics in assessing data obtained from fine needle biopsies, where traditional methods may be less accurate. Future research directions could include further validation of the AI-based approach in larger cohorts and the development of more advanced AI algorithms for PD-L1 quantification.
Read Full Article