Newswise — Upstate researcher Tamara Jamaspishvili, MD/PhD, is a second-time awardee in the past 2 years for the "Best Research Poster" Award for Faculty at the Digital Pathology Association's national conference, PathVisions 2024.

This conference brings together a multidisciplinary group of researchers, pathologists, physicians, computational scientists from both academia and industry to showcase cutting-edge advancements in digital and computational pathology, the use of artificial intelligence (AI) in research and clinical applications, and future innovations in the field.

“AI-driven diagnostics and prognostication have the potential to transform the future of healthcare practices and precision oncology,” explains Jamaspishvili, an assistant professor of Pathology and director of the Pathology Research Core Facility & Digital Pathology at Upstate (SUNY SPORE). The award recognizes Jamaspishvili for her work using AI and computational pathology to improve cancer diagnosis and treatment.

This groundbreaking project in computational immuno-oncology was conducted in collaboration with Dr. Sushant Patkar, a computational biologist and scientist from the Artificial Intelligence Resource (AIR) group at NCI/NIH, and a multi-disciplinary group of researchers and physicians at Upstate

The main goal of this study is to improve the treatment and survival of non-small cell lung cancer patients by helping doctors better predict their response to immune checkpoint inhibitors (ICI) using advanced AI tools and diagnostics. These treatments enhance the immune system's ability to fight cancer, but determining which patients will benefit the most remains a challenge.

To address this issue, researchers developed HistoTME, a cost-efficient and easily deployable AI tool. This advanced deep learning algorithm "reads" routinely stained histopathology images of tumor samples, predicting molecularly defined subtypes (derived from bulk RNA sequencing) and hence, inferring details about the tumor microenvironment (TME).

By analyzing these scanned pathology images, HistoTME can identify the presence and activity of specific cell types in the surrounding tumor tissue. This provides doctors with new insights into each patient's unique TME composition, which is crucial for predicting personalized ICI treatment responses in patients with low levels of PD-L1 expression, a commonly used companion diagnostic test.

The algorithm was successfully validated on a multi-modal dataset of over 650 lung cancer patients (>1500 images) recently developed by SUNY SPORE. Jamaspishvili and Patkar hope this method will help doctors choose personalized treatment plans more accurately and cost-efficiently and would be valuable, particularly for patients who do not have access to expensive molecular testing. Additionally, this complementary test will enhance current companion diagnostics, which face challenges in accurately identifying the right patients for the right treatment. The researchers are now planning the next phase of their study, which involves clinical validation of HistoTME. This crucial step will further assess the effectiveness of this tool in real-world clinical settings and potentially pave the way for its integration into routine cancer care.

The award-winning poster is available here.

A paper on this project was recently accepted for publication by Nature Precision Oncology journal. To read the full paper, click here.

Caption: Upstate Medical University's Tamara Jamaspishvili, MD/PhD, accepts the "Best Research Poster" Award for Faculty at the Digital Pathology Association's national conference, PathVisions 2024.

Journal Link: Nature Precision Oncology