Immediate image-based molecular profiling to empower optimized diagnosis process
Background: Cancer treatment relies on accurate diagnosis and understanding of the tumor's molecular characteristics. However, the process of molecular profiling is time-consuming, often taking weeks, which can result in delays in the ability to make treatment decisions and initiation, causing frustration for both patients and doctors. Additionally, in many cases, biopsies provide limited amounts of tissue, leading to partial molecular testing instead of comprehensive sequencing.
This study introduces an innovative approach to molecular pathology using a rapid image-based biomarkers profiling solution. The Initial diagnosis process begins with pathologists' Hematoxylin and Eosin (H&E) stained slides examination. Our solution provides the pathologist with complementary molecular profiling information within minutes using only the diagnostic slide.
Aim: To demonstrate the potential clinical utility of an immediate image-based machine-learning (ML) solution for molecular profiling, we aim to share real-world examples of cancer patients from several medical centers that required urgent treatment. Rapid molecular profiling was utilized to their benefit.
Methods: H&E whole slide images (WSIs) from several medical institutes, Imagene’s internal resources and publicly available databases were used to generate different ML-based classifiers for panels, including a wide range of biomarkers using Self Supervised and Multiple Instance Learning pipelines. Multisite validations were performed on both retrospective and prospective cohorts.
Results: In this study, we assessed several tissue biomarker panels and evaluated their accuracies using cross-validation multicenter retrospective cohorts that closely aligned with clinical implementation.
Notably, we observed several cases where using a rapid image-based solution as a complementary method to the routine workflow provided distinct benefits. It effectively identified cases that warranted further testing and optimally recommended incorporating focused molecular testing to reduce turnaround times and minimize treatment delays.
Conclusion: Here we describe several clinical cases where the use of AI-based molecular profiling emphasized the benefits it can provide for immediate, holistic and optimized cancer diagnosis and pathology workflows. This technology can be seamlessly integrated into different practice settings and assist in cases where tissue availability is limited, indicating which focused tests will most likely be informative to the patient. Moreover, it can rapidly identify patients who would likely benefit from additional, specific and faster tests, thus maximizing the potential of precision oncology for cancer patients.
- Understand the potential use of AI within the cancer diagnostic workflow.
- Understand how AI can identify cancer biomarkers and improve patient care.
- Explore how rapid molecular profiling can assist in making treatment decisions.
Nurit Paz-Yaacov, PhD
Chief Scientific Officer
Dr. Nurit Paz-Yaacov is the Chief Scientific Officer at Imagene AI - a BioMed startup company with groundbreaking technology in the field of precision oncology using Artificial Intelligence. Dr. Paz-Yaacov brings more than 20 years of multidisciplinary scientific research in human genetics and genomic sequencing and interpretation, with an emphasis on cancer research. Prior to her role at Imagene AI, Dr. Paz-Yaacov was the Head of Scientific Research at Genoox, a cloud platform that provides actionable interpretation for real-time/real-life genomic data. Dr. Paz-Yaacov holds B.Sc, M.Sc, and Ph.D. degrees from Tel-Aviv University and completed her Postdoctoral studies at Bar-Ilan University, Her studies and research in the fields of cancer, applied genomics, and epi-genomics, received recognition through numerous publications and prizes