Digital and Computational Pathology Project

Traditionally, pathologists examine tissue from a biopsy to diagnose cancer, determine the type and stage, and look for biomarkers that indicate a patient is likely to respond to certain targeted therapies. Technological advancements called digital pathology have enabled innovative approaches to these assessments. In digital pathology, slides are scanned and then digitized for storage, viewing, and analysis. Sometimes, analysis can include the use of computational pathology platforms with artificial intelligence (AI)/machine learning (ML) algorithms to aid the pathologist in tissue analysis.  

These platforms have the potential to enable greater efficiency, accuracy, reproducibility, and standardization in cancer identification and treatment. However, different computational pathology platforms may define, measure, and report biomarker status differently. As the field of computational pathology continues to advance, identifying and addressing the evidentiary needs to support their use, including analytical and clinical validation, data quality, and risk assessment, will be important to inform regulatory frameworks and successful integration into clinical trials and practice. 

Previous Friends’ efforts have assessed sources of variability in biomarker assessments using diagnostic tests and identified opportunities to harmonize methodology to support aligned measurement and use (i.e., the TMB Harmonization Project and the HRD Harmonization Project). Similar to the tests in these efforts, there may be differences between computational pathology platforms in how specific biomarkers are measured from the digitized images. These differences introduce additional complexity to the measurement of biomarkers and may lead to variability in the results generated from different computational pathology platforms. As digital and computational pathology continues to advance, it will be important to propose pathways to support the advancement of this technology, and to understand the comparability of biomarker measurements across platforms, identify factors that may contribute to any variability observed, and propose areas for alignment.  

Friends assembled a multi-stakeholder group to identify opportunities for use of digital and computational pathology in oncology drug development. The group evaluated existing regulatory frameworks and developed proposals to support robust development of these emerging technologies including a risk-based approach to assessing evidentiary needs for validation. Future work will support evaluating how comparable biomarker measurements are across computational pathology platforms to help understand the types of data needed to appropriately evaluate tests and understand how a reference set of samples may be used to evaluate concordance across platforms. This work will help to identify factors that may contribute to variability and propose alignment in the field. 

Digital and computational pathology platforms have the potential to provide greater accuracy, reproducibility, and standardization of pathology features, expedite diagnosis or pathological scoring, establish new biomarkers, and identify and select the appropriate patients for treatments – all of which can contribute to improving patient outcomes. Supporting the robust development of these platforms and identifying potential sources of variability will help to inform future use and advancements in technology that deliver more precise patient care.