Project

HYPO: Digital Pathology and Artificial Intelligence for Precision Oncology

Short Summary

The latest generation of machine learning tools, and in particular graph-based methods, provide accurate predictions that take into account complex interactions at different granularity levels. This is key to harness the wealth of information contained in whole-slide digital pathology images, e.g. specific cell-cell or cell-tissue interactions, especially when the multiplexed immunohistochemistry (mIHC) technique is used, which allows locating cells of up to 7 different types. This essential capacity will allow the identification of key patterns of the biological processes underlying response to cancer immunotherapy, and eventually the development of interpretable biomarkers that will be better understood, trusted and adopted by clinicians. Clinical outcome predictions, leading (i) to define actionable candidate biomarkers for melanoma aggressivity and immune-resistance; (ii) to recognize automatically specific cell types in the cheap Hematoxylin-Eosin (H&E) stained images that are routinely produced; and (iii) to develop next-generation H&E-based biomarkers exploiting predicted cell type information. Both arms of this proposal, namely the discovery of novel mIHC-based interpretable biomarkers in melanoma, and the potential to re-interpret existing H&E databases for biomarker discovery have strong translational potential for future patients. Indeed, with the advent of digital pathology in clinical practice, such H&E databases are bound to grow fast in the coming years.

Goals

This project promises to have a high impact on machine learning adoption by Pathologists and Oncologists. The interpretable nature of the proposed biomarkers will be key, as the absence of such explainability of results is among the main bottlenecks for a wide adoption of highly promising data science methods. Our second objective also goes in this direction by providing extra added value to existing H&E databases, which will ultimately promote the development of next-generation biomarkers across all cancer types.

Significance

The project is expected to have a high impact on patient care, as novel biomarkers of immunotherapy response, if validated clinically, will lead to better patient stratification and help proposing personalized treatment options.

Background

The recent advances in Oncology, particularly in immunotherapy with immune checkpoints inhibitors (ICI), have delivered astounding clinical successes, but the response rate remains partial, with some patients experiencing no benefit and strong adverse effects. This calls for the development of personalized medicine approaches to cancer therapy. The analysis of histological sections of tumor tissue is widely used as a diagnostic tool for various forms of cancers, but this rich source of patient-specific information can be exploited much further if combined with modern image processing and machine learning techniques. In particular, multiplexed immunohistochemistry (mIHC) reveals the localization of specific proteins that are signatures of specific cell populations. Such mIHC images hold great potential to be further analyzed in terms of complex features, such as density profiles, cell type co-localization, network connectedness, which can be subsequently fed to machine learning algorithms.

iDoc

Co-Investigators

Prof. Olivier Michielin, CHUV
Dr Dorina Thanou, EPFL
Dr Michel Cuendet, CHUV

Consortium

Status

In Progress

5th Call, iDoc