ValidAI Clinical VALIDation and implementation of the AI-based digital biomarker Histotype Px to personalize treatment in colorectal cancer

Colorectal cancer (CRC) is the third most common and the second most deadly cancer globally, with approximately 2 million new cases and 0.9 million deaths every year. Specifically, there are currently no reliable tools to guide the decision of whether or not to give chemotherapy after surgery in stage 2 and stage 3 CRC, and this leads to substantial overtreatment with chemotherapy in this patient population. In fact, up to 90% of stage 2 and stage 3 CRC patients who receive chemotherapy are only exposed to the toxic effects of this treatment without any treatment benefit. Given the substantial and often chronic side effects associated with chemotherapy use, there is currently a large need for better and more cost-effective care for colorectal cancer patients by ensuring that chemotherapy is only given to those patients that are likely to benefit from it. Recent breakthroughs in digital pathology and AI opens new opportunities to develop digital biomarkers to extract and utilize information from tumour tissue not visible to the human eye. Histotype Px, DoMore Diagnostics’ AI-driven digital biomarker, uses visual information from the tumor to predict long-term patient outcomes. In this project, we aim to validate the biomarker as predictive of chemotherapy benefit both on historical patient data and also in a gold-standard prospective, randomized trial. If successful, this project will further support the use and wide adoption of Histotype Px as a decision support tool to guide the decision of whether or not to provide chemotherapy treatment after surgery in individual stage 2 and stage 3 colorectal cancer patients.

There is a large need and potential for better and more cost-effective care for cancer patients by personalized treatment decisions. Some patients are not offered potentially beneficial treatment (7%), however most patients (+50%) suffer from toxic overtreatment. Recent breakthroughs in digital pathology and AI opens new opportunities to develop digital biomarkers to extract and utilize information from tumour tissue not visible to the human eye. Histotype Px®, DoMore Diagnostics’ AI-driven digital pathology technology platform has the potential to provide clinicians with rapid, cost-effective and highly accurate prognostic and predictive biomarkers to guide personalized treatment decisions for patients suffering from colorectal cancer, thereby improving patients’ lives and avoiding unnecessary health care costs. The objective of the current project is to document the AI-based digital biomarker Histotype Px as a predictive marker for adjuvant chemotherapy (ACT) in stage II and III colorectal cancer, by studying whether there is a differential benefit of ACT for patients classified as low, intermediate and high risk in both retrospective and prospective studies. If successful, this will strongly support personalized clinical decision making regarding offering ACT as ideally only high risk patients with predicted benefit from ACT will be offered ACT. Furthermore, this project aims to implement and validate the integration of the biomarker into Sectra Image Management System. The objective is to rearchitect the Histotype Px Platform to be a scalable, cybersecure, commercial cloud service that fulfils all regulatory requirements and is integrated into the current workflow used in the clinic.