Patients with advanced rectal cancer often receive radiotherapy before surgery. However, despite this being standard practice, this treatment only results in complete disappearance (complete response) prior to surgery in 15% of patients. Currently, there are no reliable biomarkers (a biological molecule found in blood, other body fluids, or tissues that is a sign of a normal or abnormal process, or of a condition or disease) to predict which colorectal patients will benefit from radiotherapy, meaning that many patients are unnecessarily exposed to significant side effects.
Launched in 2015, the S:CORT consortium, led by Professor Tim Maughan at the University of Oxford and driven by leading clinical and scientific expertise from the University of Aberdeen, the University of Birmingham, Universität Berne, University College London, the University of Leeds and Queen’s University Belfast, set out to identify possible predictive biomarkers, with the goal of improving and tailoring the treatment of individual rectal cancer patients.
The team has recently published three studies involving 826 patients, using advanced molecular and machine learning techniques to analyse pre-treatment biopsy samples. Patients were recruited from two distinct cohorts: the first supplied by the University of Aberdeen, and the second from the Aristotle trial, which is led by Professor David Sebag-Montefiore from the University of Leeds, funded by Cancer Research UK and coordinated by the University College London Cancer Trials Centre. The papers share promising results that suggest which patient groups are more likely to respond to radiotherapy and why.
Key Outcomes:
- Tumour Composition: The researchers found that the presence of specific stromal (supportive tissue) and immune cells within tumours is crucial for a complete response to radiotherapy.
- Gene Expression Model: Using machine learning, the researchers developed a model based on the expression of 33 genes that can predict treatment outcomes with over 80% accuracy.
- Biopsy Image Analysis: Another artificial intelligence method was used to analyse routine biopsy images and assign molecular subtypes to tumours. Molecular subtypes are categories of cancer that are defined by the unique genetic and molecular features of the tumour, helping guide more effective treatment strategies. One subtype, so-named imCMS1, was linked to a higher likelihood of complete response, while subtype imCMS4 was associated with resistance to treatment.
Further work is needed before the findings from these studies can be adopted into the clinic. Future research will focus on testing whether the addition of immunotherapy and/or chemotherapy drugs such as oxaliplatin can improve the effectiveness of radiotherapy treatment in certain rectal cancer patients.
Commenting on the findings, Professor Tim Maughan, consortium lead, said:
‘These studies provide a deeper understanding of the biological factors that influence radiotherapy response in rectal cancer. They suggest new immunotherapeutic treatment strategies and highlight the potential for using digital pathology and gene expression models to predict patient outcomes more accurately.’
In addition to trialling immune checkpoint inhibitors, the team now plans to test the value of using image-based assessments of cancer subtypes in the clinic, to design new trials for patients needing radiotherapy for rectal cancer.