A team at the University of Leeds collaborated with researchers at Roche Diagnostics to develop the technique, which could help doctors and patients to decide on the best treatment options.
They used samples from a previous trial funded by Cancer Research UK to look at the levels of two proteins, known as tumor amphiregulin (AREG) and epiregulin (EREG), which are produced by some colorectal cancers.
Algorithms driven by AI enabled the researchers to show that patients with higher levels of these proteins received significant benefit from a treatment which inhibits a different protein involved in cancer cell growth, known as EGFR. Of equal importance, patients with low levels of the proteins did not benefit from the treatment.
Currently, anti-EGFR treatments are only given to patients with advanced, incurable bowel cancers. The researchers hope their methodology could be used in the future to identify patients in the earlier stages of illness who could also benefit from the drugs.
The study
The study was funded by Innovate UK and Roche Diagnostics as well as Yorkshire Cancer Research. It was part of a program of work in this field being conducted by the National Pathology Imaging Co-operative.
The team relied on artificial intelligence-assisted immunohistochemical (IHC) evaluation of AREG and EREG expression as a combined predictive biomarker for panitumumab (Pan) therapy benefit in RAS wild-type (wt) metastatic colorectal cancer (mCRC).
According to their study, high tumor mRNA levels of the EGFR ligands, AREG and EREG are associated with anti-EGFR agent response in patients with RAS-wt mCRC, regardless of tumor location.
However, they noted that ligand RNA assays have not been adopted into routine clinical practice due to issues with analytical precision and practicality.
So they decided to test whether AREG and EREG expression assessed by IHC can predict benefit from Pan, a treatment for colorectal cancer that has spread. Pan is used to treat colon cancers that express EGFR and disease that has got worse either on or following fluoropyrimidine, oxaliplatin and irinotecan containing chemotherapy regimens.
The team relied on samples from a retrospective biomarker study within the PICCOLO trial [NCT00389870; irinotecan [Ir] ± Pan in fluoropyrimidine-resistant RAS-wt mCRC].
AREG and EREG positive tumor cells were assessed by IHC in all RAS-wt patients with available tumor tissue, they said.
Pathologists annotated tumor areas on digital images of glass slides.
AI algorithms were used to calculate the percentage of tumor cells staining positive for AREG and EREG within the tumor areas. More than 50% AREG and/or EREG tumor cell positivity was regarded as high ligand expression.
The primary endpoint was progression-free survival (PFS) and secondary endpoints were RECIST response rate (RR) and overall survival (OS).
In terms of findings the team saw that high ligand expression was associated with significant PFS benefit from IrPan compared with Ir (8 vs 3.2 months) whereas low ligand was not (3.4 vs 4.4 months).
The ligand-treatment interaction was significant and independent of mutation status and primary tumor location.
Likewise RR was significantly improved in patients with high ligand expression but not in those with low ligand expression. Lesser effect was seen on OS.
They concluded that IHC assessment of AREG and EREG identified patients who did or did not benefit from Pan, as has been previously demonstrated through mRNA quantification.
“IHC represents a more practicable technique as it can be provided at the point of care and is associated with shorter turn-around times. AREG and EREG IHC may be of use in routine practice to identify patients who would benefit from anti-EGFR therapy and those for whom alternative treatment strategies should be explored.”
Picking the right treatment option
Lead author of the report, Christopher Williams, from Leeds University's Division of Pathology and Data Analytics, said as more treatment options become available for advanced colorectal cancer, it is becoming increasingly difficult for patients and their doctors to choose the treatment that's right for them. “This test will help patients navigate this decision-making process more easily."
Though a spokesperson for the University of Leeds told BioPharma-Reporter that as the research team has just developed the algorithm, there is no prospect of the technique being applied widely yet.
The report's senior author, Kandavel Shanmugam, who is a senior director of medical innovation at Roche Diagnostics, said: "As increasing numbers of complex tests are developed to target the right cancer treatments to the right patients, developing streamlined methods for delivering test results will be essential to improve cancer care.
"By using artificial intelligence to semi-automate the test process, we anticipate it may be easier for results to be delivered to patients faster to better influence treatment decisions."
