Key Takeaways
Caris Life Sciences is challenging traditional methods for selecting brain cancer therapy with a new AI-powered signature, validated on more than 5,800 patients, that better predicts who will benefit from the frontline chemotherapy temozolomide. The move positions the company's molecular profiling platform against established, but less precise, single-gene testing methods.
"The Caris AI Insights signature for GBM showcases Caris' advanced AI capabilities in our pursuit of improving cancer patient outcomes," Caris President David Spetzler said in a statement. "We believe that this signature can complement existing testing methods to improve clinical insight for glioblastoma patients treated with TMZ."
The study, published in the peer-reviewed journal Neuro-Oncology Advances, details an AI model trained to infer O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status from comprehensive genomic data. MGMT status is a key biomarker for predicting response to temozolomide (TMZ), the standard-of-care chemotherapy for glioblastoma (GBM), but conventional pyrosequencing tests can yield variable results. The Caris model was developed using a clinico-genomic dataset of 5,841 patients and further evaluated in a prospective cohort of more than 3,400 cases.
For Caris (NASDAQ: CAI), the validated signature provides a key differentiator for its MI Cancer Seek® profiling service in the competitive precision oncology market. By offering a more reliable prognostic tool for glioblastoma—a disease where nearly 50 percent of patients do not respond to initial treatment—the company aims to increase adoption among neuro-oncologists and strengthen its position in AI-driven diagnostics against broader genomics platforms from companies like Tempus and Foundation Medicine.
Glioblastoma is the most aggressive and common form of brain cancer, with a median survival of around 12 months. A critical step in treatment is determining whether a patient will respond to TMZ. The new AI signature demonstrated high concordance with pyrosequencing-based MGMT assessment and, crucially, improved the ability to separate patients into different survival outcome groups.
According to the study, the model stratified patients into distinct survival groups based on its signature score. Higher scores were associated with significantly longer overall survival in patients who received TMZ, providing a clearer prognostic picture than the binary result from traditional tests. This allows clinicians to make more informed decisions about pursuing TMZ therapy or considering alternative treatments and clinical trials for patients predicted to have a poor response. The AI signature is available upon request with Caris's MI Cancer Seek® assay, requiring no additional tissue from the patient.
This article is for informational purposes only and does not constitute investment advice.
