Aiming to outsmart cancer

 

The MI FOLFOXai technology is designed to evaluate the likelihood of positive patient response to FOLFOX chemotherapy as a first-line treatment in combination with bevacizumab. The value of the diagnostic insights provided by MI FOLFOXai is relevant to the more than 1.8 million patients diagnosed with metastatic colorectal cancer each year.

 

“We are excited to bring the world’s first fully validated AI-based treatment algorithm to patients suffering from CRC,” says David Spetzler, president and chief scientific officer of Caris. “MI FOLFOXai will help physicians better determine the relative benefit patients will receive from the FOLFOX chemotherapy regimen and improve clinical outcomes.

 

“Precision medicine powered by AI has enormous potential in cancer profiling, as well as other areas of medicine, and we look forward to continuing to release additional AI-based molecular intelligence products in the near future.”

 

Studies show as many as 25 percent of colorectal cancer patients will present with stage IV disease, when metastatic cancer has spread to other parts of the body.

 

MI FOLFOXai was developed using a subset of results from the company’s proprietary Caris Molecular Intelligence platform, which includes next generation sequencing for DNA mutations, copy number alterations and insertions/deletions; whole transcriptome sequencing for RNA fusions and variant transcripts; and protein testing via immunochemistry.

 

According to Spetzler, “We have a team of data scientists, oncologists, pathologists and biologists that worked together to discover and validate the artificial intelligence system that we used to find the signature. We have molecular data on over 200,000 patients and real-world outcome evidence on 167,000 patients. We used these data to discover and validate the results, as well as performing a second blinded validation using samples from a clinical trial.”

 

MI FOLFOXai was validated using two independent data sets to compare the increased benefit arm to the decreased benefit arm. The first study was a blinded, prospective analysis from retrospectively tested samples from the randomized Phase 3 TRIBE2 study. This study showed a median overall survival improvement of 6.9 months. The second study involved several hundred cases with real-world evidence that showed a median overall survival increase of 11.8 months.

 

“There is a 50-percent difference in survival between the decreased benefit group and the increased benefit group for patients treated with frontline FOLFOX,” Spetzler reports. “The opposite trend is observed for patients treated with frontline FOLFIRI. All FOLFOX-treated patients received bevacizumab and most patients who received FOLFIRI did, so the differences are most likely not related to bevacizumab.

 

“Short term, we aim to provide physicians with information to help them choose between FOLFOX and FOLFIRI in the front-line setting (both are standard-of-care options),” he says. “Long term, we aim to continue to learn about the molecular signatures that drive sensitivity and resistance to therapies to help develop new therapeutic options for patients.”

 

“FOLFOX is part of an NCCN Guidelines recommended first-line treatment regimen for the roughly 35,000 patients in the U.S. newly diagnosed with metastatic colorectal cancer each year, and initiating the most appropriate therapy at this stage can be critical to successfully altering the course of disease,” reported Dr. W. Michael Korn, chief medical officer of Caris.

 

Since clinical outcomes from MI FOLFOXai validations represent a mean overall survival of nearly 50 percent, it is imperative to get this information to patients and their physicians as soon as possible, Korn added, noting that MI FOLFOXai represents Caris’ second AI-driven molecular profiling offering. In December 2019, the company launched MI GPSai, the first-ever molecular AI product used for cancer assessment in a clinical setting.

 

As an AI-driven tumor type biology similarity score to better classify tumors, MI GPSai uses both DNA and RNA analysis along with a machine-learning algorithm, which includes more than 6,500 mathematical models, to compare molecular characteristics of a patient’s tumor against Caris’ extensive database. The results provide insights to inform treatment decisions for cancer of unknown primary cases, atypical clinical presentation cases and other difficult-to-treat cancer cases.