PharmaResearch

Caris AI Study Links Treatment Selection to Longer Survival in Pancreatic Cancer

A Caris study found AI-guided treatment selection was associated with longer survival in pancreatic cancer patients using molecular data.

Answer
Caris AI Study Links Treatment Selection to Longer Survival in Pancreatic Cancer

The brief

  1. Utilize AI-guided treatment selection to enhance survival rates in pancreatic cancer patients.
  2. Identify patients likely to benefit from FOLFIRINOX over gemcitabine plus nab-paclitaxel with molecular data.
  3. Achieve a median overall survival of 16.0 months with FOLFIRINOX versus 9.9 months with gem/nab-p.

Deployment

Company
Caris Life Sciences
System
AI-based molecular signature
median overall survival (months)
16.0

Caris Life Sciences has published a study showing that its artificial intelligence (AI)-based molecular signature could help identify which patients with advanced pancreatic cancer may benefit more from specific first-line chemotherapy regimens.

The brief

  1. Utilize AI-guided treatment selection to enhance survival rates in pancreatic cancer patients.
  2. Identify patients likely to benefit from FOLFIRINOX over gemcitabine plus nab-paclitaxel with molecular data.
  3. Achieve a median overall survival of 16.0 months with FOLFIRINOX versus 9.9 months with gem/nab-p.

Deployment

Company
Caris Life Sciences
System
AI-based molecular signature
median overall survival (months)
16.0

The study, published in npj Precision Oncology, evaluated an AI approach designed to guide treatment selection between FOLFIRINOX and gemcitabine plus nab-paclitaxel (gem/nab-p). The research used molecular and clinical data to identify patterns associated with treatment benefit.

In the testing cohort, patients classified as standard molecular risk and predicted by the model to benefit from FOLFIRINOX had a median overall survival of 16.0 months when treated with FOLFIRINOX, compared with 9.9 months among those who received gem/nab-p. Caris reported a hazard ratio for death of 0.48 for the comparison.

The findings build on earlier work examining how molecular data can influence the choice between pancreatic cancer treatment regimens.

The model also identified a group of patients who could potentially receive similar or better outcomes with gem/nab-p. Among patients with the highest model scores for gem/nab-p, median overall survival was 13.3 months for those who received gem/nab-p compared with 8.3 months for those treated with FOLFIRINOX.

Approximately half of the patients in the study received a different first-line treatment from the one the model would have recommended, according to Caris. The company said this points to an opportunity to use tumor biology more directly when selecting treatment.

Molecular Data Adds Another Layer to Treatment Decisions

For advanced pancreatic ductal adenocarcinoma (PDAC), commonly used first-line options include FOLFIRINOX, gem/nab-p and NALIRIFOX. The Caris study focused on the first two regimens, which can extend survival but differ in treatment intensity and toxicity.

Rather than relying on a single biomarker, the Caris model uses machine learning to assess complex molecular patterns. The analysis incorporates Whole Exome Sequencing (WES), which examines DNA changes, and Whole Transcriptome Sequencing (WTS), which measures gene activity.

The approach builds on the broader use of molecular data to guide pancreatic cancer treatment decisions.

Caris said its validation study drew on a large clinico-genomic dataset connecting molecular profiles with real-world treatment outcomes. The resulting report provides molecular risk stratification and treatment guidance between FOLFIRINOX and gem/nab-p.

However, the study has limitations. It was based on retrospective real-world data, and the researchers noted that some clinical variables, including performance status, disease burden and CA19-9 levels, were not available for all patients. The researchers also said independent-cohort validation will be required to establish the model's broader applicability.

The model predicts treatment benefit rather than treatment toxicity, meaning physicians would still need to consider a patient's clinical condition and the adverse effects associated with each regimen.

The AI signature is included in Caris' Molecular Tumor Board Report and is built from WES and WTS data. The underlying MI Cancer Seek assay received FDA approval in November 2024 for tumor mutational profiling and specified companion diagnostic uses. The FDA approval covers the assay and its authorized indications, rather than establishing that the pancreatic cancer AI treatment-selection model itself is FDA-approved.

Written by Mukundan Sivaraj

Join the discussion

0 answers

People who run these systems answer here. Your verified role is shown, never your employer unless you choose.

Sign in with LinkedIn to answer

Read next

MachineCon USA · from AIM Media House

Where enterprise AI is decided

An invitation-led day for the CIOs, CDOs and Chief AI Officers who own enterprise AI strategy.

  • New York5th edition · 583 Park Avenue
  • Dallas1st edition

Register for MachineCon USASee MachineCon USA →

  • Keynote on the main stage, MachineCon New York 2025
  • AI150 award winners on stage, MachineCon New York 2025
  • At the tables, MachineCon New York 2025
  • Networking, MachineCon New York 2025