AI Blood Test Revolutionizes Treatment Monitoring for Pancreatic Cancer

An innovative blood test utilizing artificial intelligence is paving the way for improved treatment management in patients with pancreatic cancer. Developed by researchers at Johns Hopkins Kimmel Cancer Center, this test is designed to swiftly detect DNA fragments released by tumors into the bloodstream, which can help physicians gauge the effectiveness of ongoing cancer therapies.

The technique, referred to as ARTEMIS-DELFI, has been evaluated in blood samples from participants in two extensive clinical trials focused on pancreatic cancer treatments. Findings from these trials indicate that ARTEMIS-DELFI can effectively discern patient responses to therapy.

In comparative tests, ARTEMIS-DELFI, alongside another method known as WGMAF, demonstrated superior predictive capabilities regarding treatment outcomes compared to traditional imaging and existing clinical markers, especially two months following the initiation of treatment. However, ARTEMIS-DELFI emerged as the more favorable method due to its straightforward implementation and broader applicability.

Victor E. Velculescu, M.D., Ph.D., co-director of the cancer genetics and epigenetics program at the cancer center, emphasizes the critical nature of timely intervention in pancreatic cancer cases, as many patients often receive diagnoses at advanced stages where rapid progression is common. He advocates for the necessity of providing patients with diverse treatment options, particularly given the increasing availability of experimental therapies for pancreatic cancer.

Current clinical practices typically rely on imaging techniques to monitor the response to cancer treatments. However, these methods can yield results that are often delayed and may lack accuracy, particularly in patients undergoing immunotherapy, where treatment responses can be complex to interpret. In their research, Velculescu and his team explored two alternative strategies for assessing treatment efficacy in patients enrolled in the phase 2 CheckPAC trial of immunotherapy for pancreatic cancer.

One strategy, WGMAF, involves analyzing DNA from tumor biopsies alongside circulating cell-free DNA in blood samples to detect treatment responses. The other, ARTEMIS-DELFI, employs machine learning algorithms to analyze millions of cell-free DNA fragments exclusively from blood samples. Both methods successfully identified which patients were benefiting from their respective therapies. However, limitations arose with WGMAF, as not all patients had accessible tumor samples, and many samples contained only a minimal proportion of cancer cells amidst normal pancreatic tissue, complicating results.

ARTEMIS-DELFI proved effective across a wider patient demographic and was simpler to implement, as noted by Velculescu. Subsequent validation of ARTEMIS-DELFI's efficacy as a monitoring tool was achieved in the second clinical trial, known as the PACTO trial, which established its capability to determine patient responses as early as four weeks into treatment.

Lead study author Carolyn Hruban highlights the significance of the 'fast-fail' approach offered by ARTEMIS-DELFI, particularly for pancreatic cancer patients who may need to transition quickly to alternative therapies if initial treatments fail. This method is not only simpler and potentially more cost-effective than traditional tumor sampling techniques but also holds promise for wider applicability in various clinical settings.

The research team intends to conduct prospective studies to evaluate whether the insights gained from ARTEMIS-DELFI can facilitate clinicians in identifying effective therapies more efficiently and enhance patient outcomes. This methodology could also extend to monitoring other cancer types. Earlier this year, the team published findings in Nature Communications, demonstrating that a variation of the cell-free fragmentation monitoring technique, named DELFI-TF, was effective in assessing responses to colon cancer therapies.

Velculescu asserts that the analyses of cell-free DNA fragmentation provide real-time evaluations of patient therapy responses, enabling personalized care and improved patient outcomes.