AI Blood Test Quickly Identifies Pancreatic Cancer Therapy Effectiveness, Steers Patients Away from Ineffective Treatments

Researchers at Johns Hopkins Kimmel Cancer Center have developed an innovative artificial intelligence (AI) blood test called ARTEMIS-DELFI, designed to quickly determine the effectiveness of pancreatic cancer treatments. This test analyzes cell-free DNA (cfDNA) fragments circulating in a patient’s blood to predict therapeutic responses, offering a significant improvement over traditional imaging methods. Pancreatic cancer is notorious for its rapid progression, often diagnosed at advanced stages where effective treatment is critical. Current monitoring methods, such as imaging, can be slow and less accurate, especially for patients on immunotherapy regimens, which complicate interpretation of results. Recognizing the urgent need for faster and more reliable indicators of treatment efficacy, the team led by Victor E. Velculescu, M.D., Ph.D., co-director of the cancer genetics and epigenetics program, explored alternative approaches. In the study, ARTEMIS-DELFI and another method, WGMAF (tumor-informed plasma whole-genome sequencing), were evaluated in blood samples from patients participating in two major clinical trials. The first trial, CheckPAC, focused on immunotherapy for pancreatic cancer, while the second, PACTO, examined various treatments. Both methods aimed to detect whether patients were benefiting from their therapies. However, ARTEMIS-DELFI emerged as the superior option because it does not require tumor biopsy samples and can analyze cfDNA directly from the patient’s blood, making it simpler and potentially more widely applicable. Testing and Validation During the initial tests, ARTEMIS-DELFI demonstrated its ability to identify treatment responses more accurately and earlier than imaging or other clinical markers. Specifically, it provided actionable insights within just two months of treatment initiation, a timeline that is crucial for patients with rapidly progressing cancer. The method uses machine learning to scan millions of cfDNA fragments, identifying patterns that indicate whether a therapy is effective. To validate the findings, the team conducted further trials with the PACTO study. Results showed that ARTEMIS-DELFI could identify responsive patients as early as four weeks after starting therapy. This early detection allows clinicians to make informed decisions about continuing or adjusting treatment plans, which is particularly beneficial in a disease where time is a critical factor. Advantages Over Existing Methods The primary advantage of ARTEMIS-DELFI is its non-invasive nature. UnlikeWGMAF, which relies on tumor biopsies that are not always available, ARTEMIS-DELFI works solely with blood samples. Tumor biopsies can be difficult to obtain and often contain a mix of normal and cancerous cells, leading to less accurate results. Additionally, ARTEMIS-DELFI is logistically simpler and likely less expensive, making it more accessible to a broader range of patients. Future Prospects and Broader Applications The next phase involves prospective studies to assess whether the information provided by ARTEMIS-DELFI can indeed help clinicians select the most effective therapy faster, thereby improving patient outcomes. The team is optimistic that this approach could also be adapted to monitor treatment responses in other types of cancers. Earlier this year, they published a study in Nature Communications showing that a related method, DELFI-TF, was effective in evaluating colon cancer therapy responses. "ARTEMIS-DELFI offers a real-time, personalized assessment of therapy response, enabling healthcare providers to tailor treatments and potentially save lives," Velculescu said. The method's broad applicability and ease of use make it a promising tool in the fight against aggressive cancers like pancreatic cancer. Industry Insights and Company Profiles Industry experts have praised the development of ARTEMIS-DELFI, noting its potential to revolutionize how pancreatic and other cancers are managed. By providing early and accurate feedback, this AI-driven blood test can help avoid wasting time on ineffective treatments and guide the selection of more appropriate therapies. Johns Hopkins Kimmel Cancer Center, a leading institution in cancer research and treatment, has a robust track record of developing innovative diagnostic tools. The involvement of Delfi Diagnostics, a company specializing in AI and genomics, further underscores the potential of translating this research into practical clinical applications. The collaboration between academic researchers and industry partners is crucial for bringing cutting-edge technologies to market and ultimately to the benefit of patients.