Precision medicine clinical decision support system bests oncologists at picking cancer treatment

A prominent academic medical center, a supercomputer and a California start-up came together to determine the best regimen for a patient after doctors had tried four previous times
By Bill Siwicki
03:42 PM
Share

UCSD's supercomputer center. Credit: UCSD

The University of California San Diego Center for Personalized Cancer Therapy, working in conjunction with startup CureMatch and UCSD’s Supercomputer Center, claims that a precision medicine clinical decision support system provided treatment guidance for a patient after doctors were unsuccessful.

A 57-year old man, in fact, had undergone four previous lines of treatment. Despite surgery for several basal cell carcinomas, he developed brain, bone and liver metastases and was subsequently treated with two separate targeted therapies that are FDA-approved for metastatic basal cell carcinomas, but was non-responsive to these therapies. He also received Gamma knife radiation for brain metastases with good response, and chemotherapy was administered but that response was short-lived.

[Also: The next precision medicine breakthrough is about apps, not EHRs]

That’s when the UCSD divisions and CureMatch came together, according to Razelle Kurzrock, MD, CureMatch co-founder and chief of the division of hematology and oncology at the UCSD School of Medicine. She also is senior deputy director at the Center for Personalized Cancer Therapy.

“A liver biopsy was taken, and next-generation DNA sequencing was performed by Foundation Medicine, a FoundationOne 315 gene panel,” she said.

Those molecular diagnostic test results were then entered into the CureMatch platform, called nGENE, which was initiated at the UCSD Moores Cancer Center in collaboration with the UCSD Supercomputer Center. The system, in turn, generated a quality-controlled PreciGENE report ranking immunotherapy treatment options to guide clinicians.

CureMatch’s scoring system analyzed the individual patient’s molecular tumor profile and estimated the best therapeutic agents using proprietary algorithms and manually-curated databases, she added.

That personalized approach is when the new immunotherapy regimen began.

“After two months, scans showed remarkable partial remission with no toxicity noted,” Kurzrock said. “After 18 months the patient is in nearly complete remission.”

CureMatch recommended single-drug immunotherapy, the oncologist backed the system’s choice, and after four prior lines of treatment had failed, the patient is making considerable progress.

Kurzrock said that the algorithm recommending the successful treatment option was jointly developed by UCSD’s Cancer and Supercomputer centers.

“We trained the algorithm based on experience with genomic results in about 3,000 patients,” she added. “A subgroup of these patients received combination treatments and our analysis so far indicates that they have better outcomes.”

They also tested the algorithm on data available in the literature on super responders, about 70 patients so far.

“This was important to us because our goal is to achieve super responses,” she said. “So far, without exception, the algorithm correctly ranked the drugs that induced super responses with very high scores.”

That said, Kurzrock was both careful to point out that molecular profiling for individual cancer patients remains a novel approach and bullish that in the near future it will become a standard of care.

Twitter: @SiwickiHealthIT
Email the writer: bill.siwicki@himssmedia.com