Boston, MA—Incorporating personalized medicine into everyday oncology clinical practice will require new paradigms in an effort to match patients with cancer with the appropriate therapies, as well as attempts to treat solid tumors at an earlier stage with targeted agents, said Razelle Kurzrock, MD, Director, Center for Personalized Cancer Therapy, University of California, San Diego (UCSD) Moores Cancer Center, at the Second Global Biomarkers Consortium annual conference.
“In a very real way, it’s very hard to institute personalized medicine both in clinical trials and in the community,” Dr Kurzrock said.
A practical problem to implementing personalized medicine is that common cancers are difficult to treat, especially in the metastatic setting, because each tumor may embody more than 100 different diseases, and multiple subtypes of each tumor exist. “Even if some of these tumors have things in common, the individual landscape of each patient may be very distinct,” said Dr Kurzrock.
Molecular testing of tumors can make a difference in early clinical trials. “The old way of thinking was that we needed new drugs to treat patients,” she said. “We do need new drugs, no question about it, but very fundamentally, it doesn’t matter how good your drug is: if you don’t give it to the right set of patients, it’s not going to work. Matching the drugs with the right patient is crucial to the future of oncology, and maybe to all of medicine.”
Although the tools to match patients with target-based agents are becoming available, “in a practical way, this is hard to do.”
Even in early-phase clinical trials, matching patients to targeted agents can produce high response rates, even in difficult-to-treat populations.
The Profile-Related Evidence Determining Individualized Cancer Therapy (PREDICT) study used a histology-independent targeted approach in which multiple molecular aberrations were assessed and were used to match patients with targeted agents. When the study started in 2008, molecular profiling was primitive, and single-gene assays were used to identify 12 genetic mutations. Conventional therapy had failed in all 1144 patients who were enrolled in PREDICT.
Molecular aberrations were found in up to 73% of patients, and did not segregate by tumor histology. Most of the molecular aberrations that were identified were “actionable.”
Patients for whom a matched therapy was available in a phase 1 setting had a complete or partial response rate of 27%, whereas patients who did not have a matching therapy had a response rate of only 5%.
Progression-free survival improved with the phase 1 matched therapy but not with the unmatched therapy compared with previous conventional therapy. Overall survival improved with the matched treatment (RAF/MEK inhibitor) compared with the unmatched treatment.
Another practical problem with personalized medicine in oncology is that patients with metastatic disease usually relapse. Chronic myelogenous leukemia (CML) is a fatal disease that has been transformed with imatinib (Gleevec) therapy, with median survival extended to 20 to 25 years, but such success is rare with solid tumors. The conventional wisdom is that solid tumors are more complex than CML, and elucidating combinations of therapies to treat solid tumors is a herculean task because of this complexity, Dr Kurzrock said.
The conventional wisdom, however, does not account for the precipitous decline in response rates with advancing disease in patients with CML. In the blast crisis, median survival is only approximately 12 months, and patients develop resistance. “That sounds a little bit like solid tumors to me,” said Dr Kurzrock.
The transformation in CML outcomes was made possible by treating newly diagnosed disease. “We have not done that in solid tumors,” Dr Kurzrock said. “Comparing metastatic solid tumors to newly diagnosed CML is not an apt comparison.” A practical solution to enhancing outcomes with personalized medicine in solid tumors, therefore, is to treat newly diagnosed patients.
The future of personalized medicine in oncology is actionable cancer gene sequencing, but a practical problem is host individual differences. “The host reaction may be critical to how the tumor responds and also critical to toxicity,” Dr Kurzrock pointed out.
Tumor microheterogeneity is another practical problem. The molecular profile can differ even within a single lesion. A potential solution is liquid biopsy to obtain cell-free or circulating DNA from patients with tumors, and performing gene sequencing on this DNA. The technology is moving rapidly.
A lack of training in genomics is a barrier to personalized cancer therapy trials and treatment. “In the future, we’ll have to train oncologists in genomics,” Dr Kurzrock noted.
Another challenge is proving the concept of personalized medicine. To this end, the Worldwide Innovative Networking in Personalized Cancer Medicine trial has been initiated as a signature global trial with genomics and transcriptomics used to navigate patients with end-stage cancer to clinical trials.
