Skip to main content
Premium Trial:

Request an Annual Quote

Precision Medicine in Oncology Gets More Support With 13K-Patient Meta-Analysis


NEW YORK (GenomeWeb) – A meta-analysis of more than 340 Phase I studies involving approximately 13,200 refractory cancer patients has shown that those who received biomarker-guided treatments fared much better than those who didn't, researchers reported today.

"A biomarker-based personalized approach, which is the foundation of precision medicine, improved outcomes even in a Phase I refractory cancer population," said lead author Maria Schwaederle from the University of California, San Diego School of Medicine, previewing data that will be presented at the American Society of Clinical Oncology's annual meeting in June.

"There have been breathtaking advances in our ability to perform genomics that helps us understand the underlying biology of tumors," she said during a call hosted by ASCO. "However, the usefulness of a biomarker-driven approach for treatment selection is still a matter of debate in the community."

ASCO spokesperson Don Dizon reflected that analysis by Schwaederle's team further supports the fact that precision medicine is not only the future of cancer care, but it is the present. Around the country, cancer centers are increasingly using next-generation sequencing panels to test patients for genomic markers to personalize treatments, while advocacy organizations are pushing for comprehensive cancer profiling in lung and pancreatic cacner.

However, there are critics in the life sciences field who say there is not enough evidence yet showing that biomarker-driven approaches should become the standard in cancer, particularly since many times using these strategies patients are put on off-label drugs that haven't been studied in a patient's cancer.

The large study — a collaboration between researchers at UCSD, MD Anderson Cancer Center, the Worldwide Innovative Network for Personalized Cancer Therapy, and ASCO — was an attempt at providing further evidence for precision medicine in oncology. While this was a retrospective meta-analysis, other studies, such as NCI-MATCH and TAPUR, are testing the precision medicine hypothesis prospectively. 

Such studies have their own challenges, however, since the biomarkers of interest are rare and advanced cancer patients can't always submit tumor tissue samples for genomic analysis. For example, basket trials are one way researchers are prospectively assessing their precision oncology hypotheses, and one such 647-patient trial in advanced thoracic malignancies, called CUSTOM, was "not feasible" due to the rarity of the mutations being studied and the lack of an adaptive design that would have allowed more flexibility. 

Moreover, with the publication of the first basket study last year, Memorial Sloan Kettering Cancer Center researchers showed tumor histology is still important in molecular medicine. Looking specifically at patients with a range of cancers with BRAF V600 mutations, researchers led by MSK's Jose Baselga and David Hyman reported that certain tumor types but not all appear to respond to the BRAF inhibitor Zelboraf (vemurafenib). 

For the most recent analysis, researchers reviewed data from 346 Phase I trials published between 2011 and 2013 studying single agents. Within these studies, there were 58 arms where patients received drugs based on biomarkers, and 293 arms where patients got treatments not based on this approach. Schwaederle and her team considered trials where at least response rates were reported, but they couldn't look at overall survival because not enough trials reported this.

In treatment arms employing precision medicine approaches, 31 percent of patients saw their tumors shrink, while around 5 percent of patients responded with non-personalized approaches. Median progression-free survival was 5.7 months for those receiving precision cancer care, versus just under 3 months for those not getting such care.

Phase I studies are often used to investigate the safety and tolerability of a drug, but the data from this analysis shows that it can also provide important information about drug efficacy. "Our analysis really shows that this [view of Phase I trials] is completely outdated," Schwaederle said. "With biomarker selection and especially with genomic biomarkers, we can reach high response rates even in Phase I trials."

Patients' response rates were similar when Schwaederle and colleagues homed in on 57 trials of drugs that targeted genes or proteins implicated in cancer. However, Schwaederle highlighted that when patients received treatments based on genomic biomarkers they experienced greater tumor shrinkage than when therapy was given based on protein biomarkers — 42 percent versus 22 percent.

Schwaederle's team also made sure to account for the type of therapy influencing outcomes. "One question that is often asked is, is this simply about having a better therapy?" she posited.

In their analysis, 98 percent of personalized arms used targeted therapies, but 76 percent of arms using targeted drugs didn't employ a precision approach. When researchers compared arms involving non-personalized targeted drugs against cytotoxic agents, patient outcomes were similar — around 5 percent response rates and around 3 months median progression-free survival. "Our analysis shows that it is not just that the therapies are better but that targeted therapies must be given to the right patients," Schwaederle said.

Their analysis "revealed that a personalized strategy was an independent predictor of improved outcomes" in terms of response rate and progression-free survival, the authors concluded in an abstract. "In addition, studies that used targeted agents without a biomarker-based selection strategy had negligible response rates."

One potential source of bias in this analysis is that researchers depended on published studies and could not factor in data from Phase I studies that have not been published, which may contain, for example, more information on patients with poor outcomes on a precision treatment arm, compared to a non-personalized arm.