Not all oncologists are ready to embrace genomic testing of patients' tumor samples, although others do plan to offer this type of testing to the majority of their patients, according to a new survey conducted at Dana-Farber Cancer Institute and Brigham and Women's Hospital (BWH) and available online ahead of print in the Journal of Clinical Oncology (doi: 10.1200/JCO.2013.52.4298).
This wide variation in attitudes was partly influenced by physicians' genomic confidence, the researchers said. “A lot of people talk about personalized medicine being able to help shape care in a positive way,” said Stacy W. Gray, MD, AM, the study's first author, who is Assistant Professor of Medicine at Harvard Medical School and Associate Physician in Oncology at Dana-Farber. A DFCI/BWH personalized medicine cancer research project called Profile enables oncologists to offer comprehensive multiplex tumor genetic testing to all of their patients, she said.
That project also gave study investigators the opportunity to find out how oncologists think about genetic testing—and “surprisingly, a significant fraction hadn't planned on using the test at all, even though it was available,” she said. At the other end of spectrum were physicians who were so excited about the screening tool that they planned to use it for most, if not all, of their patients. “We didn't expect to see that spectrum of response from physicians,” she said.
Notably, in their paper, Gray and her colleagues theorized that if genomic confidence is low in a sizable minority of oncologists at their academic research institution, then the “potential for confusion over test use and for missed testing opportunities might be greater in other settings.”
STACY W. GRAY, MD, AM
Asked for his opinion, Nikhil Khushalani, MD, Section Chief of Soft Tissue and Melanoma and Director of the IL-2 Program at Roswell Park Cancer Institute, said that overall, the survey results reinforce the high variability in physician understanding and implementation of genomic testing and results.
At most academic centers, oncologists tend to be subspecialized, honing their skills on one or two cancer subtypes, he continued. While a general sense of genomic understanding may exist for most oncologists in this setting, the degree of understanding varies by subspecialty. “I suspect this will be accentuated in a community practice setting where a general medical oncologist treats patients with different cancers,” he said.
The researchers surveyed 160 medical, surgical, and radiation oncologists from September 2011 through January 2012 to obtain a baseline assessment of their attitudes prior to the rollout of Profile, offering all cancer patients OncoMap, a test for 471 alterations in 41 cancer-related genes.
Physicians were asked about their current use of somatic testing, their attitudes about multiplex testing, and their confidence in their ability to understand and use genomic data. The survey did not include a direct test of the physicians' knowledge.
Respondents said they ordered tumor genomic testing on an average of 24 percent of patients before the introduction of Profile. Twenty-two percent of oncologists reported low confidence in their knowledge about genomics. The average score on a genomic confidence scale was 3, which indicated a response of “somewhat confident.”
Fourteen percent said they lacked knowledge to explain genomic concepts to patients, and 26 percent said they doubted their ability to make treatment recommendations based on genomic data.
Forty-two percent of respondents approved of telling patients about test results even when their significance for the patient's outlook and treatment is uncertain.
The investigators also found that 18 percent of the oncologists planned to use multiplex tumor testing in 10 percent or less of patients, while 25 percent of physicians planned on testing 95 percent or more of individuals.
Higher genomic confidence was associated with wanting to test a majority of patients (adjusted OR 6.09) and anticipating using actionable (adjusted OR of 2.46) or potentially actionable (adjusted OR of 2.89) test results to inform treatment recommendations. Confidence was also associated with being a medical oncologist or a researcher.
NIKHIL KHUSHALANI, MD
Specialization at Dana-Farber is one possible explanation for why confidence was not uniformly very high, Gray speculated. Some oncologists probably use genetic testing very frequently, as is the case in the center's lung cancer practice. However, there are other malignancies such as early-stage bladder cancer where genetic testing is not routinely used to make decisions about treatment—“So, not all doctors are familiar with genomic testing,” she said.
Moreover, the survey questioned surgical and radiation oncologists, who are less likely to think about genomics in day-to-day practice, said Gray.
Other prior studies have found that general physicians' confidence in their ability to work with genomic data or genetic knowledge is lower than would be expected, she added. Consequently, the finding about oncologist confidence influencing their use of genomic testing was not entirely unexpected, she said.
In the community setting, oncologists want to distill test results down into clinically meaningful decisions about treatment, commented Steven Powell, MD, Assistant Scientist at the Cancer Biology-Sanford Research Center and Assistant Professor of Internal Medicine at the University of South Dakota Sanford School of Medicine. He said that based on feedback provided during tumor conferences at the Sanford Cancer Center, some physicians are strongly in favor of genomic testing and comfortable with it, while others are not.
Brian Lane, MD, PhD, a urologic oncologist at Spectrum Health in Grand Rapids, Michigan, said that the comfort level there is high, with multiple providers using genomic testing. “We have had a number of clinical trials open over the last seven years, being one of the first medical centers to offer such testing,” he said.
Morris D. Groves, MD, head of the Texas Oncology–Austin Brain Tumor Center and Chair of Neuro-Oncology Research for the US Oncology Network, describes the level of comfort with genomic testing at his center as above average. Moreover, when physicians get a test result back with a lot of data, “they'll spend the time necessary to understand it, read through the report, and do some background work on the function of particular genes,” he said.
At the Greenville Health System's (GHS) Cancer Institute in South Carolina, oncologists probably conduct a similar amount of genomic testing as what was described in the Dana-Farber paper, said Jeff Edenfield, MD, Medical Director of the GHS Institute for Translational Oncology Research. “We're somewhere between a community and an academic setting,” he noted, adding that the OncoMap used in the study is similar to PrecisionPath, which the GHS Cancer Institute offers to every patient who has surgery.
Within the academic setting of Oregon Health & Science University's (OHSU) Knight Cancer Institute, the individuals most plugged into genomic testing are hematologists because their understanding and treatment of various forms of leukemia are more molecularly based than in the past, commented Christopher Corless, MD, PhD, OHSU Professor of Pathology. Physicians who treat solid tumors generally vary in their comfort with molecular testing, depending on how much the treatment of one particular cancer is influenced by genetic markers versus another, he said.
Overall, oncologists in the academic setting are a little more tuned into genomic testing than those in the community setting are, noted Corless, who in his role as Executive Director and Chief Medical Officer of Knight Diagnostic Laboratories, interacts with a dozen community practices nationwide.
The comfort level at centers such as Roswell Park that have genomics programs is likely to be significantly higher than at other centers or among community providers, Khushalani agreed.
However, said Corless, within the community, some individuals are early adopters of genomic technology, and notably, those involved in clinical trials tend to understand the need for molecular markers that might be relevant to their patients.
Khushalani emphasized that in this rapidly changing landscape where eventual widespread use is inevitable, education of providers is paramount, but this is not something that is routinely incorporated into medical training curricula.
“More research is needed to help us understand where physicians have knowledge gaps and how they prefer to learn about genetic tests,” Gray said. For example, some oncologists may prefer a continuing medical education course through societies such as ASCO, while others may want the information integrated into the medical record so that they can understand the test results of particular patients.
Edenfield and his colleagues have written a proposal about providing a genomic testing translational medicine curriculum in medical school that continues throughout training and into the clinical practice setting.
At Sanford Health, Powell noted, physicians are developing collaborative research programs with companies and institutions that offer genomic testing and that provide comprehensive interpretation of results, which are then evaluated by a hospital's tumor board to then help guide treatment decisions.
And at OHSU Knight Cancer Institute, Corless said, the molecular tumor board meets about once a month to review cases and discuss potential treatment options based on genomic data.
Genetic counselors are another useful resource for oncologists because they are aware of the way genetic testing is changing and of inherited risk factors for cancer that don't commonly come to provider's attention, said Zohra Ali-Khan Catts, MS, LCGC, Director of Cancer Genetic Counseling at Christiana Care Health System's Helen F. Graham Cancer Center & Research Institute, in Newark, Delaware.
Some multiplex screenings include genes that have been inherited, and in those cases, utilizing a genetic counselor is especially important because if inherited mutations are found, the screening recommendations for the patient's family changes, she explained.
Also, Gray noted, the anticipated shortage of genetic counselors—not just in oncology but also in other specialties—may make keeping up with technology a challenge.
Standardizing best practices for genomic testing and evaluating evidence in the patient's best interest is an exciting and challenging charge for oncology societies, she added.
“The development of guidelines to shepherd the use of broad-based genomic testing will be a welcome event for all involved, and some guidelines already exist,” Edenfield said. “For example, strong evidence exists that anti-EGFR therapy is largely ineffective for colon cancers that harbor a KRAS mutation.”
Gray noted that research is now coming out at such a rapid pace that ASCO and the CDC's Evaluation of Genomic Applications in Practice and Prevention have made a concerted effort to make recommendations about new technologies—“but it's often hard to keep up.”
Another challenge, Corless said, is that oncologists currently, “have dribs and drabs of clinical data, some inconsistent, and much of it outside of the clinical trial setting. It's hard to develop treatment guidelines based on such soft data.”
Moreover, Lane added, developing guidelines based on a significant number of patients followed for a sufficient length of time will take years, and in the meantime, oncologists will continue to work in an environment with low-quality evidence to support the use of targeted molecular agents.