Dana-Farber Oncologists Differ Widely on the Use of Multiplex Tumor Genomic Testing

A new study by researchers at the Dana-Farber Cancer Institute suggests that not all doctors are ready to embrace tests that may identify hundreds of genomic changes in a patient’s tumor sample for the purpose of determining appropriate treatment.

Many cancer researchers believe that cutting-edge advances in genomics will pave the way for personalized or “precision” cancer medicine for all patients in the near future. A new study by researchers at the Dana-Farber Cancer Institute, however, suggest that not all doctors are ready to embrace tests that look for hundreds of genomic changes in a patient’s tumor sample, while others plan to offer this type of cancer genomic tumor testing to most of their patients. The study findings were published recently in the Journal of Clinical Oncology [1], along with an accompanying editorial. [2]

The wide variation in attitudes was in part determined by physicians’ “genomic confidence.” Physicians who had a lot of confidence in their ability to use and explain genomic findings were more likely to want to prescribe the test and consider using test results when making treatment recommendations. Other physicians had lower levels of genomic confidence and were more reluctant to offer such testing. These findings are particularly interesting because the survey was carried out at the Dana-Farber/Brigham and Women’s Cancer Center (DF/BWCC), which has a comprehensive research program. The DF/BWCC research program allows all consenting patients to have genomic tumor testing, which is capable of finding gene mutations and other DNA alternations that drive a patient’s cancer. In some cases, the genomic tumor profiles identify “druggable” targets that may allow doctors to use specific drugs known to be effective against particular gene mutations or alterations.

The researchers were perplexed by another key study survey finding: 42 percent of responding oncologists approved of telling patients about genomic tumor test results even when their significance for the patient’s outlook and treatment is uncertain. This issue comes with the growing use of predictive multiplex genomic testing, which can identify tens or hundreds of gene mutations simultaneously and often detects rare DNA variants that may or may not be relevant to the treatment of an individual’s cancer.

“Some oncologists said we shouldn’t return these results to the patient, and others say ‘of course we should give them to the patient’,” said Stacy W. Gray, M.D., AM, of Dana-Farber, first author of the report. “I think the fact that we found so much variation in physicians’ confidence about their ability to use genetic data at a tertiary care, National Cancer Institute-designated Comprehensive Cancer Center makes us pause and wonder about how confident physicians in the community are about dealing with this,” she said. “It begs the question at a national level, how are we going to make sure that this technology for cancer care is adequately delivered?”

The study survey was conducted in 2011 and early 2012 as a baseline assessment of physicians’ attitudes prior to the rollout of the genomic tumor testing project referred to as “Profile” (which formerly utilized a technology platform called “OncoMap“) at DF/BWCC.

For purposes of the study, a total of 160 Dana-Farber adult cancer physicians – including medical oncologists (43%), surgeons (29%), and radiation oncologists (19%) – participated in the survey. They were asked about their current use of multiplex tumor genomic testing, their attitudes about multiplex testing, and their confidence in the ability to understand and use genomic data. The survey did not include a direct test of the physicians’ knowledge.

Among the many intriguing findings of this study, a wide variability in interest in multiplex tumor genomic testing was identified—25% of respondents anticipated testing more than 90% of their patients, whereas 17% of respondents anticipated testing 10% or less. Beliefs related to the potential value of multiplex tumor genomic testing were largely positive; most expressed belief that this form of testing would increase treatment (73%) and research options (90%) for patients, as well as both physician (80%) and patient satisfaction (80%).

Despite the foregoing, less than 50% of the physicians planned to view the multiplex tumor genomic testing results routinely. Moreover, the majority of respondents planned only to “rarely” or “sometimes” use the clinically relevant results (58%), called “Tier 1” by the study authors, and potentially actionable results (88%), called “Tier 2,” to assist them in the treatment of patients. However, the respondents more often indicated that results of multiplex tumor genomic tests should be shared with patients, particularly findings revealing the presence of a Tier 1 (clinically relevant) genomic variant—87% believed that these findings should be discussed—versus a Tier 2 (potentially actionable) genomic variant (50%), or a Tier 3 (uncertain significance) genomic variant (40%). A substantial minority (39%) also disagreed with a Dana-Farber Cancer Institute policy prohibiting the disclosure of Tier 3 genomic variants to patients.

Interestingly, despite limited exposure to routine genomic tests for a large portion of the respondents, the stated “genomic confidence” of participating physicians was quite high. The majority of participants reported that they were “somewhat” or “very” confident in their (i) knowledge of genomics (78%), (ii) ability to explain genomics (86%), and (iii) ability to use genomic results to guide treatment (74%); however, a substantial minority of the Dana-Farber physicians (28%) reported genomic confidence of “not very” or “not at all confident.”

Based upon the study survey findings, Dr. Gray and her colleagues conclude that there is “little consensus” on how physicians plan to use multiplex tumor genomic testing for personalized cancer care, and they suggest the need for evidence-based guidelines to help doctors determine when testing is indicated.

“I think one of the strengths of this study is that its information comes from an institution where ‘precision cancer medicine’ is available to everyone,” commented Barrett Rollins, M.D., Ph.D., Dana-Farber’s Chief Scientific Officer and a co-author of the paper. “It highlights the fact there’s a lot of work to be done before this can be considered a standard approach in oncology.”

The senior author of the study is Jane Weeks, M.D., MSc, of Dana-Farber; additional authors include Angel Cronin, MS, of Dana-Farber and Katherine Hicks-Courant, BA, of the University of Massachusetts Medical School.

The research was supported by the Dana-Farber Cancer Institute. Dr. Gray also receives support from the American Cancer Society (120529-MRSG-11-006-01-CPPB) and the National Human Genome Research Institute (U01HG006492)

Pursuant to a new phase of Profile, initiated by Dana-Farber in 2013, a more advanced technology platform (called “OncoPanel“) utilizes “massively parallel” or “next-generation” sequencing to read the genetic code of approximately 300 genes in each patient’s tumor sample. “Massively parallel” refers to the technology’s capacity for sequencing large numbers of genes simultaneously. The 300 genes evaluated in connection with the OncoPanel were chosen because they have been implicated in a variety of cancers.

In addition to the complete DNA sequencing of more than 300 genomic regions to detect known and unknown cancer-related mutations, the OncoPanel technology can also examine those regions for gains and losses of DNA sequences and rearrangements of DNA on chromosomes. The results are entered into a database for research purposes, but, if a patient agrees, the clinically important findings can also be returned to their doctor for use in the clinic.

The OncoPanel advanced sequencing platform is an important update to Dana-Farber’s original OncoMap platform. OncoPanel can detect not only commonly known gene mutations, but also other critical types of cancer-related DNA alterations not previously identified. In contrast, OncoMap was limited to screening for known cancer-related gene mutations. The OncoPanel testing is done at the Center for Advanced Molecular Diagnostics, a CLIA-certified laboratory operated by the Department of Pathology at Brigham and Women’s Hospital.


1./ Gray SW, et al. Original Reports – Health Services and OutcomesPhysicians’ Attitudes About Multiplex Tumor Genomic TestingJ. Clin. Oncol., published online before print on March 24, 2014, doi:10.1200/JCO.2013.52.4298.

2./ Hall MJ. Conflicted Confidence: Academic Oncologists’ Views on Multiplex Tumor Genomic Testing. J. Clin. Oncol. Editorial, published online before print March 24, 2014, doi:10.1200/JCO.2013.54.8016


Dana-Farber Researchers “OncoMap” The Way To Personalized Treatment For Ovarian Cancer

Researchers have shown that point mutations – mis-spellings in a single letter of genetic code – that drive the onset and growth of cancer cells can be detected successfully in advanced ovarian cancer using a technique called OncoMap. The finding opens the way for personalized medicine in which every patient could have their tumor screened, specific mutations identified, and the appropriate drug chosen to target the mutation and halt the growth of their cancer.

Researchers have shown that point mutations – mis-spellings in a single letter of genetic code – that drive the onset and growth of cancer cells can be detected successfully in advanced ovarian cancer using a technique called OncoMap. The finding opens the way for personalized medicine in which every patient could have their tumor screened, specific mutations identified, and the appropriate drug chosen to target the mutation and halt the growth of their cancer.

Using mass spectrometry for identifying the genetic make-up of cancer cells, OncoMap can determine the point mutations in tumors by utilizing a large panel of over 100 known cancer-causing genes (referred to as “oncogenes“). In the work to be presented today (Wednesday) at the 22nd EORTCNCIAACR [1] Symposium on Molecular Targets and Cancer Therapeutics in Berlin, researchers will describe how they used OncoMap to identify oncogene mutations in tumor samples obtained from women with advanced high-grade serous ovarian cancer. [2] Earlier in the year 76 mutations in 26 different genes had been found but, since then, further work in more tumor samples has found more.

Ursula A. Matulonis, M.D., Medical Director, Gynecologic Oncology, Dana-Farber Cancer Institute; Associate Professor, Medicine, Harvard Medical School

Dr. Ursula Matulonis, director/program leader in medical gynecologic oncology at the Dana-Farber Cancer Institute located in Boston, Massachusetts (USA) and Associate Professor of Medicine at Harvard Medical School, will tell the meeting:

“Epithelial ovarian cancer is the most lethal of all the gynecologic malignancies, and new treatments are needed for both newly diagnosed patients as well as patients with recurrent cancer. The success of conventional chemotherapy has reached a plateau, and new means of characterizing ovarian cancer so that treatment can be personalized are needed.

We know that many human cancers have point mutations in certain oncogenes, and that these mutations can cause cancer cells to have a dependence on just one overactive gene or signalling pathway for the cancer cell’s growth and survival – a phenomenon known as ‘oncogene addiction’. If the mutation that causes the oncogene addiction can be inhibited, then it seems that this often halts the cancer process. Examples of mutations that are successfully inhibited by targeted drugs are HER2 (for which trastuzumab [Herceptin®] is used in breast cancer), EGFR (erlotinib [Tarceva®] in lung cancer) and c-kit (imatinib [Gleevec®] in chronic myeloid leukemia). So if we know the status of specific genes in a tumor, then this enables us to choose specific treatments that are likely to work successfully against the cancer.”

Dr Matulonis and her colleagues used OncoMap to investigate the mutation status of high-grade serous ovarian tumors that were known not to be caused by inherited mutations in the BRCA 1 and BRCA 2 genes. They found mutations previously identified to be involved in ovarian cancer: KRAS, BRAF, CTNNB1 and PIK3CA. The KRAS and PIK3CA mutations were the most common, while BRAF was more rare. The researchers also identified a low frequency of mutations in many other different oncogenes.

Dr. Matulonis further noted:

“This study shows that it’s feasible to use OncoMap to identify whether a patient’s tumor has a mutation in an oncogene for which a known drug is available to target that specific gene, so as to enable us to place her on a clinical study of that drug; for instance, XL147 or GDC-0941 are inhibitors for the P13kinase mutation that are in clinical trials at present.  In addition, someone’s cancer could harbor a mutation (such as ALK) that is not known to be associated with ovarian cancer or has not yet been studied in ovarian cancer – these patients could be matched with a drug that inhibits that protein too. As new drugs get developed, this information would be used to match future drugs with patients and their cancers.”

The researchers hope that OncoMap will become a clinical test for all cancer patients at the Dana-Farber Cancer Institute before long, so that the genetic information obtained can be used to choose the best treatment for them.

Dr. Matulonis said:

“At present, only a few targeted therapies are being used for newly diagnosed ovarian cancer and most are being used to treat recurrent ovarian cancer, but this will change eventually. I have already referred several of our patients who are either newly diagnosed or have recurrent cancer and who have mutations (one with KRAS and one with PIK3CA) to our phase I program for drugs studies specific to these mutations.  For ovarian cancer, understanding mutational analysis is one piece of the genetic puzzle. Our group will also start looking for chromosomal and gene amplifications and deletions in patients’ tumors, which we know are important for ovarian cancer.”

Matulonis believes that OncoMap and other similar analytical tools will become mainstream practice in all cancer clinics before long. Tools for detecting genes with the incorrect numbers of copies or abnormal expression will also help doctors to choose the best treatment for individual patients.”

Source: Researchers map the way to personalised treatment for ovarian cancer, Abstract no: 35. Oral presentation in plenary session 2.  22nd EORTC-NCI-AACR Symposium on Molecular Targets and Cancer Therapeutics, Berlin, Germany, November 16- 19, 2010.


[1] EORTC [European Organisation for Research and Treatment of Cancer, NCI [National Cancer Institute], AACR [American Association for Cancer Research].

[2] The study was funded by the Madeline Franchi Ovarian Cancer Research Fund, twoAM Fund and the Sally Cooke Ovarian Cancer Research Fund.

Related Information: