Skip to main content
Premium Trial:

Request an Annual Quote

UCSD Molecular Tumor Board Aids Patient Care Decisions in Age of NGS Tests


NEW YORK (GenomeWeb) – Academic medical centers that have begun incorporating next-generation sequencing-based tests for cancer patients often rely on molecular tumor boards to help interpret the results in the context of other clinical features, such as the patient's medical history and pathology results. Tumor boards are not new in cancer care and are a common tool used to discuss complicated cases, patients with rare cancer, or patients without an obvious course of treatment. But now that next-gen sequencing tests are more frequently being offered to cancer patients, tumor boards have evolved to now also include genomics experts and bioinformaticians.

Researchers at the University of California, San Diego's Moores Cancer Center published a study this month in The Oncologist describing their experience with such a tumor board to determine its impact on patient management, including how often the tumor board is able to match a patient with a therapy, the challenges of incorporating such testing into patient care, and how the process may be improved in the future.

Molecular tumor boards "may be crucial for patient care because most oncologists have not been trained in the genomics field," the authors concluded.

"We have very well-trained, experienced cancer physicians, but they're not trained in genomics," Razelle Kurzrock, senior author of the study and head of UCSD's Center for Personalized Cancer Therapy, told Clinical Sequencing News. The molecular tumor board can "help the physician discuss the patients and come to a consensus about the best course of action."

In the study, the UCSD researchers evaluated 34 patients with an NGS-based assay and the results, along with other clinical information, were discussed at a molecular tumor board to determine whether a treatment course could be decided on based on the sequencing results. Of the 34 patients, 12 were treated based on their molecular profile. For an additional seven patients, the board found an aberration that indicated a treatment, but the patient was either ineligible for the appropriate clinical trial, could not travel to the center conducting the trial, or did not have the means to pay for the matched therapy. Thirteen patients continued to do well on their current therapy. One patient passed away before the tumor board discussed the case and one patient did not have any known actionable results.

Kurzrock said that the study revealed a couple of trends. First, she said, "the molecular tumor board really helped physicians gain comfort with the molecular diagnostic."

One way in which this was evident, she said, was how physicians' use of the test changed over the course of the study. Initially, physicians ordered the test when the patient failed therapy. But because turnaround time from when the test was ordered, and including sample acquisition, testing, and results averaged 27 days, this was often too long to wait.

"More and more, the physicians started to order testing while the patient was doing well. They were anticipating, based on what we know about cancer, that the patient wouldn't do well indefinitely," Kurzrock said. "And the physician wants to have the results at the time the patient is no longer doing well."

The UCSD team tested 34 patients from 10 different oncologists. The patients had a variety of cancer types, including 16 with breast cancer; eight with gastrointestinal cancer; four with head and neck cancer; two with lung cancer; one patient each with epithelioid sarcoma, myoepithelial carcinoma, paranganglioma; and one patient with a tumor of unknown origin.

The vast majority of the patients, 33, were tested with Foundation Medicine's cancer panel. Response Genetics' ResponseDx panels, Caris Life Science's Molecular Intelligence test, and Champions Oncology's exome sequencing test were also ordered for two, three, and one patient, respectively. Four patients received multiple tests.

Turnaround time from sample order to results took an average of 27 days, but ranged from 14 days to 77 days. Simply acquiring the sample took an average of 11 days, the authors reported.

Testing identified at least one aberration in each patient, with an average of four per patient. In one patient, previous testing via Life Technologies' Ion AmpliSeq Hotspot panel identified no mutations, but subsequent testing with the FoundationOne test, which analyzed 236 genes, found 14 mutations. Thirteen of the mutated genes were not included in the AmpliSeq panel and the 14th gene was amplified in the patient, which is not detectable via the AmpliSeq Hotspot panel, which only analyzes point mutations.

No two patients had identical mutations. Even when patients had an altered gene in common, the gene either had a different class of mutation or was mutated at a different location between patients. The most commonly mutated genes were TP53, MYC, PIK3CA, KRAS, PTEN, CDKN2A, and APC.

So far, 12 patients have been treated based on the results of the NGS test, one of which was not evaluable at the time of the study because treatment had just been started. Three of 11 patients achieved a partial response, four had stable disease, and another four continued to progress.

Kurzrock said that although at the time of the study, only 12 of 34 patients had been treated based on the molecular profiling results, she expects that additional patients that were still doing well on their currently therapy will eventually receive a therapy indicated by the sequencing results.

Nevertheless, she said that approximately half of the patients will not be able to be treated based on the results. Of those, she said, "the most important category is patients where we know they have an abnormality that perhaps we could act on, but we can't get the drugs." Kurzrock estimated that between 25 percent and 40 percent of patients would fall into this category. Often patients cannot be treated because the aberration points to a drug or clinical trial that is not open to the patient's tumor type. Or the patient is unable to travel to the site of the clinical trial. "Getting the drugs is a very big obstacle," she said.

Other reasons testing doesn't lead to a treatment may be because the test reveals only mutations of unknown significance or does not find any mutations at all.

Since doing this study, Kurzrock said that UCSD now offers an NGS-based test in-house out of its CLIA certified laboratory that analyzes between 40 and 50 genes.

In addition, she said the study has spurred the team to set up clinical trials to test the impact of NGS testing on patient outcomes. Several organizations are also setting up patient registries to track outcomes, she said. "The idea is to try, with the power of electronic medical records, to capture the data about how a doctor is treating a specific patient and figure out what happens to that patient."

Kurzrock said that while tumor profiling is becoming increasingly used and showing to provide useful information, a number of hurdles still remain.

One major hurdle will be cost. "Who will pay for the testing? What portion will insurance pay for?," Kurzrock questioned.

In addition, she said, organizations looking to offer molecular tumor profiling tests will need to possess a wide range of expertise in order to put together a tumor board that is able to discuss the tests and make appropriate recommendations. Molecular tumor boards need to include experts from a range of different fields on both the clinical and basic science sides, she said, which could be a challenge for some organizations.

Kurzrock said that molecular tumor boards will continue to play an important role in guiding treatment for those cancer patients that receive genomic testing. It is unlikely that software could provide the same expertise as such boards, at least in the near future. Even with the molecular tumor board, Kurzrock said that UCSD relies heavily on automation. All commercial NGS tests come with a "very extensive automated report with treatment suggestions," she said. "A broad report that gives you the aberrations with nothing else would be completely bewildering. Automation is an essential step, but it's not enough at this point in time," Kurzrock said.