NEW YORK (GenomeWeb) – Sequencing cancer patients' tumors has gained momentum with the availability of reimbursable commercial tests and the launch of many clinical research programs at academic cancer centers. Thousands of patients have now had their tumors sequenced with the goal of identifying treatments that target their specific genomic alterations. And while in many cases, researchers are demonstrating that molecular profiling can be done in clinically relevant time frames and produce actionable information, one challenge that has not been overcome is getting patients enrolled in clinical trials or access to targeted therapies.
At two recent conferences — the Advances in Genome Biology and Technology meeting in Hollywood Beach, Florida and the Molecular Medicine Tri-Conference in San Francisco — researchers from various institutions discussed the successes and challenges of their cancer sequencing programs, with nearly all concluding that the availability of targeted therapies and clinical trial enrollment are still challenges, regardless of whether patients' tumors are profiled by small gene panels or comprehensive whole-genome and transcriptome sequencing.
Potentially actionable results are found in between 39 percent and 95 percent of patients tested with a targeted panel or more comprehensive sequencing, according to Stacy Gray, an associate clinical professor at City of Hope, who evaluated published data from four US-based academic cancer centers and one community group. Despite this wide range in so-called actionable results, however, the various projects all have similarly low rates, between 5 percent and 16 percent, for matching patients to genomically guided therapies, whether through approved or off-label drugs or in a clinical trial.
"Many people are not benefitting from precision cancer care, who could be," Gray said.
In a presentation at the Tri-Conference, Kenna Shaw, executive director of the Sheikh Khalifa Bin Zayed Al Nahyan Institute for Personalized Cancer Therapy at MD Anderson Cancer Center, described the efforts her group is undertaking to reduce the gap between patients who receive an actionable result from the genomic test and those who ultimately receive a genomically guided therapy.
Highlighting the scope of the problem, she cited a retrospective observational study of 2,000 consecutive patients who received a targeted 50-gene panel test at MD Anderson. The study, published in the Journal of Clinical Oncology in 2015, found that around 40 percent of tested patients, or 789, had a potentially actionable mutation.
Shaw said that the team had deemed 35 genes on the panel to be actionable, based on the availability of approved and investigational targeted drugs or clinical trials.
Nonetheless, only 83 out of the 789 patients, or 11 percent, who had an actionable result were ever enrolled into a genotype-matched clinical trial. In the retrospective JCO study, the researchers wanted to understand why this was happening. They looked at the subset of patients who had mutations in the PIK3CA, BRAF, and KRAS pathways, since those mutations were well known and MD Anderson also had available clinical trials. A positive result in those pathways should represent the "best-case scenario," Shaw said. Nonetheless, of 429 patients with a mutation in those genes, only 61, or 14 percent, were ultimately enrolled in a trial.
Several reasons account for this lack of enrollment. After testing, a subset of patients did not return to the clinic. Another subset of patients never had their results recorded in their medical chart, and for others, there is no record of the result being discussed with the physician.
"We're losing about one third of patients at every checkpoint," Shaw said in her presentation. She said one major problem is how genomic information is communicated to the treating physician. Doctors are left to make sense of a complicated genomic test report that may simply list the genes tested, with circles around those in which an actionable mutation was found, Shaw said. Now, however, her group is in the midst of overhauling the way in which genomic test results are reported, which will hopefully improve the success rate of enrolling patients on trials.
"When a physician realizes there is an important gene and there is a clinical trial available, we do reasonably well at putting the patient on a trial," she said.
Shaw said that the group is studying the impact of implementing a decision support tool on patient outcomes, and is hopeful that it will help. Nonetheless, even the best decision support tool will not increase the number of targets for which drugs are available or the availability of clinical trials, she said.
The National Institutes of Health's basket trial, NCI-Match, which seeks to stratify patients into clinical trials based on results of a targeted gene panel, has also had lower than expected rates of trial enrollment, at 23 percent, versus its originally anticipated 33 percent. But, perhaps because the genomic testing is linked specifically to clinical trials, they do better than some other centers.
One benefit of genomic testing that is sometimes not appreciated is that test results can often inform the physician what not to do, according to Andrew Kung, chair of the Department of Pediatrics at Memorial Sloan Kettering Cancer Center who presented on the institution's Precision in Pediatric Sequencing (PIP-seq) program at AGBT.
In about 3 percent of cases, sequencing results inform the physician that a bone marrow transplant, which would have been the next step if sequencing had not been done, would not be effective, Kung said. "That's a significant clinical impact."
Kung described one patient, who at two months of age appeared to have hemophagocytic lymphohistiocytosis, a hyperinflammatory blood disorder that is treated by bone marrow transplantation. In some cases though, the inflammation is caused because of infection, Kung said. An initial 14-gene panel that included the genes responsible for about 70 percent of HLH cases was negative. However, the child was still enrolled in PIP-seq, where participants receive exome sequencing of their tumor and normal tissue, as well as tumor transcriptome sequencing.
Sequencing identified a de novo mutation in the MLL2 gene that was associated with Kabuki syndrome, which can lead to immunological abnormalities. This finding helped the team determine that the infant had secondary HLH in response to a respiratory infection. Rather than subjecting the infant to a bone marrow transplant, he was able to be treated with low levels of chemotherapy. He overcame the infection and his HLH never re-occurred, Kung said.
Another class of important findings are cancer predisposition mutations in the germline. Kung said that about 13 percent of participants in PIP-seq have a germline cancer predisposition mutation, which can have implications not only for the patient, but for the patient's family, even though it might not point to a therapeutic drug.
For instance, he described one patient with acute myeloid leukemia whose sister was a perfect HLA match and could potentially be a donor for bone marrow transplantation. However, the sister had received some abnormal blood work results, so the researchers decided to dig deeper. Exome sequencing of the patient uncovered a germline mutation in RUNX1, indicating a risk for blood count abnormalities with a 30 percent chance of transforming to leukemia. When they tested the sister, they found she had the same heritable mutation. These results spurred the team to find an unrelated donor for the transplantation and to begin a surveillance program to catch early signs of leukemia developing in the sister.
Another clinical research program that uses comprehensive sequencing is the Personalized Oncogenomics Project at the British Columbia Cancer Agency. The BCCA team launched a pilot program in 2012 to perform whole-genome and transcriptome sequencing on tumors from metastatic cancer patients. Martin Jones, who leads the clinical informatics team at the Michael Smith Genome Sciences Centre at BCCA, said that so far, researchers have sequenced 524 patient tumors and analyzed results of 509. Of those patients, around 60 percent have been deemed to have an actionable result.
Nonetheless, he said, for 36 percent of patient s with an actionable result, , no action is taken — typically because the patient has passed away, is too sick to undergo treatment, or has withdrawn from the study for other reasons. Another 26 percent of patients are either responding well to their current treatment or are in remission and doctors are in a "wait and see mode," Jones said. That leaves just 38 percent who receive new treatment as a result of the genomic test. Of those, about half are treated with standard of care therapies., about one third are treated with off-label drugs indicated by their genomic profiles, and 14 percent are enrolled in clinical trials.
Like others in the field, Jones noted that the project is generating copious amounts of data and oncologists are increasingly seeing value in genomic testing. Over 82 percent of clinical oncologists within the BCCA have interacted with the project, he said.
Performing comprehensive sequencing as opposed to a targeted gene panel has both advantages and disadvantages, according to researchers. While the advantages are a more complete understanding of the tumor makeup, including gene expression, a better ability to detect complex alterations, and the ability to reanalyze the data for druggable targets as therapies become available, the most notable challenges are the higher cost and longer turnaround time.
To shorten the turnaround time, Jones said that the team has implemented a two-tiered reporting system. The first analysis is a drug target analysis, focusing on the genes and alterations that would indicate a therapy or clinical trial. That initial report is delivered in three to four weeks, he said. A second, more comprehensive report is completed within about 45 days.
Going forward, he said, the group is working to address the cost issue as well. Currently, the Personalized Oncogenomics Project is funded through philanthropic support, and as sequencing prices continue to drop, the program will become less expensive. "But, we don't yet have a good answer for how we'll fund it if not through philanthropy," he said.
Whether a targeted panel or comprehensive sequencing, the two types of testing strategies share one major problem — access to drugs. Jones agreed with others that this is a major limitation, and one of the main causes for low rates of genomically guided treatment.
"There's a gap between finding potentially actionable alterations and getting patients targeted therapies," Gray said. "In order to really realize precision medicine, we have to make sure that every time there's a chance to act on information that could make a patient's treatment more precise, it is quickly acted upon," she added.