NEW YORK (GenomeWeb) – Sequencing the genomes of cancer patients to identify mutations that can point to potential treatments or management is becoming more common, although the number of available targeted treatments is limited and access to them based on a molecular profile remains challenging.
At last week's Individualizing Medicine conference hosted by the Mayo Clinic in Rochester, Minnesota, a number of researchers discussed their experiences from clinical cancer sequencing programs. Access to targeted treatments is a major challenge, researchers said, and they noted several ways in which they are trying to overcome the challenge: through more innovative clinical trial designs, individualized mouse tumor models to quickly test drugs, and drug repurposing.
During a presentation, Alan Bryce, medical director of the genomic oncology clinic at the Mayo Clinic Arizona, discussed the Mayo's experience with three years of genomic testing of cancer patients. Between October 2012 and 2015 a genomic tumor board, which consisted of clinicians, oncologists, surgeons, pathologists, geneticists, bioinformaticists, and ethicists, considered 165 patients. Genomic testing, which consisted of gene panels, exome sequencing, or even whole-genome sequencing, delivered internally or via commercial entities, was ordered for 141 patients.
Of 141 tests, which were primarily Foundation Medicine's FoundationOne panel, Bryce said, 92, or 65 percent, had actionable results. Forty-six patients' test results were not actionable, two had no reportable results, and one patient's test failed.
Of the 92 patients with actionable results, 31 patients received targeted treatment, 29 of which were driven by the results of the genomic test. The other two received treatment based on conventional testing. Another 21 patients received non-genomically targeted treatment, while for six patients, treatment was unknown. A total of 34 patients did not receive treatment after getting results of the genomic test, usually because their disease has progressed too quickly, Bryce said, although some patients had also achieved remission.
Of the 29 patients whose treatment was guided by the genomic test, 13, or 45 percent, had a clinical response.
Bryce cited these statistics as a way of describing the difficulty and complexity of trying to implement genomic testing clinically. When considering the likelihood of a genomic test making a difference in patient outcome, "at the time the test is ordered, there is only about a 7 percent chance that the patient will respond to a targeted test," Bryce said. That number is about the same for the percent of patients who will respond to a phase I clinical trial.
Bryce said that there are a number of factors that explain why that number is not higher, "especially since two-thirds of patients have an actionable mutation."
Thus far, he said that one of the main challenges has been getting patients on treatment. The way that drugs have historically been approved has been for a specific indication defined by the body site where the tumor is, not the mutational target. Insurance providers, regulators, and drug companies "are not ready for the paradigm shift saying a patient has a mutation, so we want to treat them with this drug that targets this mutation," he said.
In addition, he said, there is a large percentage of patients who end up not getting a drug despite having a molecular test indicating they would be a good candidate for the drug because by the time the test is done and results are delivered, the patient has progressed.
Bryce said that throughout the three-year study, patients were increasingly sequenced earlier and earlier in the course of their disease, which helped increase the number of patients that received drugs.
However, Edwin Cuppen, director of the Hartwig Medical Foundation in Amsterdam, noted that it can be challenging to make the case that sequencing patients with primary disease is cost-effective, since around half are cured with standard of care.
"Ideally, we would sequence every cancer patient," he said, but given limited resources, he said that his institution decided to focus on those with metastatic disease. "It's a large unmet need, clinically," he said, and "many novel and expensive drugs are entering the market for which there are a lack of actionable biomarkers."
The Netherlands Center for Personalized Cancer Treatment was founded in 2010 by University Medical Center Utrecht, Erasmus MC/Daniel den Hoed Clinic in Rotterdam, and the Netherlands Cancer Institute in Amsterdam and began offering whole-genome sequencing to metastatic cancer patients. In 2014, it expanded the service to all academic hospitals in the Netherlands; in 2015, it expanded to all teaching hospitals; and next year any hospital in the Netherlands will have the option of ordering whole-genome sequencing for metastatic cancer patients. All the sequencing is performed on an Illumina HiSeq X Ten system at a centralized sequencing laboratory at the Hartwig Medical Foundation.
Nonetheless, Cuppen said that getting a good biopsy from a metastatic tumor is also challenging. From a retrospective analysis of 500 biopsies, he said that between 1 to 2 percent of patients experience a "severe complication" from the biopsy procedure itself. Sequencing is successful in 74 percent of cases and most often fails because of a lack of tumor cells in the biopsy material.
Like Bryce, Cuppen noted that patients need access to targeted treatments. To address this issue, the CPCT recently launched a drug repurposing study to see if they could identify off-label uses of drugs in various cancer types. The study is open at 12 centers in the Netherlands, and the researchers are collaborating with pharmaceutical companies to get access to drugs.
Researchers at the Mayo Clinic are taking a slightly different approach toward drug discovery, looking to identify drugs that would target a patient's molecular alterations by first creating a mouse model of the patient's tumor. Other groups are also testing drugs on mice avatars to identify the right therapy for a patient, including at Columbia University and Mount Sinai. At last week's conference, Aleksandar Sekulic, associate director for the Center of Individualized Medicine at the Mayo Clinic, discussed a program where researchers there are creating mouse models of melanoma patients' tumors.
Sekulic said that for the melanoma study researchers are targeting patients without a BRAF mutation, since that subgroup has the most unmet clinical need. Patients who meet that criteria and who consent to the trial are randomly assigned to either receive molecularly guided treatment or "physician's choice," where a physician can opt to give the patient standard-of-care or molecularly guided drugs.
The researchers perform low-coverage whole-genome sequencing, transcriptome sequencing, and exome sequencing. As of June, 44 patients had enrolled and 23 had been treated. In addition, the development of patient avatars is built in, and as of July, 34 patients had xenograft models, Sekulic said. "We have cohorts of mice for each patient that we can treat," he said.
Importantly, Sekulic said, the researchers can test combinations of drugs based on the molecularly altered pathways in the tumors. "We see this as a potential way to accelerate personalized drug development," he said.