Broad treatment guidelines may be based on relatively narrow trial data.
Targeted treatment regimens have certainly enjoyed considerable success, but Michael Kolodziej, MD, FACP, sees major systemic problems in the ways we analyze tumors, choose therapies, and collect the data that should drive improvement.
Kolodziej is Aetna’s national medi­cal director for oncology strategies. However, as he took the stage on the second day of the 4th annual meeting of Patient-Centered Oncology Care to discuss current problems with and possible fixes for precision medicine, he noted the views pre­sented were his own and not those of his employer.
Problems with the cur­rent system begin the mo­ment a physician decides to send a patient’s tumor out for genetic testing, Kolodziej said. Different companies of­ten use different techniques to evaluate mutations in any given gene, and nei­ther the physicians nor anyone else has much idea which methodology is best: are lightly-regulated testing companies executing their chosen methodologies competently or do third-party tests pro­vide the same information as companion diagnostics? Worse, he continued, when results indicate a mutation in a relevant gene, they often fail to provide context that could help caregivers make wise de­cisions. Test results often fail to specify if a particular mutation is actually in the tumor or how common it is in the tumor. Moreover, gaps in research often leave physicians no way of knowing whether a particular mutation alters tumor function or whether they should use a different regimen if the muta­tion is expressed in 2% versus 82% of all tumor cells.
Kolodziej lamented a ten­dency to base broad treatment guidelines on relatively nar­row trial data and illustrated his point by asserting that the National Comprehensive Cancer Network (NCCN) rec­ommends targeted therapy for too many patients with HER2 mutations. “Outside of breast and GE [gastroesophageal] junction cancer, amplification is not the most com­mon abnormality in HER2; it’s actually a point mutation. And actually if you look at the point mutations that occur in HER2, some of them are activating, and some of them are not. So, guess what? If you give Herceptin to a HER2-mutated patient with a nonactivating mutation, it has zero chance of working. Zero,” he said.
“I could say that about just about every­thing on this list,” he added, pointing to an NCCN guideline that broadly recom­mended particular targeted therapies for patients with particular genetic altera­tions. “Part of our problem is we’ve dumb­ed this down too much. We think that if you got a mutation that represents a lock and we got that key and we’re going to open it, but that actually turns out not to be the case.”
Kolodziej also lamented the unsustain­able growth in treatment costs before turning his attention to large studies that could make precision medicine far more precise. He argued, however, that efforts like TAPUR (Targeted Agent and Profil­ing Utilization Registry) and NCI-MATCH (Molecular Analysis for Therapy Choice) will only provide maximum benefits if their sponsors work together now to en­sure that the information they produce is compatible enough for pooled analyses.
Kolodziej hopes that research organi­zations everywhere will agree on stan­dards that will produce comparable data from similar studies anywhere around the world. He also hopes that regulators will create and enforce stan­dards for data collection and presen­tation that make every diagnostic test and treatment regimen for patients in a clinical setting available for study.
“Here’s what I think the president can do,” said Kolodziej. “The testing no­menclature needs to be standardized. Proficiency testing is absolutely manda­tory. We need to develop a methodol­ogy to curate and validate the catalogue of mutations. We need to put enough money behind basket-and-umbrella tri­als, which are a good way to try to an­swer some of these questions. We need a transparent, public access national registry, a rapid learning system. We need to learn from every patient. And most importantly, that data needs to be democratized and transparent. Every­body needs to get access to it. It’s not something that you can sell, or at least it shouldn’t be.”