Pathways Have the Potential to Deliver Personalized Medicine
Washington, DC—The institution of cancer treatment pathways is not incompatible with personalized medicine, but rather it has the potential to offer access to a rapid learning system that can promote personalized therapy, said Michael Kolodziej, MD, National Medical Director, Oncology Solutions, Aetna, at the Fifth Annual Conference of the Association for Value-Based Cancer Care.
Pathways can enable personalized care, going beyond just the genome. A patient with lung cancer who is aged 70 years is different from a patient with lung cancer who is aged 40 years, irrespective of their mutational structure.
The view of cancer as a linear disease, in which treatment is chosen based on a known mutation, does not necessarily lead to improved patient outcomes, he argued. One example is the use of EGFR mutation status to select treatment for patients with NSCLC. Patients with an EGFR mutation—deletion 19 or substitution—have better progression-free survival than patients without an EGFR mutation.
A retrospective data analysis showed greater overall survival associated with afatinib (Gilotrif) treatment than with chemotherapy in patients with deletion 19, but with a hint that overall survival is worse in patients with a point mutation receiving afatinib.
“The 2 mutations that we would consider to be indicative of a positive response, ultimately had a different impact on the patient,” said Dr Kolodziej.
There are numerous EGFR variants, including more than 250 point mutations, and each mutation is clinically distinct and responds differently to afatinib or to erlotinib (Tarceva), or does not respond at all in the case of point mutations. Unless the clinician can know the specific variant, rational treatment selection is not possible. “We need a construct by which we actually learn something,” he said, pointing to the American Society of Clinical Oncology’s CancerLinQ (Figure).
In CancerLinQ, patients undergoing genomic profiling and receiving treatment based on the results are entered into a national registry and their outcomes are being tracked.
“We need to think about expression assays, we need to think about proteomics, we need to think about immunological markers,” he said. “If it were really an N of 1, we would be paralyzed by indecision. What we need is the ability to sort a population into just the right size…put in molecular markers, and we make an intervention and measure what happens,” Dr Kolodziej said. “We need to be able to take the population and identify appropriate subpopulations that allow us to predict in a probabilistic fashion the likelihood that that next patient who is similar will respond, and identify what the clinically different elements are that predict a lack of response.”
Pathways have the potential to promote personalized therapy, Dr Kolodziej said. “The answer might be in the cloud, but it doesn’t make decisions at the bedside, and that’s why we need pathways to actually deliver personalized medicine.”