Lisandra E. West, PhD, Senior Scientific Knowledge Engineer, CollabRx.
Q: Is there enough benefit to justify sequencing all patients’ tumors? – Perspective from a scientific knowledge engineer.
A: As a scientist working as a scientific knowledge engineer in molecular oncology, part of my job is to collect, organize, and deliver data that correspond to aberrations present in cancer that could inform clinical decision making.
I recently read with great interest, the opposing viewpoints on “Is there enough benefit to justify sequencing all patients tumors?”, published in JAMA Oncology (April 14, 2016). From the Yes side: sequencing technology provides invaluable intelligence to diagnose, treat, and defeat cancer (Universal Genomic Testing Needed to Win the War Against Cancer by Razelle Kurzrock, MD, and Vivek Subbiah, MD). From the No side: rigorously tested superiority and outcome data are lacking and the cost to patients is still prohibitive (No Solid Evidence, Only Hollow Argument for Universal Tumor Sequencing by Howard Jack West, MD). Both of these are compelling arguments, in spite of their contradictory nature, that deserve ongoing consideration in the oncology community. What we are increasingly coming to understand is that beyond a relatively small set of driver aberrations, whose testing and treatment are specified in standard treatment guidelines, there are a wealth of other molecular abnormalities in cancer that may be highly useful in informing clinical decision making. A few examples to illustrate the point:
1) Biomarker overexpression suggests selection of one targeted therapy drug may be superior to the other targeted drugs recommended for the diagnosis in treatment guidelines (example: high AXL expression in EGFR-mutated NSCLC is associated with acquired and de novo resistance to first generation EGFR inhibitors).
2) Identification of an oncogenic fusion responsive to one class of targeted therapy approved for a cancer and resistant to another (example: BRAF fusions in pan-negative melanoma are responsive to MEK inhibitors, but resistant to BRAF inhibitors).
3) Biomarker loss of expression is predictive of poor prognosis suggesting more intensive therapy is needed (example: loss of expression of MET and RON receptor tyrosine kinases is an independent prognostic factor in DLBCL patients receiving R-CHOP).
4) Gene amplification is predictive of clinical benefit from targeted therapy suggesting the patient may want to consider joining a clinical trial (example: CDK4 amplification in liposarcoma is associated with favorable progression-free survival for liposarcoma patients treated with a CDK4/6 inhibitor).
Before the cost of genetic testing for all patients’ tumors becomes universally accessible thanks to increased affordability, how can we unite the best quality emerging data with each physician-patient pair at the crucial treatment decision points? One piece of the solution may lie in development of focused, cancer specific panels that include the highest strength of evidence markers. These markers will likely span biomarker modalities, including not just sequence variants (detected by NGS), but also protein expression (IHC), copy number variation (FISH), and gene fusions (FISH). Strength of evidence and frequency of occurrence in respective cancers may inform aberration selection for such panels. Physician demand could drive both panel creation and availability, while the evidence for treatment optimization and patient benefit could provide justification for insurance reimbursement to help cover the cost for patients. I hope such advancements in panel design are not too far in the future.
Meanwhile, to the extent that molecular testing is currently available today, how can physicians, and patients, access the highest value evidence that can guide clinical decision making over the course of cancer treatment and tumor evolution? Every physician should have easy access to the highest quality data in a digestible format. Every patient has the right to know, not just what the options are, but that they have been considered by both their personal oncologist, and expert physicians in the community. To me, these are at least in part, problems of knowledge engineering, and thus problems that I would like to work on solving. To this end, I hope to collaborate with some of you on tools and solutions to enable this process in the near future.
Copyright: This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.