Curious Dr. George | Plumbing the Core and Nibbling at the Margins of Cancer

A Plea for Gold Standards in Precision Oncology

George Lundberg, MS, MD, ScD, MASCP, FCAP, Chief Medical Officer and Editor in Chief, CollabRx, a Rennova Health Company; Editor at Large, Medscape; Executive Adviser, Cureus; Consulting Professor of Pathology and Health Research and Policy, Stanford University; President and Chair, The Lundberg Institute; @glundberg

Q: Is Precision Oncology Accurate?
A: Precision oncology in 2017 is neither accurate nor precise.
From my clinical pathology background, accurate means correct, as close to a “Gold Standard” as you can get, regarding sensitivity and specificity. Precise means reproducible. Prasad and Galehave recently demonstrated that not even the use of the term “precision oncology” is precise. It has changed often and dramatically.
We need to determine current (they would float with new information) “Gold Standards” for every step (and many sub-steps) in the Brain to Brain loop of molecular testing and clinical actions in oncology.
For example:
WHICH cancers should be subjected to some sort of molecular analysis? All, or all except non-melanoma skin cancers? Only those cancers for which there are FDA approved therapies matched by molecular definition? Cancers for which there are open relevant clinical trials?
WHEN should NGS be performed on a cancer? At initial diagnosis, or not until after spread, or both, or after other treatment failures, or depends on histopathologic diagnosis?
WHERE to sample the tumor? Primary, and/or one or many metastatic sites, or depends on tumor diagnosis?
Should SAMPLE be liquid biopsy and/or solid tumor biopsy or fine needle aspiration cytology or whole tumor segments after surgical removal, or paraffin block, or perhaps other sources?
WHERE should testing be done: local pathology lab, regional reference lab, nearest academic or comprehensive cancer center, or large commercial lab company? Perhaps this should in part depend on lab accreditation, TAT, and price.
HOW TO CALL the variants consistently via bioinformatics and experienced professional judgment, reproducible from lab to lab.
HOW is the mutation(s) identified? As a passenger or driver? Relevant or irrelevant?
DOES a particular mutation, CNV or fusion confer “actionability”?
Would ACTIONABILITY include a clinical trial match as a criterion?
WHAT FDA-approved, or investigational, single drug or combination of drugs, concurrently or in sequence, would be the best choices?
WHEN/WHETHER should clinical trials be large and randomized, or at the level of an n-of-1 after a fully informed consent and expanded access (compassionate use)?
WHO/WHAT should pay for the costs? Patient, insurance company, government, drug company, medical laboratory, or medical treating institution?
HOW MUCH COST for the NGS testing and for the drug (which can cost $100,000- 200,000 or more per drug, per patient/cancer) can be justified by QALY or research?
WHAT THERAPEUTIC STRATEGY might make scientific sense for that patient’s cancer?
WHETHER there is ANY currently available drug that is a reasonable choice (estimated 10% of cancers).
HOW to ascertain and how to report frequency and severity of unanticipated adverse clinical effects?
WHAT outcome can be anticipated and communicated to the patient? Is there a possible cure? A few weeks or months of average/median extended life seems to me a low “bar” to affirm cost-effective “success.”
Acceptable QUALITY OF LIFE during treatment? Meaningful? Useful? Worthwhile? Pain free? Pain tolerable? Desired? Comfortable? Connected? Freely chosen?
Roughly 1.7 million Americans a year are diagnosed with potentially lethal malignancies. About 600,000 die. Establishing GOLD STANDARDS, consistent with current knowledge, is really important for many patients, the field, the public health and many micro-economies.
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.