9 January 2014
Happy New Year. This week continues my Med Into Grad clinical exposure, and last night I was privileged to attend Molecular Tumor Board at the UCSD Moores Cancer Center.
Tumor boards are weekly meetings where clinical oncologists discuss their patients and solicit opinions of their peers and others in attendance. Meetings usually cluster cases together by organ type, i.e. breast cancer. To address the challenges of the recent flood of new molecularly targeted chemotherapeutics to the clinic, about a year ago the Molecular Tumor Board began, drawing attendance of about 15-20 clinical physicians as well as pathologists, radiologists, genetic counselors, and a few research scientists.
The cases are presented with information like age, date of diagnosis, treatment history, and for molecular tumor board, molecular profile as well. Yesterday the case was presented for a woman with malignant colon cancer. Unlike every other case I’ve seen presented, the patient had molecular profiling performed by no less than five different diagnostic centers: Foundation Medicine, Caris, Clarient, Consultive Proteomics, and Oncopath.
The tests range from immunohistochemical to whole exome sequencing, so variability between the tests would be expected. Also, sample biopsies are from colon, peritoneum and lymph nodes over a two year period, representing differing geographical and temporal snapshots of what is already (assumed to be) a heterogeneous, evolving disease. Perhaps it is not surprising that among the 15 actionable targets indicated by these tests combined, only three of these targets were reported by more than one test.
After a few minutes of discussion, another shared “target” was revealed with similar mRNA elevation consistent between two tests, but was included in only one of the two official reports because of different thresholds of significance used by different companies.
So, what are the actual actionable targets for this patient?!? It would be easy to dismiss this as gibberish, but what this does represent, however, are five example clinical information scenarios of the same patient. Depending on the time of diagnosis, previous therapy, location of biopsy, and preference of diagnostic center, any of these five test results could have reasonably found their way to this patient’s oncologist’s iPad.
Instead, these are all presented for one patient, underscoring the tremendous challenge posed to the clinical translation of this knowledge for personalized medicine, likely representing not only highly heterogeneous disease, but highly heterogeneous therapy courses, and heterogeneous outcomes as well.
So, why not get this five dimensional testing done for all cancer patients? There are a few barriers. The first is cost: each test is upwards of $5000. The second is ethical: tissue biopsies are invasive, and multiple biopsies are only feasible for patients with highly accessible disease.
How will we standardize this knowledge? Should we? To what degree? How should it be presented?
Ryon