In a recent issue of the Journal of the American Medical Association, a team led by researchers at Washington University in St. Louis published two articles reporting its use of whole-genome sequencing in two difficult cancer cases. Genome Technology's Christie Rizk recently spoke with Washington University's Elaine Mardis to find out more about the studies and how whole-genome sequencing could eventually be used in the treatment of cancer patients.
Genome Technology: Why did you decide to use whole-genome sequencing and not another sequencing method?
Elaine Mardis: Philosophically speaking, our group has always championed whole-genome sequencing over exome just because of the complexities in terms of chromosomal aberrations and other large-scale structural events that tend to occur in cancer. In the case of the patient sample with acute promyelocytic leukemia, we felt that it was really important to use whole-genome sequencing for exactly these reasons.
Typically in APL patients, you have a large-scale structural alteration, which is the fusion of chromosomes 15 and 17 — and, reciprocally, 17 and 15 — that brings together PML and RAR alpha in the 15-17 translocation. We were already studying that in a large number of APL patients, so we were well aware of that structural variation … and we felt that whole-genome sequencing was the only real approach to use here because of the likelihood that we would find something that was a structural variation that couldn't be picked up by exome sequencing.
GT: Under what circumstances would be it be most useful to perform whole-genome sequencing for cancer?
EM: I think at this particular point in time, because the approach is so new and because we're still working out a lot of the details about best practices, validation, and then really how to take validated mutations … and inform the physician about the best way that they might consider treating the patient — it seems that best answer right now are two classes of cancer patients. One [class] are end-stage patients, so these are typically patients where experimental therapies are tried anyway because all the conventional approaches to their disease have been exhausted. The other class that I would suggest would be … difficult-to-diagnose patients.
GT: Could whole-genome sequencing be done quickly enough to help patients?
EM: That's really the crux of the matter, quite frankly. With the APL patient, we set
ourselves a time frame of six weeks in order to complete the sequencing and the analysis of the data so that we could return the information back to that patient's treating physician. Six weeks essentially fits with the clinical paradigm by which these patients are traditionally diagnosed — namely there's the pathology examination of the white-blood cell population that allows for the determination that in fact acute promyelocytic leukemia is likely. But that's always followed by a cytogenetic and FISH analysis, which is the most time-consuming part. To get the cells growing and the chromosomes dividing in synchrony, and do the FISH preparations, that's typically in the clinic about a six-week time frame. I would say moving forward, whatever the time frame is for a patient between their initial diagnosis and the onset of therapy, the whole-genome sequencing and analysis really has to fit into that. I think for most patients with solid tumors, it's typically on the order of four to five weeks, and so we're roughly there already.
GT: Putting aside the issue of speed, do we have enough researchers and clinicians with the expertise to analyze this data?
EM: Undoubtedly we do not. I think this is the biggest challenge. I've long said that generating the data is the most straightforward part of the equation. Analyzing the data is significantly more complex.