American Society of Hematology


"The “Roadmap” paper provides an important overview of what’s needed to accelerate progress against lymphoma. As I understand it, the primary need and objective described here is to capture quality biospecimens that includes clinical annotation – that supports rigorous correlative analysis and additional interrogation of the tumor features by other scientists.

I would add that there’s an urgent need for novel trial designs that makes participation in such trials an acceptable/preferred treatment decision.

It seems evident that the design needs to be a good fit with the primary objective of the study - be it to identify the best dose (single-arm), confirmatory studies (RCT), or exploratory biospecimen-based studies (?) to capture efficacy signals for study protocols -- specific to the underlying biology of the disease.

My understanding is that the value of a predictive biomarker will be specific to the treatment received. For example, immune signature X (or mutation Y) might predict response to Rituxan monotherapy but not to chemo- immunotherapy.

Assuming that the responses to treatments are driven by the biology of the disease, it should not matter (for such trials) that a treatment was assigned by a coin flip, or by physician-guided patient choice. A hybrid design that allows physician-guided patient choice – to choose random assignment or a preferred study arm – should accrue very rapidly. The added expense of larger trials that allow participants to choose randomization or a preferred study arm might be mitigated by charging user fees for studies on the annotated high-quality biospecimens acquired in such study – and by the knowledge gained.

Randomized trial design is an essential research tool … but such studies may not achieve individual equipoise unless the study arms and controls are similar in approach (comparing apples to apples with x). The patient guided by his or her physician will decline participation if any of the study arms are deemed an undesirable or poor choice – particularly if a preferred treatment can be used off study – or if the control arm is similar to what failed to benefit the patient previously in the relapse setting.

For more information on this hybrid design proposal see the Observation-enriched RCT see (Again this design (already used in a few studies) is not meant to replace RCT … but to be applied in select cases)."

-Karl Schwartz

back to top

"While we have learned a great deal from genomic approaches, by definition this will miss signaling alterations that occur due to altered post-translational modifications - unless one of the members of the modification or de-modification enzymes is mutated. While this is in part captured through the concept of epigenomics, this leaves a number of PTMs un-tapped. For this reason, I would recommend inclusion of a functional genomic approach to identification of non-mutated proteins that are required for disease pathogenesis."

-Paul Galardy

back to top