Patients and families of patients with mesothelioma do not ask for much. They have no time to waste on therapies that have proven useless for extending survival when confronted with the incredibly small portfolio of efficacious treatment options in this disease. For this reason, Wieneke Buikhuisen and colleagues’ long awaited, superbly designed and executed switch maintenance trial with thalidomide1 is a landmark achievement in mesothelioma, despite the fact that it failed to show improvement in time to progression compared with best supportive care (median time to progression 3·6 months in the thalidomide group vs 3·5 months in the active supportive care group; hazard ratio [HR] 0·95, 95% CI 0·73—1·20, p=0·72). This 5 year randomised trial in 222 patients represents an international collaboration and involved a novel approach built on previous data, thoughtful stratification of the two cohorts, central reading of imaging, and an attempt to correlate contemporary biomarkers and inflammatory or angiogenic-related molecules with time to progression and survival.
So what are the implications that no benefit was reported for yet another attempt to control the angiogenic pathway in mesothelioma? The outcome means that despite our greatest efforts to understand the biology of this disease, and to target promising pathways that seem to contribute to mesothelioma progression, we have no idea which of these pathways is most important, and specifically in which patients we should target that pathway. Within the data presented in this trial there are hints for future work, albeit limited by lack of power. Interleukin 6 seems to be prognostic in a small subset of patients, and non-epithelial histology classically has a worse prognosis than epithelial. Careful scrutiny of the data reveals that the hazard ratio for use of thalidomide seems to favour improved time to progression in individuals with raised interleukin 6 (HR 0·7, 95% CI 0·3—1·4) and non-epithelial histology (0·8, 95% CI 0·3—2·2), which is the very group that we would like to see results in. Although the statistical chance that this is a real finding might be small, the point is that thalidomide, like other new drugs, must be used only in patients for whom a benefit will be seen, and until we are able to match predictive markers with carefully done randomised trials, we could be harming some patients. Buikhuisen and colleagues point this out in their analyses, which showed shorter median overall survival in the thalidomide group (10·6 months) than in the active supportive care group (12·9 months), although this difference was not significant (HR 1·2, 95% CI 0·9—1·6, p=0·21). Certainly logic would dictate that it is high time to use specific agents to hit specific targets in mesothelioma.
But do we have any reasonably specific targets for mesothelioma? Many of the ongoing trials, unlike thalidomide therapy, are based on targeting a molecule that seems to be specific to the disease. Unfortunately, however, despite overexpression of EGFR in mesothelioma,2, 3 most studies with the exception of two4—6 have not described the common mutations seen in non-small-cell lung cancer, and no efficacy has been noted for gefitinib7 or erlotinib.8 By contrast, the glycoprotein mesothelin9 is overexpressed in mesothelioma as well as selected other tumours, and is being targeted for therapy with recombinant immunotoxins10 or monoclonal antibodies.11 Both therapies have had encouraging preliminary responses in combination with pemetrexed and cisplatin. Finally, WT1 peptide vaccination has been shown to elicit immune responses in mesothelioma patients who overexpress WT1,12 and is the basis for an ongoing post-chemotherapy vaccination trial (NCT01265433).
So why do we struggle and why is the pace of success in mesothelioma so slow? In this orphan disease, as illustrated by this trial, patient accrual takes a long time, and when we do accrue patients, we really must study the disease using a coordinated, multiplatform, and novel technology based approach. We need to have sufficient archives of longitudinally collected serum, plasma, mononuclear cells, and paraffin blocks so that we can define genomic, proteomic, microRNA, or metabolomic differences between responders and non-responders for a given therapy. This discussion leads into another emerging issue, which is our interpretation of response to therapy considering the complex geometry of the abdomen and the pleura. We need trials that are large enough to allow accurate definitions of response, possibly using tumour volumes13,14 or metabolic volumes.15
There is no question that we need more randomised trials in mesothelioma to help us to discard useless therapies and pursue promising ones; however, we need to pick up the pace for all inclusive phase 2 ventures that have enough power not only to define promising efficacy to move to phase 3 but also to define molecular or biochemical reasons for this efficacy, as well as to validate more quantitative ways of recording responses. International efforts must therefore not only plan for central interpretation of clinical endpoints, but also prepare for validation of novel discoveries from the archives of such trials. Obviously, funding for these efforts is crucial, but without proper insightful and forward-looking trial planning, funding support will certainly be as difficult as improving survival for these patients.