A few more answers (and questions) about the PARP inhibitor trials

The Feb. 28 posting regarding PARP inhibitors drew several comments I thought should be addressed. There has been considerable interest in PARP inhibitors because of the early activity cancers associated with BRCA1 and BRCA2 mutations, as well as triple-negative (estrogen, progesterone and HER2 receptor-negative) breast cancer. The fact that the follow-up and more definitive phase 3 trial comparing chemotherapy with or without the PARP inhibitor iniparib (also known as BSI-201) did not confirm the dramatic survival benefits seen in the earlier and smaller phase 2 trial raised a lot of questions. Our readers have responded with queries and some degree of dismay, including those who are currently on the expanded access trial with this drug. The evolution of descriptors from "spectacular" to "exciting" to "not meeting its endpoint" is a repeated theme in science as we go from preliminary data, where conclusions are shakier (but tantalizing because they are new) to more definitive studies, where the results are more reliable, but may not always confirm the earlier findings. However, there are some details in this particular case that are worth mentioning, some of which were alluded to in comments from our readers. The phase 2 and 3 trials had equivalent treatment arms, but the endpoints were not the same. The phase 2 trial was initially designed to see if cancer could be stabilized in more patients receiving iniparib added to chemotherapy compared with chemo alone. There was an unexpected improvement in survival with the addition of iniparib - greater than has been seen with any other drug. This led to the phase 3 trial being designed with very high expectations – that both survival and disease-free survival would be improved. Therefore, both these endpoints were built into the statistical analysis plan. When the result fell short of the "p-value," or the level of statistical significance, the trial was declared to have not met its endpoint. However, it is important to note that there still could be an important benefit for all patients. Just not by the definition set for this trial, which was expected to win FDA approval. It is also possible that patients receiving first-line therapy were not the same as those receiving second or third line therapy, since those who have very aggressive cancer initially might not qualify for later lines of therapy. This type of selection bias could result in a different biological profile such that PARP inhibitors might work better in those who are able to make it to second/third line treatment, as was demonstrated in the Phase III trial. Of course, these are all speculative reasons for these results, and only diligent completion of the planned trials will ultimately answer these questions. In the meantime, we should recognize the elegance of targeting DNA repair, and those who are on clinical trials with PARP inhibitor should certainly continue, with the same hope that this drug can make a difference. The truth will eventually emerge, and while it is possible that we might be disappointed, the data so far suggests that PARP inhibitors will be effective drugs – it remains to be seen which inhibitor (since they all seem to be different) and which patients will benefit (since we don't know which tumors really rely on this enzyme). Science is not only about cold facts – it involves drama, emotions and contradictions. But ultimately, truth prevails with well-designed studies and patients – just not as soon as we would like.