August 30, 2011

This disappointment was almost palpable; the promise of iniparib, combined with gemcitabine and carboplatin, was not realized in the large phase III trial performed to confirm the strikingly positive data from the randomized phase II trial. In a twist of coincidence, the data from the phase II trial were published in the New England Journal of Medicine 2 weeks before a press release announced that the phase II study had not met its primary end points.

Hope S. Rugo

The detailed data, presented at ASCO this year, supports a number of hypotheses that will likely lead to a better understanding of triple-negative disease, and more appropriate use of targeted agents in the long run. The trial, randomizing more than 500 women with metastatic triple-negative breast cancer, had two primary end points with shared P values. Although progression-free survival was improved in patients receiving the combination chemotherapy with iniparib, compared with those receiving chemotherapy alone, the difference did not meet the prespecified criteria for statistical significance.

A breakdown of the data into first or greater line of therapy demonstrated improved outcome only in patients treated in the second- and third-line settings. This does not make much sense in terms of what we understand about tumor biology, and what we see in clinic each day. A multivariate analysis failed to identify specific factors that predicted response or resistance, such as time from initial diagnosis to diagnosis of metastatic disease.

Although the addition of iniparib did not add to the toxicity of the chemotherapy combination, the chemotherapy itself led to significant cytopenias. Nonetheless, gemcitabine and carboplatin is a reasonably active regimen in this patient population, although not clearly superior to other standard therapies.

What happened? How could we see such remarkable data in a 120-patient phase II trial, and then not be able to duplicate it in a confirmatory, much-larger trial? The most likely hypothesis is that the definition of the disease subgroup was faulty. In other words, triple-negative disease is not really a biologically identifiable disease subset, but instead actually represents a very heterogeneous group of diseases with differential response to therapy. This suggestion is further encouraged by preliminary data suggesting significant variability in intrinsic disease subtype in the study population, with only 20% of the patients having basal-like disease and an unknown number with the highly resistant claudin-low subtype. These data are currently being confirmed and will be critical for planning future trials.

Several other considerations include the mechanism of action of iniparib, and its dosing schedule. Although initially thought to be a PARP inhibitor, it is now clear that at clinical doses, iniparib is not inhibiting PARP, although it clearly has a number of biologic effects. Its target may indeed be a pathway more specific to a subtype of triple-negative disease. These agents have a short half-life and, with little toxicity, it may be that the dose and schedule of iniparib could be further optimized.

Where do we go from here? A number of agents that do inhibit PARP are in clinical trials, both in sporadic triple-negative disease and in cancer associated with BRCA mutations. Further studies with iniparib are planned, and a nonrandomized neoadjuvant study will present data later this year. Clearly, we need to understand what disease we are actually targeting, as we move forward in this complex area of study; this will require tissue and intense collaboration. New targets and new agents are under active study as well.

<- Back to: What Matters