Patients with BRCA mutations comprise a relatively small fraction of breast cancer cases, and the economic calculus of drug development likely plays a role with respect to PARP inhibitor development in this population. Attempts are underway to expand the target population beyond mutation carriers with efforts to target BRCA-ness in sporadic tumors.¹¹ The majority of BRCA1-associated breast cancers are of the basal-like intrinsic subtype, as are the majority of triple-negative breast cancers (TNBCs, which are estrogen receptor negative, progesterone receptor negative, and human epidermal growth factor receptor 2 negative). BRCA1 seems to be silenced in a subset of sporadic TNBCs, either by promoter hypermethylation or by other means, and there are suggestions that other TNBCs may have defects in homology-directed repair that may not be directly related to BRCA1 deficiency. These findings have generated considerable interest in developing PARP inhibitors for TNBC.¹² A randomized phase II study of carboplatin with gemcitabine chemotherapy compared with the doublet with iniparib was conducted in unselected TNBC with strikingly positive results, including an overall survival benefit.¹³ On the basis of these promising data, a phase III study of nearly identical design quickly followed. Unfortunately, this study failed to meet its primary objective of improvement in the combined end point of progression-free and overall survival.¹⁴ This demonstrates that attempts to expand the role of PARP inhibitors beyond BRCA1/2 mutation–associated cancers are not guaranteed to be successful, no matter how compelling the preclinical rationale. Data regarding germline BRCA1/2 status are not available at this time to determine whether mutation carriers preferentially benefited from the addition of iniparib in this study. One must note, however, that iniparib was recently reported to be a poor inhibitor of PARP1 and PARP2; therefore, this trial may not be an adequate test of PARP inhibition as a therapeutic strategy in nonhereditary TNBC.

Although proportionally small, BRCA mutation carriers still constitute a significant absolute number of patients worldwide who could benefit from these agents, particularly if they came to be used in adjuvant treatment. A targeted therapeutic agent would presumably lead to widespread BRCA testing at diagnosis to identify those who would benefit, similar to current hormone receptor and human epidermal growth factor receptor 2 tumor analysis for patients with breast cancer, and thus would additionally expand the potential number of patients who are eligible for treatment. A number of agents are approved for numerically small populations. For example, imatinib is approved in the United States for a combined population of approximately 9,000 new cases per year of GI stromal tumors, chronic myelogenous leukemia, and dermatofibrosarcoma protuberans. Only 3% of non–small-cell lung cancers are ALK mutation positive, and yet crizotinib is being rapidly developed. Vandetanib has been approved for medullary thyroid cancer, another uncommon disease. Therefore, one hopes that population size will not be an insurmountable barrier to drug development; if it is, then this obstacle will plague the development of many targeted therapies.

If one accepts that individuals with BRCA1/2 mutation–associated cancers are the most logical target population for PARP inhibitors, one must still grapple with the seemingly simplistic question of how, from a regulatory standpoint, one defines the population. In the United States, the US Food and Drug Administration requires an approved companion diagnostic test that will define the population of interest before approval is granted for an agent directed toward that population. There is presently no US Food and Drug Administration– approved diagnostic test for determining germline BRCA status, although mutation results have been used for more than a decade to make major decisions about preventive surgeries. The regulatory approval of such a diagnostic is hampered by ongoing uncertainty with respect to the status of the BRCA testing patent held by Myriad Genetics (Salt Lake City, UT). To facilitate development of molecularly targeted therapies, regulatory authorities worldwide should provide clear clinical guidance with respect to the requirements for diagnostics to be used in this setting.

A number of challenges confront those designing studies to obtain regulatory approval for targeted therapies. Of particular concern is selection of an appropriate comparator. Because standard alternative treatments exist for patients with BRCA mutation–associated cancers, the current so-called orphan drug mechanism is likely not appropriate. In addition, there have been few treatment trials for women with BRCA-associated breast or ovarian cancers, so there are no data on the performance of standard treatments in this specific population. For the most successful targeted therapeutics, such as imatinib in chronic myelogenous leukemia and GI stromal tumors and PLX4032 (vemurafenib) in melanoma, the standard treatment comparator had limited activity. In breast and ovarian cancer, of course, there are many more active agents. The challenge is exacerbated if regulatory approval requires an overall survival advantage compared with standard treatment. Cross-over designs are likely to confound survival end points, but the lack of a cross-over may be poorly received by study participants, advocates, and some physicians. Noninferiority designs that incorporate quality of life end points may be an option, but are less attractive if they require a larger sample size. Recruitment to large phase III studies of specialized populations with numerous standard options will require international studies with enthusiastic support from patients, advocates, and investigators. Such studies are expensive and may take longer to complete than is desirable. For all of these reasons, it would be beneficial to develop alternative regulatory pathways that employ a more refined approach than basing approval on an overall survival advantage in a phase III study with a conventional cytotoxic comparator.

Pending regulatory approval, what are physicians and patients to do about targeted agents that manifested enormous promise in early development and generated hope and enthusiasm, but have run afoul of the challenges described here? PARP inhibitor development in BRCA mutation–associated breast cancer has slowed, with no clear path to approval. There is particularly limited availability of PARP inhibitor studies for mutation carriers who have estrogen receptor–positive breast cancer. Compassionate use and extended access programs are attractive to patients, but the management and funding of such programs present challenges, not the least of which is that they may undermine the ability to conduct trials for regulatory approval. However, the current lack of access, especially after the loud and promising publicity that greeted the initial PARP trials, has the potential to undermine public faith in the concept of personalized drug development for breast cancer. This is particularly true if the challenges described are not communicated and the perception takes root that the lack of development is a result of only economic considerations.

The PARP inhibitor story is in many ways a cautionary tale for personalized medicine. To proceed requires commitment and collaboration from academics, pharmaceutical companies, and the regulatory agencies. One point is clear: we cannot allow the challenges delineated here to halt development of these compounds.