The Importance of Context

Abstract

1237 Journal of Thoracic Oncology ® • Volume 10, Number 9, September 2015 Therapeutic strategies directed against defined molecular subsets of non-small cell lung cancer (NSCLC) has underscored the transformative potential of biomarker directed drug development in oncogene-driven solid tumors. Riding on the success of targeting epidermal growth factor receptor (EGFR) mutations and anaplastic lymphoma kinase (ALK) rearrangements, small molecule inhibitors of the phosphatidyl-inositol 3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) pathway have engendered much excitement due to its critical role as a central signaling axis downstream of several receptor tyrosine kinases implicated in NSCLC (e.g., EGFR and MET). Although published data of novel agents targeting the PI3K/AKT/mTOR pathway in NSCLC have only demonstrated modest clinical activity to date (Table 1), the majority of studies have been conducted in unselected patient populations. In this issue of the journal, Vansteenkiste et al report the final efficacy results from a phase II trial of buparlisib, a pan-class I PI3K inhibitor in selected “PI3K-activated” NSCLC patients. In a commendable effort, 1242 patient samples underwent molecular prescreening for PIK3CA mutations, PTEN mutations, or PTEN loss by immunohistochemistry, identifying 63 subjects (13.5%) that were enrolled into two cohorts (squamous and nonsquamous). In terms of efficacy analysis, this trial used a conservative futility criterion—defined as 12-week progression-free survival rate less than 50%—that was unfortunately met in the first stage, terminating the trial early. Despite being a negative study from an efficacy standpoint, this trial has provided unique insights into the current challenges with targeted therapies directed against genetic alterations. Underpinning the success of contemporary trials evaluating targeted therapeutics are patient enrichment strategies using high-precision predictive biomarkers. Are PTEN and PIK3CA alterations sufficiently specific to select for tumors most likely to respond to PI3K inhibitors? Preclinical data certainly support this hypothesis, where PIK3CA-mutant transgenic lung cancer mouse models clearly respond to PI3K/mTOR inhibitors, and conditional ablation of PIK3CA or PIK3CB (through which PTEN signals) in cell-based models result in a significant reduction in cell growth. Yet, despite selecting for the relevant alterations in this trial, tumor responses were disappointingly low at 3% in both nonsquamous and squamous cell carcinoma, and there was no difference in progressionfree survival regardless of whether patients’ tumors were PIK3CA mutant or wild type. Furthermore, a subgroup of 23 patients with sufficient archival tissue subjected to broader sequencing of 600 genes failed to yield any notable coalterations that may have limited the efficacy of PI3K inhibitors. Similarly, responders were not differentiated by a specific mutational profile, with one patient harboring PTEN loss in the squamous group while another patient in the nonsquamous group had a PIK3CA and a KRAS mutation. Thus, somatic alterations alone cannot explain response or resistance to buparlisib.