CTNI-07. PHASE I STUDY OF PAXALISIB AND RADIOTHERAPY FOR CNS DISEASE HARBORING PI3K PATHWAY MUTATIONS: PILOT ANALYSIS OF CIRCULATING TUMOR DNA FOR PATIENT ELIGIBILITY CONFIRMATION AND POST-TREATMENT RESPONSE

Abstract

Abstract INTRODUCTION Radiotherapy (RT) is effective for brain metastases (BM) or leptomeningeal disease (LMD), but disease control is poor for tumors harboring PI3K pathway alterations (PI3Kalt). We hypothesized that combining RT with paxalisib, a CNS-penetrant PI3K/mTOR inhibitor, offers therapeutic synergy. The role of plasma circulating tumor DNA (ctDNA) to confirm study eligibility and response monitoring is unclear. METHODS This is a multi-institutional, phase I trial of concurrent paxalisib and whole brain RT (30 Gy in 10 fractions, NCT04192981) for CNS disease. To establish eligibility, patients had documented PI3Kalt performed using institutional tumor targeted sequencing (MSK-IMPACT) of primary or metastatic tumors (eligibility sequencing). Plasma ctDNA was collected at baseline and 3-, 6- and 9-months post-treatment and sequenced using institutional plasma ctDNA targeted sequencing platform (MSK-ACCESS). Our objectives were to a) determine PI3Kalt concordance between eligibility sequencing and baseline ctDNA and b) analyze quantitative post-treatment plasma ctDNA dynamics. RESULTS Between 3/2020-5/2023, 15 patients were enrolled to receive daily paxalisib (45mg, n = 11 or 60mg, n = 4) and WBRT. Twelve were treated for BM and 3 for LMD. Eligibility sequencing was determined a median of 8.1months (range: 2.4-38.6) pre-study enrollment and the most common PI3Kalt were PIK3CA (n = 7), INPP4B (n = 2), INPPL1 (n = 1) and PIK3CG (n = 1), and PTEN copy number alterations (homdel n = 2). Baseline ctDNA was sequenced from 13/15 patients (18 samples: baseline = 13, 3-months = 4 and 6-months = 1). We confirmed 100% of PIK3CA mutations using baseline ctDNA. One breast cancer BM patient demonstrated significant variant allele frequency reduction from 50% (baseline) to 0.1% (3-months post-treatment) for two tracked PIK3CA mutations (p.E545K and p.A1066V), which correlated with a partial response per RANO-BM. CONCLUSIONS For patients on genomically-driven targeted therapy and RT combination trials, alterations detected using plasma ctDNA, especially PIK3CA, may be used to rapidly assess eligibility. Serial ctDNA collection may serve as a potential response biomarker with further validation.