Phase I trial to determine the safety and efficacy of paclitaxel, sapanisertib, and serabelisib. Patients with previously treated advanced solid tumors were eligible for this open label, cohort study of sapanisertib (TAK-228) and serabelisib (TAK-117) with weekly paclitaxel. A traditional 3 + 3 dose escalation design with 5 dosing cohorts was used. Patient reported outcomes were also evaluated. 19 heavily pretreated patients were enrolled (10 ovarian, 3 breast, and 6 endometrial cancers). All patients received comprehensive genomic profiling prior to enrollment. RP2D is sapanisertib 3 or 4 mg, serabelisib 200 mg on days 2-4, 9-11, 16-18 and 23-25 with paclitaxel 80 mg/m2 on days 1, 8 and 15 every 28 days. All patients in Cohort 5 required dose reductions and one patient experienced a DLT. The most frequent grade 3 or 4 adverse events were decreased WBCs (20%), nonfebrile neutropenia (12%), anemia (9%), elevated liver enzymes (4%), and hyperglycemia (11%). 3 patients had a CR, 4 had a PR, and 4 patients had SD > six months. ORR was 47% and CBR was 73% in 15 evaluable patients. Including all 19 enrolled patients, the PFS was 11 months and OS is still ongoing at 17 months. The combination of sapanisertib, serabelisib, and paclitaxel was safe and generally well tolerated. Preliminary efficacy was remarkable in an area of unmet need, especially for patient with PI3K/AKT/mTOR pathway aberrations. Positive effects and sustained clinical benefit were even seen in patients that were refractory to platinum and had failed taxane, everolimus, or temsirolimus. ClinicalTrials.gov, NCT03154294.

Breast cancer affects 12% of females in the United States and is the leading cause of cancer death in the female population. Personalized therapy is being used in clinical practice to treat breast cancer based on tumor molecular profiling, which can be obtained from tissue biopsy or plasma liquid biopsy as circulating tumor deoxyribonucleic acid (ctDNA). The available ctDNA tests provide a non-invasive way to monitor the cancer genome in a real-time manner. In this case report, a 38-year-old female with recurrent estrogen receptor (ER) positive breast cancer is treated with letrozole, everolimus, and palbociclib. The drugs target the hormonal signaling pathway, phosphoinositide 3-kinase (PI3K)-RAC-alpha serine/threonine-protein kinase (AKT) pathway, and cyclin D1 (CCND1)-CDK4/6 pathway, based on the patient’s estrogen-receptor-positive (ER+) disease and phosphatidylinositol -4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) mutation, as well as PIK3CA and CCND1 amplification. After 11 months of treatment, retinoblastoma protein transcriptional corepressor 1 (RB1) mutation was caught in ctDNA, which suggests an acquired resistance to palbociclib. Pazopanib was then used instead of palbociclib, targeting the fibroblast growth factor 3/4/19 (FGF3/4/19) amplification that was initially observed in her molecular profiling. Trametinib was also suggested recently due to the increasing allele frequency of B-Raf proto-oncogene, serine/threonine kinase (BRAF) mutation in ctDNA, following the treatment of letrozole + everolimus + pazopanib. The patient has no evidence of disease after five months of treatment initiation and has remained disease-free for over 16 months. In conclusion, the analysis of ctDNA is an effective way to monitor the real-time changes in a patient's tumor genome, which is a great supplement to the molecular profile from the tissue biopsy. The combination of these two tests provides an efficient strategy to make more informed treatment decisions, which greatly adapt along disease development.

2512 Background: Precision medicine has been studied in patients (pts) with advanced, heavily-treated cancers by administering molecularly matched monotherapies. With increasing availability of large gene assays and cognate agents, we hypothesized that offering customized combination therapies to treatment-naïve tumors would be feasible and improve response rates. Methods: Investigation of Profile-Related Evidence Determining Individualized Cancer Therapy (I-PREDICT, NCT02534675) targeted metastatic and/or unresectable, untreated lethal cancers ( > 50% 2-yr mortality). Comprehensive genomic profiling (CGP, Foundation Medicine; 315 genes), and, if possible, PD-(L)1 IHC, tumor mutational burden (TMB) and circulating tumor DNA were performed. A molecular tumor board discussed results immediately upon receipt, and emphasized customized combinations. Final decisions were the treating physician’s choice. Results: CGP was evaluable in 40/47 treatment-naïve pts (85.1%); 22 (46.8%) were treated [17 matched (36.2%); 5 unmatched (10.6%); 11 different diagnoses). The other 25 pts (53.2%) are awaiting therapy (8, 17%) or could not be treated (17, 36.2%), mainly due to patient deterioration or payor limitations. Each tumor had a unique genomic portfolio. The median (range) of genomic alterations/patient was 5 (1-12). TMB was available in 17 pts (12 low; 4 intermediate; 1 high). The median (range) Matching Score [(matches (#)/characterized genomic alterations (#)] was 33% (14-100%; 100% designated immunotherapy match or all alterations matched to targeted agents) [Reference PMID 2719717]. Nine/17 matched pts (53%) achieved SD > 6 months (N = 2) or CR (1)/PR (6). The median progression-free survival (PFS) for matched vs. unmatched pts was 4.7 vs. 1.0 months ( P= 0.0019). There were no drug-related deaths. Conclusions: With the use of broad-based DNA sequencing assays, inclusion of pts earlier in their disease course, timely mandated molecular tumor board discussions, and increasing availability of cognate drugs for customized combinations, we report: 1) high molecular matching rates (~36%); 2) high rates of SD > 6 months/CR/PR (~53%); and 3) improved PFS. Study expansion is ongoing. Clinical trial information: NCT0253467.

Objectives: Phase 1 trial to determine the safety and tolerability of everolimus and niraparib in patients with advanced ovarian and breast malignancies.Results: Fourteen heavily pretreated patients were enrolled (12 high‐grade serous ovarian cancer, 1 clear cell ovarian cancer, and 1 triple negative breast cancer). All patients were PARP naïve and received comprehensive genomic profiling prior to enrollment. Two DLTs were experienced in cohort 2 (niraparib 200 mg daily and everolimus 5 mg 3 days per week) with one patient experiencing prolonged thrombocytopenia and the other experiencing severe hypertension. Four additional patients were enrolled after dose de‐escalation with one patient again experiencing severe hypertension leading to conclusion of the study. The most frequent grade 3 or greater adverse events were thrombocytopenia, hypertension, anemia, fatigue, neutropenia, and elevated alkaline phosphatase. Two patients had a PR and five patients had SD. ORR was 18% and the CBR was 45% in 11 evaluable patients. Median PFS was 6 months, and median OS is approximately 18 months with three patients still alive at the data cutoff.Conclusions: The combination of everolimus and niraparib demonstrated significant toxicity at lower doses and is not feasible due to rapid onset and severe hypertension. This limitation possibly blunted the efficacy of the combination as PFS was modest, but OS was surprisingly robust due to three patients with ovarian cancer remaining alive with platinum refractory disease. Further investigation of multiagent blockade of the PI3K pathway combined with PARP is warranted.

Personalized treatment vs. standard of care is much debated, especially in clinical practice. Here we investigated whether overall survival differences in metastatic colorectal cancer patients are explained by tumor mutation profiles or by treatment differences in real clinical practice. Our retrospective study of metastatic colorectal cancer patients of confirmed European ancestry comprised 54 Americans and 54 gender-matched Germans. The Americans received standard of care, and on treatment failure, 35 patients received individualized treatments. The German patients received standard of care only. Tumor mutations, tumor mutation burden and microsatellite status were identified by using the FoundationOne assay or the IDT Pan-Cancer assay. High-risk patients were identified according to the mutational classification by Schell and colleagues. Results: Kaplan–Meier estimates show the high-risk patients to survive 16 months longer under individualized treatments than those under only standard of care, in the median (p < 0.001). Tumor mutation profiles stratify patients by risk groups but not by country. Conclusions: High-risk patients appear to survive significantly longer (p < 0.001) if they receive individualized treatments after the exhaustion of standard of care treatments. Secondly, the tumor mutation landscape in Americans and Germans is congruent and thus warrants the transatlantic exchange of successful treatment protocols and the harmonization of guidelines.