A case for the use of receiver operating characteristic analysis of potential clinical efficacy biomarkers in advanced renal cell carcinoma.

AIM: Assess patient-level utility of suggested pretreatment biomarkers of sunitinib in advanced renal cell carcinoma. PATIENTS & METHODS: Kaplan-Meier analysis of data from a randomized, Phase II study (n = 292) suggested baseline predictive value for circulating soluble Ang-2 and MMP-2 and HIF-1α percentage of tumor expression. Using this dataset, the sensitivity, specificity and area under the curve (AUC) were calculated, using receiver operating characteristic (ROC) curves. RESULTS: Based on a ROC (sensitivity vs 1 - specificity) threshold AUC value of >0.8, neither Ang-2 (0.67) nor MMP-2 (0.65), nor HIF-1α percentage of tumor expression (0.65), performed appropriately from a patient-selection standpoint. CONCLUSION: To properly assess potential biomarkers, sensitivity and specificity characteristics should be obtained by ROC analysis.

Genotype Correlations With Blood Pressure and Efficacy From a Randomized Phase III Trial of Second-Line Axitinib Versus Sorafenib in Metastatic Renal Cell Carcinoma.

BACKGROUND: In the phase III axitinib second-line (AXIS) trial, axitinib significantly prolonged progression-free survival (PFS) versus sorafenib in patients with previously treated metastatic renal cell carcinoma (mRCC). Analyses of associations between germline single-nucleotide polymorphisms (SNPs) and outcomes are reported. PATIENTS AND METHODS: DNA samples from blood were genotyped using TaqMan allelic discrimination. Logistic/Cox regression analyses were used to evaluate association of 15 SNPs in vascular endothelial growth factor (VEGF)-A, VEGF receptor (VEGFR)1, VEGFR2, or hypoxia-inducible factor (HIF)-1α with outcomes for blood pressure (BP; Grade ≥ 3 hypertension, diastolic BP > 90 mm Hg, and increase ≥ 15 mm Hg from baseline) and efficacy (independent review committee-assessed objective response rate and PFS, and overall survival [OS]). Multivariate analyses assessed SNPs and baseline characteristics as potential predictors of PFS and OS. RESULTS: Genotype data were available for 305 (42.7%) of 714 patients; 159 received axitinib and 146 sorafenib. After Bonferroni adjustment, no SNP was associated with BP outcomes. In axitinib-treated patients, VEGF-A rs699947 (A/A vs. C/C) and rs833061 (C/C vs. T/T) were associated with longer OS (27.0 vs. 13.4 months; hazard ratio [HR], 0.39; Padjusted = .015). In sorafenib-treated patients, VEGFR2 rs2071559 (G/G vs. A/A) was associated with longer OS (26.8 vs. 13.8 months; HR, 0.41; Padjusted = .030). In multivariate analyses, no SNP predicted axitinib efficacy; VEGFR2 rs2071559 predicted PFS (P = .0053) and OS (P = .0027) for sorafenib. Sensitivity/specificity of VEGFR2 rs2071559 for OS was < 80%. CONCLUSION: No SNP predicted axitinib outcomes. Although VEGFR2 rs2071559 predicted sorafenib efficacy in patients with mRCC, sensitivity/specificity limitations preclude its use for selecting individual patients for sorafenib treatment.

A Phase I, dose-finding study in patients with advanced solid malignancies of the oral γ-secretase inhibitor PF-03084014.

PURPOSE: To estimate the maximum tolerated dose (MTD) for continuous oral administration of the γ-secretase inhibitor PF-03084014, determine the recommended phase II dose (RP2D), and evaluate safety and preliminary activity in patients with advanced solid tumors. EXPERIMENTAL DESIGN: This open-label, phase I study consisted of a dose-finding portion based on a 3+3 design, followed by an expansion cohort. PF-03084014 was administered orally, twice daily (BID) for 21 continuous days. Tested doses ranged from 20 to 330 mg BID. In the expansion cohort, patients were to receive the estimated MTD or a lower dose of PF-03084014. RESULTS: A total of 64 patients received treatment. The MTD was estimated to be 220 mg BID. The RP2D was determined to be 150 mg BID, based on the better safety profile versus the 220-mg BID dose, given comparable NOTCH-related target inhibition. The most common treatment-related adverse events were diarrhea, nausea, fatigue, hypophosphatemia, vomiting, rash, and decreased appetite, which were generally mild to moderate in severity. One patient with advanced thyroid cancer had a complete response, and five of seven response-evaluable patients with desmoid tumor achieved a partial response (71.4% objective response rate). Tumor responses were mostly durable, ranging from 1.74+ to 24+ months. PF-03084014 demonstrated a generally dose-dependent pharmacokinetic profile at doses ranging from 20 to 330 mg BID. Consistent downmodulation of NOTCH-related HES4 gene expression was observed in peripheral blood from all evaluable patients. CONCLUSION: Further development of PF-03084014 for the treatment of patients with advanced solid tumors is warranted and currently under evaluation.

Investigation of novel circulating proteins, germ line single-nucleotide polymorphisms, and molecular tumor markers as potential efficacy biomarkers of first-line sunitinib therapy for advanced renal cell carcinoma.

PURPOSE: Sunitinib is a first-line advanced renal cell carcinoma (RCC) standard of care. In a randomized phase II trial comparing sunitinib treatment schedules, separate exploratory biomarker analyses investigated the correlations of efficacy with selected serum, germ line single-nucleotide polymorphism (SNP), or tumor markers. METHODS: Advanced RCC patients received first-line sunitinib 50 mg/day on the approved 4-week-on-2-week-off schedule (n = 146) or 37.5 mg/day continuous dosing (n = 146). The following correlation analyses were performed: (1) response evaluation criteria in solid tumors-defined tumor response with serum soluble protein levels via two distinct multiplex (n < 1,000) platforms; (2) response and time-to-event outcomes with germ line SNPs in vascular endothelial growth factor (VEGF)-A and VEGF receptor (VEGFR)3 genes; and (3) response and time-to-event outcomes with tumor immunohistochemistry status for hypoxia-inducible factor 1-alpha (HIF-1α) and carbonic anhydrase-IX or tumor Von Hippel-Lindau (VHL) gene inactivation status. RESULTS: Lower baseline angiopoietin-2 (Ang-2) and higher baseline matrix metalloproteinase-2 (MMP-2) levels were identified by both platforms as statistically significantly associated with tumor response. There were no significant correlations between VEGF-A or VEGFR3 SNPs and outcomes. Progression-free survival was longer for HIF-1α percent of tumor expression groups 0-2 (HIF-1α low) versus 3-4 (HIF-1α high; p = 0.034). There were no significant correlations between outcomes and each VHL inactivation mechanism [mutation (86% of VHL-inactive patients), methylation (14%), and large deletion (7%)] or mechanisms combined. CONCLUSIONS: Serum Ang-2 and MMP-2 and tumor HIF-1α were identified as relevant baseline biomarkers of sunitinib activity in advanced RCC, warranting further research into their prognostic versus predictive value.

Phase I, open-label, multicentre, dose-escalation, pharmacokinetic and pharmacodynamic trial of the oral aurora kinase inhibitor PF-03814735 in advanced solid tumours.

This phase I study (ClinicalTrials.gov ID: NCT00424632) evaluated the safe dose, pharmacokinetics, and pharmacodynamics of the aurora kinase A and B inhibitor, PF-03814735. Patients with advanced solid tumours received oral, once-daily (QD) PF-03814735 on Schedule A: days 1-5 (5-100mg); or Schedule B: days 1-10 (40-60mg) of 21-day cycles. Fifty-seven patients were treated: 32 and 25 on Schedules A and B, respectively. Dose-limiting toxicities were: febrile neutropenia (Schedule A); and increased levels of aspartate amino transferase, left ventricular dysfunction, and prolonged low-grade neutropenia (Schedule B). Maximum tolerated doses were 80mg QD (Schedule A) and 50mg QD (Schedule B). Common treatment-related adverse events were mainly mild to moderate and included diarrhoea, fatigue, nausea, and vomiting. Nineteen patients achieved stable disease, which was prolonged in four cases. PF-03814735 was rapidly absorbed and demonstrated linear pharmacokinetics up to 100mg QD; mean terminal half-life ranged from 14.4 to 23.6h. Aurora B activity, assessed by histone H3 phosphorylation in mitotic cells, decreased in tumour tissue from 10/12 patients evaluated (range: -70% to -3%). (18)F-fluorodeoxyglucose positron emission tomography demonstrated metabolic responses in only 1/21 patients. PF-03814735 was generally well tolerated with manageable toxicities, and a recommended phase II dose could be established for both schedules. Aurora B activity was inhibited in tumour tissue, but clinical or metabolic antitumour activity was limited.