Determination of an optimal response cut-off able to predict progression-free survival in patients with well-differentiated advanced pancreatic neuroendocrine tumours treated with sunitinib: an alternative to the current RECIST-defined response

BACKGROUND: Sunitinib prolongs progression-free survival (PFS) in patients with advanced pancreatic neuroendocrine tumours (pNET). Response Evaluation Criteria in Solid Tumors (RECIST)-defined partial responses (PR; classically defined as ⩾30% size decrease from baseline) are infrequent. METHODS: Individual data of pNET patients from the phase II [NCT00056693] and pivotal phase III [NCT00428597] trials of sunitinib were analysed in this investigator-initiated, post hoc study. The primary objective was to determine the optimal RECIST (v.1.0) response cut-off value to identify patients who were progression-free at 11 months (median PFS in phase III trial); and the most informative time-point (highest area under the curve (AUC) by receiver operating characteristic (ROC) analysis and logistic regression) for prediction of benefit (PFS) from sunitinib. RESULTS: Data for 237 patients (85 placebo; 152 sunitinib (n=66.50 mg ‘4-weeks on/2-weeks off’ schedule; n=86 ‘37.5 mg continuous daily dosing (CDD)’)) and 788 scans were analysed. The median PFS for sunitinib and placebo were 9.3 months (95% CI 7.6–12.2) and 5.4 months (95% CI 3.5–6.01), respectively (hazard ratio (HR) 0.43 (95% CI 0.29–0.62); P<0.001). A PR was seen in 19 patients (13%) on sunitinib; the median change in the sum of the lesions (vs baseline) was −12.8% (range −100 to +36.4). Month 7 was the most informative time-point (AUC 0.78 (95% CI 0.66–0.9); odds ratio 1.05 (95% CI 1.01–1.08), P=0.002). Reduction of 10% (vs baseline) achieved the highest sensitivity (50%) and specificity (82%), amongst cut-offs tested. A 10% reduction in marker lesions was associated with improved PFS in the whole sunitinib population (HR 0.55 (95 CI 0.3–0.9); P=0.04); mostly in patients on sunitinib CDD (HR 0.33 (95% CI 0.2–0.7); P=0.005). A 10% reduction in marker lesions (P=0.034) and sunitinib treatment (P=0.012) independently impacted on PFS (multivariable analysis). CONCLUSIONS: A 10% reduction within marker lesions identifies pNET patients benefiting from sunitinib treatment with implications for maintenance of dose intensity and future trial design.