Modified IPCW model: A method for adjusting for bias in the estimation of overall survival due to the use of second and third line therapies in locally advanced or metastatic pancreatic cancer patients.

Abstract

e15787 Background: Nimotuzumab, a unique and affinity differentiated anti-EGFR antibody had been used in combination with gemcitabine on the treatment of pancreatic cancer patients. The aim of the study was to evaluate overall survival. Methods: Patients with newly diagnosed, locally advanced or metastatic pancreatic cancer, KPS ≥ 70 %, 18-72 years old, with adequate renal and liver function were included. Pts received gemcitabine 1000 mg/m2and nimotuzumab or placebo fixed dose of 400 mg once a wk, for 3 wks, followed by a 1-wk rest (d1, 8, 15, q28) until disease progression or unacceptable toxicity. The primary endpoint was OS and secondary PFS, ORR, CBR, safety and QoL. For OS determination, a KM log-rank test was used and a modified IPCW with a cox regression as a secondary analysis. On this evaluation using a modified IPCW model, 41.7% of pts from treatment arm and 42.7% from control arm who received 2nd and 3rd line treatment were censored after progression, while pts that did not receive 2nd and 3rd line treatment were weighted to compensate for the bias created by censoring switchers to 2nd and 3rd line treatment. Results: 192 pancreatic cancer pts were recruited. Ninety-six pts (62 male and 34 female) with a median age of 67 years, range (31, 83) were randomized to treatment arm and 96 pts (57 male and 39 female) with a median age of 64 years, range (41, 82) were randomized to control arm. In the primary analysis, median OS [95% CI] in the treatment arm was 8.57 mo [5.93, 10.90] vs 6.03 mo [4.97, 7.60] in the control arm. The HR [95% CI], 0.83 [0.62, 1.12] and p = 0.23 and when a modified IPCW model as a secondary analysis was used to remove the effect of 2nd and 3rd line therapies, the median OS was statistically significant with a HR [95% CI], 0.81 [0.67, 0.98] and a p = 0.030. The median PFS [95% CI] was 4.43 mo [3.67, 6.00] in the treatment arm vs 3.47 mo [2.60, 4.03] in the control arm with a HR [95% CI] 0.68 [0.51, 0.92] and p = 0.012. Conclusions: A modified IPCW model had proven that addition of nimotuzumab to gemcitabine increases median overall survival of newly diagnosed chemotherapy-naïve locally advanced or metastatic pancreatic cancer patients. Clinical trial information: NCT00561990.