Infusion-related reactions (IRR) during NGR-hTNF therapy as potential predictors of clinical outcome

e13593 Background: NGR-hTNF, a selective antivascular agent, transiently modulates the systemic release of cytokines/chemokines. Intravenous infusion of NGR-hTNF is distinguished by an early on-target effect, IRR, which mostly consists of short-lived grade 1/2 chills. Methods: Incidence, predictors of development and relationships of IRR with treatment outcome were assessed in a pooled analysis of individual patient data from 5 phase II trials with NGR-hTNF. Global data set comprised 205 previously treated patients (pts) who had received NGR-hTNF 0.8 µg/m2 every 3 weeks (q3w) or weekly (q1w) alone in mesothelioma, colon and liver cancers or in combination with doxorubicin in small cell lung and ovarian cancers. In all trials, response to therapy (RECIST) was measured q6w until progression. Regression models estimated the effect size of IRR on response rate (RR; complete and partial response), disease control rate (DCR; RR plus stable disease) and progression free survival (PFS). Results: 137/205 pts (67%) experienced IRR on treatment, while 68 pts (33%) did not. In the IRR group, 63% of pts had grade 1 and 37% grade 2. By time to onset, 88% of pts developed IRR during first 6 weeks of therapy and 12% later. According to dosing schedule, IRR rates were higher with q1w than with q3w (85% vs 62%; p=.008). Baseline characteristics were (IRR vs non-IRR groups): median age: 66 vs 64; male: 50% vs 56%; PS ≥ 1: 34% vs 35%, range of prior lines: 1-4 vs 1-5. Among baseline characteristics, logistic regression analysis retained only a lower number of prior lines associated with higher IRR rates (OR=1.4; p=.02). Onset of IRR was related to better treatment outcome. RR was 12% in the IRR and 3% in the non-IRR group (OR=4.4; p=.05), while DCR was 50% in pts who had IRR and 34% in pts who did not (OR=2.0; p=.02). In pts with or without IRR, 6-month PFS was 21% vs 7% (HR=0.62; p=.001), respectively. Improvement in PFS time remained significant in a landmark analysis set at the 6-week time point (p=.004). On multivariate analyses adjusted for baseline variables, IRR independently predicted higher odds of tumor response (p=.05) and lower risk of progression or death (p=.009). Conclusions: Onset of IRR on NGR-hTNF treatment may predict subsequent clinical outcome. Clinical trial information: NCT00483080-NCT00483509-NCT00484211-NCT00484276-NCT00484432.