487P - Correlation of infusion-related reactions (IRR) and outcome in patients receiving NGR-HTNF treatment

Date 01 October 2012
Event ESMO Congress 2012
Session Poster presentation III
Topics Clinical Research
Complications/Toxicities of Treatment
Basic Scientific Principles
Presenter Gilda Maria Rossoni
Authors G.M. Rossoni1, V. Gregorc2, A. Bulotta1, M.G. Viganò2, G. Todisco2, A. Lambiase3, C. Bordignon3
  • 1IRCCS San Raffaele, IT-20132 - Milano/IT
  • 2Oncologia, IRCCS San Raffaele, IT-20132 - Milano/IT
  • 3Clinical Development, MolMed, 20132 - Milan/IT



NGR-hTNF (asn-gly-arg-human tumor necrosis factor) induces systemic release of cytokines and intratumoral infiltration of effector T cells. Intravenous infusion of NGR-hTNF is characterized by onset of IRR, mostly consisting of transient grade 1-2 chills. Incidence, predictors of development, and relationships with outcome of IRR were assessed across 5 phase 2 single-arm trials of NGR-hTNF.


205 patients (pts) with solid tumors received NGR-hTNF 0.8 µg/m2 every 3 weeks (q3w) given either alone in colon cancer (n = 45), liver cancer (n = 40), and mesothelioma (n = 55), or with doxorubicin in small-cell lung cancer (n = 28) and ovarian cancer (n = 37). Tumor assessment by RECIST was done q6w until progressive disease (PD). Logistic and Cox proportional-hazards regression models were used to analyze associations between IRR and outcome, in terms of response rate (RR, complete and partial response), disease control rate (DCR, rate of pts without PD at 6 weeks), and progression-free survival (PFS).


Overall, 137/205 pts (67%) experienced IRR on treatment, while 68 pts (33%) did not report IRR. In the IRR group, 63% of pts had grade 1 and 37% grade 2. By time of onset, 88% of pts developed IRRs within first 6 weeks and 12% later. Baseline characteristics in the IRR v non-IRR groups were: median age 66 v 64; male 50% v 56%; PS 1-2 34% v 35%, range of prior lines 1-4 v 1-5. Among baseline factors, only a low number of prior regimens predicted for IRR (odds ratio, OR 1.4 p = .02). The RR was 12% (95% CI 7-18) in the IRR group and 3% (1-10) in the non-IRR group (OR 4.4 p = .05), while DCR was 50% (42-59) in pts who had IRR and 34% (23-46) in pts who did not (OR 2.0 p = .02). On landmark analysis set at 6-week time point, PFS was significantly improved in pts with IRR (hazard ratio, HR 0.63 p = .007). Six-month PFS rates were 17% (11-23) for the IRR group and 6% (1-12) for the non-IRR group (log-rank p = .005). In multivariable analyses (adjusting for age, sex, PS, prior lines, and tumor types), the onset of IRR remained independent predictor of higher likelihood of response (OR 4.8 p = .05) and lower risk of progression or death (HR 0.62 p = .009).


Occurrence of IRR may identify pts who are more likely to gain benefit from NGR-hTNF treatment.


A. Lambiase: Employment- MolMed.

C. Bordignon: Employment - MolMed.

All other authors have declared no conflicts of interest.