With interest we read the editorial of Robbins et al,1 which questioned the conclusion of a prospective randomized trial2 in which intra-arterial (IA) cisplatin chemoradiation was not found to be superior to the current standard of intravenous cisplatin chemoradiation.
Bilateral IA infusion was stricter than prescribed in Radiation Therapy Oncology Group (RTOG) trial 9615 (RTOG-9615). If the tumor exceeded >1 cm over the anatomical midline, double-sided infusion according to the flow ratio between the 2 feeding arteries was recommended. The lower rate of double-sided infusions in earlier phase 2 investigations by Robbins et al may therefore be attributed to a different patient selection. The difference in mean tumor volume of 18 mL versus 30 mL between the 2 series is indicative in that respect.
Second, we performed an unplanned subset analysis and scored for all patients the eligibility of unilateral infusion and tumor volume. Local control was found to be significantly better in the IA versus the intravenous arm among patients whose tumor did not extend across the midline (hazard ratio, 0.17; 95% confidence interval [95% CI], 0.05-0.60 [P = .0025]), whereas it was nonsignificantly worse in the other patients (HR, 1.43; 95% CI, 0.66-3.09). The benefit from IA treatment was limited to 14 of the 26 patients with a large (>30 mL) tumor not extending across the midline (HR, 0.14; 95% CI, 0.03-0.72), whereas no significant benefit was observed in the remaining 3 groups.
Another concern of Robbins et al is the potential influence of human papillomavirus (HPV) infection. They speculate that benefit from IA versus intravenous treatment may be limited to the small subgroup of HPV-negative patients. In this case, even an analysis accounting for HPV status would likely not have been able to detect an overall difference. Unfortunately, HPV status was not determined within the trial, but because assays are currently being performed, we will eventually be able to evaluate whether HPV status is predictive of benefit from IA treatment.
Therefore, we conclude that the jury has come to a conclusion: for patients with advanced, inoperable head and neck cancer, IA is not superior to intravenous chemoradiation.