Phase II study of pemetrexed (P) and gemcitabine (G) in patients with advanced head and neck cancer (SCCHN)

Abstract

6052 Background: Unresectable/metastatic squamous cell cancer of the head and neck (SCCHN) has a poor prognosis despite standard therapy with cisplatin. Given the substantial toxicities of cisplatin and its widespread use in combination with RT in LA- SCCHN, we investigated an alternative regimen of gemcitabine (G) and pemetrexed (P). A phase II study of pemetrexed in SCCHN had a response rate (RR) of 26.5%. A phase I study of the P and G combination showed the doses and schedule below to be feasible. Methods: Eligibility stipulated measurable, incurable, unresectable recurrent/metastatic SCCHN; Karnofsky PS of > 60%, and adequate bone marrow, renal and hepatic function. One prior therapy for recurrent disease was allowed. Patients were treated with P 500 mg/m2 and G 1,500 mg/m2 every 2 weeks. Each cycle was 28 days. Primary endpoint was RR. Secondary endpoints included overall survival and toxicity. Patients were re-imaged every 2 cycles. In a 2-step design, 16 patients were to be accrued during stage 1, and 16 were projected for stage 2. If 1 or fewer responses were observed during the first stage, then the trial was to be stopped and the regimen deemed insufficiently active. Results: 17 patients were enrolled in stage 1. Median age was 64 years (range, 49–82); 10 males; 13 smokers (including 4 current smokers). 15 patients previously received definitive therapy with surgery and/or chemo-RT, and 9 had prior systemic therapy for recurrent/metastatic disease. Median number of cycles administered was 2 (range, <1 to 10). Reasons for treatment discontinuation included progressive disease in 15 patients and decreased PS in 1. Grade 3 toxicities included neutropenia (1), anemia (3), thrombocytopenia (1), dysphagia (11), fatigue (4), hypertension (4), and hyperglycemia (4). 1 pt had grade 4 anemia. Best response: 1 confirmed PR, 1 unconfirmed PR, 3 SD, 10 PD, 2 not evaluable. 7 patients received subsequent treatment. Estimated median overall survival is 5.5 months. Conclusions: The combination of P and G for advanced HNSCC was well tolerated, but its clinical efficacy was marginal in an unselected population. Due to early stopping rules, the study was closed to accrual after stage 1 for lack of efficacy. Further study of this regimen should be considered in patients enriched for TS and RRM1 status. [Table: see text]