Predictive factors of a complete response to and adverse effects of a CDDP-5FU combination as primary therapy for head and neck aquamous carcinomas

Abstract

Retrospective analysis of detailed patient and tumour factors associated with a complete response to combination inductive chemotherapy with CDDP-5FU (96 or 120 hour continuous infusion) was performed using data from 147 patients with a previously untreated squamous cell carcinoma of the oral cavity, oropharynx or pharyngo-larynx following completion of two (29 patients) or three (118 patients) cycles. Adverse reactions to chemotherapy were documented for all 164 patients included in the study. Eight drug-related deaths occurred due to: acute myocardial infarction (five patients), peptic ulcer disease (two patients) and severe neutropenia with sepsis (one patient). Severe non-lethal complications included marrow depletion (14 patients), peptic ulcer (two patients), thrombophlebitis (seven patients), angina pectoris (two patients), stroke (one patient), pulmonary oedema (one patient) and convulsions (one patient). Six patients refused further treatment because of untoward side effects and tumoral progression was observed in three cases. Separate response rates for the primary site and nodes were determined and analysis of respective predictive factors of response was performed. Complete response was obtained in 31 per cent at the primary site versus 18 per cent for the nodes (p<0.05). The combined (primary site + nodes) overall complete response rate was 22 per cent. Among 11 factors studied (age, sex, performance status, primary site, tumour differentiation, initial resectability, 5FU dosage per cycle, number of cycles, T, N and TN stages), only performance status, N stage, resectability and number of cycles were associated with a combined complete response. Multivariate analysis showed performance status, N stage, TN stage and resectability to be significant predictive factors of a combined complete response. Optimum use of inductive chemotherapy will ultimately rely upon a better knowledge of predictive factors in order to avoid drug-related complications in patients not expected to respond.