The predictive value of tumor regression rates during chemoradiation therapy in patients with advanced head and neck squamous cell carcinoma.

OBJECTIVE: The value of tumor regression rates in predicting survival outcome during chemoradiation therapy was prospectively evaluated. METHODS AND MATERIALS: Sixty-two patients diagnosed with locally advanced stage III/IV unresectable head and neck squamous cell carcinoma underwent weekly clinical and endoscopic serial assessment of primary and nodal tumor sizes during chemoradiation therapy between July 1993 and September 1995. Chemoradiation therapy consisted of protocol treatment using supradose intra-arterial targeted cisplatin (SIT-P) at 150 mg/m2 four times at weekly intervals along with intravenous sodium thiosulfate at 9 g/m2 and concurrent conventionally fractionated radiotherapy at 1.8 to 2.0 Gy/fraction (fx) to a total dose of 68 to 74 Gy. Tumor reduction was serially measured as a percentage of the original pretreatment size at weekly intervals by the same team of surgical and radiation oncologists. Correlations were then made between tumor regression rates and survival.
RESULTS: Complete or near complete regression of disease during chemoradiation therapy as compared with nonresponsive/partially responsive disease was associated with better survival outcome (P = 0.001 and P = 0.013, respectively). Among patients exhibiting complete or near complete regression of disease, rapid tumor reduction (median = 4.2 weeks) was associated with inferior survival outcome when compared with slower disease regression (median = 6.4 weeks, P = 0.007).
CONCLUSIONS: Our findings fail to support the "traditional" hypothesis that rapid tumor regression during treatment is predictive of an improved survival outcome. Treatment strategies that alter ongoing therapy based upon initial tumor regression rates should be avoided.