The tumor volume and clonogen number relationship: tumor control predictions based upon tumor volume estimates derived from computed tomography

Abstract

Purpose: While tumor volume is an important parameter predicting clinical outcome, its relationship to clonogen number remains uncertain. This uncertainty is related to many factors, among them treatment response heterogeneity, which obscures the influence of patients and treatment-related parameters. In this study, we analyze the effect of tumor volume on local and regional recurrence in a setting tightly controlled for dose, treatment time, and patient selection. The hypothesis that changes in clonogen number scale directly with changes in tumor volume is tested.

Methods and materials: Using digital reconstruction of diagnostic computed tomography (CT) scans, primary and total tumor volumes were estimated in 51 cases of advanced squamous cell carcinoma of the head and neck. All patients were managed with a concomitant boost accelerated superfractionated schedule to a median dose of 70.2 Gy. Clinical data were fitted to a mixture model to relate tumor volume parameters to control probability where volume and clonogen number were related by the relationship m = a.Vb, where m is initial clonogen number, a is a proportionality constant, V is tumor volume, and b is the volume exponent.

Results: Tumor volume estimates for primary tumor ranged from 3-196 cm3 and for total tumor volume 5-196 cm3. Actuarial local-regional control is 63%. The estimated volume exponent b is 0.85 (95%, confidence interval (c.i.): 0.40-1.29) for primary tumor volume and 1.1 (95%, c.i.: 0.33-1.85) for total tumor volume.

Conclusion: This study quantifies the adverse influence of tumor volume on local-regional disease control in advanced head and neck cancer. The derived volume exponent approximates to one, the theoretical expectation if the growth fraction is roughly constant and clonogen number increases linearly with volume. Finally, these results suggest that radiobiological parameters are more reliably estimated from clinical data with narrowly defined strata.