The required number of treatment imaging days for an effective off-line correction of systematic errors in conformal radiotherapy of prostate cancer--a radiobiological analysis.

BACKGROUND AND PURPOSE: To use radiobiological modelling to estimate the number of initial days of treatment imaging required to gain most of the benefit from off-line correction of systematic errors in the conformal radiation therapy of prostate cancer.

MATERIALS AND METHODS: Treatment plans based on the anatomical information of a representative patient were generated assuming that the patient is treated with a multi leaf collimator (MLC) four-field technique and a total isocentre dose of 72 Gy delivered in 36 daily fractions. Target position variations between fractions were simulated from standard deviations of measured data found in the literature. Off-line correction of systematic errors was assumed to be performed only once based on the measured errors during the initial days of treatment. The tumour control probability (TCP) was calculated using the Webb and Nahum model.

RESULTS: Simulation of daily variations in the target position predicted a marked reduction in TCP if the planning target volume (PTV) margin was smaller than 4 mm (TCP decreased by 3.4% for 2 mm margin). The systematic components of target position variations had greater effect on the TCP than the random components. Off-line correction of estimated systematic errors reduced the decrease in TCP due to target daily displacements, nevertheless, the resulting TCP levels for small margins were still less than the TCP level obtained with the use of an adequate PTV margin of approximately 10 mm. The magnitude of gain in TCP expected from the correction depended on the number of treatment imaging days used for the correction and the PTV margin applied. Gains of 2.5% in TCP were estimated from correction of systematic errors performed after 6 initial days of treatment imaging for a 2 mm PTV margin. The effect of various possible magnitudes of systematic and random components on the gain in TCP expected from correction and on the number of imaging days required was also investigated.

CONCLUSIONS: Daily variations of target position markedly reduced the TCP if small margins were used. Off-line correction of systematic errors can only partly compensate for these TCP reductions. The adequate number of treatment imaging days required for systematic error correction depends on the magnitude of the random component compared with the systematic component, and on the size of PTV margin used. For random components equal to or smaller than the systematic component, 3 consecutive treatment imaging days are estimated to be sufficient to gain most of the benefit from correction for current clinically used margins (6-10 mm); otherwise more days are required.