Chest-wall irradiation by kilovoltage x-rays: Observed local control rate versus predicted radiobiological tumor control probability

Introduction: The high local control rates observed in patients with breast cancer who had been treated by modified radical mastectomy, chemotherapy and chest-wall irradiation using 120 kV X-rays (50 Gy in 25 fractions) motivated us to investigate their observed tumor control probability (TCP) and test a well-established radiobiological TCP model against our data. Materials and Methods: To evaluate and compare the two TCPs (predicted vs. observed), we used the Royal Marsden TCP model and compared its results with our observed TCP, which was derived from the clinical data of 126 patients treated during a period of 10 years. To obtain typical dose distributions, we selected a representative sample of 10 patients’ CT images with different thicknesses and anatomies of the chest wall and computed the dose distributions by the MCNP Monte Carlo code using a validated model of the 120 kV x-ray tube. Then, the dose volume histogram for each patient’s chest wall target volume was computed and entered into the TCP model in the Biosuite software. Various breast cancer model parameters were then adjusted and the corresponding TCPs were recorded. Clonogenic cell density in the chest wall after mastectomy has not been reported and, therefore, we tested a wide range of values (1, 106 or 109 cells in the whole chest wall). Results: The observed TCP in patients was 80.7%. However, the dose distributions in the target volumes were inhomogeneous and large volumes of the targets received lower doses. The mean dose received by 50% of the target volume was 72% of the prescription. Assuming just one cell on the total volume of the chest wall, TCP was in the range 85.9%-99.7%, while for 106 or 109 cells, TCP was always zero irrespective of other model parameters. Conclusion: The high local control rate observed despite the highly inhomogeneous dose distributions obtained with this radiotherapy technique is similar to the rate for modern methods, which deliver a homogeneous dose to the whole target volume. This questions the hypothesis that the whole chest wall must receive 50 Gy. Our results add support to the suggested suitability of smaller target volumes for such patients. The model predicted TCPs comparable to the observed rate only when extremely low clonogenic cell densitiy values were used, which requires further investigation.