Overall and GTV subvolumes tumour control probability (TCP) for head and neck cancer treated by 3D-IMRT with inhomogeneous dose distribution

Introduction.  In this study, an original model has been developed to estimate the real TCP that is a product of the TCPs calculated for GTV subvolumes of head and neck cancer based on 3D-IMRT dose planning. Material and methods.  Retrospective pilot group consist of 16 cases of oropharyngeal cancer in stage T1–2N0 previously treated with 3D-IMRT with at least 3-year follow-up. The total dose (TD) was 60–70 Gy in 2.0 Gy fractions delivered over 42–49 days. Within GTV two subvolumes were marked out: SVA with the planned 100% TD, and underdosed (90–95%) SVB. The TCP for both was calculated using the original formula developed by Withers and Maciejewski. Results.  During 3-year follow-up, 8 local recurrences (LR) occurred. In about 70% of SVB “dose cold spots” encompassed more than 50% GTV volume. This resulted in the TCPSVB decrease to 60%. Thus, the real overall TCP was much lower than a priori predicted, and in these cases local recurrences occurred. Discussion. Both cold spot SVB volumes and their dose deficit strongly correlated with a high risk of LR. Conclusions. In conclusion the magnitude of dose deficit and the size of cold subvolume within GTV have an indepen­dent negative impact on real TCP and demand dose re-planning.