Inter-observer variability in contouring the penile bulb on CT images for prostate cancer treatment planning

Several investigations have recently suggested the existence of a correlation between the dose received by the penile bulb (PB) and the risk of erectile dysfunction (ED) after radical radiotherapy for clinically localized prostate carcinoma. A prospective multi-Institute study (DUE-01) was implemented with the aim to assess the predictive parameters of ED. Previously, an evaluation of inter-observer variations of PB contouring was mandatory in order to quantify its impact on PB dose-volume parameters by means of a dummy run exercise. Fifteen observers, from different Institutes, drew the PB on the planning CT images of ten patients; inter-observer variations were analysed in terms of PB volume variation and cranial/caudal limits. 3DCRT treatment plans were simulated to evaluate the impact of PB contouring inter-variability on dose-volume statistics parameters. For DVH analysis the values of PB mean dose and the volume of PB receiving more than 50 Gy and 70 Gy (V50 and V70, respectively) were considered. Systematic differences from the average values were assessed by the Wilcoxon test. Seven observers systematically overestimated or underestimated the PB volume with deviations from the average volumes ranging between -48% and +34% (p < 0.05). The analysis of the cranial and caudal borders showed a prevalence of random over systematic deviations. Inter-observer contouring variability strongly impacts on DVH parameters, although standard deviations of inter-patient differences were larger than inter-observer variations: 14.5 Gy versus 6.8 Gy for mean PB dose, 23.0% versus 11.0% and 16.8% versus 9.3% for V50 and V70 respectively. In conclusion, despite the large inter-observer variation in contouring PB, a large multi-centric study may have the possibility to detect a possible correlation between PB % dose-volume parameters and ED. The impact of contouring uncertainty could be reduced by "a posteriori" contouring from a single observer or by introducing Magnetic Resonance Imaging (MRI) in the planning procedures and/or in improving the skill of observers through post-dummy run tutoring of those observers showing large systematic deviations from the mean.