Abstract
BACKGROUND AND PURPOSE: Daily target re-delineation in online adaptive radiotherapy (oART) introduces uncertainty. The aim of this study was to evaluate artificial intelligence (AI) generated contours and inter-observer target variation among radiotherapy technicians in cone-beam CT (CBCT) guided oART of bladder cancer. MATERIALS AND METHODS: For each of 10 consecutive patients treated with oART for bladder cancer, one CBCT was randomly selected and retrospectively included. The bladder (CTV-T) was AI-segmented (CTV-T(AI)). Seven radiotherapy technicians independently reviewed and edited CTV-T(AI), generating CTV-T(ADP). Contours were benchmarked against a ground truth contour (CTV-T(GT)) delineated blindly from scratch. CTV-T(ADP) and CTV-T(AI) were compared to CTV-T(GT) using volume, dice similarity coefficient, and bidirectional local distance. Dose coverage (D(99%)>95 %) of CTV-T(GT) was evaluated for treatment plans optimized for CTV-T(AI) and CTV-T(ADP) with clinical margins. Inter-observer variation among CTV-T(ADP) was assessed using coefficient of variation and generalized conformity index. RESULTS: CTV-T(GT) ranged from 48.7 cm(3) to 211.6 cm(3). The median [range] volume difference was 4.5 [-17.8, 42.4] cm(3) for CTV-T(ADP) and -15.5 [-54.2, 4.3] cm(3) for CTV-T(AI), compared to CTV-T(GT). Corresponding dice similarity coefficients were 0.87 [0.71, 0.95] and 0.84 [0.64, 0.95]. CTV-T(GT) was adequately covered in 68/70 plans optimized on CTV-T(ADP) and in 6/10 plans optimized on CTV-T(AI) with clinical margins. The median [range] coefficient of variation was 0.08 [0.05, 0.11] and generalized conformity index was 0.78 [0.71, 0.88] among CTV-T(ADP). CONCLUSIONS: Target re-delineation in CBCT-guided oART of bladder cancer demonstrated non-isotropic inter-observer variation. Manual adjustment of AI-generated contours was necessary to cover ground truth targets.