Interobserver variation in postimplant computed tomography contouring affects quality assessment of prostate brachytherapy

Abstract

Purpose: Permanent seed implants are accepted treatment of early stage prostate cancer. Implant quality is assessed by post implant CT-based dosimetry but prostate contours on CT images are obscured by metallic seed artefact and edema. Outcome depends on implant quality, but perceived implant quality depends on accurate prostate contouring. This study documents inter observer variation in prostate contouring on post implant CT scans. Methods and Materials: Ten patients had implant dosimetry calculated on 4 copies of the post implant CT scan. Prostate contours from MRI-CT fusion were the gold standard for prostate edge identification. CTs were contoured by an experienced prostate brachytherapist matching CT images to the pre implant TRUS, and by 2 GU radiation oncologists experienced in conformal radiotherapy planning. Dosimetry was compared to that obtained using MRI-CT fusion in terms of D90 and V100. Results: Contours and dosimetry were not reproducible among the 3 observers. The V100's of the experienced brachytherapist differed from that of MRI-CT fusion by a mean of 2.4% compared to 9.1% and 4.4% for observers 1 and 2, and the D90 by a mean of 9.3 Gy compared to 30.3 and 14.4 Gy for observers 1 and 2. Conclusions: Quality assessment of prostate brachytherapy based on 1 month post implant CT is difficult. This may obscure the dose-response relationship in brachytherapy as well as create problems for quality assurance in multicentre trials evaluating brachytherapy against standard modalities. Whenever possible, MRI-CT fusion should be employed to verify prostate contours post implant.