Validation study of ultrasound-based high-dose-rate prostate brachytherapy planning compared with CT-based planning.

PURPOSE The use of transrectal ultrasound (TRUS) to both guide and plan high-dose-rate (HDR) brachytherapy (BT) for prostate is increasing. Studies using prostate phantoms have demonstrated the accuracy of ultrasound (US) needle tip reconstruction compared with CT imaging standard. We have assessed the in vivo accuracy of needle tip localization by TRUS using cone-beam CT (CBCT) as our reference standard. METHODS AND MATERIALS Needle positions from 37 implants have been analyzed. A median of 16 needles (range, 16-18) per implant were inserted, advanced to the prostate base, and their tips identified using live TRUS images and real-time planning BT software. Needle protrusion length from the template was recorded to allow for reverification before capturing images for planning. The needles remained locked in the template, which was fixed to the stepper, while a set of three-dimensional TRUS images was acquired for needle path reconstruction and HDR-BT treatment planning. Following treatment, CBCT images were acquired for subsequent needle reconstruction using a BT Treatment Planning System. The coordinates of each needle tip were recorded from the Treatment Planning System for CT and US and compared. RESULTS A total of 574 needle tip positions have been compared between TRUS and CBCT. Of these, 59% agreed within 1 mm, 27% within 1-2 mm, and 11% agreed within 2-3 mm. The discrepancy between tip positions in the two modalities was greater than 3 mm for only 20 needles (3%). CONCLUSIONS The US needle tip identification in vivo is at least as accurate as CT identification, while providing all the advantages of a one-step procedure.