Reconstruction errors in clinical intraoperative TRUS-based prostate HDR-BT detected using electromagnetic tracking

To investigate the occurrence of errors in transrectal ultrasound (TRUS)-based implant reconstructions for high-dose-rate brachytherapy (HDR-BT) in prostate cancer using an afterloader-integrated electromagnetic tracking (EMT) system.

Fourteen patients were treated with one TRUS-based treatment fraction in an intraoperative setting while under general anesthesia, as part of their prostate HDR-BT (2×13.5 Gy) treatment. EMT measurements were performed before the start of the treatment in all implanted needles at dwell positions (DPs) with an interval of 5 mm. The Euclidean distances (EDs) between clinically reconstructed and EMT-measured DPs after registration were calculated. Errors were evaluated per needle (minimum ED of 2mm) and stratified into 4 severity levels (minor, moderate, major and severe). Error causes were investigated through retrospective inspection of TRUS imaging.

The median (range) ED between EMT-measured and clinically reconstructed DPs was 1.0 (0.1-9.4) mm. Higher EDs were observed in the anterior and lateral regions of the prostate. From 265 evaluated needle reconstructions, 23% (61/265) had minor errors or higher, while 9% (24/265) had major or severe errors. Severe errors were mostly caused by incorrect needle or depth selection. Major, moderate and minor errors were mostly caused by artifact, shadowing, and user errors, respectively.

This study found that a quarter of needle reconstructions contained errors >2mm, and that high and severe errors were not uncommon. EMT can play an important role in detecting and preventing these reconstruction errors without disrupting the clinical workflow.