Click here to


Are you sure ?

Yes, do it No, cancel

An Image Processing Technique for Identification of Interstitial Catheters in High- and Low-Field MR-Guided HDR Brachytherapy

K Singhrao*, S Park , T Wong , Y Gao , M Guo , J Fu , P Hu , Y Yang , A Chang , J Lewis , UCLA School of Medicine, Los Angeles, CA


(Monday, 7/30/2018) 9:30 AM - 10:00 AM

Room: Exhibit Hall | Forum 4

Purpose: To develop an image processing technique that enables easy identification of contrast-filled catheters for high- and low-field MR guided HDR brachytherapy.

Methods: Ten contrast-filled MR catheters were constructed by filling 4 french (Fr) gauge catheters with a water-based carrageenan solution. The catheters were sealed with bone wax to prevent leakage. The catheters were inserted into 6 Fr gauge needles. The needles were implanted into an anthropomorphic prostate phantom (CIRS, Virginia). CT images were acquired on a Siemens Somatom CT scanner and two sets of T2-weighted MR images were acquired using a 0.35T Viewray, and a 3T Siemens Skyra. CT images were acquired with a metal guide-wire placed inside each needle. MR images were acquired with and without MR catheters. The catheter ground truth position was defined by applying a 200 HU threshold to the CT image. MR catheter reconstruction was performed by histogram-based normalization and image subtraction between the filled and unfilled catheter images.

Results: For the 0.35T images, the maximum positional difference between the CT and reconstructed MR catheters in the Anterior-Posterior (AP), Left-Right (LR) and Superior-Inferior (SI) directions was 0.6mm, 0.3mm and 3mm respectively. The mean absolute needle tip location difference between the CT and reconstructed MR catheters was 0.1 mm. For the 3T images the maximum positional difference between the CT and reconstructed MR catheters in the AP, LR and SI directions was 0.2mm, 1.2mm and 1.6mm respectively. The mean needle tip location difference between the CT and MR reconstructed catheters was 0.4mm.

Conclusion: We have identified a relatively simple image processing technique to identify the position of contrast-filled catheters for both high- and low-field MR images. The maximum catheter positional error was 3mm using 0.35T MR images and 1.6mm using 3T MR images. Future work will include development of an automatic MR-guided catheter identification workflow.

Funding Support, Disclosures, and Conflict of Interest: Funded partially through a master research agreement with Varian Medical Systems.


Brachytherapy, MRI, Treatment Planning


TH- Brachytherapy: prostate brachytherapy

Contact Email