Premium
SU‐E‐T‐412: What Is the Benefit of Fiducial Marker Implantation for Pancreatic SBRT?
Author(s) -
Jones B,
Gan G,
Schefter T,
Miften M
Publication year - 2013
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4814846
Subject(s) - fiducial marker , medicine , nuclear medicine , pancreatic cancer , cone beam computed tomography , radiology , pancreas , cancer , computed tomography , endocrinology
Purpose: Localization of pancreatic tumors during Stereotactic Body Radiation Therapy (SBRT) is difficult due to the poor soft‐tissue contrast of abdominal cone‐beam CT (CBCT). The purpose of this study was to measure the increased accuracy of pancreas localization when radiopaque fiducial markers (RFM) are implanted into the tumor and peri‐pancreatic tissue to facilitate a reduction of margins and escalation of dose. Methods: 15 CBCT scans were acquired of patients implanted with RFM enrolled in a prospective study involving SBRT for pancreatic cancer, receiving 30 Gy in five fractions with PTV margins of 3–5 mm. In order to quantify the effect of RFM, matching sets of markerless CBCT images were created. An automated tracking algorithm was developed to identify the location of markers in each CBCT projection, and 2D Gaussian fitting was used to filter these markers from the images. The 3D dataset was re‐reconstructed using these filtered projections, yielding a CBCT image of the same patient geometry with no RFM. These markerless images were aligned to the reference CT images using the anterior vertebral surface/aorta and the pancreatic stent as a guide. Using the RFM images as a gold standard, the differences in position represent the error in localization due to the lack of RFM information. Results: The automated algorithm successfully yielded CBCT images without RFM, star artifacts due to markers, or artifacts due to the filtration process. When the vertebral bodies and aorta were used for localization, the average error (compared to RFM images) was 15.0±7.4 mm. When a pancreatic stent was used for localization, the average error was 3.9±2.1 mm. Conclusion: Implantation of RFM increases localization accuracy by 4–15 mm compared to conventional techniques. Marker implantation is necessary when using tight margins and/or increasing tumor dose, and can be used to increase the therapeutic ratio.