Correlation of liver and pancreas tumor motion with normal anatomical structures determined with deformable image registration. (6th January 2017)

• Record Type: Journal Article
• Authors: Kaderka, Robert; Paravati, Anthony J; Sarkar, Reith; Tran, Josephine; Fero, Katherine E; Panjwani, Neil; Simpson, Daniel; Murphy, James D; Atwood, Todd F
• Is Part Of: Biomedical physics & engineering express. Volume 3:Number 1(2017)
• Publication Date: 2017-01-06
• Subjects: liver tumor -- pancreas cancer -- correlation model -- fiducials -- deformable image registration -- organ motion; Medical physics -- Periodicals; Biophysics -- Periodicals; Biomedical engineering -- Periodicals; Medical sciences -- Periodicals; 610.153
• DOI: 10.1088/2057-1976/aa54d0
• Languages: English
• ISSNs: 2057-1976

Abstract

Abstract: Purpose. The inherent difficulty of identifying liver and pancreas tumors without intravenous contrast creates the need for implanted metal fiducials to visualize tumor position and motion in stereotactic body radiation therapy (SBRT). Unfortunately, the invasive procedure of implanting fiducials carries a risk of toxicity, introduces a treatment delay, and creates streak artifacts on treatment planning images, which can hinder tumor identification. A fiducial-less motion management strategy would improve the safety, tolerability, and availability of abdominal SBRT. We hypothesized that upper abdominal tumor motion would correlate with the motion of nearby organs and could thereby serve as a fiducial-less proxy for tumor motion. Methods. We retrospectively identified fifteen patients with pancreatic cancer or liver cancer treated with SBRT. The liver, superior mesenteric artery, and celiac artery were delineated on 4DCT images and used to predict tumor position. The correlation with tumor motion was quantified with Pearson correlation coefficients ( r ), and accuracy of the tumor position prediction was expressed as the mean absolute error. Results. The majority of motion with respiration occurred in the superior–inferior (SI) direction with an average of 6.4 mm (range 2.4–11.3 mm) for pancreatic and 13.0 mm (range 6.4–21.2 mm) for liver tumors. In the SI direction we found a tight correlation between liver and tumor motion in pancreas cancer patients ( r = 0.92 ± 0.10), and liver tumor patients ( r = 0.97 ± 0.02). Using the liver as surrogate, predicted tumor location was on average 0.5 mm from the actual position and not greater than 3.0 mm. Conclusions. This study demonstrates a potential correlation of normal organ and tumor motion which could serve as a fiducial-less surrogate for SBRT in the upper abdomen as on-site 4D volumetric imaging becomes available during treatment. Moving this motion management strategy into the clinic requires additional research to optimize 4D image quality and understand inter-fraction reproducibility.