Comparison of In-Room Biplane Radiography and Tomosynthesis in Simulation.

Abstract

PURPOSE/OBJECTIVE(S)\nMany real-time guidance technologies for lung SBRT use surrogates of tumor motion. Direct imaging of the lung tumor may improve safety and efficacy of treatments, but the visibility of lung tumors with in-room biplane radiography is insufficient. We used simulations to assess the potential of in-room biplane tomosynthesis as an alternative.\n\n\nMATERIALS/METHODS\nBiplane tomosynthesis is envisioned as an in-place upgrade to biplane radiography. The envisioned system can be used either for radiography with a large field of view, or tomosynthesis with a small field of view. When tomosynthesis is desired, nine x-ray sources are energized simultaneously. Each x-ray source is positioned and collimated so that the detector receives all nine images in adjacent, non-overlapping sectors of the detector. We assume a 40 cm flat panel detector and 1.5x magnification, yielding an 8.8 cm square field of view. We selected 50 CT cases from the NSCLC Radiomics dataset with Stage I cancer and GTV diameters less than 5 cm as representative of lung SBRT cases. Simulated images of both in-room tomosynthesis and in-room radiography were produced using forward projection with dose of 0.3 mAs. Two thoracic radiation oncologists evaluated these images on a 4-point scale as follows: 1, tumor not visible; 2, visualization inadequate for motion management; 3, visualization adequate for motion management; 4, visualization excellent for motion management. The primary endpoint was the fraction of cases under each modality with scores of 3 or above. The cases were read in randomized order in two sessions spaced one week apart. Tomosynthesis also has potential for real-time imaging because all sources are energized simultaneously, but this capability could not be tested in static datasets.\n\n\nRESULTS\nOn average, the two readers found 28% (range, 18-38%) of radiography images and 50% of tomosynthesis images (range, 48-52%) adequate or excellent for motion management. The average scores for radiography and tomosynthesis were 2.0 and 2.4 out of 4, respectively. The two readers agreed exactly for 54% of the images and disagreed by 1 point for 43% of the images. Tomosynthesis did not show a selective advantage for tumors larger or smaller than the median tumor volume of 7 cc.\n\n\nCONCLUSION\nIn-room tomosynthesis has the potential to provide a fiducial-free mechanism for visualizing lung tumors while breathing. Nearly twice as many tumors were rated adequate for motion management using tomosynthesis when compared to radiography, but half of the tumors were rated inadequate using either modality.\n\n\nAUTHOR DISCLOSURE\nS.S. Hsieh: None. L.W. Ng: None. M. Cao: Research Grant; NIH. Honoraria; Varian Medical Systems Pacific Inc., ViewRay Inc, Varian Medical System Inc. Consultant; Varian Medical System, Guidepoint Inc. Travel Expenses; Varian Medical Systems Pacific Inc., Varian Medical System Inc.; Medical Dosimetrist Certification Board. Chair; AAPM.P. Lee: Research Grant; Viewray, Varian Inc., AstraZeneca, Inc. Honoraria; Viewray, Varian Inc., AstraZeneca, Inc. Consultant; Viewray, Varian Inc. Speaker s Bureau; AstraZeneca, Inc. Advisory Board; AstraZeneca, Inc. Travel Expenses; Viewray, Varian Inc.; Committee Co-Chair.