M. Sullivan1, M. Hemmati2, T. D. Malouff3, and R. Boopathy4; 1Stephenson cancer center University of Oklahoma, Oklahoma city, OK, 2university of oklahoma, norman, OK, 3Creighton University Medical Center, Omaha, NE, 4University of Oklahoma Health Sciences Centre, Oklahoma City, OK
Purpose/Objective(s): Stereotactic body radiation therapy (SBRT) for larger palliative tumor is made possible with Spatially fractionated radiation therapy (SFRT) technique because of its promising tumor response and minimal toxicity. To study the efficacy of the treatment outcome a methodical and consistent placement/contouring of grid spheres inside the tumor volume is vital. In this study the inter-planner variation in sphere placement is studied. Materials/
Methods: Five SFRT patients that was treated before were retrospectively chosen whose clinical target volume (CTV) sizes were ranging between 350 cc to 4580cc. Seven planners which include physicists and physicians were asked to contour/place the grid spheres by following the same grid contouring guidelines. Once the spheres were contoured, the ratio of grid volume to CTV volume were calculated. This grid to CTV ratio nullified the impact of CTV size thus allowing the comparison between patients and planners normalized. Results: The Grid placements were very inconsistent in terms of co-ordinates/positioning and the total volumes between planners. The standard deviation in volumes between each planner for same patients varied from 3.5 cc to 21.4 cc. The spread of the ratios was computed for each case along with the quartiles. Both positive (n=2) and negative (n=2) skewness were observed with one case suffering from highly negative skewness. The interquartile range (IQR) for the planning ratio was computed for each patient, ranging from 0.32% to 1.23%, which indicates a wide variation in IQR values even in the presence of GRID contouring guidelines. Using the same analysis, one contouring was marked as an outlier. Conclusion: The study shows significant inconsistencies in grid placements by manual methods. To address the inconsistencies in grid contouring a stringent training for personnel for grid contouring is recommended. Developing automated optimization algorithm for grid placements could be an ideal solution.
