PURPOSE: Intensity modulated radiation therapy (IMRT) treatment planning can have wide variation among different treatment centers.
We propose a system to leverage the IMRT planning experience of larger institutions to automatically create high-quality plans for outside clinics. We explore feasibility by generating plans for patient datasets from an outside institution by adapting plans from our institution.
METHODS AND MATERIALS: A knowledge database was created from 132 IMRT treatment plans for prostate cancer at our institution. The outside institution, a community hospital, provided the datasets for 55 prostate cancer cases, including their original treatment plans. For each "query" case from the outside institution, a similar "match" case was identified in the knowledge database, and the match case's plan parameters were then adapted and optimized to the query case by use of a semiautomated approach that required no expert planning knowledge. The plans generated with this knowledge-based approach were compared with the original treatment plans at several dose cutpoints.
RESULTS: Compared with the original plan, the knowledge-based plan had a significantly more homogeneous dose to the planning target volume and a significantly lower maximum dose. The volumes of the rectum, bladder, and femoral heads above all cutpoints were nominally lower for the knowledge-based plan; the reductions were significantly lower for the rectum. In 40% of cases, the knowledge-based plan had overall superior (lower) dose-volume histograms for rectum and bladder; in 54% of cases, the comparison was equivocal; in 6% of cases, the knowledge-based plan was inferior for both bladder and rectum.
CONCLUSIONS: Knowledge-based planning was superior or equivalent to the original plan in 95% of cases. The knowledge-based approach shows promise for homogenizing plan quality by transferring planning expertise from more experienced to less experienced institutions.
Good D, Lo J, Lee WR, Wu QJ, Yin FF, Das SK. Are you the author?
Medical Physics Graduate Program, Duke University, Durham, North Carolina.
Reference: Int J Radiat Oncol Biol Phys. 2013 Apr 25. pii: S0360-3016(13)00324-6.