Room: AAPM ePoster Library
Purpose: is a commercial treatment planning system newly available to the market, which includes a novel automated treatment planning tool. We evaluated its performance by comparing the automatically generated plans with the manual plans in intensity modulated radiotherapy (IMRT) for patients with cervical cancer.
Methods: patients were selected for this study. For each patient, one manual (UIH-M) and one automated (UIH-AP) static SIB-IMRT plans (2-dose level: 50Gy and 45Gy) were generated using uRT-TPS, respectively. For UIH-AP, a list of critical clinical goals was derived from our institutional requirements and was used as input to the tool. Additionally, another manual plan had been generated with the widely-used TPS Monaco (Monaco-M) for each patient to compare with UIH-M. Monte Carlo dose calculation algorithm was applied to all plans. Dose endpoints of targets and organs at risk were calculated for plan quality comparison. The effective planning time were also recorded for each method.
Results: All plans fulfilled the clinical goals set for targets and organs at risk, and were acceptable for treatment. UIH-M achieved preferred PTV dose conformity (CI), homogeneity (HI) and OAR sparing, while Monaco-M achieved higher PTV mean dose. Compared to UIH-M, mean dose of bladder, rectum and femoral heads were improved by 3.33% (p=0.01), 1.93% (p=0.16), 13.03% (p<0.01, left) and 13.69% (p<0.01 right) with UIH-AP. PTV dose conformity and homogeneity of UIH-AP were worse, but the differences were not significant. The average effective planning time was 13.4±3.1 minutes using UIH-AP, compared to 23.3±1.3 minutes in UIH-M.
Conclusion: generated clinically acceptable IMRT plans with significantly improved OAR sparing and comparable target coverage, in comparison with the manual plans for cervical cancer radiotherapy. Moreover, the effective planning time was substantially reduced.
Cross Validation, Statistical Analysis, Treatment Planning
TH- External Beam- Photons: Treatment planning using machine learning/Knowledge Based Planning/automation