ABSTRACT

To propose a novel plan difficulty index (PDI) and scoring model and evaluate its application in intensity-modulated radiation therapy (IMRT) plans for lung cancer. Thirty-four optimal IMRT plans of lung cancer were retrospectively analyzed. These plans have been clinically implemented to treat patient. Four scoring models, segmented uniform (SU), segmented non-uniform (SNU), non-segmented uniform (NSU) and non-segmented non-uniform (NSNU), containing 12 metrics, were designed for plan evaluation. A novel PDI was defined to quantify how easy or difficult a treatment plan is to make and verify the rationality of scoring models, followed by the evaluation of the relationship between PDI and scoring value of plan. Significant differences were found between different scoring models. These scores were 72.34 ± 10.10, 69.40 ± 12.06, 53.11 ± 10.44 and 50.30 ± 11.67 for SU, SNU, NSU and NSNU models, respectively. The pass rate (64.7%) of plans with scores of ≥70 was superior in the SU model compared to that of other models (SNU: 38.2%, NSU: 5.9%, and NSNU: 2.9%). The mean PDI of 34 plans was 4.20 for the uniform scoring model and 4.47 for the non-uniform scoring model. Negative correlations were shown between PDI and scores [SU: −0.52 (P < 0.05), SNU: −0.49 (P < 0.05), NSU: −0.59 (P < 0.05), and NSNU: −0.53 (P < 0.05)]. PDI is feasible for IMRT plan evaluation for lung cancer, and the SU scoring model was superior to NSU, SNU, and NSNU models in clinical practice.

摘要

提出一种新的计划难度指数（PDI）和评分模型，并评估其在肺癌调强放射治疗（IMRT）计划中的应用。回顾性分析了 34 种肺癌最佳 IMRT 计划。这些计划已在临床上实施以治疗患者。为计划评估设计了四个评分模型，分段均匀 (SU)、分段非均匀 (SNU)、非分段均匀 (NSU) 和非分段非均匀 (NSNU)，包含 12 个指标。定义了一种新的PDI来量化治疗计划制定的难易程度，并验证评分模型的合理性，然后评估PDI与计划评分值之间的关系。不同评分模型之间存在显着差异。这些分数为 72.34 ± 10.10、69.40 ± 12.06、53.11 ± 10.44 和 50。SU、SNU、NSU 和 NSNU 模型分别为 30 ± 11.67。与其他模型（SNU：38.2％，NSU：5.9％和NSNU：2.9％）相比，SU模型中得分≥70的计划的通过率（64.7％）优于其他模型。对于统一评分模型，34 个计划的平均 PDI 为 4.20，对于非统一评分模型为 4.47。PDI 和分数之间显示负相关 [SU: -0.52 (P < 0.05)，SNU：-0.49 ( P < 0.05)，NSU：-0.59 ( P < 0.05)，和 NSNU： -0.53 ( P < 0.05)]。PDI在肺癌IMRT方案评估中是可行的，SU评分模型在临床实践中优于NSU、SNU和NSNU模型。