食管癌调强计划三维剂量验证中γ通过率最佳阈值的研究

目的 基于三维剂量验证系统EDose^TM，寻找食管癌调强放疗计划验证中利用γ分析方法判别计划是否通过的最佳阈值。方法 回顾性地选取25例经临床批准的食管癌7野调强计划，使用Edose^TM在计划CT中重建三维剂量，并进行γ分析和剂量体积直方图（DVH）评价。然后，按照以下两个准则对计划分类：DVH特定剂量学指标绝对百分剂量差异<5%视为临床可接受计划。3%/3 mm基准下，γ通过率>90%，认为计划通过。根据分类结果绘制出受试者工作特征（receiver operating characteristic，ROC）曲线，通过计算约登系数的方式寻找全局γ在5%/3 mm、3%/3 mm和2%/2 mm基准下的最佳通过率阈值，并对最佳阈值在计划判别中的敏感性和特异性进行分析。结果 全局5%/3 mm、3%/3 mm和2%/2 mm基准下的最佳阈值分别为98.66%、94.84%和78.56%。在常规阈值90%处，全局3%/3 mm基准下的敏感性和特异性分别为0.17和0.84，而最佳阈值同样条件下的敏感性和特异性分别为0.85和0.27；全局5%/3 mm和2%/2 mm基准下最佳阈值的敏感性分别为0.89和0.65，特异性分别为0.23和0.47。结论 在3%/3 mm基准下最佳阈值较常规阈值（90%）敏感性得到显著提升，提高了阻止不可接受计划通过的能力；在2%/2 mm基准下，最佳阈值的敏感性和特异性更为平衡（相对3%/3 mm基准），一定程度上减少了不可接受计划通过的概率，同时也降低了可接受计划被误判的机会。

The optimal gamma passing rate thresholds of IMRT dosimetric verification in the treatment of esophageal cancer

Objective To investigate the optimal gamma passing rate of intensity-modulated radiotherapy (IMRT) dosimetric verification in the treatment of esophageal cancer using a three-dimensional dose verification system EDose^TM. Methods Twenty five esophageal cancer patients treated by 7-field IMRT were retrospectively reviewed. Measured dose distribution were reconstructed on CT image and evaluated by gamma analysis and DVH metrics using the EDose^TM system. Plans with DVH metrics dose difference<5% or with gamma passing > 90% under 3%/3 mm criteria were accepted. The optimal gamma passing rate for criteria of 5%/3 mm, 3%/3 mm, 2%/2 mm were investigated by drawing the receiver operating characteristic(ROC) curves and calculating the Youden Index. The sensitivity and specificity of the these optimal thresholds in the plan verification were also analyzed. Results The optimal thresholds for global gamma indices with 5%/3 mm, 3%/3 mm, 2%/2 mm were 98.66%, 94.84%, 78.56%, respectively. In the 90% common threshold, The sensitivity and specificity for common 90% threshold and optimal threshold under 3%/3 mm criteria were 0.17 vs. 0.85 and t 0.84 vs. 0.27, respectively. The sensitivity and specificity were 0.89, 0.65 and 0.23, 0.47 for optimal thresholds under 5%/3 mm and 2%/2 mm criteria, respectively. Conclusions The sensitivity of optimal threshold gamma passing rate improved significantly compared with the common threshold (90%) at 3%/3 mm criteria.,The sensitivity and the specificity were more balanced at the 2%/2 mm criteria compared with those at 3%/3 mm criteria.