解剖结构对直肠癌IMRT计划膀胱受量影响

Relationship between anatomical factors and dosimetric sparing of the bladder in IMRT plans for rectal cancer

目的：建立基于解剖结构定量分析直肠癌静态IMRT计划膀胱受量的统计模型。方法选择2012—2013年100例直肠癌放疗患者，制定AP布野方式的7个野逆向IMRT计划。患者解剖结构信息用膀胱与PTV和PTV外扩0．5 cm重叠区域的大小来定量分析。利用DVH对膀胱受量分析，建立膀胱受量与解剖结构信息之间数学模型，并在20例新直肠癌计划上验证所建立的模型是否准确。结果膀胱V50与膀胱和PTV重叠区域占膀胱百分体积（ x％）呈线性关系（ V50＝0．89x－0．99）；V40与膀胱和PTV外扩0．5 cm重叠区域占膀胱的百分体积（ y％）也近似存在线性关系；平均剂量取决于x％和y％。模型预测20例直肠癌膀胱V50与V40偏差绝对值范围分别为（－3．13％～3．78％）和（－5．30％～5．66％），平均剂量相对偏差范围（－3．94％～3．76％）。结论这个模型提供了一种定量预估直肠癌IMRT膀胱受量方法。

Objective To establish a statistical model that can quantitatively analyze the dosimetric sparing of the bladder based on individual patient’ s anatomy in the static intensity-modulated radiotherapy （IMRT） plans for rectal cancer.Methods Static IMRT plans （7 AP fields） for 100 rectal cancer patients were used to train the model from 2012 to 2013.The anatomical features were quantitatively analyzed by the sizes of overlap regions of bladder-planning target volume （PTV） and bladder-PTV＋0.5（0.5 cm margin around the PTV） .The mathematic relationship between anatomical features and dosimetric sparing of the bladder was evaluated after the bladder sparing dose was analyzed using dose-volume histogram.The established model was verified in the IMRT plans for additional 20 rectal cancer patients.Results Bladder V50 was linearly correlated with the ratio of bladder-PTV overlap size to bladder volume （ denoted as x%） , with an equation of V50=0.89x-0.99.Bladder V40 showed an approximately linear correlation with the ratio of bladder-PTV＋0.5 overlap size to bladder volume （denoted as y%）.The mean dose depended on both x%and y%.For the additional 20 plans, the absolute deviation between predicted and actual values for V50 and V40 were （-3.13%-3.78%） and （-5.30%-5.66%） , respectively, and the relative deviation for the mean dose was （-3.94%-3.76%） .Conclusions The model obtained in this work provides an effective method for quantitatively estimating the bladder sparing dose in IMRT plans for rectal cancer.