乳癌患者接受放射治療後發生放射性肺炎之萊曼正常組織併發症機率模型參數分析

LYMAN KUTCHER-BURMAN NORMAL TISSUE COMPLICATION PROBABILITY MODEL PARAMETERS ANALYSIS OF RADIATION PNEUMONITIS IN BREAST CANCER PATIENTS AFTER RADIOTHERAPY

丁慧敏(Hui-Min Ting) ； 郭仕賢(Shih-Sian Guo) ； 藍仁鴻(Jen-Hong Lan) ； 李曉飛(Hsiao-Fei Lee) ； 趙珮如(Pei-Ju Chao) ； 黃郁傑(Yu-Je Huang) ； 李財福(Tsair-Fwu Lee)

《放射治療與腫瘤學》 24卷1期 (2017/03) Pp. 25-34

https://doi.org/10.6316/TRO/201724(1)25

摘要

目的：探討乳癌患者接受放射治療後，其同側肺臟劑量與放射性肺炎併發症（radiation pneumonitis, RP）之影響，本研究建立萊曼 Lyman Kutcher-Burman（LKB）正常組織併發症機率（Normal Tissue Complication Probability, NTCP）模型，分析併發症對應之相關劑量關係。材料與方法：本研究針對97位乳癌患者，使用混合型強度調控放射治療技術（hybrid intensitymodulated radiotherapy, Hybrid IMRT）進行50.4 Gy共28次的分次放射治療。患者於放射治療後三個月進行併發症評估，最終併發症定義根據Arriagada分級標準，主要觀察胸腔電腦斷層掃描影像變化，患者於 CT 影像中有第一級以上肺炎併發症之影像變化，定義為具有放射性肺炎併發症。經過同側肺臟平均劑量與放射性肺炎併發症相關性評估後，計算患者同側肺臟之等效均勻劑量（equivalent uniform dose, EUD），建立 LKB NTCP 預測模型，同時進行模型預測效力評估檢定。NTCP 預測模型主要用以探討劑量對於放射性肺炎併發症之S 型反應曲線（sigmoid curve），建立本地乳癌患者併發症發生率之劑量參數。結果：同側肺臟平均劑量與放射性肺炎併發症具有統計顯著性，建立同側肺臟劑量之LKB NTCP 模型參數：50% 併發症發生率所對應的同側肺臟劑量 TD50 = 21.26 Gy（95% CI, 19.78 - 22.97），斜率參數 m = 0.32（95% CI, 0.19 - 0.81），模型之受測者特徵之曲線下面積（area under the receiver operating characteristic curve, AUC）為 0.65，且通過白氏得分數、Hosmer-Lemeshow test 及校準曲線之模型效能評估。結論：本研究建立乳癌放射性肺炎 Lyman Kutcher-Burman（LKB）正常組織併發症機率模型，定義本地患者之相關模型參數，可應用於乳癌患者放射治療計畫評估，避免產生放射性肺炎，提升患者治療後之生活品質。

關鍵字

乳癌 ； 放射性肺炎 ； LKB正常組織併發症預測模型

並列摘要

Purpose : To investigate the correlation of the ipsilateral lung dose and radiation pneumonitis of breast cancer after radiotherapy, we established Lyman Lyman Kutcher-Burman (LKB) normal tissue complication probability (NTCP) model to study the corresponding dose of complication. Materials and Methods : In total 97 patients with breast cancer were included, they were treated with hybrid intensity-modulated radiotherapy (Hybrid IMRT) for 50.4 Gy in 28 fractions. Patients were evaluated at three months after treatment through the observation of the change of chest computed tomography image. The final endpoint of complication was according to Arriagada’s classification. Patients with grade 1+ change on CT images were defined as a complication of radiation pneumonitis. We evaluated the correlation of radiation pneumonitis and mean ipsilateral lung dose. We established LKB NTCP model through calculated the equivalent uniform dose (EUD) of ipsilateral lung and validated the model performance. The NTCP model was mainly used to investigate the dose-response curve for complications of radiation pneumonitis and provided dose parameters to prevent complication of endemic breast cancer patients. Result : Radiation pneumonitis was significantly correlated with mean ipsilateral lung dose. The fitted LKB NTCP parameters were: TD50 = 21.26 Gy (95% CI, 19.78 - 22.97), m = 0.32 (95% CI, 0.19 - 0.81). The area under the receiver operating characteristic curve (AUC) of the NTCP model was 0.65. The model was passing through model performance validations, including Brier score, Hosmer- Lemeshow test and the calibration curve. Conclusion : We established Lyman Kutcher Burman normal tissue complication probability model for breast cancer radiation pneumonitis in this study, and defined the correlated parameters for endemic patients. We can apply these parameters to breast cancer radiation treatment planning evaluation, avoid radiation pneumonitis, and improve the patients’ quality of life after radiotherapy.

並列關鍵字

Breast cancer ； Radiation pneumonitis ； Lyman Kutcher-Burman normal tissue complication probability