Risk factors for radiation pneumonitis in stageⅢnon-small-cell lung cancer treated with three-dimensional conformal radiotherapy
ZHU Xiang-zhi,WANG Lv-hua,WANG Ying-jie,JI Wei,WANG Xiao-zhen,ZHANG Li,CAO Jian-zhong,ZHANG Zhong,ZHANG Ke,YIN Wei-bo,朱向帜,王绿化,王颖杰,姬巍,王小震,张莉,曹建中,章众,张可,殷蔚伯
DOI: https://doi.org/10.3760/j.issn:1004-4221.2007.06.004
2007-01-01
Abstract:Objective To evaluate the clinical and dosimetric factors for radiation pneumonitis (RP) in stageⅢnon-small cell lung cancer(NSCLC) treated with three dimensional conformal radiotherapy (3DCRT). Methods From November 2001 to March 2005, 107 patients with pathologically diagnosed stageⅢNSCLC were treated with curative intent. Forty-two patients received radiotherapy alone, 65 patients received sequential chemoradiation. All patients were treated with computed tomography based three dimensional planning and had dose volume histogram data available. The total target dose was between 26 Gy and 74 Gy ( median 60 Gy). The following clinical parameters were considered: gender, age, use of chemotherapy, presence of chronic obstructive pulmonary disease, Karnofsky performance status, pre-radiotherapy forced vital capacity (FVC) , Pre-RT diffusion capacity of carbon monoxide, Pre-RT forced expiratory volume in 1 s, Pre-RT hemoglobin, tumor location, radiation dose, smoking history and pathological type. The following dosimetric parameters were taken from the dose-volume histogram ( DVH) : total lung volume, mean lung dose( MLD) , involved ipsilateral lung MLD( IMLD) , contralateral lung MLD( CMLD) , planning target volume( PTV) , % of lung receiving≥5 ,10,15 ,20 and 60 Gy (respectively V5-60、involved ipsilateral lung IV5-60 and contralateral lung CV5-60, in increments of 5 Gy). The endpoint of the study was the incidence of RP(grade≥2 radiation pneumonitis) and SRP(grade≥3 radiation pneumonitis). The relationships between clinical or dosimetric factors and the incidence of RP were analysed using the Fisher's exact test for categorical variables, and the nonparametric Mann-Whitney exact test for quantitative variables. A logistic regression analysis, including those factors having univariate significance, were performed using a backward regression procedure with a P value≤0. 10 for entry. Pearson correlation analysis was used to determine correlations between dosimetric factors. Results Thirty nine and thirteen of the 107 patients developed RP and SRP, respectively. Grades 2,3 and 5 RP was observed in 26 patients, 10 patients and 3 patients, respectively. For≥2 grade RP, Univariate analysis revealed that the MLD, IMLD, CMLD, IV20-35, CV10, CV15, CV40, CV50, and V10-35 were statistical significant relative to the incidence of RP. Multivariate analysis revealed that V20 was the single independent predictor of RP(P =0. 005). For SRP, Univariate analysis revealed that the MLD, IV20, IV25, IV30 , CV5, CV10 ,V10 , V15, V20 and V30 were statistical significant risk factors. Multivariate analysis identified that CV10 was the single independent predictor of SRP( P = 0.028). Dosimetric factors were found to be strongly associated with the incidence of≥2 grade RP in stageⅢNSCLC. V20 seemed to be the most sensitive one. Conclusion Ten Defivite statistical significant risk factors of radiation pneumonitis are found in stageⅢnon-small cell lung cancer treated with three-dimensional conformal radiotherapy.