| Objective:This study aimed to analysis the level of cytokines in serum, clinical factors and dosimetry parameters in lung cancer patients with radiation therapy, and to explore correlation between radiation induced lung injury and those risk factors. To build an effective prediction model for radiation induced lung injury.Methods:1 The patients treated with radical three dimensional conformal radiotherapy(3D-CRT) or intensity-modulated radiotherapy(IMRT) from the radiotherapy department in People’s Liberation Army 251 th Hospital and People’s Liberation Army 254 th Hospital, were recruited according to the standard from 2014.04.01 to 2015.08.20. And the follow up time is to 2016.02.01.2 All of the patients were collected detail medical history. To record the clinical data, including: age, gender, PS score, pathological type and stage, tumor location and diameter, lung surgery and history of chemotherapy, pulmonary function, complications, etc. According to the dose volume histogram(DVH) record the V5, V10, V15, V20, V30, V40, mean lung dose(MLD), and the total dose and planning target volume(PTV). Blood sample were collected at the time of before radiotherapy, in radiation at 20 Gy, 40 Gy and the end of radiation therapy, to detect the level of cytokines(IL-6, IL-8, TGF-β1, MMP-2).3 All the patients were followed up once a week during radiotherapy, to record any fever, cough, chest tightness they had. All patients had received CT scans at the fourth week and 1th month, 3th month, 6th months after radiotherapy, combined clinical symptoms to evaluate of the grade of RILI according to the RTOG acute lung pneumonia criteria. Grade 2 or more radiation pneumonia and pulmonary fibrosis was used as end point.4 The results were analyzed with the statistic method by SPSS19.0 statistic software.Results:1 A total of 103 patients were enrolled. The follow-up time was from 6 months to 15 months, the radiotherapy dose for all patients in the groups was above 56 Gy, with a mean dose of 61±4 Gy(56-70 Gy).There were 27 cases got acute radiation induced lung injury greater than or equal to grade 2. Among the total, 19 cases were grade 2, 8 cases were grade 3, no case was 4 or 5 grade. The incidence of RILI was 26.2%.2 Analysis of clinical factors, including gender, age, smoking history, pathological types, TNM stage, whether or not the patient received radiochemotherapy, chemotherapy drugs, pulmonary function, and radiotherapy dose, showed that except radiotherapy dose, there was no significant difference in other factors between the two groups(P>0.05). Among the patients with a total radiotherapy dose>64 Gy, the incidence of RILI was 36.7%. The incidence of RILI was 16.7% among the patients with a total radiotherapy dose<64 Gy.(chi-square value of 5.349, P=0.021).3 In this study planning target volume(PTV)was not significantly associated with occurrence of RILI, P=0.492. Univariate analysis of dosimetric parameters showed that all lung V5 and MLD, ipsilateral lung V5 and V10, contralateral lung V5 were significant difference between group with RILI and group without RILI(P=0.010, 0.042, 0.041, 0.001 and 0.007, respectively). Other dosimetric parameters values were not significantly different(both P>0.05).4 In the group with RILI and the group without RILI, only pre-radiotherapy serum IL-8 and TGF-β1 in radiation at 20 Gy levels were significant difference(P=0.000, 0.044, respectively) in different time point. The other levels of cytokines at each time point were not significantly different(both P>0.05). The result of hypotheses testing about area under the curve(AUC) of the level of TGF-β1 in radiation at 20 Gy was not statistically significant(P>0.05) by using the receiver operating characteristic(ROC) curve, which showed it has no predictive value for RILI.5 Multivariate analysis identified radiotherapy dose, contralateral lung V5 and pre-radiotherapy IL-8 level were independent risk factor for symptomatic RILI. Among the patients with the contralateral lung V5≥43.6%, the incidence of RILI was 44.4%. The incidence of RILI was 12.1% among the patients with a total radiotherapy dose<43.6%.(chi-square value of 13.732, P=0.000). When the contralateral lung V5≥43.6% was used as the threshold for predicting occurrence of RILI, the sensitivity was 74.1%, the specificity was 67.1%, and the accuracy was 73.7%. In the patients with pre-radiotherapy IL-8 level<998.79 ng/L and≥998.79 ng/L, the incidence of RILI was 58.1% and 12.5% respectively(chi-square value of 23.260, P=0.000). When IL-8< 998.79 ng/L before radiotherapy was used as the threshold for predicting occurrence of RILI, the sensitivity was 66.7%, the specificity was 82.9%, and the accuracy was 76.2%.6 Radiotherapy dose>64 Gy, contralateral lung V5 and pre-radiotherapy IL-8 level were included in the model finally, its matching effect is preferable(P=0.602), the positive predictive value was 55.6%, the negative predictive value was 94.7%, and the accuracy was 84.5%.Conclusions:1 Many factors affected the occurrence and progress of radiation induced lung injury. Radiotherapy dose, contralateral lung V5 and pre-radiotherapy IL-8 level were independent risk factor for RILI.2 The risk of RILI would increase when radiotherapy dose>64 Gy, contralateral lung V5≥43.6%, pre-radiotherapy IL-8 level<998.79 ng/L. When we designing a radiotherapy plan, not only to qualify V20, V30 and MLD, and give full consideration to the clinical factors, the small dose spread to the whole lung especially the contralateral lung are also worthy of attention. The patient who has low level of IL-8 before radiotherapy indicates a greater risk of RILI.3 Prediction mode: logit(P)=-15.754+1.631X1+0.067X2-0.011X3. This mode improve a certain degree of prediction ability for RILI, which will be helpful in predicting the course of RILI. This prediction mode should be proofed at the following research. |
PDF Full Text Request