| Summary: | 碩士 === 國立高雄應用科技大學 === 電子工程系碩士班 === 103 === Purpose : To investigate the relationship between lung dose and radiation pneumonitis in breast cancer patients after radiotherapy.
Materials and methods : We built and compared five normal tissue complication probability (NTCP) models through outcomes of radiation pneumonitis in breast cancer patients, and defined the best predictive NTCP model for local population in this study.
87 patients with breast cancer were evaluated and 5 outlier samples were excluded. In total, 82 patient data were used in this study. The patients were treated by intensity-modulated radiotherapy or hybrid intensity-modulated radiotherapy techniques. The patients were evaluated by chest computed tomography (CT) at 3 months after completion of radiation therapy. Density changes on chest CT were evaluated by comparing with the CT image prior to radiation therapy for radiation therapy treatment planning. Clinically complication was defined according to the modified Common Toxicity Criteria of the National Cancer Institute (CTC-NCIC). We used the sample data to build five NTCP models. The five models were LKB (Lyman Kutcher-Burman), Logistic, Schultheiss, Poisson and Kallman-s model, respectively. The five NTCP models were compared by different model performance validation tools. We also built LKB - Veff model based on the LKB model. LKB - Veff model provided the correlation of effective volume and dose at the same complication probability for clinical phycisian.
Results : The fitted parameters of five NTCP models were (1) LKB model :
TD50 = 21.42 Gy (95% CI, 20.13 - 22.83), m = 0.27 (95% CI, 0.18 - 0.56)；(2) Logistic model : TD50 = 21.41 Gy (95% CI, 20.12 - 22.86), γ= 1.48 (95% CI, 0.71 - 2.35)；(3) Schultheiss model : TD50 = 21.26 Gy (95% CI, 19.89 - 22.74), k = 5.65 (95% CI, 2.67 - 9.07)；(4) Poisson model : TD50 = 21.21 Gy (95% CI, 19.83 - 22.68), γ= 1.46 (95% CI, 0.74 - 2.21) ；(5) Kallman-s model : TD50 = 21.66 Gy (95% CI, 20.25 - 23.15), γ= 1.46 (95% CI, 0.74 - 2.20), s = 1.01. Overall performance Akaike's Information Criterion (AIC) of Kallman-s model was better than the other four models, but other performance validation were equal in five models.
Conclusions : Reducing lung radiation dose in breast cancer patients can effectively reduce the probability of radiation pneumonitis. The dose of 50% probability of complications of radiation pneumonitis was 21.66 Gy in Kallman-s model, which was the best model in our study. Reducing the effective volume of irradiated lung could improve the quality of life of breast cancer patients.
|
|---|
