We use dose-volume factors to predict the risk of radiation-induced hepatic toxicity (RIHT) complications in patients with hepatocellular carcinoma (HCC) for controlling the low tolerance of liver organs to radiation and reducing the incidence of radiation-induced hepatic toxicity complications. This study retrospectively collected 114 patients who underwent Intensity Modulation Radiation Therapy (IMRT) for hepatocellular carcinoma between 2014 and 2017. The total number of patients was 69 after excluding normal liver organs whose volume did not reach 700 cc and extreme data. A total of 138 experimental samples were generated using the bootstrap method. All patients were evaluated using the Common Terminology Criteria for Adverse Events (CTCAE) during treatment to determine the degree of increase in blood draws and to judge for radiation-induced hepatic toxicity complications. The patient received dose-volume parameters were uniformly adjusted using a bioequivalent dose conversion of 2 Gy/fraction. The study data were divided into normal and total liver received dose-volume. Least absolute shrinkage and selection operator (LASSO) was used to select predictors and logistic regression (LR) was used to establish the performance model. LASSO was used to select the patient dose-volume parameter predictor. The risk factors of normal liver received dose-volume were age and NLV30 Gy. The risk factors of total liver received dose-volume were age and TLV35 Gy. For patients with hepatocellular carcinoma receiving radiation therapy (RT), this study recommends that a normal liver receiving a dose volume of 30 Gy should be less than 54.75%, so that the probability of RIHT can be less than 50%. A total liver receiving a dose volume of 35 Gy should be less than 54.75% so that the probability of RIHT can be less than 50%. It can control the low tolerance of liver organs to radiation and reduce the incidence of hepatotoxic complications induced by radiotherapy techniques.

No abstract

Radiation therapy is an essential part of the comprehensive breast cancer treatment strategy, and radiation dermatitis is the inevitable side-effect. According to either patient-related or treatment-related factors, patients will experience different degrees of acute radiation dermatitis. This study proposes a machine learning architecture based on image and time series features. Using the skin image of the irradiated part during radiotherapy, the image feature is extracted with a gray-level co-occurrence matrix (GLCM) and color space, combined with the time series feature with gradient boosting decision trees (GBDT) to predict the severity of dermatitis after seven days of treatment. The results show that, through the combination of image and time series features, the predicted accuracy (ACC) and area under the curve (AUC) can be effectively improved to 0.8 and 0.85 respectively. The results of GBDT show higher prediction accuracy and robustness than AdaBoost algorithm. This framework can be used as an auxiliary diagnostic tool to assist doctors in making appropriate treatments before severe dermatitis occurs, in order to reduce the radiotoxicity caused by radiotherapy of patients.