Pairwise machine learning-based automatic diagnostic platform utilizing CT images and clinical information for predicting radiotherapy locoregional recurrence in elderly esophageal cancer patients.

OBJECTIVE: To investigate the feasibility and accuracy of predicting locoregional recurrence (LR) in elderly patients with esophageal squamous cell cancer (ESCC) who underwent radical radiotherapy using a pairwise machine learning algorithm. METHODS: The 130 datasets enrolled were randomly divided into a training set and a testing set in a 7:3 ratio. Clinical factors were included and radiomics features were extracted from pretreatment CT scans using pyradiomics-based software, and a pairwise naive Bayes (NB) model was developed. The performance of the model was evaluated using receiver operating characteristic (ROC) curves and decision curve analysis (DCA). To facilitate practical application, we attempted to construct an automated esophageal cancer diagnosis system based on trained models. RESULTS: To the follow-up date, 64 patients (49.23%) had experienced LR. Ten radiomics features and two clinical factors were selected for modeling. The model demonstrated good prediction performance, with area under the ROC curve of 0.903 (0.829-0.958) for the training cohort and 0.944 (0.849-1.000) for the testing cohort. The corresponding accuracies were 0.852 and 0.914, respectively. Calibration curves showed good agreement, and DCA curve confirmed the clinical validity of the model. The model accurately predicted LR in elderly patients, with a positive predictive value of 85.71% for the testing cohort. CONCLUSIONS: The pairwise NB model, based on pre-treatment enhanced chest CT-based radiomics and clinical factors, can accurately predict LR in elderly patients with ESCC. The esophageal cancer automated diagnostic system embedded with the pairwise NB model holds significant potential for application in clinical practice.

Predicting the effects of radiotherapy based on diffusion kurtosis imaging in a xenograft mouse model of esophageal carcinoma.

The aim of the present study was to assess the predictive value of diffusion kurtosis imaging (DKI) on the effects of radiotherapy in a xenograft model of esophageal cancer. A total of 40 tumor-bearing mice, established by injection of Eca-109 cells in nude mice, were used. The experimental group (n=24) received a single dose of 15 Gy (6 MV by X-ray), and the control group (n=16) did not receive any treatment. Tumor volume, apparent diffusion coefficient (ADC), mean kurtosis (MK) and mean diffusivity (MD) of the two groups were compared, and the expression of aquaporin (AQP) 3 and necrosis ratio at matched time points in xenografts were also observed. There was a significant difference between the two groups from the 7th day of radiotherapy onwards; the xenograft volume of the experimental group was significantly smaller compared with the control group (P<0.05). On the 3rd day, the ADC and MD of the experimental group was significantly higher compared with the control group, and MK was significantly lower compared with the control group (P<0.05). On the 3rd day, AQP3 expression in the experimental group was lower compared with the control group, and the proportion of necrotic cells was higher compared with the control group (P<0.05). Single large fraction dose radiotherapy inhibited the growth of a xenografted esophageal tumor. Changes in ADC, MK and MD were observed prior to morphological changes in the tumor. The change in AQP3 expression and necrosis ratio was in also agreement with the DKI parameters assessed. DKI may thus provide early predictive ability on the effect of radiotherapy in esophageal carcinoma.

Survival Comparision of Three-dimensional Radiotherapy Alone vs. Chemoradiotherapy for Esophageal Squamous Cell Carcinoma.

BACKGROUND AND OBJECTIVE: To compare the survival of esophageal squamous cell carcinoma (ESCC) patients who received chemoradiotherapy (CRT) or radiotherapy (RT) alone. METHODS: A total of 753 well-matched patients were enrolled. A total of 299 patients were treated with CRT, and 454 patients were treated with RT alone. Propensity score matching (PSM) was performed with the R project. The Kaplan-Meier method was used to calculate survival rates, and the log-rank test was used to assess differences in survival. RESULTS: The response rate was 99.0% with CRT and 98.3% with RT alone (p = 0.651). The 1-, 3-, 5- and 10 year overall survival (OS) rates were as follows: 72.2, 40.1, 30.7 and 13.9% with CRT, 68.1, 35.2%, 23.3 and 12.5% with RT alone (p = 0.033); 73.4, 40.1, 31.0 and 16.1% with concurrent chemoradiotherapy (CCRT); and 68.1, 35.2, 23.3 and 12.5% with RT alone (p = 0.028). There was no significant difference in OS between the CCRT group and the sequential chemoradiotherapy (SCRT) group (p = 0.527). Consolidation chemotherapy (CCT) after CCRT led to a significant increase in the OS rate compared with no CCT after CCRT (p = 0.003). Compared with the OS of patients who received 1∼2 cycles of CCT, the OS of patients who received 3∼4 cycles of CCT was significantly improved (p = 0.011). Acute toxic effects were more severe in the CRT, but no significant differences in late reactions. CRT exhibited more appetite loss and fatigue symptoms than RT alone, and dysphagia of CRT relief more obviously. The CRT group had a significantly lower rate of local control failure than the RT alone group (p = 0.019). CONCLUSIONS: For patients with ESCC, CRT led to a significantly improved OS compared to RT alone, and this trend was more obvious with CCRT. CCT after CCRT prolonged OS, especially in patients who received at least 2 cycles of CCT. CRT can reduce the deaths due to local control failure compared to RT alone.