Guidelines for Radiotherapy of Esophageal Carcinoma (2020 Edition)

Deep Learning for Automatic Gross Tumor Volumes Contouring in Esophageal Cancer Based on Contrast-Enhanced Computed Tomography Images: A Multi-Institutional Study

PURPOSE

To develop and externally validate an automatic artificial intelligence (AI) tool for delineating gross tumor volume (GTV) in patients with esophageal squamous cell carcinoma (ESCC), which can assist in neo-adjuvant or radical radiation therapy treatment planning.

METHODS AND MATERIALS

In this multi-institutional study, contrast-enhanced CT images from 580 eligible ESCC patients were retrospectively collected. The GTV contours delineated by 2 experts via consensus were used as ground truth. A 3-dimensional deep learning model was developed for GTV contouring in the training cohort and internally and externally validated in 3 validation cohorts. The AI tool was compared against 12 board-certified experts in 25 patients randomly selected from the external validation cohort to evaluate its assistance in improving contouring performance and reducing variation. Contouring performance was measured using dice similarity coefficient (DSC) and average surface distance. Additionally, our previously established radiomics model for predicting pathologic complete response was used to compare AI-generated and ground truth contours, to assess the potential of the AI contouring tool in radiomics analysis.

RESULTS

The AI tool demonstrated good GTV contouring performance in multicenter validation cohorts, with median DSC values of 0.865, 0.876, and 0.866 and median average surface distance values of 0.939, 0.789, and 0.875 mm, respectively. Furthermore, the AI tool significantly improved contouring performance for half of 12 board-certified experts (DSC values, 0.794-0.835 vs 0.856-0.881, P = .003-0.048), reduced the intra- and interobserver variations by 37.4% and 55.2%, respectively, and saved contouring time by 77.6%. In the radiomics analysis, 88.7% of radiomic features from ground truth and AI-generated contours demonstrated stable reproducibility, and similar pathologic complete response prediction performance for these contours (P = .430) was observed.

CONCLUSIONS

Our AI contouring tool can improve GTV contouring performance and facilitate radiomics analysis in ESCC patients, which indicates its potential for GTV contouring during radiation therapy treatment planning and radiomics studies.

A Diagnostic Quandary: Carboplatin-Paclitaxel-Induced Stevens-Johnson Syndrome in a Rare Case of Carcinosarcoma of the Esophagus and Review of the Literature

Sarcomatoid carcinoma of the esophagus, a mixed tumor comprising both carcinomatous and sarcomatoid components and known as carcinosarcoma, is a rare malignancy. Clinically and radiologically, it presents like other esophageal cancers. Here we discuss the case of a 69-year-old male patient with sarcomatoid carcinoma of the esophagus who developed Stevens-Johnson syndrome (SJS) after chemotherapy with carboplatin and paclitaxel. The patient was evaluated for dysphagia and odynophagia. He was initially misdiagnosed to have an esophageal polyp and underwent excision for the same. He presented with recurrent growth at the local site, with histopathological examination showing sarcomatoid carcinoma of the esophagus. After the development of paclitaxel-carboplatin-induced SJS, the patient was subsequently treated with palliative radiotherapy at the primary site for symptomatic relief. He underwent feeding gastrostomy as a supportive nutritional measure and was on best supportive care after a multidisciplinary tumor board discussion. Paclitaxel-carboplatin-induced SJS poses numerous diagnostic conundrums, on account of there being only one reported incident prior to this in literature, to the best of our knowledge. In this report, we explore the diagnostic and therapeutic predicaments associated with a rare disease that is under-reported and understudied in literature and delve into the various treatment modalities that can benefit the patients. The case also demonstrates the delicate balance between cancer chemotherapeutics and their Pandora's box of adverse effects.

Development and internal validation of a nomogram for predicting survival of nonoperative EGFR-positive locally advanced elderly esophageal cancers

Purpose

This study aims to develop and validate a prediction model for non-operative, epidermal growth factor receptor (EGFR)-positive, locally advanced elderly esophageal cancer (LAEEC).

Methods

A total of 80 EGFR-positive LAEEC patients were included in the study. All patients underwent radiotherapy, while 41 cases received icotinib concurrent systemic therapy. A nomogram was established using univariable and multivariable Cox analyses. The model's efficacy was assessed through area under curve (AUC) values, receiver operating characteristic (ROC) curves at different time points, time-dependent AUC (tAUC), calibration curves, and clinical decision curves. Bootstrap resampling and out-of-bag (OOB) cross-validation methods were employed to verify the model's robustness. Subgroup survival analysis was also conducted.

Results

Univariable and multivariable Cox analyses revealed that icotinib, stage, and ECOG score were independent prognostic factors for LAEEC patients. The AUCs of model-based prediction scoring (PS) for 1-, 2-, and 3-year overall survival (OS) were 0.852, 0.827, and 0.792, respectively. Calibration curves demonstrated that the predicted mortality was consistent with the actual mortality. The time-dependent AUC of the model exceeded 0.75, and the internal cross-validation calibration curves showed good agreement between predicted and actual mortality. Clinical decision curves indicated that the model had a substantial net clinical benefit within a threshold probability range of 0.2 to 0.8. Model-based risk stratification analysis demonstrated the model's excellent ability to distinguish survival risk. Further subgroup analyses showed that icotinib significantly improved survival in patients with stage III and ECOG score of 1 (HR 0.122, P<0.001).

Conclusions

Our nomogram model effectively predicts the overall survival of LAEEC patients, and the benefits of icotinib were found in the clinical stage III population with good ECOG scores.

Utilization and Outcomes of Radiation in Stage IV Esophageal Cancer

Introduction

For patients with stage IV esophageal cancer, esophageal radiation may be used selectively for local control and palliation. We aimed to understand patterns of radiation administration among patients with stage IV esophageal cancer and any potential survival associations.

Methods

In this retrospective cohort study, the National Cancer Database was queried for patients with metastatic stage IV esophageal cancer diagnosed between 2016 and 2019. Patterns of radiation use were identified. Survival was determined through Kaplan-Meier analysis of propensity score-matched pairs of patients who did and did not receive radiotherapy and time-to-event models.

Results

Overall, 12,088 patients with stage IV esophageal cancer were identified, including 32.7% who received esophageal radiation. The median age was 65 (interquartile range [IQR]: 58-73) years, and 82.6% were male. Among the irradiated patients, the median total radiation dose was 35 (IQR: 30-50) Gy administered in a median of 14 (IQR: 10-25) fractions given in 22 (IQR: 14-39) days. Overall, esophageal radiation was not associated with better survival (log-rank p = 0.41). When stratified by radiation dose, a survival advantage (over no radiation) was found in the 1144 patients (29% of the irradiated patients) who received 45 to 59.9 Gy (time ratio = 1.28, 95% confidence interval: 1.20-1.37, p < 0.001) and the 88 patients (2.2%) who received 60 to 80 Gy (time ratio = 1.37, 95% confidence interval: 1.11-1.69, p = 0.003).

Conclusions

One-third of the patients with metastatic stage IV esophageal cancer in the National Cancer Database received esophageal radiation. Most received a radiation dose that, although consistent with palliative regimens, was not associated with a survival advantage. Further study is warranted to understand the indications for radiation in stage IV esophageal cancer and potentially reevaluate the most appropriate radiation dose for palliation.

Optimization of Minimum Segment Width Parameter in the Intensity-Modulated Radiotherapy Plan for Esophageal Cancer

Purpose

This study was designed to explore the optimal minimum segment width (MSW) in the intensity-modulated radiotherapy (IMRT) plan for esophageal cancer.

Patients and Methods

The imaging data of 20 esophageal cancer patients were selected for this study. Four IMRT plans were designed for each patient with MSWs of 0.5, 1.0, 1.5, and 2.0 cm. The conformity index (CI) and homogeneity index (HI) of the planning target volumes (PTV), organs at risk (OARs), control points (CP), monitor units (MU), plan delivery time (DT), and gamma passing rates (GPR) were collected and compared to appraise the treatment plan quality and delivery efficiency.

Results

Lower-MSW plans had larger CI and smaller HI values, and lower dose parameters of OARs and PTVs. The HI, CI, and dose parameter of OARs in the 0.5 and 1.0 cm MSW groups were similar and much better than those of the 1.5 and 2.0 cm MSW groups. Meanwhile, the plan in the 0.5 cm MSW group had significantly higher MUs, CPs, and DTs, and a significantly lower relative dose of GPR with a 3% dose difference and 3 mm distance to agreement criteria than the other three groups.

Conclusion

The 0.5 and 1 cm MSW groups had better dosimetric parameters and IMRT plan quality than the other groups. However, plans with 0.5 cm MSW had worse delivery accuracy and efficiency than the other three groups. Thus, MSW of 1.0 cm was the optimal choice to ensure good quality, delivery accuracy, and treatment efficiency in IMRT plans for esophageal cancer.