Predicting Tumor Response Using Combo of Peritumoral Radiomics Features Appears Effective

A combination of peritumoral radiomics features appeared to improve the predictive performance of intratumoral radiomics to estimate pathological complete response after neoadjuvant chemoradiation in patients with esophageal squamous cell carcinoma.

Findings published in JAMA Network Open suggest that a combination of peritumoral radiomics features appears to improve the predictive performance of intratumoral radiomics to estimate pathological complete response following neoadjuvant chemoradiation in patients with esophageal squamous cell carcinoma (ESCC).1

Specifically, microenvironmental immune components were found to be most likely to be associated with both the intratumor and peritumoral radiomics prediction.

“This study underlines the significant application of peritumoral radiomics to assess treatment response in clinical practice,” the authors wrote.

Researchers included a total of 231 patients with ESCC who underwent baseline contrast-enhanced computed tomography and received neoadjuvant chemoradiation followed by surgery at 2 institutions in China. Single-institution data between April 2007 and December 2018 was used to extract radiomics features from intratumoral and peritumoral regions and established intratumoral, peritumoral, and combined radiomics models using different classifiers. Using independent data collected from another hospital during the same period, researchers also conducted external validation.

Patients with stage III disease accounted for the majority (173 patients [74.9%]). The pCR rates of the training and test sets were 46.0% and 44.3%, respectively.

Eight classifiers were used to construct radiomics models with intratumoral or peritumoral features. Of the total cohort, the optimal intratumoral and peritumoral radiomics models yielded similar areas under the receiver operating characteristic curve of 0.730 (95% CI, 0.609-0.850) and 0.734 (95% CI, 0.613-0.854), respectively.

Combining intratumoral and peritumoral features for all 8 classifiers significantly improved performance. The combined model was composed of 7 intratumoral and 6 peritumoral features, with an area under the receiver operating characteristic curve of 0.852 (95% CI, 0.753-0.951), accuracy of 84.3%, sensitivity of 90.3%, and specificity of 79.5% in the test set.

“Although a wide range of performance across different classifiers was achieved in the test set, the combination of intratumoral and peritumoral features always improved the classification accuracy regardless of the choice of classifier, confirming the additional predictive value of peritumoral radiomics features,” explained the authors.

In an editorial written by Ruijiang Li, PhD, of the Stanford University School of Medicine, Li suggested that the ability to reliably predict treatment response may help determine which patients are most likely to benefit from neoadjuvant chemoradiotherapy while sparing others from the toxic effects associated with the treatment.2 Further, Li added that the radiomic signature developed in the current study provides a promising direction to follow in pursuing this goal.

However, Li also explained that before the implementation of such a method can occur, several issues and questions should be addressed. For example, Li indicated that the generalizability of the imaging signature across different computed tomography scanners and imaging protocols needs to be rigorously assessed. Moreover, because the current study evaluated multiple machine learning classifiers with different levels of accuracy, it will be important that a final model is chosen.

“Despite the need for further validation, the study by Hu et al represents a step forward toward an individualized approach to the treatment of esophageal cancer.”


1. Hu Y, Xie C, Yang H, et al. Assessment of Intratumoral and Peritumoral Computed Tomography Radiomics for Predicting Pathological Complete Response to Neoadjuvant Chemoradiation in Patients With Esophageal Squamous Cell Carcinoma. JAMA Network Open. doi: 10.1001/jamanetworkopen.2020.15927

2. Li R. Peritumoral Radiomics and Predicting Treatment Response. JAMA Network Open. doi: 10.1001/jamanetworkopen.2020.16125