A Novel Quantitative Model Based on Gross Tumor Volume Corresponding to Anatomical Distribution Measured with Multidetector Computed Tomography to Determine Resectability of Non-distant Metastatic Esophageal Squamous Cell Carcinoma: A Case Control Study

Abstract

Abstract Background It is essential to accurately determine resectability of thoracic esophageal squamous cell carcinoma (ESCC) for treatment decision making. Previous studies revealed that CT-derived gross tumour volume (GTV) associates well with T category of ESCC, nodal metastases and N stage, treatment failure rate, and disease survival rate. This study aimed to explore whether anatomical distribution-based GTV of non-distant metastatic thoracic ESCC measured with multidetector computed tomography (MDCT) could quantitatively determine the resectability. Methods 473 consecutive patients with biopsy-confirmed non-distant metastatic thoracic ESCC underwent contrast-enhanced CT were randomized into the training (TC, 376 patients) and validation (VC, 97 patients) cohorts. GTV was retrospectively measured on MDCT. Univariate and multivariate analyses were performed to identify risk factors of non-distant metastatic ESCC resectability in TC. Subsequently, Mann-Whitney U test was applied to compare GTV based on different anatomic distributions between patients of resectable and unresectable ESCCs. Receiver operating characteristic (ROC) analysis was to clarify if anatomical distribution-based GTV could help quantitatively determinate resectability. Unweighted Cohen’s Kappa tests in VC were to assess the performance of the previous models. Results Univariate analysis demonstrated that gender, anatomic distribution, cT stage, cN stage and GTV were related to resectability of non-distant metastatic ESCC (all P-values < 0.05) in TC. Multivariate analysis showed that GTV and anatomic distribution (P < 0.001 and = 0.027; odds ratio = 1.158 and 1.924, respectively) were independent risk factors of resectability. The ROC analysis showed that GTV cutoffs of the upper, middle and lower thoracic portions were 23.57 cm3, 22.89 cm3 and 22.58 cm3 to determine resectability with areas under the ROC curves of more than 0.9, respectively. Unweighted Cohen’s Kappa tests showed an excellent performance of the ROC models in the upper, middle and lower thoracic portions with Cohen k-values of 0.913, 0.879 and 0.871, respectively. Conclusions GTV and anatomic distribution of non-distant metastatic thoracic ESCC could be independent risk factors of resectability, and anatomical distribution-based GTV could well quantitatively determine resectability.