Gross tumor volume of adenocarcinoma of esophagogastric junction corresponding to cT and cN stages measured with computed tomography to quantitatively determine resectabiliy: A case control study.

PURPOSE

To determine whether gross tumor volume (GTV) of adenocarcinoma of esophagogastric junction (AEG) corresponding to cT and cN stages measured on CT could help quantitatively determine resectability.

MATERIALS AND METHODS

343 consecutive patients with AEG, including 279 and 64 randomly enrolled in training cohort (TC) and validation cohort (VC), respectively, underwent preoperative contrast-enhanced CT. Univariate and multivariate analyses for TC were performed to determine factors associated with resectability. Receiver operating characteristic (ROC) analyses were to determine if GTV corresponding to cT and cN stages could help determine resectability. For VC, Cohen's Kappa tests were to assess performances of the ROC models.

RESULTS

cT stage, cN stage and GTV were independently associated with resectability of AEG with odds ratios of 4.715, 4.534 and 1.107, respectively. For differentiating resectable and unresectable AEG, ROC analyses showed that cutoff GTV of 32.77 cm in stage cTN with an area under the ROC curve (AUC) of 0.901. Particularly, cutoffs of 27.67 and 32.77 cm in stages cT and cT obtained AUC values of 0.860 and 0.890, respectively; and cutoffs of 27.09, 33.32 and 37.39 cm in stages cN, cN and cN obtained AUC values of 0.852, 0.821 and 0.902, respectively. In VC, Cohen's Kappa tests verified that the ROC models had good performance in distinguishing between resectable and unresectable AEG (all Cohen's K values > 0.72).

CONCLUSIONS

GTV, cT and cN stages could be independent determinants of resectability of AEG. And GTV corresponding to cT and cN stages can help quantitatively determine resectability.