G astric cancer is the fifth most common malignancy in the world, and about half of the cases in the world occur in Eastern Asia, mainly in China (1-3). Complete tumor excision of the tumor is the first-line therapy for gastric cancer (4). However, even after potentially curative resection with satisfactory safety margins, the prognosis of patients with locally advanced gastric cancer is still worrying. Consequently, adjuvant and neoadjuvant therapies are now increasingly used in conjunction with surgery for locally advanced gastric cancer, which can significantly downstage the tumor, improve R0 resection rate, progression-free survival, and overall survival (5, 6).The accurate evaluation of preoperative TNM staging of gastric cancer is essential for selecting the optimal treatment method and predicting prognosis (7). Lu et al. (8) reported that depth of invasion, lymph node metastasis stage, metastatic lymph node ratio, lymphatic invasion, and tumor size were independent predictors of prognosis in gastric cancer patients who underwent radical surgery (R0 resection). Shiraishi et al. (9) demonstrated that serosal invasion, extragastric lymph node metastasis, and liver metastasis were independent prognostic factors in patients with large gastric cancer. Endoscopic ultrasonography (EUS) and CT are currently the main methods of staging gastric cancer. EUS has been in use since the 1980s and is reported to have high T staging accuracy. CT has a great advantage on the evaluation of T stage and is considered the best modality for the staging of gastric cancer, as it can perform 257 From the Department of Radiology (S-X.R. raoxray@163.com), Zhongshan Hospital, Fudan University, Shanghai, China.
METHODSThis study retrospectively evaluated 42 patients diagnosed with gastric cancer, who underwent chemotherapy followed by surgery. Pre-and post-treatment CT tumor volumes (V T ) were measured in portal venous phase and volume reduction ratios were calculated. Correlations between pre-and post-treatment V T , reduction ratio, and pathologic stages were analyzed. Receiver operator characteristic (ROC) analyses were also performed to assess diagnostic performance for prediction of downstaging to T0-2 stage and N0 stage.
RESULTSPretreatment V T , post-treatment V T , and V T reduction ratio were significantly correlated with T stage (r s =0.329, r s =0.546, r s = -0.422, respectively). Post-treatment V T and V T reduction ratio were significantly correlated with N stage (r s =0.442 and r s = -0.376, respectively). Pretreatment V T , post-treatment V T , and V T reduction ratio were significantly different between T0-2 and T3,4 stage tumors (P = 0.05, P < 0.001, and P = 0.002, respectively). The differences between N0 and ≥N1 groups were also statistically significant (P = 0.005 for post-treatment V T , P = 0.016 for V T reduction ratio, respectively). The area under the ROC curve (AUC) for identification of T0-2 groups was 0.70 for pretreatment V T , 0.88 for post-treatment V T , and 0.82 for V T reduction ratio, respective...
