Low tumor blood flow assessed with perfusion CT correlates with lymphatic involvement in patients with stage T1b non-small cell lung cancer.

BACKGROUND: To investigate the correlation of computed tomography (CT) perfusion parameters and lymphatic involvement in patients with stage T1b non-small cell lung cancer (NSCLC). METHODS: Forty-six patients (30 men and 16 women; age range, 36-73 years; mean age, 57 years), with stage T1b non-small cell lung cancer, underwent perfusion CT before surgery. The correlations between CT perfusion parameters (blood flow, blood volume, peak enhancement intensity), tumor angiogenesis (microvessel density and maturity of microvessels of surgical specimens) and lymphatic involvement were retrospectively investigated. Receiver operator curve (ROC) analysis was used to identify the parameter threshold at which tumors had or did not have lymph node metastasis, and the corresponding sensitivity and specificity were calculated. RESULTS: A significant tendency for tumors with low blood flow and high density of immature microvessels to show lymphatic involvement was found (all P < 0.001). High correlation (r =-0.769, P < 0.001) was observed between tumor blood flow and immature microvessels. The area under ROC curves (AUC) for blood flow to detect lymph node metastasis was 0.866 (95% confidence interval, 0.766-0.966). For blood flow, the sensitivity, specificity, and accuracy of predicting lymph node metastasis were 88.9, 64.3, and 73.9% respectively, if the cutoff point was set at 43.05 mL/100 g/minute. CONCLUSIONS: Blood flow may be useful to predict lymphatic involvement before surgery in stage T1b NSCLC.

Lymphatic microvessel density combined with CT used in the diagnosis of mediastinal and hilar lymph node metastasis of non-small cell lung cancer.

BACKGROUND AND AIMS: Lymphatic microvessel density (LMVD) has been demonstrated to correlate with tumor metastasis. The purpose of this study is to determine whether the criteria combining LMVD with computed tomography (CT) could improve the diagnostic accuracy of lymph node (LN) metastasis in non-small cell lung cancer (NSCLC). METHODS: Ninety four patients with NSCLC who had chest CT scans preoperatively and LMVD tested by immunohistochemistry postoperatively were randomized into two groups: the training set (n = 66) and the test set (n = 28). Cut-off point of LMVD was selected to separate the LN metastasis-predictive positive and negative groups. On the basis of LMVD levels, chest CTs of the training set were re-analyzed and hypothetical criteria for LN metastasis diagnosis were established. Diagnostic characteristics for LN metastasis were tested by using the combined criteria in the test set as compared to those of CT alone. RESULTS: There was a significantly positive correlation between LMVD and LN metastasis (p <0.01). For sensitivity, specificity, positive predictive value (PPV) and negative predictive values (NPV), accuracy was 67, 81, 75, 81 and 79% for the combined criteria, respectively. Diagnostic efficacy of the combined criteria was significantly higher than that of CT only (p <0.05). CONCLUSIONS: Diagnosis of LN metastasis using a combination of LMVD and CT is superior to the CT-only diagnosis. In future clinical trials, it is necessary to evaluate the efficacy of adjuvant therapy for the selection of patients according to the combined criteria.