Abstract
OBJECTIVE: We wanted to determine the clinical feasibility of using non-breath-hold real-time MR-echo imaging for the evaluation of mediastinal and chest wall tumor invasion. MATERIALS AND METHODS: MR-echo imaging was prospectively applied to 45 structures in 22 patients who had non-small cell lung cancer when the tumor invasion was indeterminate on CT. The static MR imaging alone, and the static MR imaging combined with MR-echo examinations were analyzed. The surgical and pathological findings were compared with using the Wilcoxon-signed rank test and McNemar's test. RESULTS: The accuracy, sensitivity and specificity of the combined MR-echo examination and static MR imaging for determining the presence of invasion were 84%, 83% and 85%, respectively, for the first reading session and they were 87%, 83% and 87%, respectively, for the second reading session (there was substantial interobserver agreement, k = 0.74). For the static MR imaging alone, these values were 62%, 83% and 59%, respectively, for the first reader and they were 69%, 67% and 74%, respectively, for the second reader (there was moderate interobserver agreement, k = 0.49). The diagnostic confidence for tumor invasion was also higher for the combined MR-echo examination and static MR imaging than that for the static MR imaging alone (p < 0.05). CONCLUSION: The combined reading of a non-breath-hold real-time MR-echo examination and static MR imaging provides higher specificity and diagnostic confidence than those for the static MR imaging reading alone to determine the presence of mediastinal or chest wall tumor invasion when this was indeterminate on CT scanning.