Recurrence assessment is very important for management of postoperative non-small cell lung cancer (NSCLC) patients. For several decades, detection of recurrent disease using conventional radiological method was made difficult by the often extensive anatomic abnormalities that exist after definitive treatment. For this reason, FDG-PET/CT has been found more effective than standard methods for diagnosis in this setting (1). In addition, it has been suggested in the literature (2, 3) that whole body magnetic resonance imaging (WB-MRI) with and without diffusion-weighted MR imaging (DWI) at 1.5T and 3T MR systems are at least as accurate as, or more so, as a diagnostic tool for TNM staging and recurrence assessment in NSCLC patients as compared with FDG-PET/CT. In addition, MRI fused with FDG-PET (MR/PET) is recently proposed as a new promising method for TNM staging evaluation in NSCLC patients (4). Under these circumstances, no direct comparison for postoperative recurrence assessment capability has been made among whole-body MRI with and without DWI, MR/PET, PET/CT and conventional radiological examinations in NSCLC patients.

We hypothesized that whole-body PET/MRI and MRI with DWI have a potential for improving recurrence assessment as compared with whole-body PET/CT with contrast-enhanced (CE-) brain MRI and conventional radiological examinations in NSCLC patients. The purpose of this study was to directly compare the diagnostic performance for postoperative lung cancer recurrence assessment among whole-body PET/MRI, MRI with and without DWI, PET/CT and conventional radiological examination in NSCLC patients.

96 consecutive postoperative NSCLC patients (52 males, 44 females; mean age 72 years) prospectively underwent WB-MRI with and without DWI, FDG-PET/CT, conventional radiological examination, pathological examination and more than 2-year follow-up examinations. All WB-MRI with and without DWIs were obtained by using a 3T MR system (Vantage Titan 3T, Toshiba Medical Systems) with multi-channel whole-body coil as having parallel imaging capability (Atlas SPEEDER coil, Toshiba). All co-registered MR/PETs were generated by means of our proprietary software. Final diagnosis of distant metastasis and/or recurrence in each patient was determined according to the results of conventional radiological, pathological and follow-up examinations. According to the final diagnosis, all patients were divided into two groups as follows: distant metastasis and/or recurrence group (n=17) and non-metastasis and recurrence group (n=79). All WB-MRIs with and without DWI, MR/PET, PET/CT and conventional radiological examinations were prospectively assessed by five different reader groups, which were consisted by two readers. Then, probability of recurrence on each method was evaluated by using 5-point visual scoring systems on a per patient basis. Final diagnosis in each patient was made by consensus of two readers.

To evaluate interobserver agreement on each method, weighted kappa statistic was used on a per-patient basis. To compare capabilities for recurrence assessment among all methods, ROC analysis was used on a per-patient basis. This was followed by a statistical comparison of sensitivity, specificity and accuracy by means of McNemar’s test. A p value less than 0.05 was considered statistically significant for all statistical analyses.

Representative case is shown in Figure 1. Interobserver agreement on each method was assessed as almost perfect (0.74<kappa<0.87). Area under the curve (Az) of WB-MRI with DWI (Az=0.99) and MR/PET (Az=0.99) were significantly higher than that of PET/CT (Az=0.92, p<0.05) and conventional radiological examination (Az=0.91, p<0.05). Results of comparison for recurrence assessment among all methods are shown in Figure 2. Specificities and accuracies of WB-MRI with and without DWI and MR/PET were significantly higher than those of PET/CT and conventional radiological examination (p<0.05). Whole-body MR/PET and MRI with and without DWI are suggested as having better capability for postoperative recurrence assessment than PET/CT and conventional radiological examination in NSCLC patients.