Increasing the accuracy of ¹⁸F-FDG PET/CT interpretation of "mildly positive" mediastinal nodes in the staging of non-small cell lung cancer

Introduction The aim of this study was to identify radiological factors that may reduce false-positive results and increase diagnostic accuracy when staging the mediastinum of patients with non-small cell lung carcinoma (NSCLC). Methods This was a retrospective, interdisciplinary, per-node analysis study. We included patients with NSCLC and mediastinal nodes with an SUV max in the range of 2.5-4.0 on PET-CT. We hypothesized that the greatest number of false positive cases would occur in this cohort of patients. Results A total of 92 mediastinal lymph nodes were analyzed in 44 patients. Mediastinal disease (N2/N3) was histologically confirmed in 15 of 44 patients and in 34 of 92 lymph nodes; positive predictive value of 37% and false positive rate of 63%. Lymph node SUV max, tumor size, ratio of node SUV max to tumor SUV max (SUVn/SUVp), and ratio of node SUV max to node size (SUV n/SADn) were significantly higher in true positive cases. Using a threshold of 0.3 for SUV node/tumor and 3 for SUV node/size yielded sensitivities of 91% and 71% and specificities of 71% and 69% respectively for the detection of mediastinal disease. Using both ratios in combination resulted in a sensitivity of 65% and a specificity of 88%. Concurrent benign lung disease was observed significantly more frequently in false-positive cases. Conclusion SUVn/SUVpt and SUVn/SADn may be complimentary to conventional visual interpretation and SUV max measurement in the assessment of mediastinal disease in patients with NSCLC.