Abstract
1349
Purpose: To investigate the sensitivity and specificity of HDCT versus PET/CT for the diagnosis of lymphangitic carcinomatosis (LC) using the gold standard of histology. Materials and
Methods: One hundred six patients addressed for initial staging of lung cancer were retrospectively included. Using double blind analysis, we assessed the presence of signs favoring LC on HRCT (smooth lines or nodular septal lines, subpleural nodularity, peribronchial thickening, satellite nodules, lymph node enlargement and pleural effusion). 18F-FDG PET/CT were reviewed to assess the quantify tracer uptake in peritumoral areas within a range of 3 cm with comparison to normal background uptake. Histology served as goldstandard in all patients. The study was approved by the ethics committee. Results: Among 106 included patients, 71% (76/106) had histologically confirmed LC. Mean metabolic variables including lesional SUVmax, SUVmean, TLG and MTV as well as perilesionnal SUVmax, SUVmean and their respective ration to background were significantly higher in LC group versus the non - LC group (p=0.0024, p=0.0026, p=0.0026, p=0.0026, p=0.001, p=0.001, p=0.001 and p=0.001 respectively) using a t - test. Bronchovascular thickening and lymph node involvement had a sensitivity, specificity, and ROC area of 66%, 83% and 0.75 (0.66-0.83) and 64%, 69% and 0.66 (0.56-0.76), respectively, while perilesional SUVmax and SUVmean had a significantly higher sensitivity, specificity, and ROC area of 96%, 83% and 0.97 (0.94-1) and 93%, 90% and 0.95 (0.91-1) for detecting LC (p<0.05). Conclusion: Standard FDG - PET/CT using perilesional SUVmax and SUVmean outperforms HRCT for the detection of LC using in a large patient group with histology as gold standard. Combining resultant morphologic and metabolic criteria may help to establish a powerful tool for the initial diagnosis of LC.