Low Accuracy of Computed Tomography and Positron Emission Tomography to Detect Lung and Lymph Node Metastases of Colorectal Cancer.

Abstract

Minimally invasive surgery, stereotactic radiotherapy, and radiofrequency ablation are commonly proposed in the case of pulmonary colorectal-metastasis as alternatives to conventional open surgery. Preoperative imaging assessment by computed tomography (CT) scan and fluorodeoxyglucose positron emission tomography (FDG-PET) are critical to guide oncologic radical treatment. Our aim was to investigate the accuracy of CT and FDG-PET for the evaluation of the number of pulmonary colorectal metastases and thoracic lymph nodal involvement (LNI). Patients who underwent lung surgical resection for pulmonary colorectal metastases from 2004 to 2014 were analyzed. Concordance between histology, CT scan, and FDG-PET findings were assessed. Data of 521 patients were analyzed. Of those, FDG-PET was performed in 435 (83.5%). A moderate agreement between both CT scan (kappa index: 0.42) and FDG-PET (kappa index: 0.42) findings and the histologically proven number of metastases was observed. The number of histologically proven metastases was correctly discriminated in 61.7% of cases with CT scan and in 61.8% of cases with FDG-PET. Multiple metastases were discovered in 20.9% of clinical single metastasis cases with CT scan, and in 24.4% of those cases with FDG-PET. One hundred fifty patients (29.1%) presented with pathologic LNI. A poor agreement was observed between LNI and CT scan findings (kappa index: 0.02), and a weak agreement was observed concerning LNI and FDG-PET findings (kappa index: 0.39). Computed tomography and FDG-PET have limitations if the objective is to detect all malignant nodules and to discriminate the LNI in cases of pulmonary metastases of colorectal cancer.