Background: This study assesses the accuracy and interaction of clinical suspicion and positron emission tomography (PET) scans in diagnosis of suspected thoracic malignancy.
Material and Method: 313 patients evaluated in a University Hospital lung cancer evaluation center who underwent PET scanning and subsequently had tissue confirmation or a course of stability by CT scans over a period of two years. At the time of the initial visit, clinical suspicion based on history, physical exam and computerized tomography (CT) characteristics was assigned as low, intermediate or high probability of malignancy. Subsequently PET results were classified as negative, intermediate or positive, based on standardized uptake value [SUV] of 0, between 0 and 2.5, greater than 2.5 respectively.
Results: ROC analysis showed similar results for clinical suspicion (0.762) and PET (0.779). High clinical suspicion and positive PET had a PPV of 96.6%; low clinical suspicion and negative PET had a NPV of 100%. When PET and clinical suspicion were fully discordant, clinical suspicion was accurate in 80%, PET in 20%. When PET scan or clinical suspicion was intermediate, 35% were malignant.
Conclusions: Clinical suspicion and PET are both accurate in diagnosing thoracic malignancy. When suspicion and PET are concordant, diagnostic accuracy is very high; when discordant, clinical suspicion was more accurate. When clinical suspicion or PET were intermediate, there is a significant likelihood for cancer.

Keywords: Risk Factors, Physician's Role, Positron-Emission Tomography, Male, Middle Aged, Lung Neoplasms - radionuclide imaging, Humans, Female