Researchers involved in a large, multi-institutional study comparing the accuracy of positron emission tomography (PET) and computed tomography (CT) in the characterization of lung nodules found that PET was far more reliable in detecting whether or not a nodule was malignant.
“CT and PET have been widely used to characterize solitary pulmonary nodules (SPNs) as benign or malignant,” said James W. Fletcher, professor of radiology at Indiana University School of Medicine in Indianapolis, Ind. “Almost all previous studies examining the accuracy of CT for characterizing lung nodules, however, were performed more than 15 years ago with outdated technology and methods, and previous PET studies were limited by small sample sizes,” he noted.
“Detecting and characterizing SPNs is important because malignant nodules represent a potentially curable form of lung cancer. Identifying which SPNs are most likely to be malignant enables physicians to initiate the proper therapy before local or distant metastases develop,” said Fletcher.
In a head-to-head study addressing the limitations of previous studies, PET and CT images on 344 patients were independently interpreted by a panel of experts in each imaging modality, and their determination of benign and malignant nodules were compared to pathologic findings or changes in SPN size over the next two years.
The researchers found that when PET and CT results were interpreted as “probably” or “definitely” benign, the results were “strongly associated with a benign final diagnosis”—in other words, the modalities were equally good at making this determination. PET's superior specificity (accuracy in characterizing a nodule as benign or malignant), however, resulted in correctly classifying 58 percent of the benign nodules that had been incorrectly classified as malignant on CT. Furthermore, when PET interpreted SPNs as definitely malignant, a malignant final diagnosis was 10 times more likely than a benign.