Stephen J.LurieMD, PhD, Senior EditorIndividualAuthorJody W.ZylkeMD, Contributing EditorIndividualAuthor
To the Editor: I have several concerns about the meta-analysis by Dr Gould and colleagues1 about positron emission tomography (PET) imaging for diagnosis of pulmonary nodules. First, the prevalence of malignancy in the articles that they included was extremely high (55%-100%; mean, 72%), suggesting that the types of lesions that have been evaluated thus far with PET imaging are those with a very high likelihood of malignancy. The accuracy of PET imaging may be far lower in a population in whom the risk of malignancy is lower. The sensitivity might be lower because of a milder spectrum of disease in patients with a lower risk of malignancy, and the specificity might be lower because of more overlapping PET findings in small lesions of varying etiologies.2 Thus, the results of their review are only applicable to a population of patients with a very high prevalence of cancer. Until additional studies provide evidence that PET imaging is accurate in a population with a low prevalence of cancer, it is premature to suggest application of PET imaging in this group, as the authors have done.
Smith-Bindman R. Positron Emission Tomography to Evaluate Lung Lesions. JAMA. 2001;285(21):2711-2712. doi:10.1001/jama.285.21.2711