Accuracy of PET/CT in characterization of solitary pulmonary lesions

J Nucl Med. 2007 Feb;48(2):214-20.


Characterization of a pulmonary lesion is a well-established indication for metabolic imaging with 18F-FDG. There is extensive literature on the use of PET and CT in the characterization of a solitary pulmonary nodule (SPN). The performance of dual-modality imaging with PET/CT for characterizing SPNs was investigated in a clinical referral setting.

Methods: We performed a retrospective study involving patients referred for SPN characterization with PET/CT between September 2002 and June 2004, for whom a pathologic diagnosis was available. The group consisted of 12 men and 30 women whose age ranged from 35 to 84 y (mean age +/- SD, 67 +/- 11 y). A dual-slice CT/lutetium oxyorthosilicate PET system was used for imaging. CT images were acquired without intravenous contrast. Blinded interpretation was performed by 1 chest radiologist for CT and 2 nuclear medicine physicians for PET. PET/CT images were read in consensus. Lesions were analyzed by location, texture, axial dimension, and metabolic activity and visually scored on a 5-point scale from benign to malignant; the maximum standardized uptake value (SUVmax) was measured.

Results: Lesion diameter varied from 7 to 30 mm (mean +/- SD, 15 +/- 6 mm). The SUVmax ranged from 0.5 to 17.2 (mean +/- SD, 3.0 +/- 3.0). SUVmax corrected for lean body mass was 0.4-12.1 (mean +/- SD, 2.1 +/- 2.0). Comparison of CT versus PET versus PET/CT yielded accuracies of 74%, 74%, and 93%, respectively. PET and CT correctly characterized 31 and PET/CT correctly characterized 39 of the 42 lesions as malignant or benign. The sensitivity and specificity for CT, PET, and PET/CT was 93%/31%, 69%/85%, and 97%/85%, respectively. There were significant differences (P < 0.05) between PET/CT and PET for accuracy, sensitivity, and specificity. Accuracy did not improve by quantitative analysis using an SUVmax cutoff of 2.0 for malignancy. Lean body mass correction of the SUVmax did not change accuracy.

Conclusion: PET/CT demonstrates an excellent performance in classifying SPNs as benign or malignant. The combination of anatomic and metabolic imaging is synergistic by maintaining the sensitivity of CT and the specificity of PET, resulting in an overall significantly improved accuracy. Visual interpretation is sufficient for characterizing an SPN. Quantitative analysis does not improve accuracy of PET/CT for SPN characterization.

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Female
  • Fluorodeoxyglucose F18
  • Humans
  • Image Processing, Computer-Assisted
  • Lung Diseases / diagnostic imaging*
  • Lung Diseases / pathology
  • Lung Neoplasms / diagnostic imaging
  • Lung Neoplasms / pathology
  • Male
  • Middle Aged
  • Positron-Emission Tomography
  • ROC Curve
  • Radiopharmaceuticals
  • Retrospective Studies
  • Tomography, X-Ray Computed


  • Radiopharmaceuticals
  • Fluorodeoxyglucose F18