The diagnostic accuracy of integrated positron emission tomography/computed tomography in the evaluation of pulmonary mass lesions in a tuberculosis-endemic area

  • R du Toit
  • JA Shaw
  • EM Irusen
  • F von Groote-Bidlingmaier
  • JM Warwick
  • CFN Koegelenberg

Abstract

Background. Integrated positron emission tomography/computed tomography (PET-CT) is a  well-validated modality for assessing pulmonary mass lesions and specifically for estimating risk of malignancy. Tuberculosis (TB) is known to cause false-positive PET-CT findings.
Objective. To investigate the utility of PET-CT in the evaluation of pulmonary mass lesions and nodules in a high TB prevalence setting.
Methods. All patients referred for the evaluation of a solitary pulmonary nodule or mass and who  underwent PET-CT scanning over a 3-year period were included. The PET-CT findings, including maximum  standardised uptake value (SUVmax), were compared with the gold standard (tissue or microbiological diagnosis). The sensitivity, specificity, positive and negative predictive values and diagnostic accuracy for malignant disease were calculated according to the SUVmax cut-off of 2.5 and a proposed cut-off  obtained from a receiver operating characteristic (ROC) curve.
Results. Forty-nine patients (mean (standard deviation) age 60.1 (10.2) years; 29 males) were included, of whom 30 had malignancy. Using an SUVmax cut-off of 2.5, PET-CT had a sensitivity, specificity,  positive and negative predictive value and diagnostic accuracy for malignancy of 93.3%, 36.8%, 70.0%, 77.8% and 71.4%, respectively. After a ROC curve analysis, a suggested SUVmax cut-off of 5.0 improved the specificity to 78.9% and the diagnostic accuracy to 86.7%, with a small reduction in sensitivity to 90.0%.
Conclusions. The diagnostic accuracy of PET-CT in the evaluation of pulmonary mass lesions using the conventional SUVmax cut-off of 2.5 was reduced in a TB-endemic area. An SUVmax cut-off of 5.0 has a higher specificity and diagnostic accuracy for malignancy, with a comparable sensitivity.
Published
2016-03-02
Section
Articles

Journal Identifiers


eISSN: 0256-95749
print ISSN: 2078-5135