In vivo repeatability of automated volume calculations of small pulmonary nodules with CT

AJR Am J Roentgenol. 2009 Jun;192(6):1657-61. doi: 10.2214/AJR.08.1825.


Objective: The objectives of our study were to evaluate the in vivo reproducibility of automated volume calculations of small lung nodules with both low-dose and standard-dose CT and to assess whether repeatability within each technique varies according to the diameter, site, or morphology of the nodule or to percentage of emphysema.

Subjects and methods: Sixty-six subjects with 83 solid pulmonary nodules between 5 and 10 mm in diameter were enrolled in this prospective study. Four consecutive MDCT data sets, two low dose and two standard dose, were obtained for each nodule on separate breath-holds during the same session. The volume of each nodule was calculated by automated software. Repeatability was evaluated by Bland-Altman's approach and the coefficient of repeatability. Associations of the percentage of volume variation between two measurements with nodule diameter, emphysema percentage, nodule site, and nodule morphology were assessed by Spearman's correlation coefficient and the Kruskal-Wallis test. A p value of < 0.05 was considered statistically significant.

Results: The range of variation of the volumes of pulmonary nodules between two subsequent measurements was -38% +/- 60% for low-dose CT and -27% +/- 40% for standard-dose CT. No significant statistical association was found between variation in volume measurements and nodule site, nodule diameter, nodule morphology, or emphysema percentage by semiautomated calculation of lung density.

Conclusion: Automated volume calculations of small pulmonary nodules can significantly differ between two subsequent breath-holds with both low-dose and standard-dose CT techniques; in clinical practice we recommend that a volume variation of greater than 30% for nodules between 5 and 10 mm should be confirmed by follow-up CT to be sure that a nodule is actually growing.

MeSH terms

  • Algorithms
  • Artificial Intelligence*
  • Female
  • Humans
  • Imaging, Three-Dimensional / methods*
  • Lung Neoplasms / diagnostic imaging*
  • Male
  • Middle Aged
  • Pattern Recognition, Automated / methods*
  • Radiographic Image Enhancement / methods
  • Radiographic Image Interpretation, Computer-Assisted / methods*
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Solitary Pulmonary Nodule / diagnostic imaging*
  • Tomography, X-Ray Computed / methods*