Performance of a computer-aided program for automated matching of metastatic pulmonary nodules detected on follow-up chest CT

AJR Am J Roentgenol. 2007 Nov;189(5):1077-81. doi: 10.2214/AJR.07.2057.

Abstract

Objective: The purpose of this study was to evaluate the performance of a computer-aided program that allows automated matching of metastatic pulmonary nodules imaged with two serial clinical chest CT studies.

Materials and methods: The cases of 30 patients with metastatic pulmonary nodules depicted on two serial clinical MDCT scans (16- or 64-MDCT, 5-mm section thickness) were studied. The number of nodules per patient varied from a minimum of two to innumerable. A maximum of 10 well-defined solid nodules per patient, a total of 210 nodules, were selected from each baseline CT scan and were evaluated for matching detection in follow-up CT by means of an automated program. Substantial changes in lung findings and lung volumes between serial scans were visually assessed. The effects on matching rate of interval lung changes and location, size, and total number of nodules in the lung were analyzed with contingency tables. Chi-square tests were used to evaluate patterns for statistical significance.

Results: The nodule-matching rate per patient ranged from 0 to 100% (median, 87.5%). By nodule, the overall matching rate was 140 of 210 (66.7%). Matching rate was highly associated with changes in lung quality between serial studies. Matching of 122 of 148 nodules (82.4%) occurred in 23 patients with relatively unchanged lung findings, compared with 18 of 62 nodules (29.0%) in seven patients with substantial interval changes (p < 0.001). The matching rate decreased with an increased total number of nodules per lung. For 10 or fewer nodules per lung, matching was successful for 31 of 36 nodules; for 11-50 nodules per lung, 60 of 73 nodules; for 51-100 nodules per lung, 33 of 47 nodules; and for more than 100 nodules per lung, 16 of 54 nodules (p < 0.001). The matching rate was not significantly different with location or size of nodules.

Conclusion: The rate of automated matching of metastatic pulmonary nodules on clinical serial CT scans was high (82.4%) when the lung findings and lung expansion between the serial scans were relatively unchanged. The rate decreased significantly, however, with substantial interval changes in the lung and a larger number of nodules.

Publication types

  • Evaluation Study

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Algorithms
  • Artificial Intelligence*
  • Female
  • Humans
  • Image Enhancement / methods
  • Image Interpretation, Computer-Assisted / methods*
  • Lung Neoplasms / diagnostic imaging*
  • Lung Neoplasms / secondary*
  • Male
  • Middle Aged
  • Pattern Recognition, Automated / methods*
  • Radiography
  • Reproducibility of Results
  • Sensitivity and Specificity
  • Solitary Pulmonary Nodule / diagnostic imaging*
  • Solitary Pulmonary Nodule / secondary*