CT-based weight assessment of lung lobes

PURPOSE We aimed to evaluate the validity of lung lobe weight assessment via computed tomography (CT) by comparing CT-derived and ex vivo measurements. MATERIALS AND METHODS Unenhanced CT scanning was performed in 30 consecutive patients before lobectomy for lung cancer. The CT images were analyzed using research software after allowing for lobar weight quantitation. The lobar weight estimated by CT was then compared with that measured after surgery using a precision scale (ex vivo measurement). Comparisons as well as assessment of intra- and interoperator variability were conducted using the Bland-Altman method and the coefficient of repeatability (CR). Correlations were examined using Pearson's correlation analysis. RESULTS Comparison analyses were feasible for 28 cases. The ex vivo lobe weight was 186.2 +/- 57.3 g, whereas the weights measured by the two operators by CT were 190.0 +/- 55 and 182.4 +/- 58.2 g, respectively. As compared with ex vivo weights, the CR was 36.4 for operator 1 and 50.4 for operator 2; the mean differences were 3.8 and -3.8 for operators 1 and 2, respectively. The intraoperator and interoperator CR were 20.9 and 36.6, respectively. The mean differences for the intra- and interoperator analysis were -1.5 and -7.5, respectively. The correlation was very high between CT-based and ex vivo measurements (r=0.95 and r=0.90 for operators 1 and 2, respectively; P < 0.001). CONCLUSION Estimation of lung lobe weight by semi-automated CT analysis is sufficiently reproducible and in agreement with ex vivo measurements.