Marjolein Anne Heuvelmans BSc, Presenter: Nothing to Disclose

Rozemarijn Vliegenthart MD, PhD, Abstract Co-Author: Nothing to Disclose

Pim A. De Jong MD, PhD, Abstract Co-Author: Nothing to Disclose

Willem P. Mali MD, PhD, Abstract Co-Author: Nothing to Disclose

Gonda Jasmijn de Jonge MD, PhD, Abstract Co-Author: Nothing to Disclose

Geertruida H. De Bock, Abstract Co-Author: Nothing to Disclose

Matthijs Oudkerk MD, PhD, Abstract Co-Author: Nothing to Disclose

To determine the agreement of diameter and volume measurements for different types of nodules found in low-dose computed tomography lung cancer screening, using data of the NELSON trial.

The study was institutional review board approved. Data of 2,240 solid nodules with volume of 50-500mm3 (intermediate-sized) detected at baseline in 1,498 participants were used. Volume based on semi-automatically (SA) derived maximal transversal (MT) diameter and mean of MT and perpendicular diameter were compared to SA-derived volumes by Bland-Altman plots; both for the total group of nodules, per margin (smooth, lobulated, spiculated and irregular) and per shape (spherical or non-spherical). Diameters in a random sample of 100 nodules were measured manually by two independent radiologists, and compared to the SA-derived diameters in a prospective validation study. Moreover, the implications for referral rates were evaluated for the use of a diameter-based or a volume-based protocol.

Median participant age was 59 years (interquartile range:8), and 212/1,498 (14.2%) were women. Using SA-derived mean or MT diameter to assess nodule volume lead to a mean volume overestimation of 47.2% (95%-confidence interval (CI):44.7-49.7%) and 85.1% (95%-CI:81.2-89.0%), respectively, compared to SA-derived volume. For irregular and non-spherical nodules, the mean overestimation was even 161.7% (95%-CI:131.7%-191.8%) and 168.9% (95%-CI:155.2%-182.5%), respectively. Manual diameter measurement overestimated SA-derived MT diameter by ≥10% in 44% (44/100) and underestimated in 18% (18/100) of the nodules. Compared to a 10-mm criterion for referral, using SA-derived MT diameter, 7.9% (177/2240) of the volume-based indeterminate nodules would have led to direct referral. Manual measurements would even have led to 31% (31/100) referrals.

The agreement between manual and SA-derived diameter, as well as between SA-derived diameter-based volume and SA-derived volume is poor. Applying manual and SA-derived diameter measurement in CT lung cancer screening leads to a substantial shift in nodule classification compared to SA volume measurements. 

Applying manual and SA-derived diameter measurement in CT lung cancer screening leads to a substantial shift in nodule classification compared to SA volume measurements.

Heuvelmans, M, Vliegenthart, R, De Jong, P, Mali, W, de Jonge, G, De Bock, G, Oudkerk, M, Agreement of Diameter- and Volume-based Pulmonary Nodule Management in Lung Cancer Screening.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14010556.html