Jane P. Ko MD, Abstract Co-Author: Nothing to Disclose

Manmeen Kaur MD, Abstract Co-Author: Nothing to Disclose

Erika Jacobs MD, Presenter: Nothing to Disclose

Elan Bomsztyk, Abstract Co-Author: Nothing to Disclose

James S Babb PhD, Abstract Co-Author: Nothing to Disclose

Alissa K Greenberg MD, Abstract Co-Author: Nothing to Disclose

David Paul Naidich MD, Abstract Co-Author: Nothing to Disclose

Henry Rusinek PhD, Abstract Co-Author: Nothing to Disclose

et al, Abstract Co-Author: Nothing to Disclose

To assess the precision in measuring pulmonary nodules in vivo and the distribution of nodule growth rates.

Reports of all subjects enrolled over 2 years in a lung cancer CT screening trial were reviewed. Nodules reported as clinically stable for at least 2 years on follow-up chest CTs or proven to be malignant were included. CT protocol consisted of 1.25 x 1 mm sections (120 kVp, 20 -120 mAs, 0.5s rot., high-frequency algorithm, 25-35 cm FOV). A phantom-validated semi-automated 3D volume measurement method that accounts for volume averaging was applied to CTs of nodules from 2 or more time points, yielding absolute rates of change (ARC), relative rates of change (RRC), and volume measurement error (VME). Precision was expressed in terms of standard deviation (SD) and confidence intervals (CI).

We identified 125 clinically stable nodules (mean 5.4 mm, 1.6-19 mm) in 60 subjects meeting criteria. Mean follow-up was 2.6 years with a mean 4 CTs per subject (range 2-8). 102 nodules were solid, 12 pure ground-glass (GG), and 11 mixed attenuation. Nodules were central (n=23), peripheral (n=102), abutting vessels (n=56), and irregular (n=15). Stable solid nodules had ARC and RRC of(median +/- SD) -0.6 +/- 22.0 mm3 and -2.4 +/- 35%, respectively; non-solid stable nodules had ARC of -0.1 +/- 84 mm3 and RRC of 0.2 +/- 46%; all stable nodules had ARC of -2.0 +/- 38 mm3 and RRC of -2.4 +/-35%. Therefore solid nodules with ARC greater than 18 mm3 or RRC greater than 47% could be assessed as growing using a 90%CI, respectively (36 mm3 ARC or 56% RRC, 95%CI). Mean (95%CI) VME was 0.47% (-47%, +48%) for stable solid nodules, 0.35% (-50%, 51%) for all nodules. Inter-reader variability was VME -15.1% (17.3%, -47.5%). Lower precision (P<0.05) was associated with: irregular borders, adjacent vessels, central location. For 11 pathologically confirmed malignant nodules, RRC were of mean 140%

A 3D volume method is a promising tool for detecting changes in nodule volumes although should be applied with knowledge of factors associated with lower precision, such as GG and irregular nodules.

Ko, J, Kaur, M, Jacobs, E, Bomsztyk, E, Babb, J, Greenberg, A, Naidich, D, Rusinek, H, et al, , Precision of 3D Computer-assisted Nodule Volume Assessment in Vivo.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4411724.html