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

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

To evaluate the impact of radiologists’ expertise on test result decisions made in a CT lung cancer screening trial.

In the NELSON randomized lung cancer screening trial, the baseline CT screen result was based on the lung nodule with largest volume. According to the protocol, nodule volume<50mm3, 50-500mm3 and >500mm3 led to a negative, indeterminate and positive screen result, respectively. However, the NELSON protocol allowed radiologists to manually adjust the screen result in case of high suspicion on benign or malignant nature. In this study, all participants whose baseline screen result was based on a solid nodule were included. The screen result that should have been made accordingly to the nodule management protocol and the final baseline CT screen result were compared, to see whether the screen result was manually adjusted. Histology was the reference for diagnosis, or, to confirm benignity, stability of the nodule volume on subsequent CT scans for at least two years after baseline.

In 3268 participants (2759 male, median age 58.0 years), the baseline result was based on a solid nodule. In 189 participants (5.8%) the initial baseline screen result for the largest nodule was adjusted by the radiologist. The screen result was adjusted downwards from positive or indeterminate to negative in two and 118 participants, respectively, and from positive to indeterminate in 64 participants. None of these nodules turned out to be malignant, two years after baseline. The result was adjusted upwards from negative to indeterminate in one participants. No lung cancer was diagnosed in this nodule. In four participants the screen result was adjusted upwards from indeterminate to positive; two nodules were malignant (50%). 

In baseline lung cancer screening, readers adjusted screen results in about one in twenty cases (97.4% downwards), leading to reduction of false-positive and false-negative test results. 

In lung cancer screening, radiologists’ expertise can improve nodule classification in addition to a general nodule management protocol. 

Heuvelmans, M, Vliegenthart, R, De Jong, P, Mali, W, Oudkerk, M, Effect of Reader Adjustments on Lung Nodule Classification in CT Lung Cancer Screening.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14045586.html