Objective To evaluate the impact of radiological expertise on screen result decisions in a CT lung cancer screening trial. Methods In the NELSON lung cancer screening trial, the baseline CT result was based on the largest lung nodule's volume. The protocol allowed radiologists to manually adjust screen results in cases of high suspicion of benign or malignant nodule nature. Participants whose baseline CT result was based on a solid or part-solid nodule were included in this study. Adjustments by radiologists at baseline were evaluated. Histology was the reference for diagnosis or to confirm benignity and stability on subsequent CT examinations. Results A total of 3,318 participants (2,796 male, median age 58.0 years) were included. In 195 participants (5.9 %) the initial baseline screen result was adjusted by the radiologist. Adjustment was downwards from positive or indeterminate to negative in two and 119 participants, respectively, and from positive to indeterminate in 65 participants. None of these nodules turned out to be malignant. In 9/195 participants (4.6 %) the screen result was adjusted upwards from negative to indeterminate or indeterminate to positive; two nodules were malignant. Conclusion In one in 20 cases of baseline lung cancer screening, nodules were reclassified by the radiologist, leading to a reduction of false-positive screen results. Key Points • The NELSON study allowed radiologists to manually adjust the screen result • At baseline, radiologists adjusted the result in about one in 20 cases (95.4 % downwards) • Radiologists' adjustments led to a 22 % reduction of falsepositive screen results • Radiologists' expertise can improve nodule classification in addition to a nodule protocol