Missed Pulmonary Nodules on CT: Determinants and Implications for Radiologic Practice

Undetected or unreported lung nodules are a significant cause of failure-to-diagnose errors in radiology. The available literature indicates consistent characteristics of nodules that radiologists find difficult to detect. For example, small nodules and pure ground-glass nodules are both more likely to be missed. Studies also indicate that adjacent structures influence nodule detection, as nodules are more likely to be missed in anatomically complex regions, including endobronchial and perivascular locations. Missed nodules additionally occur secondary to biases and other cognitive and perceptual limitations in conjunction with contextual factors such as reader fatigue. Although advances in CT technology and computer-aided detection can improve radiologists' performance, significant limitations remain, and an understanding of the contributors to missed nodules may help radiologists adopt mitigation strategies. Current medicolegal frameworks rely on subjective notions of obviousness; however, more objective characterizations of missed nodules could provide the evidence needed to refine these frameworks. This Review discusses factors influencing nodule detection on chest CT, focusing on three broad categories of factors that may result in a nodule being overlooked: technical limitations, intrinsic nodule characteristics and characteristics of adjacent structures, and reader factors. We additionally explore the influence of artificial intelligence and comment on medicolegal ramifications.