Missed or delayed breast cancer diagnoses on mammography remain a persistent challenge in breast imaging, with significant implications for patient outcomes and medico-legal risk. A substantial proportion of cancers that later become clinically apparent are retrospectively visible on earlier mammograms, highlighting the complexity of perceptual and interpretive processes involved in image reading. Errors arise not from a single cause but from the interaction of human cognition, lesion biology, patient characteristics and technical constraints inherent to imaging systems. Understanding how these factors contribute to false-negative interpretations is essential for improving diagnostic performance. A systematic approach that combines structured reading strategies, continuous education, quality assurance and supportive technologies has been shown to reduce variability and enhance accuracy in mammography interpretation, supporting earlier detection and more timely treatment.
Must Read:DL Pipeline Advances Breast Cancer Diagnosis
Reader-Related Factors and Cognitive Bias
Radiologist-related factors play a central role in missed breast cancers, particularly perceptual and interpretive errors. Cognitive biases influence judgement and decision-making, often under conditions of high workload, time pressure or distraction. Satisfaction of search can lead to premature termination of image review after identification of an initial abnormality, reducing vigilance for additional findings, including contralateral disease. Anchoring and confirmation bias may result in undue reliance on prior reports or initial impressions, even when new or discordant features are present. Availability bias can skew interpretation towards recently encountered diagnoses, while blind spot bias reflects inattention to peripheral or less routinely scrutinised regions of the breast.
These cognitive tendencies are compounded by fatigue and interruptions, which reduce sustained attention during screening interpretation. Structured search patterns, diagnostic checklists and consistent review of the entire breast, including known blind spots, are key mitigation strategies. Double reading, peer review and systematic feedback help identify individual patterns of error and support reflective practice. Integration of artificial intelligence-based computer-aided detection systems can provide a consistent second reader, flagging subtle abnormalities that may otherwise be overlooked.
Lesion Characteristics That Obscure Detection
Intrinsic features of breast lesions significantly affect their detectability on mammography. Some cancers grow slowly or infiltrate tissue diffusely, producing minimal interval change and a false impression of stability. Low-grade ductal carcinoma in situ and invasive lobular carcinoma are notable examples, often lacking a discrete mass or clear progression over short timeframes. Conversely, high-grade tumours may present with smooth or circumscribed margins that mimic benign morphology, increasing the risk of misclassification.
Lesions visible on only one mammographic view present another diagnostic challenge and are frequently dismissed as tissue overlap or benign asymmetry, despite a meaningful proportion representing malignancy. Isodense asymmetries, particularly in dense breasts, can blend into surrounding fibroglandular tissue and require meticulous comparison with multiple prior examinations to identify subtle evolution. Architectural distortion and faint microcalcifications are among the most commonly missed findings, especially when they lack associated masses or are located at the periphery of the image. Digital breast tomosynthesis, magnification views and targeted ultrasound or magnetic resonance imaging play an important role in clarifying these subtle abnormalities and guiding further assessment.
Patient and Technical Influences on Image Quality
Patient-specific variables further complicate interpretation. Dense breast tissue both increases cancer risk and reduces mammographic sensitivity by masking lesions through tissue overlap. Anatomical and postural challenges, such as scoliosis, limited shoulder mobility or obesity, can lead to suboptimal positioning and incomplete visualisation of breast tissue. Prior surgery, implants or augmentation materials introduce additional complexity, as scarring, distortion and foreign substances may obscure or mimic pathology.
Technical factors are equally critical. Inadequate positioning, improper compression, motion artefacts and suboptimal exposure settings can conceal early-stage cancers that present with subtle radiographic signs. Excessive or insufficient compression and poorly maintained or outdated equipment degrade image quality and diagnostic confidence. Adherence to established image quality criteria, routine equipment calibration and robust quality assurance programmes are essential to minimise technical sources of error. Well-trained technologists and a willingness to repeat inadequate images are fundamental to ensuring comprehensive breast coverage and optimal contrast.
Missed breast cancers on mammography result from the interplay of cognitive, biological, patient-related and technical factors rather than isolated failures. While the consequences for patients and healthcare systems are substantial, these errors are not inevitable. Structured reading approaches, awareness of cognitive bias, continuous education and peer learning form the foundation of improved interpretive performance. Advances in imaging technology, including tomosynthesis and artificial intelligence tools, further support radiologists by enhancing lesion conspicuity and providing consistent secondary review. Together, these strategies promote earlier detection, more accurate diagnosis and improved outcomes for individuals undergoing breast cancer screening.
Source: Insights into Imaging
Image Credit: iStock
