Breast ultrasound (US) does not always show lesions as discrete, three-dimensional masses. Some findings appear as a distinct area of altered echotexture without a definable shape or margins, commonly described as nonmass lesions (NMLs). These findings are increasingly encountered in clinical practice, yet standardised terminology and reporting remain less mature than for masses. Variability in how radiologists recognise and label NMLs can affect communication, follow-up and biopsy decisions. A single-centre retrospective analysis of biopsy-proven breast US findings assessed how consistently radiologists classified lesions as masses or NMLs after targeted training. It also examined which US characteristics were linked to greater diagnostic confidence when identifying NMLs, with attention to how these features related to more consistent reader agreement.
How Nonmass Lesions Fit into Breast US Reporting
A mass at US is defined as a three-dimensional, space-occupying lesion visible in at least two planes, typically described by shape, margins and internal echogenicity. By contrast, an NML is characterised as a discrete abnormal area distinguishable from surrounding tissue but lacking a mass-like margin and lacking a specific shape. Professional guidance outside BI-RADS has addressed NMLs for years, and an updated US lexicon has been under review, reflecting growing recognition that nonmass presentations require clearer description.
In the analysis, breast radiologists classified each lesion as either a mass or an NML using a shared working definition supported by dedicated training with example cases. The cohort consisted of 1067 biopsy-proven lesions in 912 women, with a mean age in the early fifties. Readers were blinded to pathology but had clinical indication and lesion location. Lesions were then grouped by how consistently readers agreed: those called masses by most readers were treated as definite masses, those called NMLs by strong consensus were treated as definite NMLs, and those with split opinions were labelled controversial NMLs. This framework allowed comparison between confidently recognised NMLs and those that repeatedly created ambiguity.
Must Read:Super-Resolution Ultrasound in Breast Imaging
Agreement Between Readers and What Was Most Often Classified
Most lesions in this biopsy-selected cohort were classified as masses. Definite masses accounted for roughly four in five cases, while definite NMLs represented a smaller minority, with an additional group of controversial NMLs where readers divided between mass and NML categories. Overall, agreement for the mass-versus-NML decision was moderate to substantial. Pairwise concordance was high, and κ values fell in the mid range across reader pairs, with an overall k in the same band.
Importantly, malignancy was not confined to one classification group. The malignancy rate for definite masses was in the mid-thirties percentage range, and the malignancy rate across NMLs was similar. Within the NML category, the malignant proportion was comparable between definite and controversial NMLs. These findings indicate that disagreement about whether a lesion is a mass or an NML does not, by itself, separate lower-risk from higher-risk findings in a population already selected for biopsy. For decision-makers, this reinforces the operational importance of consistent description rather than assuming that uncertainty in lesion type implies reduced clinical significance.
Ultrasound Features Linked to Greater Confidence in NML Classification
Differences between definite and controversial NMLs were not explained by symptoms or by broad categories of mammographic or MRI findings in this dataset. Instead, the clearest signals were specific US characteristics that either anchored readers towards the NML category or pushed them towards disagreement.
Smaller lesions and those showing only subtle architectural distortion were less consistently identified as NMLs. In modelling that compared NML subtypes to lesions without distinctive characteristics, this subgroup was associated with a lower likelihood of consensus classification, with an odds ratio well below 1. In practice, these subtle presentations appear more vulnerable to being interpreted as small masses by some readers and as NMLs by others.
By contrast, certain associated features strengthened consensus. Echogenic foci that corresponded to mammographic calcifications were linked to more consistent classification as NMLs, with an odds ratio above 2. The presence of microcysts within an abnormal area showed an even stronger association with definite NML classification, with an odds ratio above 3. Ductal abnormalities and intraductal lesions also tended to align with NML interpretation, but the analysis did not provide evidence of a clear independent association for this category. Taken together, the results suggest that when an abnormal area includes identifiable internal markers such as microcysts or calcification-linked echogenic foci, radiologists are more likely to converge on a shared interpretation of the lesion as nonmass.
Several contextual factors shape how far these findings can be extended. The analysis was retrospective, performed at a single centre and limited to lesions that underwent biopsy. Readers were experienced breast radiologists working in the same environment and all received focused instruction on NML appearance, which may have improved consistency compared with general practice. The assessments were based on recorded imaging rather than real-time scanning, and examinations were acquired using different systems, which may influence visibility of subtle features.
The analysis shows that, after targeted training, radiologists can achieve moderate to substantial agreement when classifying breast US findings as masses or nonmass lesions, even though NMLs remain less common and less standardised in routine reporting. Disagreement was most evident for smaller or subtly distorted abnormalities, while the presence of microcysts and echogenic foci aligned with calcifications supported stronger consensus. Similar malignancy rates across masses and NMLs, including those that were controversial, underline the need for consistent terminology and careful feature-based description rather than reliance on lesion type alone. The findings support ongoing efforts to refine NML definitions and descriptors to improve reporting consistency and downstream decision-making in breast imaging pathways.
Source: Radiology
Image Credit: iStock
References:
Eom HJ, Cha JH, Cho SM et al. (2025) Interreader Agreement and Diagnostic Confidence in Discriminating Masses and Nonmass Lesions at Breast US. Radiology; 317:3.
