The Power of the Probe: Targeted US Alone for Evaluation of Breast Symptoms

See also the article by Appelman et al in this issue.

Download as PowerPoint

Focal breast complaints are a common malady in the United States, accounting for 3% of family physician office visits and experienced by 16% of women over any 10-year period (22.8 presentations per 1000 person-years) (1). Most of these symptoms are not caused by cancer, but the likelihood of cancer is high enough that focal clinically suspicious symptoms should be evaluated with diagnostic imaging, especially in the case of palpable breast lumps, where the positive predictive value for cancer ranges from 8.1% to 24.6% (2). This combined frequency and urgency compels us as radiologists to assess these patients in the most accurate manner, but also in a way that maximizes efficiency of workflow and resource utilization.

In the United States, the currently accepted imaging approach for evaluating focal breast symptoms has generally included stratification by age. For patients who are younger than 30 years, US is the primary assessment tool, due in part to the low incidence of breast cancer in this age group. For those aged 40 years and older (screening age in the United States), diagnostic mammography is usually followed by targeted US. For those aged 30–39 years, the approach is variable, with targeted US or mammography seen as reasonable primary evaluation modalities (3,4).

In this issue of Radiology, Appelman et al (5), in their elegant study, evaluated whether patients with focal breast complaints can be accurately evaluated with US as a stand-alone tool, removing the need to also use digital breast tomosynthesis (DBT) and thereby decreasing cost and increasing imaging efficiency. Their multicenter prospective Breast US Trial (BUST), carried out in the Netherlands, evaluated 1961 women with a broad and inclusive variety of focal breast complaints, most commonly a palpable lump (78%). The mean age of the participants was 47 years (range, 30–90 years; 73% of patients were aged ≥40 years). All participants were evaluated in a “flipped” manner, with targeted US (and biopsy completed as indicated) performed before DBT. The readers had 3–30 years of experience, and five trainees were involved in the study. The main goal was to determine the frequency of cancer detected only with DBT at the site of clinical concern when the US findings were negative (ie, false-negative US findings). The authors also assessed the accuracy of US alone, how often cancer was found with DBT elsewhere in the breasts (incidental cancers), and the overall sensitivity of DBT plus US in finding cancer in this cohort. One year follow-up was used to establish truth.

Ten percent of participants were found to have cancer at the symptomatic site. Targeted US as a solo tool had 98.5% sensitivity (vs 87% for DBT alone), 90.8% specificity, and 90% overall accuracy. The negative predictive value (NPV) for US was 99.8%. US failed to depict three cancers at the site of concern that were then found with DBT. The addition of DBT to the work-up increased the symptomatic cancer detection rate (CDR) by one per 1000 women. DBT also depicted eight incidental cancers in women in whom the focal complaint had a benign underlying cause (one ipsilateral, seven contralateral), for an additional CDR of 4.5 per 1000 women. None of these malignant additional (nonsymptomatic) DBT findings occurred in women younger than 40 years. All three symptomatic cancers missed at US were visible in retrospect at second-look US after DBT depicted a suspicious finding. Of note, during the follow-up period, seven cancers were subsequently diagnosed in the area of the original symptoms, changing the sensitivity of US alone to 95% (vs 96.5% for US plus DBT). The authors concluded that US is accurate as a stand-alone tool in the evaluation of focal breast complaints.

Recent studies support this conclusion. Holtz and colleagues (6) showed that targeted US used to evaluate symptoms in pregnant women of all ages (most aged 40 years or older) had 100% sensitivity and NPV for cancer detection. In lactating women, Chung et al (7) noted US to have sensitivity and NPV of 100%, and they also found that adding mammography did not lead to supplemental cancer detection but did lead to false-positive findings. A previous study by Appelman and associates (8) showed low added value of mammography over US alone (which had an NPV of 99.8%) in the evaluation of focal complaints, especially if the target site was benign. A review of the literature by Dodelzon and Katzen (9) showed that for evaluation of palpable breast abnormalities, the performance of US was superior to that of mammography across each included study, although they also noted that the highest NPV was achieved when US and mammography were used in combination. Leung et al (10) outlined lack of value added by mammography in patients presenting with a new lump who had had a negative mammogram within the prior 12 months. US depicted all interval cancers present. Only three cases showed mammographic change at the clinical site not also detected at US, with all three of these findings shown to be benign.

Appelman et al (5) note that bias may have been introduced into their study because they performed biopsies before obtaining DBT images, possibly resulting in obscuration of potential findings. However, it is also possible the presence of a marking clip alerted readers to findings that may have otherwise been overlooked, an opposite bias. Of important note, US scanning in this trial was performed exclusively by breast imaging physicians, rather than technologists. This highlights the need for close oversight and meticulous technique if US is to be relied on as the sole interrogation method for patients with focal breast problems. This is further emphasized by the fact that all three missed symptomatic cancers were seen in retrospect upon repeat scanning of the area after DBT showed a suspicious finding.

So how should symptomatic patients potentially be imaged based on the results of this study and the other accumulating literature? The data suggest the following as a possible approach: Patients younger than 40 years can be safely evaluated with US alone. Women aged 40 years and older can potentially undergo US alone unless they are due for screening mammography, in which case bilateral mammography should also be performed. The literature shows that the incidental CDR (cancers detected remote from area of clinical concern) in this age group is similar to the CDR for routine screening. Women of screening age who have had a negative mammogram within a year can most likely be evaluated with US alone, removing a potentially superfluous diagnostic mammographic examination. For all ages, mammography can be added if suspicious US findings are identified, cancer is diagnosed (to determine extent of disease), or the clinical finding is highly suspicious and US findings are negative.

Some might ask why we would not routinely perform bilateral mammography in symptomatic women in their 30s as an “opportunistic” screening encounter; although this study showed no incidental cancers in women younger than 40 years, other studies (9) have shown an incidental CDR in the range of two to eight per 1000 women. As always, additional research, including more data around the incidental CDR in women younger than 40 years, as well as a prospective trial to assess for difference in interval cancer rates in symptomatic patients in whom diagnostic DBT is excluded from the protocol (as suggested by the authors), will help us continue to triangulate toward the Polaris of true best practices.