On pathological examinations, the specimen sizes ranged from 10 mm to 53 mm (mean size, 31 mm), and the lesions ranged in size from 3 mm to 25 mm (mean size, 12 mm). The pathological diagnoses of the 72 lesions were a benign lesion in 34 cases (47%) (Fig. 2), a high-risk lesion in 30 cases (42%) (Fig. 3) and a malignant lesion in eight cases (11%) (Figs. 4, 5) (Table 2). In the eight malignant lesions and nine high-risk lesions, re-excision was performed after US-guided localization with focal excision, and follow-up US images were obtained before re-excision. There was no residual lesion seen at an excision site on a follow-up US examination for these 17 lesions. All eight malignant lesions were treated surgically with the use of a modified radical mastectomy in four cases and with breast conserving surgery in four cases. Four of the eight malignant were identified as residual carcinomas by a pathological examination after surgery. These four malignant lesions with residual carcinomas were further characterized as three ductal carcinomas in situ and one invasive carcinoma, with sizes ranging from 3 to 9 mm (mean size, 5 mm). Eleven high-risk lesions (seven atypical ductal hyperplasic lesions, two radial scars and two intraductal papillomas) that underwent re-excision showed no residual lesions as seen on US or pathological examinations during follow-up.

All 72 lesions were identified by US in the excised specimens. Five (7%) of the 72 lesions were deemed to have been incompletely excised by immersion-US and prompt re-excision was undertaken. A pathological examination was performed with both the initial and re-excised specimens. In two of these five lesions, the initial specimens were benign lesions; however, high-risk lesions were found on the re-excised specimens (one atypical ductal hyperplasia and one sclerosing adenosis).

In terms of image qualities, two (3%) lesions were assessed as grade 1, six (8%) lesions were assessed as grade 2, and 64 (89%) lesions were assessed as grade 3. Satisfactory or excellent images were obtained for the majority of lesions (70/72, 97%).

On postoperative follow-up breast US examinations, there was no residual mass in the surgical field seen in any patient. Thus, the miss rate was 0%.

Precise localization under imaging guidance is necessary before excising nonpalpable breast lesions. Mammography has been widely used as an imaging-guide for needle localization of nonpalpable breast lesions. However, when breast lesions are identified by preoperative US, US-guided needle localization is desirable. Furthermore, when a breast lesion is detected by mammography and US, US is the preferred modality for imaging guidance as the use of breast US has advantages over mammography in terms of needle localization or skin marking (3-5). For example, US-guided needle localization allows the needle to be observed in real time and allows insertion of the needle directly through the lesion. Moreover, US-guided needle localization is faster; mammography-guided needle localization requires two pairs of check mammograms and the overall procedure for US-guided needle localization rarely lasts less than 20 minutes, even if digital mammography or a stereotactic device is used. In addition, the distance between the skin and a breast lesion is usually less for US-guided needle localization than for mammography-guided needle localization. Therefore, US-guided needle localization can be performed more rapidly and produce better cosmetic results (4, 5).

After US-guided needle localization, US of the specimen is needed to verify successful excision. However, US of specimens is not widely used and few reports exist on the topic (4-7). Nevertheless, the studies performed have demonstrated that US of the specimens is effective to identify the presence of a lesion within a specimen. In previous studies, specimen US of specimens was performed in the operating room, and radiologists wore aseptic dress and had to wait until the procedure was required. However, in the present study, we performed US in an ultrasound room in the radiology department. Fortunately, the operating room and the ultrasound room at the Korea University Ansan Hospital are relatively close, and thus, delivery times averaged only 6.4 minutes.

Studies by Fornage et al. (5) and Feld et al. (8) used saline to perform US of specimens as in the present study. However, in the study by Fornage et al. (5), a relatively small number of cases (18 cases) were evaluated. Moreover, the study by Feld et al. (8) included palpable and nonpalpable breast masses. In the present study, we included 72 nonpalpable breast lesions, which was a sufficient number to assess the feasibility of the use of immersion-US.

For US of specimens, the devised immersion-US technique offers advantages versus the in-air conventional US technique in terms of image quality and efficiency. First, it allows better quality US images to be obtained. The immersion technique involves the use of saline, which fills entrapped air pockets between the transducer and specimen, and thus improves image quality. In the present study, satisfactory or excellent images were obtained for 97% of the lesions, and good quality images were obtained for small and superficial lesions. Furthermore, in terms of decision-making for complete excision, we believe that immersion-US is a better method to utilize than surgical bed US examinations as the image qualities are superior and the procedure times are shorter. Surgeons can perform surgical bed US after excision in the operating room; however, surgical disruptions of soft tissue planes, edema, gas, and hemorrhage are all likely to reduce image quality, and under such circumstances, even an expert radiologist can overlook the presence of a residual lesion. The mean time required for surgical bed US found was reported by Feld et al. (8) as 24.2 minutes. The mean time required for immersion-US in the present study was only 5.2 minutes. A second reason for a preference of immersion-US over conventional in-air specimen US is that immersion-US is performed easily and quickly. Furthermore, radiologists do not have to wear aseptic dress or require the need for special tools. In the present study, we used a small plastic container filled with saline and preformed examinations in 5.2 minutes as opposed to the 22 minutes required by Feld and colleagues (8) for the conventional in-air technique.

In the present study, the miss rate was determined by postoperative US imaging. Follow-up postoperative US images were obtained for all 72 lesions, and no residual lesion was found in any case. Initially, five (7%) of these lesions had been incompletely excised as determined by immersion-US and these lesions were re-excised and reexamined. In two of these five lesions, initial specimens were benign lesions; however, high-risk lesions were found on the re-excised specimens. Thus, if immersion-US had not been performed, surgeons might have missed or underdiagnosed these five breast lesions. The failure rate of excision with US needle localization has been reported as 5-22% (1). Localizing needles are usually inserted at an angle and at considerable distances from lesions, which can cause confusion regarding precise lesion locations. In particular, when lesions are small, a radiologist may fail to place a needle within a lesion. Furthermore, if the localizing needle moves or is not properly placed through a lesion, there is a concern that the surgeon will be unable to locate the lesion without an imaging device during surgery. Therefore, we recommend that immersion-US be conducted after the excision of nonpalpable breast lesions to prevent surgical misses.

To assess tumor margins in malignant breast lesions, a study by Mesurolle et al. (6) showed that the use of US of specimens is limited by false negative and false positive results. In that study, US of specimens had a 21% false positive rate and 23% false negative rate for the assessment of tumor margins in malignant lesions based on the pathological findings. Our results indicated that 11 high-risk lesions (seven atypical ductal hyperplasia, two radial scars and two intraductal papillomas) showed no residual lesion on a pathological examination after second wide excision. However, four (50%) of eight malignant lesions had residual carcinomas as determined on a pathological examination after second surgical excision. Three (75%) of these four malignant lesions with residual carcinomas were ductal carcinomas in situ. Silverstein et al. (9) have noted that 45% of cases of ductal carcinoma in situ that were considered as adequately excised had residual disease, as determined either by re-excision or by mastectomy. Moreover, the thicknesses of free margins were found to be directly related to local recurrence-free survival. Therefore, although a malignant breast lesion may have been deemed as successfully excised based on US of the specimen after needle localization, a second wide excision should be performed to achieve a safe free margin.

In conclusion, US of specimens is required after the surgical excision of US-guided needle localized nonpalpable breast lesions to assess surgical success. In this context, the described immersion technique was demonstrated as a straightforward and effective method. In particular, immersion-US also provided good quality US images of small and superficial breast lesions.