Search

Magnetic resonance imaging (MRI)-detected lesions are often category 2 or 3 lesions on initial ultrasound examination. In addition, in the case of new non-mass lesions detected on MRI, one would expect to find lesions with ductal dilatation with minimal secretory accumulation, single short lesions with ductal dilatation, cyst-like lesions less than 5 mm in size, mammary gland-like lesions less than 8 mm in size, and very indistinct lesions. Detection is expected to be even more difficult. Currently, there are no clear uniform criteria for the indication of second-look ultrasonography (US) for MRI-detected lesions, so it is not possible to make a general comparison, but recent studies have indicated that the ratio of mass to non-mass MRI-detected lesions is 7:3. And it has been pointed out that the percentage of malignancy is about 30% for each. Before about 2012, the US detection rate was about 70%, and MRI-guided biopsies of undetected lesions showed a small percentage of malignant lesions. Therefore, some observers believe that lesions not detected on US should be followed up, while others believe that MRI-guided biopsy should be performed. Recently, however, the use of surrounding anatomical structures as landmarks for second-look US has increased the detection rate to as high as 87–99%, and the percentage of malignancy remains the same. In addition, recent surveillance of high-risk breast cancer requires careful management of MRI-detected lesions. In this review, we will discuss the literature on MRI-detected lesions and describe ultrasound techniques to accurately detect small lesions and reliably reveal pale lesions based on their structural differences from their surroundings.

Purpose : Ultrasound-guided breast tissue biopsy is an essential technique for diagnosing breast disease, but sample errors reduce its accuracy. This study investigated whether the histopathological results can be inferred from the macroscopic findings for Ultrasound-guided breast Vacuum Assisted Biopsy (VAB) specimens. Methods: Biopsy specimens from 101 patients who underwent mammary gland VAB were photographed with a smartphone, and the relationships between the macroscopic findings and the pathological results were examined. Results : A significant difference was observed with regard to the presence / absence of turbidity: malignancy was detected in 33 / 37 (89%) specimens with turbidity and in 2 / 47 (4%) cases without turbidity (p < 0.001). A significant difference was also observed regarding the surface properties : malignancy was detected in 14 / 70 (19%) smooth specimens and in 24 / 29 (83%) rough specimens (p < 0.001). Also, malignancy was detected in 11 / 13 (85%) specimens with white spots, and the difference was significant (p < 0.001). In addition, the characteristics of intraductal papilloma, fibroadenoma, and mastopathy could be confirmed by macroscopic findings. Conclusions : When needle–biopsy of a lesion that is targeted for resection yields macroscopic findings that match the predicted histopathological findings, it can be thought that the biopsy had been properly performed. This means that false–negatives due to poor specimens can be prevented.

Purpose Strain elastography for imaging lesion stiffness is being used as a diagnostic aid in the malignant/benign discrimination of breast diseases. While acquiring elastography in addition to B-mode images has been reported to help avoid performing unnecessary biopsies, intraductal lesions are difficult to discriminate whether they are malignant or benign using elastography. An objective evaluation of strain in lesions was performed in this study by measuring the elasticity index (E-index) and elasticity ratio (E-ratio) of lesions as semi-quantitative numerical indicators of the color distribution of strain. We examined whether ductal carcinoma in situ (DCIS) and intraductal papilloma could be distinguished using these semi-quantitative numerical indicators. Methods In this study, 170 ultrasonographically detected mass lesions in 162 cases (106 malignant lesions and 64 benign lesions)—in which tissue biopsy by core needle biopsy and vacuum-assisted biopsy, or surgically performed histopathological diagnosis, was performed—were selected as subjects from among 1978 consecutive cases (from January 2014 to December 2016) in which strain elastography images were acquired, in addition to standard B-mode breast ultrasonography, by measuring the E-index and E-ratio. Results The cut-off values for E-index and E-ratio in the malignant/benign discrimination of breast lesions were determined to be optimal values at 3.5 and 4.2, respectively, based on receiver operating characteristic (ROC) curve analysis. E-index sensitivity, specificity, accuracy, and AUC value (area under the curve) were 85%, 86%, 85%, and 0.860, respectively, while those for E-ratio were 78%, 74%, 74%, and 0.780, respectively. E-index yielded superior results in all aspects of sensitivity, specificity, accuracy, and AUC values, compared to those of E-ratio. The mean E-index values for malignant tumors and benign tumors were 4.46 and 2.63, respectively, indicating a significant difference (P Conclusion E-index in strain elastography yielded better results than E-ratio in the malignant/benign discrimination of breast diseases. On the other hand, E-index has a high false-negative rate and false-positive rate for intraductal lesions, a factor which should be taken into account when making ultrasound diagnoses.

Deep learning techniques have achieved remarkable success in lesion segmentation and classification between benign and malignant tumors in breast ultrasound images. However, existing studies are predominantly focused on devising efficient neural network-based learning structures to tackle specific tasks individually. By contrast, in clinical practice, sonographers perform segmentation and classification as a whole; they investigate the border contours of the tissue while detecting abnormal masses and performing diagnostic analysis. Performing multiple cognitive tasks simultaneously in this manner facilitates exploitation of the commonalities and differences between tasks. Inspired by this unified recognition process, this study proposes a novel learning scheme, called the cross-task guided network (CTG-Net), for efficient ultrasound breast image understanding. CTG-Net integrates the two most significant tasks in computerized breast lesion pattern investigation: lesion segmentation and tumor classification. Further, it enables the learning of efficient feature representations across tasks from ultrasound images and the task-specific discriminative features that can greatly facilitate lesion detection. This is achieved using task-specific attention models to share the prediction results between tasks. Then, following the guidance of task-specific attention soft masks, the joint feature responses are efficiently calibrated through iterative model training. Finally, a simple feature fusion scheme is used to aggregate the attention-guided features for efficient ultrasound pattern analysis. We performed extensive experimental comparisons on multiple ultrasound datasets. Compared to state-of-the-art multi-task learning approaches, the proposed approach can improve the Dice’s coefficient, true-positive rate of segmentation, AUC, and sensitivity of classification by 11%, 17%, 2%, and 6%, respectively. The results demonstrate that the proposed cross-task guided feature learning framework can effectively fuse the complementary information of ultrasound image segmentation and classification tasks to achieve accurate tumor localization. Thus, it can aid sonographers to detect and diagnose breast cancer.

Background Extensive vaccination programs are being implemented worldwide for coronavirus disease 2019 (COVID-19). With the spread of vaccination, swelling of the lymph nodes after vaccination is frequently seen. We encountered a patient who developed left axillary lymphadenoma following vaccine administration. Case presentation The patient was a Japanese woman in her 80 s who had previously undergone surgery for right breast cancer. She received two injections of the Pfizer-BioNTech COVID-19 vaccine in her left arm. Approximately 3 months later, she complained of left axillary swelling, and imaging resulted in a diagnosis of left axillary lymphangioma. In accordance with the patient’s wishes, we performed axillary mass resection. The pathological diagnosis was lymphangioma. Conclusion Our examination findings indicated that congestion of the axillary lymph vessels might have been caused by upper-arm injections of the COVID-19 vaccine.

Background: If the histopathological results could be inferred from the macroscopic findings for breast biopsy specimens, false-negative needle biopsies due to poor specimens could be prevented. However, no research on this has been reported to date.Methods: Biopsy specimens from 101 patients who underwent mammary gland Vacuum Associated Biopsy (VAB) were photographed with a smartphone, and the relationships between the macroscopic findings (presence/absence of turbidity, surface properties, presence/absence of white spots, and characteristic findings) and the pathological results (benign/malignant) were examined.Results: A significant difference was observed with regard to the presence/absence of turbidity: malignancy was detected in 33/37 (89%) specimens with turbidity and in 2/47 (4%) cases without turbidity (p

Background Second-look ultrasonography (US) is commonly performed for breast lesions detected using magnetic resonance imaging (MRI), but the identification rate of these lesions remains low. We investigated if US methods using anatomical breast structures can improve the lesion identification rate of MR-detected lesions and evaluated the diagnostic performance of fine-needle aspiration cytology (FNAC) of the second-look US using the above-mentioned method. Methods We retrospectively assessed 235 breast lesions (hereinafter, “targets”) subjected to second-look US following MRI between January 2013 and September 2015. US was employed using the conventional methods, and this assessment measured the positional relationships of lesions with regard to surrounding anatomical breast structures (glandular pattern, Cooper’s ligaments, adipose morphology, and vascular routes). Associations were assessed among the following variables: the MRI findings, target size, identification rate, and main US indicators that led to identifying the target; FNAC results and MRI findings; MRI findings and histopathological findings; and FNAC results and histopathological findings. Moreover, the sensitivity and specificity of FNAC were determined. Results The identification rate was 99%. The main US indicators leading to identification were a glandular pattern (28–30% of lesions) and other breast structures (~ 25% of lesions). FNAC was performed for 232 targets with the following results: sensitivity of 85.7%, specificity of 91.6%, PPV of 94.1%, NPV of 92.9%, false-negative rate of 14.3%, false-positive rate of 2.1%, and accuracy of 89.7%. Conclusions Second-look US using anatomical breast structures as indicators and US-guided FNAC are useful for refining the diagnosis of suspicious breast lesions detected using MRI.

The use of color Doppler ultrasound (CD) for distinguishing between benign and malignant breast lesions remains controversial. This study (JABTS BC-04 study) was aimed at confirming the usefulness of our CD diagnostic criteria. We evaluated ultrasound images of 1408 solid breast masses from 16 institutions in Japan (malignant: 839, benign: 569). Multivariate analysis indicated that vascularity (amount of blood flow), vascular flow pattern ("surrounding marginal flow" or "penetrating flow") and the incident angle of penetrating flow were significant findings for distinguishing between benign and malignant lesions. However, the sensitivity and specificity of B-mode alone did not improve significantly with CD addition (97.6% → 97.9%, 38.3% → 41.5%, respectively). We explored the causes of these negative results and found that age should have been considered when evaluating vascularity. Simulation experiments suggested that specificity is significantly improved when age is taken into consideration (38.3% → 46.0%, p0.001) and we thereby improved our diagnostic criteria.

This paper presents a method of segmenting mammary gland tissue for breast ultrasound images. In order to improve the quality of diagnoses for breast ultrasound screening, Computer Aided Detection (CADe) systems have been developed by a number of researchers. However, such systems suffer in terms of making many false positives for non-mammary gland tissue where no cancer exists. Accordingly, it is necessary to accurately determine the area of mammary gland tissue within ultrasound images for CADe systems. In this study, we propose a method of segmenting mammary gland tissue from ultrasound images for breasts. Within ultrasound images, brightness is attenuated as ultrasound waves move from the ultrasound probe to internal tissue. Accordingly, we correct intensity (brightness) using zero-phase component analysis (ZCA) whitening in order to reduce such brightness attenuation. Moreover, we apply a Conditional Random Field (CRF) which can render as smooth the region boundary of ultrasound images. The results of an experiment, conducted with five target patients to verify the effectiveness of the proposed methods, demonstrate that the method is capable of segmenting mammary gland tissue from breast ultrasound images.

The indication for postmastectomy radiotherapy (PMRT) in breast cancer patients with one to three positive lymph nodes has been in discussion. The purpose of this study was to identify patient groups for whom PMRT may be indicated, focusing on varied locoregional recurrence rates depending on lymphatic invasion (ly) status.Retrospective analysis of 1,994 node-positive patients who had undergone mastectomy without postoperative radiotherapy between January 1990 and December 2000 at our hospital was performed. Patient groups for whom PMRT should be indicated were assessed using statistical tests based on the relationship between locoregional recurrence rate and ly status.Multivariate analysis showed that the ly status affected the locoregional recurrence rate to as great a degree as the number of positive lymph nodes (p0.001). Especially for patients with one to three positive nodes, extensive ly was a more significant factor than stage T3 in the TNM staging system for locoregional recurrence (p0.001 vs. p = 0.295).Among postmastectomy patients with one to three positive lymph nodes, patients with extensive ly seem to require local therapy regimens similar to those used for patients with four or more positive nodes and also seem to require consideration of the use of PMRT.

Background With the recent advances in modalities for diagnostic imaging of the breast, it is now essential to detect isoechoic masses and small nonmass lesions, to which little attention has so far been paid using ultrasound (US) of the breast. It will be possible with the observation method to understand normal breast structural images and anatomy. We elucidated the detailed histological architecture of the normal breast, information indispensable for diagnostic US of the breast. Methods Verification of the above hypotheses was carried out using the breasts of 11 patients who underwent total mastectomy at our clinic. Results Isoechoic structures with fat are lobules, all ducts, and surrounding stroma that support the ducts; intervening hyperechoic areas are edematous stroma and fat-containing stroma that support the breast. By taking an isoechoic structure that reflects the course of the ducts as the basic structure for observation, the boundary between the lobes can be inferred. Conclusions Detection of deviations from the normal structure using the method for interpreting three-dimensional ultrasound images of mammary lobes is a radical new approach for diagnosing breast cancer. This technique is very simple and amenable to standardization once one understands the underlying theory. Furthermore, it is useful as a screening method as well as for easy detection of faint minute lesions that can only be detected by magnetic resonance imaging or second-look targeted US.

Background With the increasing use of high-resolution ultrasound (US) examination, many breast carcinomas that cannot be identified by mammography (MMG) alone have been detected. Many of these carcinomas are ductal carcinoma in situ (DCIS) and small-sized invasive carcinomas. Until date, DCISs have often been described as palpable masses with calcifications on MMG, but what are the characteristics of DCISs that are detectable by US alone? Methods One hundred fifty cases with DCIS that we experienced at our clinic from 2003 to 2007 were classified into 47 cases (echo group) diagnosed by US alone and 103 cases (MMG/PE group) diagnosed by MMG or clinically. Results US findings of the echo group showed cystic or solid lesions in 37 cases (79%). The mean age of the echo group was significantly higher than that of the MMG/PE group (59.6 vs. 51.2 years; P

Ductal carcinoma in situ (DCIS) has been detected primarily by palpation, calcification on mammogram (MMG), and nipple discharge. Ultrasonographic investigation of DCIS has been performed, but has not up to now revealed any form of DCIS that could not be discovered by the above three methods. In the present report, we describe a form of DCIS that has not been detected by earlier methods, and propose to refer to this new DCIS as 3 non-DCIS to distinguish it from conventional DCIS: 3 non-DCIS represents nonpalpable DCIS without calcification on MMG and without nipple discharge. Only the method of ultrasound screening described in this chapter can detect 3 non-DCIS. We diagnosed 23 patients as 3 non-DCIS at this institution between May 1997 and March 2003. There was neither calcification, mass, nor distortion on MMG in these patients. On ultrasonography, a small mass measuring 10mm or less was revealed. Thorough examination by fine-needle aspiration cytology showed 90% of the patients were positive. On analysis of subtype of DCIS, 22 of the 23 cases of DCIS were classified into pure noncomedo type and 1 was mixed type. Lesions in 3 non-DCIS were frequently less extensive than those in conventional DCIS.