Combined Use of Cell Block and Smear Improves the Cytological Diagnosis of Malignancy in Non-Palpable Breast Lesions Screened by Imaging.

Background: Currently, core needle biopsy is replacing fine needle aspiration biopsy (FNAB) for pathological diagnosis of breast lesions. However, FNAB is extensively used for diagnosing breast lesions, including screened lesions, at our hospital. Furthermore, direct smears as well as cell blocks (CBs) from the FNAB specimens have been used. To prepare the CBs, hematoxylin and eosin (HE) staining as well as immunostaining with a mixture of p63 and cytokeratin 5/6 antibodies are routinely used. Therefore, in the current study, we sought to assess the efficacy of diagnosing breast lesions using conventional smears and CB immunostaining. Methods: Breast FNAB reports of direct smears and CBs from The Nagoya Medical Center between December 2014 and March 2020, were reviewed. The efficiency of diagnoses made with direct smears and CBs were compared using histology-based diagnoses. Results: Among the 169 histologically confirmed malignant lesions, 12 lesions that were reported as unsatisfactory, benign, or atypia probably benign, using direct smears were diagnosed as malignant using CB. Histologically, these lesions were diagnosed as carcinomas with mild atypia or papillary structures. Ten (83.3%) of the twelve lesions were non-palpable and only detected upon imaging. Conclusion: Combined use of CB and conventional smear leads to the detection of more malignant lesions in breast FNAB specimens, particularly in lesions detected by imaging alone. Immunostaining of CB sections using a mixture of p63 and cytokeratin 5/6 antibodies provides more information than HE staining alone. Breast FNAB with CB preparation can be successfully applied for evaluation of breast lesions in developed countries.

Validity of computed mean compressed fibroglandular tissue thickness and breast composition for stratification of masking risk in Japanese women.

BACKGROUND: The volumetric measurement system for mammographic breast density is a high-precision objective method for evaluating the percentage of fibroglandular tissue volume (FG%). Nonetheless, FG% does not precisely correlate with subjective visual estimation (SVE) and shows poor evaluation performance regarding masking risk in patients with comparatively thin compressed breast thickness (CBT), commonly found in Japanese women. We considered that the mean compressed fibroglandular tissue thickness (mCGT), which incorporates the CBT element into the evaluation of breast density, may better predict masking risk. METHODS: Volumetric measurements and SVEs were performed on mammograms of 108 breast cancer patients from our center. mCGT was calculated as the product of CBT and FG%. SVE was classified using the Breast Imaging-Reporting and Data System classification, 5th edition. Subsequently, the performance of mCGT, SVE, and FG% in predicting masking risk was estimated using the AUC. RESULTS: The AUC values of mCGT and SVE were 0.84 (95% confidence interval, 0.71-0.92) and 0.78 (0.66-0.86), respectively (P = 0.16). The AUC of the FG% was 0.65 (0.52-0.77), which was significantly lower than that of mCGT (P < 0.001). The sensitivity and specificity of mCGT in predicting negative detection were 89% and 71%, respectively; of SVE 83% and 61% (versus 72% and 57% with FG%), suggesting that mCGT was superior to FG% in both sensitivity and specificity, and comparable with SVE. CONCLUSIONS: Objective mCGT calculated from the volumetric measurement system will highly likely be useful in evaluating breast density and supporting visual assessment for masking risk stratification.

Improving breast mass classification by shared data with domain transformation using a generative adversarial network.

Training of a convolutional neural network (CNN) generally requires a large dataset. However, it is not easy to collect a large medical image dataset. The purpose of this study is to investigate the utility of synthetic images in training CNNs and to demonstrate the applicability of unrelated images by domain transformation. Mammograms showing 202 benign and 212 malignant masses were used for evaluation. To create synthetic data, a cycle generative adversarial network was trained with 599 lung nodules in computed tomography (CT) and 1430 breast masses on digitized mammograms (DDSM). A CNN was trained for classification between benign and malignant masses. The classification performance was compared between the networks trained with the original data, augmented data, synthetic data, DDSM images, and natural images (ImageNet dataset). The results were evaluated in terms of the classification accuracy and the area under the receiver operating characteristic curves (AUC). The classification accuracy improved from 65.7% to 67.1% with data augmentation. The use of an ImageNet pretrained model was useful (79.2%). Performance was slightly improved when synthetic images or the DDSM images only were used for pretraining (67.6 and 72.5%, respectively). When the ImageNet pretrained model was trained with the synthetic images, the classification performance slightly improved (81.4%), although the difference in AUCs was not statistically significant. The use of the synthetic images had an effect similar to the DDSM images. The results of the proposed study indicated that the synthetic data generated from unrelated lesions by domain transformation could be used to increase the training samples.

Polygon method: A systematic margin assessment for breast conservation.

BACKGROUND: Radiation therapy (RT) for women with ductal carcinoma in situ (DCIS) undergoing breast-conserving surgery (BCS) may be overtreatment for some, especially for those in which DCIS is eradicated, and ipsilateral breast tumor recurrence (IBTR) risk approaches the contralateral breast cancer (CBC) level. The aim of this study was to clarify whether the polygon method, a new systematic method of en face (tangential, shaved) margin assessment, can identify a subset of DCIS that can be safely treated by BCS alone. METHODS: A key tool of the polygon method is an adjustable mold that prevents the "pancake phenomenon" (flattening) of breast tissue after surgical removal so that the specimen is fixed in the shape of a polygonal prism. This preanalytical procedure enables us to command a panoramic view of entire en face margins 3-5-mm deep from the real peripheral cut surfaces. Competing risk analysis was used to quantify rates of IBTR and CBC and to evaluate risk factors. RESULTS: From 2000 to 2013, we identified 146 DCIS patients undergoing BCS with a contralateral breast at risk. In 100 DCIS patients whose margin was negative by the polygon method, 5 IBTR (3 DCIS and 2 invasive ductal carcinoma [IDC]) and 10 CBC (6 DCIS and 4 IDC) cases were identified during a median follow-up of 7.6 years (range, 0.9-17.4). Five- and 10-year cumulative incidence rates were 3.0% and 5.3% for IBTR, and 7.1% and 13.3% for CBC, respectively. Thus, patients with a negative margin consistently showed at least twofold lower IBTR than CBC despite omission of RT. CONCLUSIONS: Japanese women classified with a negative margin by the polygon method show a very low risk of IBTR and account for approximately half of CBC cases. In this subset of DCIS patients, additional RT is not beneficial.

Can quantitative evaluation of mammographic breast density, "volumetric measurement", predict the masking risk with dense breast tissue? Investigation by comparison with subjective visual estimation by Japanese radiologists.

BACKGROUND: Sensitivity to detect breast cancer (BC) is not high in a dense breast due to masking in mammography. To evaluate the breast density, a volumetric measurement system has been recently developed that measures the percent fibroglandular tissue volume (percent FGV, hereafter termed as "FG%") to the breast volume (BV). This study was designed to investigate whether evaluation using FG% can accurately predict the masking risk by comparing with the current standard method of subjective visual estimation (SVE). METHODS: Using pre-biopsy mammograms of 114 cases histopathologically diagnosed with BC in our facility, SVE based on BI-RADS (5th edition) and volumetric measurements of FG% were conducted. Performance to predict the masking risk was evaluated using the area under the receiver operating characteristic curve (AUC). Relationship between these parameters and the masking risk was evaluated by the adjusted multivariate linear regression analysis. RESULTS: The AUC of SVE values was 0.742 (95% CI 0.641-0.822), while that of FG% was as significantly low as 0.560 (95% CI 0.427-0.685) (P = 0.0014). The SVE values correlated with the detection of BC in mammography (P = 0.0035), but there was no significant relationship with FG% (P = 0.74). The median BV and FGV were 313 cm3 (IQR 191-440) and 63 cm3 (IQR 44-102), respectively. The FGV was comparable to the data for Caucasian women reported in previous studies, but the BV was one-half of the previous data. CONCLUSION: The current volumetric measurement system to evaluate FG% to BV was found to be insufficient in the performance to predict the masking risk in Japanese women with relatively small-sized breasts.

Diagnostic performance of digital breast tomosynthesis and full-field digital mammography with new reconstruction and new processing for dose reduction.

BACKGROUND: The main barrier to adoption of digital breast tomosynthesis (DBT) plus full-field digital mammography (FFDM) is radiation exposure dose. The purpose of this study was to evaluate the diagnostic performance of DBT plus FFDM, both with newly developed technology (nd), at a dose comparable to that of the conventional FFDM alone. METHODS: Nine hundred and thirteen participants were recruited from May 2014 to January 2016 consecutively. For each subject, the exposure setting for DBT(nd) + FFDM(nd) was also used for the conventional FFDM alone. Retrospective reader studies were performed: DBT(nd) + FFDM(nd) (142 cases, including 42 cancer cases) and conventional FFDM (258 cases, including 87 cancer cases). Eight radiologists provided Japanese categorizations and probability of malignancy independently. Diagnostic performance was assessed by comparing sensitivity, specificity, and area under the receiver operating characteristic curve (AUC). Two-sided P values were calculated. RESULTS: DBT(nd) + FFDM(nd) showed sensitivity and AUC significantly increased over the conventional FFDM (85.4 vs. 80.3%, P = 0.015 and 90.9 vs. 88.3%, P = 0.049) and specificity did not significantly increase (89.6 vs. 88.4%, P = 0.52). The mean glandular dose (MGD) difference of DBT(nd) + FFDM(nd) and conventional FFDM was not significant (difference - 0.11 mGy, P = 0.08). CONCLUSIONS: In this study population, DBT plus FFDM, both with newly developed technology, provided diagnostic performance improved over the conventional FFDM alone, even at comparable MGD.

Detectability comparison of modes in dual-mode digital breast tomosynthesis.

BACKGROUND: The objectives of this study were: (1) to evaluate the detectability of full-field digital mammography (FFDM) plus dual-mode digital breast tomosynthesis (DBT) and compare it with that of FFDM alone and (2) to compare the detectability of high-resolution-mode (HR mode used with 40°-angle imaging, 100-µm pixel size, and higher dose) DBT with that of standard-mode (ST mode used with 15°-angle imaging, 150-µm pixel size, and lower dose) DBT for diagnostic evaluation. MATERIALS: The local Institutional Review Board approved this retrospective study of two different sets of cases. All participants gave written informed consent. FFDM and DBT images of 471 women who were recalled were acquired between August 2013 and October 2014. HR mode and ST mode were applied to 155 breasts and 157 breasts, respectively. The cases of both modes were selected randomly. Eight radiologists interpreted the images. The detectability for recall cases and for follow cases, and area under the receiver operating characteristic curve (AUC) were calculated. RESULTS: Adding DBT to FFDM significantly increased the detectability for recall cases and AUC relative to those of FFDM alone (HR mode 8.9 %; 95 % confidence interval (CI) 5.7, 15.0 %; P = 0.013 and 4.9 %; 95 % CI 2.1, 7.7 %; P = 0.001; ST mode 8.3 %; 95 % CI 4.1, 12.1 %; P = 0.007 and 2.9 %; 95 % CI 0.5, 5.3 %; P = 0.02), whereas the detectability for follow cases did not significantly differ. The AUC increase was significantly higher in HR mode than in ST mode (1.5 %; 95 % CI 0.5, 3.7 %; P = 0.023). CONCLUSION: Adding HR-mode or ST-mode DBT to FFDM significantly improved the detectability for diagnostic evaluation case.

Sclerosing adenosis as a predictor of breast cancer bilaterality and multicentricity.

Breast cancer is occasionally complicated by sclerosing adenosis (SA). Although both lesions usually originate in the terminal duct lobular unit, their pathogenetic relationship has not yet been elucidated. The present study analyzed 63 breast cancer patients with SA (involving a total of 75 breasts) to clarify if coexisting SA increased the frequency of multicentric breast cancer or not. Using the topographical classification proposed in our previous study, breast cancers with SA were classified into the following three types: type A (n = 22), cancer area was completely surrounded by the SA; type B (n = 26), cancer area partially overlapped the SA; and type C (n = 27), cancer area was located separate from the SA. Breast cancers with SA had a significant (P < 0.001) increase in frequency of harboring bilateral and multicentric cancers [17 of 63 (27%) and 15 of 63 (24%), respectively] when compared to breast cancer patients without SA, regardless of topographical type. Breast cancers with SA were less invasive (P < 0.001), of lower histological grade (P = 0.034), and had similar frequency of estrogen receptor-positive (P = 0.21) and HER2-positive (P = 0.74) tumors. In conclusion, contralateral and ipsilateral multicentric breast cancers occurred at a higher frequency in those with SA. Our data suggest that SA is, in addition to lobular neoplasia, a predictor of multicentric breast cancers.

Uniqueness of ductal carcinoma in situ of the breast concurrent with papilloma: implications from a detailed topographical and histopathological study of 50 cases treated by mastectomy and wide local excision.

AIMS: To clarify the diagnostic clues of ductal carcinomas in situ (DCIS) associated with papilloma and optimal clinical management of papilloma diagnosed on core needle biopsy (CNB). METHODS AND RESULTS: A total of 50 surgically resected cases were examined histopathologically and topographically. Thirty-nine cases (78%) spread in segmental fashion. Papilloma and DCIS were intermingled closely in 44 cases (88%), occupying the same areas in varying proportions from DCIS-predominant to papilloma-predominant. The two components occupied discrete areas and collided focally in six cases (12%). Most were non-high-grade. Cribriform and solid architectures with fibrovascular stroma were frequent. The cribriform pattern was unique, consisting of fused tubules separated by fibrovascular stroma. Intraductal myoepithelial cells were present to varying degrees in 40 cases (80%). In 38 cases (76%), points were identified where papilloma and DCIS coexisted or collided within a single lumen (CC point). Forty-eight cases (96%) had either intraductal myoepithelial cells or CC points, implying that DCIS and papilloma existed in the same duct system. Radiology showed segmental abnormalities in 83% of the available cases. CONCLUSIONS: Intraductal myoepithelial cells do not always guarantee benignity. Surgical resection is recommended for papilloma in CNB when radiology shows segmental abnormalities.

Representation of lesion similarity by use of multidimensional scaling for breast masses on mammograms.

Presentation of similar reference images can be useful for diagnosis of new lesions. A similarity map which can visually present the overview of the relationship between the lesions with different types may provide the supplemental information to the reference images. A new method for constructing the similarity map by multidimensional scaling (MDS) for breast masses on mammograms was investigated. Nine pathologic types were included; three regions of interests each from the nine groups were employed in this study. Subjective similarity ratings by expert readers were obtained for all possible 351 pairs of masses. Using the average ratings, MDS similarity map was created. Each axis of the MDS configuration was fitted by the linear model with 13 image features to reconstruct the similarity map. Dissimilarity based on the distance in the reconstructed space was determined and compared with the subjective rating. The MDS map consistently represented the similarity between cysts and fibroadenomas, invasive lobular carcinomas and scirrhous carcinomas, and ductal carcinomas in situ, solid-tubular carcinomas, and papillotubular carcinomas with the experts' data. The correlation between the average subjective ratings and the dissimilarities based on the distance in the reconstructed feature space was much greater (-0.87) than that of the dissimilarities based on the distance in the conventional feature space (-0.65). The new similarity map by MDS can be useful for visualizing the relationship between breast masses with different pathologic types. It has potential usefulness in selecting the similarity measures and providing the supplemental information.

[Optimization of exposure conditions for amorphous selenium direct conversion DR-based mammography system].

A new direct-conversion detector for DR mammography has improved the detectability of microcalcifications and masses. Each optimized exposure condition (target/filter combination and tube voltage) was defined through comparison of physical values and visual evaluation on breast specimens using the innovative DR mammography. The contrast-to-noise-ratios (CNRs) of PMMA phantoms of various thicknesses were obtained under a variety of exposure conditions whose average glandular doses (AGDs) were made consistent. Fifty breast specimens were irradiated under these combinations. Visual evaluation was conducted on the images, whose histograms were controlled for consistency. In the phantoms with thicknesses of 20 mm or more, tungsten/rhodium had the highest CNRs of the targets/filters such as molybdenum/molybdenum and molybdenum/rhodium. For visualizing microcalcifications and masses on breast specimens of thicknesses of 35 mm and below, molybdenum/molybdenum was the best. Nevertheless, to obtain better image quality, molybdenum/rhodium was superior for 35-55 mm thickness, and tungsten/rhodium was superior for 55 mm and above under the same AGD, enabling accurate and efficient diagnosis. The study showed that the exposure conditions differ for obtaining the highest CNR using phantoms and those under which breast specimen images allow the most accurate and efficient diagnosis. In addition, image evaluations of the breast specimens allowed optimization of exposure conditions that are closer to those of the actual diagnosis using mammography.

Intracytoplasmic lipid accumulation in apocrine carcinoma of the breast evaluated with adipophilin immunoreactivity: a possible link between apocrine carcinoma and lipid-rich carcinoma.

Although apocrine carcinoma is a distinct histologic entity, there is no immunohistochemical marker to confirm apocrine differentiation with high sensitivity and specificity, and its differential cytologic characteristics are still not fully clarified. Despite the foamy cytoplasm of some apocrine carcinomas and the existence of lipid in the normal apocrine gland, intracytoplasmic lipid in apocrine carcinomas has not been fully explored. By using immunohistochemistry for adipophilin, which is a specific marker of lipid accumulation that can be applied to paraffin sections, we examined intracytoplasmic lipid in apocrine carcinomas. Twenty-four of 26 (92%) apocrine carcinomas and 38 of 116 (33%) nonapocrine carcinomas contained intracytoplasmic lipid. The frequency of adipophilin-positive cases was significantly higher in apocrine carcinomas compared with nonapocrine carcinomas (P<0.01). The positive cell rate per tumor ranged from 10% to 70% (mean, 29%) for apocrine carcinomas. The staining density was heterogeneous from cell to cell. There was no difference in the staining pattern of adipophilin between apocrine ductal carcinoma in situ and invasive apocrine carcinoma or between eosinophilic cells and foamy cells. Sporadic or mosaic distribution of adipophilin-positive cells throughout the tumor and microvesicular or fine granular cytoplasmic staining with heterogeneous density were characteristic features of apocrine carcinoma. Although intracytoplasmic lipid was identified in most apocrine carcinomas, none of the apocrine carcinomas contained prominent intracytoplasmic lipid in >90% of the tumor cells; thus, the criteria for lipid-rich carcinoma was not fulfilled. However, the immunohistochemical study suggests that lipid-rich carcinomas are closely related to apocrine carcinomas.

Topographical, morphological and immunohistochemical characteristics of carcinoma in situ of the breast involving sclerosing adenosis. Two distinct topographical patterns and histological types of carcinoma in situ.

AIM: To examine the histopathological features of 24 surgically resected carcinoma in situ (CIS) involving sclerosing adenosis (SA), with special reference to the topographical relationship between CIS and SA. METHODS AND RESULTS: In 13 (54%) lesions, CIS was entirely surrounded by SA (type A) and in 11 (46%), CIS involved SA at least focally but was not confined to the SA area (type B). The mean size of CIS in type B (30.45 mm) was significantly larger than in type A (18.00 mm). The mean size of SA in type A (39.46 mm) was significantly larger than in type B (19.54 mm). Most type A CIS were non-high-grade, and the oestrogen receptor (ER)(+)/progesterone receptor (PgR)(+)/HER2(-) immunophenotype predominated. Most type B CIS were high-grade and six (54%) were ER(-)/PgR(-). Most type A were bcl-2(+)/p53(-) in both SA and CIS areas, but two (18%) apocrine ductal CIS of type B were bcl-2(-)/p53(+) in both SA and CIS areas. Expression of ER and cyclin D1 in SA was not different from that of SA unassociated with cancer. CONCLUSIONS: Most CIS involving SA arises within SA and high-grade DCIS tends to grow beyond SA. Occasional CIS may arise outside SA and secondarily involve SA.

Measurement of human epidermal growth factor receptor type-2 extracellular domain and cancer antigen 15-3 levels in needle washout fluid: a potential adjunct to the cytological diagnosis of breast cancer.

To determine whether the measurement of biomarkers in the needle washout fluid after fine needle aspiration cytology could strengthen cytological diagnosis, 29 benign lesions, 26 ductal carcinomas in situ (DCIS), and 95 invasive carcinomas of the breast were analyzed. Immediately after the aspiration samples were expelled onto glass slides, the needle was rinsed out with 2 ml of saline. Human epidermal growth factor receptor two extracellular domain (HER2-ECD) and cancer antigen 15-3 (CA 15-3) in the washout fluid were measured using chemiluminescence. The measuring limit for HER2-ECD (0.5 ng/ml) and that for CA 15-3 (4 U/ml) were used as the cutoff values, respectively. The proportion of patients with positive HER2-ECD values was 6.9% in cases of benign lesions, 26.9% in DCIS, and 10.5% in invasive carcinomas. The proportion of patients with positive CA 15-3 values was 6.9% in cases of benign lesions, 42.3% in DCIS, and 10.5% in invasive carcinomas. The proportion of patients with any biomarker elevation was 6.9% in cases of benign lesions, 46.2% in DCIS, and 15.8% in invasive carcinomas. Thus, biomarker elevation was most frequent in cases of DCIS, followed by invasive carcinomas (P < 0.01). HER2-ECD values over 6 ng/ml or CA 15-3 values over 25 U/ml were seen exclusively in cases of DCIS or invasive ductal carcinomas with an extensive intraductal component. If the biomarker status is included, 5 of 12 cases of DCIS with equivocal cytology categories would be upgraded. Further clinicopathological validation is necessary before applying this new approach to the clinical setting.

Serous papillary adenocarcinoma of the female genital organs and invasive micropapillary carcinoma of the breast. Are WT1, CA125, and GCDFP-15 useful in differential diagnosis?

Serous papillary adenocarcinoma of the female genital organs and invasive micropapillary carcinoma of the breast have close histologic similarities. Thus, when these cancers occur synchronously or metachronously in the same patient, it is difficult to determine the primary site. We examined 23 serous papillary adenocarcinomas (16 ovarian, 5 endometrial, and 2 peritoneal) and 37 invasive micropapillary carcinomas of the breast (12 pure and 25 mixed types) on immunohistochemical expression of Wilm's tumor antigen-1 (WT1), CA125, and gross cystic disease fluid protein-15 (GCDFP-15), which have been reported to be useful in the differential diagnosis of primary ovarian carcinomas versus metastatic breast cancer to the ovary. The positive rates of WT1, CA125, and GCDFP-15 in serous papillary adenocarcinomas were 78%, 78%, and 0%, respectively, and the corresponding rates in invasive micropapillary carcinomas were 3%, 40%, and 38%. The CA125-positive rate of invasive micropapillary carcinoma was higher than the rate reported for other types of breast carcinomas. We consider CA125 to be not always useful in the differential diagnosis of serous papillary adenocarcinoma and invasive micropapillary carcinoma. Although the positive rate of WT1 was significantly higher in serous papillary adenocarcinoma than in invasive micropapillary carcinoma, WT1 expression in endometrial serous papillary adenocarcinoma was infrequent (20%). WT1 and GCDFP-15 could be useful markers for the differential diagnosis of ovarian and peritoneal serous papillary adenocarcinoma versus breast invasive micropapillary adenocarcinoma. However, the availability of GCDFP-15 is limited because of the low positive rate of GCDFP-15 in invasive micropapillary carcinomas.

Evaluation of new image processing conditions for digital mammograms from Fuji computed radiography.

PURPOSE: A new processing parameter (T-type parameter) setting for gray scale was developed to improve the image quality of digital mammograms. To clarify the usefulness of this parameter setting, we have evaluated the image quality of digital mammograms (hard copy images) processed with this parameter, and compared it with S/F mammography. MATERIAL AND METHODS: Mammograms were made under the same radiographic conditions by the S/F and FCR systems (type 1, S/F; type 2 and 3, FCR with new T-type parameters; type 4 and 5, FCR with conventional parameters). A total of 49 images from 10 cases was selected for evaluation testing. Evaluation items were the contrast visibility of mammary glands and adipose tissues together with their granularity and sharpness. Eleven medical doctors participated in evaluating the images. RESULTS: The FCR hard copy images processed with the T-type parameter settings were significantly preferred over the conventional S/F images for the contrast visibility of mammary glands and adipose tissue. As for the other items (except for granularity), the FCR hard copy images processed with the T-type parameter settings were subjectively evaluated as slightly better than or equal to the S/F images. In contrast, the conventional S/F images were significantly preferred over the FCR hard copy images processed with the conventional parameter settings. CONCLUSION: The image quality of FCR hard copy images processed with the T-type parameter settings was preferred over that of conventional S/F images as evaluated by medical doctors who specialized in mammography interpretation.