BI-RADS Research Articles

Breast density changes over time: Frequency, patterns, and practice implications.

e18772 Background: Dense breast tissue is known to obscure malignancy on mammography and is considered an independent risk factor for breast cancer. Because of its clinical significance, breast density has been incorporated into cancer risk assessment tools, supplemental screening recommendations, and patient notification laws. Changes in density—whether due to intrinsic changes in breast tissue over time or to inter-radiologist variability in interpreting mammographic results—can therefore have significant implications for patients. The purpose of this study was to quantify longitudinal patterns of breast density change. Methods: This retrospective cohort study tracked breast density changes among 37,156 patients who received multiple mammograms at a large academic medical center between October 1, 2015 and June 29, 2021. Breast Imaging-Reporting and Data System (BI-RADS) density categories A (least dense) through D (most dense), visually determined by radiologists at the time of screening, were abstracted from electronic medical records and dichotomized into one of two density statuses: non-dense (categories A and B) and dense (categories C and D). A sequence analysis of longitudinal changes between the two density statuses was performed using SQL. Results: The majority (91.3%, n = 33,920) of patients maintained the same density status (dense or non-dense) throughout the six-year study period, while 8.7% of participants (n = 3,236) experienced between one and six density status changes. Among patients who experienced any density status change, the vast majority (96.7%, n = 3,131) moved exclusively between BI-RADS categories B and C, the intermediate density categories (Table). Frequency of density status change was positively correlated with average number of mammograms performed per patient. Conclusions: Although most patients maintain a consistent breast density status, some will experience multiple status changes even over this relatively short timeframe (< 6 years), with more frequent changes being associated with more frequent mammograms. The overwhelming majority of patients whose density status changed moved exclusively between BI-RADS categories B and C, suggesting a role for inter-radiologist variability in visual assessments of intermediate-density tissue. These findings have implications for how density status changes are incorporated into patient education, risk assessment, and supplemental screening recommendations.[Table: see text]

MR imaging properties of breast cancer molecular subtypes.

In this study, we aim at investigating the magnetic resonance (MR) imaging features of molecular subtypes, according to the BIRADS Atlas. The preoperative MRI examinations of 104 breast cancer patients were reviewed retrospectively using the 5th Edition Breast Imaging-Reporting and Data System (BI-RADS) Atlas. According to BI-RADS, cases were classified as mass or non-mass enhancement. Background parenchymal enhancements of the cases were evaluated. The population was examined for shape, contour, enhancement characteristics of masses, distribution and internal enhancement patterns of non-mass enhancements, background parenchymal enhancement, multifocal/multicentric (MFMS) status, presence of axillary LAP, ADC values, and kinetic aspects. The Kruskal-Wallis and Chi-square tests were used to explore the connection between molecular subtypes and MR data. The link between molecular subtypes and mass/non-mass enhancement was discovered to be statistically significant (p=0.007). The shape (p=0.001) and contour (p=0.001) properties of the masses were observed to differ depending on the molecular subtypes. The Luminal types were usually irregularly shaped with irregular/spiculated contours, whereas the HER-2 (+) and Triple (-) subtypes were mostly oval/round with smooth contours. The subtype with the highest non-mass enhancement rate (70%) was HER-2 (+). Axillary lymphadenopathy was most common (64.3%) in the Triple (-) subtype (p=0.033). According to the BIRADS Atlas, molecular subtypes exhibit a wide range of imaging properties in MR in our study.

A breast cancer (BC) risk model incorporating Tyrer-Cuzick version 8 (TCv8) and a polygenic risk score (PRS) for diverse ancestries.

557 Background: BC risk assessment is important for guiding personalized screening and risk-reducing interventions. TCv8 is used to estimate BC risk based on age, breast density, family cancer history and other clinical factors. Accuracy may be improved by combining TCv8 with a PRS. We developed and validated a PRS for diverse ancestries based on 149 common genetic variants (PRS-149) comprised of 56 ancestry-informative and 93 BC-associated variants. Here, we describe a BC risk model that combines PRS-149 with TCv8. Methods: Subjects had multigene panel testing for hereditary cancer and were negative for pathogenic variants in known BC susceptibility genes. A combined risk score (CRS), incorporating PRS-149 and TCv8, was developed based on 189,230 women, including 43,444 (23%) with a history of BC. Breast Imaging Reporting and Data System (BI-RADS) breast density measurements were available for 12,363 women. We used multivariable logistic regression to test breast density and PRS-149 for association with risk of BC. An independent test cohort of 6,030 BC-unaffected women with BI-RADS assessment was used to evaluate the effect of PRS-149 on risk stratification. Relative contributions of family history, breast density, other clinical factors in TCv8 and PRS-149 were examined by adding terms sequentially to an ANOVA model. We compared differences in classification of women as high (20%) or low/moderate (20%) remaining lifetime risk according to TCv8 versus CRS. Results: In the development cohort, increased breast density was significantly associated with higher risk of BC (p=3.0x10-6) with an effect size consistent with TCv8. PRS-149 improved BC risk prediction over age, breast density and family history (OR per unit standard deviation: 1.41, 95% CI: 1.37 – 1.46; p: 1.8x10-105). PRS-149 was weakly but significantly correlated with both family history (r=0.09) and breast density (r=0.01). After adjusting for multiple testing, no other factors were correlated with PRS-149. In the independent test cohort, PRS-149 explained 27% of CRS variability after accounting for family history, breast density and other clinical factors. Adding PRS-149 to TCv8 significantly altered risk estimates, with 16.3% (983/6,030) of patients classified differently by CRS versus TCv8. By TCv8 alone, 38.0% (2,289/6,030) of patients were classified as high-risk. Among patients who were high-risk by TCv8, 25.2% (576/2,289) were downgraded by CRS. Conclusions: This is the first BC risk model that includes breast density, family history, and a PRS based on genetically determined ancestry that is validated for diverse populations. Addition of PRS-149 improved risk prediction and substantially modified risk stratification compared to TCv8 alone. Implementation of CRS may therefore lead to improved identification of women who are likely to benefit from increased surveillance and preventive medications.

The clinical value of conventional ultrasound combined with contrast-enhanced ultrasound in the evaluation of BI-RADS 4 lesions detected by magnetic resonance imaging.

To determine the value of conventional ultrasound combined with contrast-enhanced ultrasound (CEUS) in Breast Imaging Reporting and Data System (BI-RADS) Category 4 lesions as detected by MRI. A total of 176 breast lesions from 171 patients were detected by MRI and categorised as BI-RADS 4. All patients also underwent ultrasound and CEUS scans. The combination of ultrasound-BI-RADS and CEUS 5-point scoring system created the Rerated BI-RADS (referred to as CEUS-BI-RADS). The diagnostic performances of ultrasound and CEUS-BI-RADS were then compared. A χ2 test was used to compare the CEUS features of mass-like and non-mass-like enhancement types of MRI-BI-RADS 4 lesions. There were 167 (167/176) breast lesions detected by ultrasound, with a detection rate of 94.89%, while all were subsequently detected by "second-look" ultrasound combined with CEUS, with a detection rate of 100%. The areas under the receiver operating characteristic curves for ultrasound and CEUS-BI-RADS were 0.810 and 0.940, respectively. The diagnostic efficiency of CEUS-BI-RADS was significantly higher than that of ultrasound alone (z = 3.264, p = 0.001). For both mass-like and non-mass-like enhancement types of MRI-BI-RADS 4 lesions, CEUS-BI-RADS demonstrated satisfactory sensitivity and accuracy. Moreover, 29 (29/176) category 4 lesions were downgraded to 3 by CEUS-BI-RADS. Ultrasound combined with CEUS can allow reclassification, reduce biopsy rates, and facilitate pre-surgical localisation for biopsy or surgery for MRI-BI-RADS 4 lesions. For MRI-BI-RADS Category 4 lesions with a wide range of malignancies, ultrasound combined with CEUS is a promising diagnostic approach with high clinical utility.

Factors associated with upgrade, malignant upgrade, and subsequent breast cancer occurrence of papillary breast lesions diagnosed on core needle biopsy.

In this study, we analyzed the upgrade rate and associated factors for upgrade, malignant upgrade, and subsequent breast cancer occurrence of papillary breast lesions diagnosed on core needle biopsy (CNB). One hundred sixty-nine patients who underwent surgery for the treatment of papillary breast lesions diagnosed on CNB were included in this study. Medical records including radiological and pathological reports were retrospectively reviewed. The overall upgrade rate was 29.6%, and upgrade rate to malignancy was 16.6%. Age over 45 years, preoperative tumor size≥0.7cm on breast ultrasound, pathologic tumor size≥0.4cm, breast imaging reporting and data system (BIRADS) category 4b or 4c, and personal history of breast cancer were associated with upgrade. In addition, age over 45 years, preoperative tumor size≥0.9cm, pathologic tumor size≥0.6cm, atypia in CNB, and BIRADS category 4b or 4c were associated with malignancy. The risk of subsequent breast cancer occurrence was increased in preoperative tumor size≥0.8cm, pathologic tumor size≥0.5cm, multiple and recurrent lesions. Our study showed high upgrade rate of papillary breast lesions diagnosed on CNB. Our findings suggest that surgical excision is recommended for papillary breast lesions diagnosed on CNB in selected patients.

Incidence of Occult Breast Cancer in Carriers of BRCA1/2 or Other High-Penetrance Pathogenic Variants Undergoing Prophylactic Mastectomy: When is Sentinel Lymph Node Biopsy Indicated?

This study sought to determine the likelihood of occult malignancy during risk-reducing mastectomy in high-penetrance pathogenic variant carriers to help refine axillary staging recommendations. The authors performed a retrospective cohort study analyzing all female carriers of pathogenic variants in BRCA1/2, PALB2 or other genes who underwent prophylactic surgery at their institution between 2006 and 2021. Occult breast cancer was defined as the unanticipated presence of in situ or invasive malignancy on pathologic evaluation of prophylactic mastectomy specimens. Of 523 women, 243 carriers met the inclusion criteria for the study including 124 BRCA1 (51.0%), 108 BRCA2 (44.4%), and 11 PALB2, TP53, CDH1, or PTEN (4.6%) carriers. The median age was 44 years (interquartile range, 37-52 years). Overall, 128 women (52.7%) underwent bilateral prophylactic mastectomies, and 115 (47.3%) underwent contralateral prophylactic mastectomy. In the 371 mastectomies performed, 16 (4.3%) occult malignancies were diagnosed. Most of the occult malignancies were ductal carcinoma in situ (13 mastectomies, 3.5%), whereas 3 mastectomies (0.8%) contained invasive breast cancer. If Breast Imaging Reporting and Data System (BIRADS) 1-2 or BIRADS 3 findings were reported on preoperative magnetic resonance imaging (MRI), the rate of occult malignancy decreased to 3.0 and 2.8%, respectively, per mastectomy. The patient-level factors associated with a likelihood of occult breast cancer greater than 10% included a history of prior breast cancer, age exceeding 60 years, and BIRADS 4 findings on preoperative imaging. Occult invasive malignancy was detected in less than 1% of the risk-reducing mastectomies performed for women with BRCA1/2 or PALB2 pathogenic variants. Sentinel lymph node biopsy can be safely avoided when BIRADS 1-3 findings are reported on preoperative MRI.

Combined Use of Shear Wave Elastography, Microvascular Doppler Ultrasound Technique, and BI-RADS for the Differentiation of Benign and Malignant Breast Masses.

ObjectiveTo evaluate the value of the combined use of Breast Imaging Reporting and Data System (BI-RADS), qualitative shear wave elastography (SWE), and AngioPLUS microvascular Doppler ultrasound technique (AP) for distinguishing benign and malignant breast masses.Materials and MethodsA total of 210 pathologically confirmed breast lesions in 210 patients were reviewed using BI-RADS, qualitative SWE, and AP. The sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), accuracy, and area under the receiver operating characteristic curve (AUC) of BI-RADS and the combination of qualitative SWE and/or AP with BI-RADS were compared, respectively.ResultsCompared with using BI-RADS alone, the use of combined qualitative SWE and/or AP with BI-RADS had higher AUC values (P < 0.001). Besides this, the combination of qualitative SWE and AP with BI-RADS had the best diagnostic performance for differentiating between benign and malignant masses. When AP and SWE were combined with BI-RADS, 49/76 benign masses were downgraded from BI-RADS category 4a into BI-RADS category 3, while no benign masses were upgraded from BI-RADS category 3 into BI-RADS category 4a. Three sub-centimeter malignant masses were downgraded from BI-RADS category 4a into BI-RADS category 3, while three malignant masses remain in BI-RADS category 3 due to a benign manifestation in both AP and qualitative SWE. Moreover, 5/6 of them were sub-centimeter masses, and 4/6 of them were intraductal carcinoma. The sensitivity, specificity, PPV, NPV, accuracy, and AUC were 91.0%, 81.1%, 69.3%, 95.1%, 84.3%, and 0.861 (95% confidence interval, 0.806–0.916; P < 0.001), respectively. Compared with BI-RADS alone, the sensitivity slightly decreased, while the specificity, PPV, NPV, and accuracy were significantly improved.ConclusionCombination of qualitative SWE and AP with BI-RADS improved the diagnostic performance in differentiating benign from malignant breast lesions, which is helpful for avoiding unnecessary biopsies. However, we should be careful about the downgrading of sub-centimeter BI-RADS 4a category lesions.

Physiological background parenchymal uptake of 18F-FDG in normal breast tissues using dedicated breast PET: correlation with mammographic breast composition, menopausal status, and menstrual cycle

This study aimed to quantitatively evaluate the effects of age, mammographic density, menopausal status, and menstrual cycle on background parenchymal uptake (BPU) using ring-shaped dedicated breast positron emission tomography (dbPET). This study included 186 adult women who underwent mammography and dbPET on the same day and had no abnormalities classified as Breast Imaging Reporting and Data System (BI-RADS) category 1 on both examinations. The volume of interest (VOI) was placed in the glandular tissue of both breasts, and the maximum standardized uptake value (SUVmax), mean standardized uptake value (SUVmean), and metabolic breast volume (MBV) were measured as indicators of BPU. We analyzed the correlation between BPU and age, mammographic density, menopausal status, and menstrual cycle. The SUVmax and SUVmean for normal breast tissue were inversely correlated with age (both p < 0.001). The SUVmax, SUVmean, and MBV of mammographically dense breast tissues were significantly higher than those of non-dense breast tissues (all p < 0.001). The SUVmax, SUVmean, and MBV of normal breast tissue in premenopausal women were significantly higher than those in postmenopausal women (p < 0.001, p < 0.001, p = 0.002, respectively). In the study, 59 premenopausal women, the SUVmax of normal breast tissue in the menstrual-follicular phase was significantly lower than that in the periovulatory-luteal phase (p = 0.02). When we sorted the premenopausal women by mammographic breast composition, the SUVmax and SUVmean of normal breast tissues in the menstrual-follicular phase were significantly lower than those in the periovulatory-luteal phase in the 44 premenopausal women with dense breasts (p = 0.007, and p = 0.038, respectively), whereas no statistically significant difference was found between the menstrual-follicular phase and the periovulatory-luteal phase in the 15 premenopausal women with non-dense breasts. BPU in normal breast tissues assessed using ring-shaped dbPET was associated with mammographic density, menopausal status, and women's menstrual cycle. The menstrual cycle was significantly associated with BPU in premenopausal women with dense breasts but not in women with non-dense breasts.

Volumetric breast density evaluation using fully automated Volpara software, its comparison with BIRADS density types and correlation with the risk of malignancy

BackgroundMammography is currently the modality of choice for mass screening of breast cancer, although its sensitivity is low in dense breasts. Besides, higher breast density has been identified as independent risk factor so it has been conceptualized that women with dense breasts should be encouraged for supplemental screening. In this study, we aimed to estimate the distribution of volumetric breast density using fully automated Volpara software and to analyze the level of agreement between volumetric density grades and Breast Imaging Reporting and Data System (BI-RADS) density grades. We also aim to estimate the distribution of breast cancer in different VDG and to find a correlation between VDG and risk of malignancy.ResultsVDG-c was most common followed by VDG-b and BIRADS grade B was commonest followed by grade C. The density distribution was found inversely related to the age. Level of agreement between VDG and BIRADS grades was moderate (κ = 0.5890). Statistically significant correlation was noted between VDG-c and d for risk of malignancy (p < 0.001).ConclusionDifficulties associated with the use of BI-RADS density categories may be avoided if assessed using a fully automated volumetric method. High VDG can be considered as independent risk factor for malignancy. Thus, awareness of a woman’s breast density might be useful in determining the frequency and imaging modality for screening.

MRI for Evaluation of Complications of Breast Augmentation.

Breast augmentation is one of the most common aesthetic procedures performed in the United States. Several techniques of breast augmentation have been developed, including the implantation of breast prostheses and the injection of autologous fat and other materials. The most common method of breast augmentation is to implant a prosthesis. There are different types of breast implants that vary in shape, composition, and the number of lumina. The rupture of breast implants is the leading cause of implant removal. The rupture rate increases substantially with the increasing age of the implant. Most implant ruptures are asymptomatic. Implant complications can be grouped into two categories: local complications in the breast and adjacent soft tissue, and systemic complications associated with rheumatologic or neurologic symptoms. The onset of local complications may be early (infection and periprosthetic collections including seromas, hematomas, or abscesses) or late (capsular contraction, implant rupture, gel bleed, or breast implant-associated anaplastic large cell lymphoma). Although mammography is the imaging modality for breast cancer screening, noncontrast breast MRI is the imaging modality of choice for evaluation of the integrity of breast implants and the complications of breast augmentation, for equivocal findings at conventional imaging, and as a supplement to mammography in patients with free injectable materials. The fifth edition of the Breast Imaging Reporting and Data System (BI-RADS) provides a systematic outline for MRI evaluation of patients with breast implants. Silicone- and water-selective sequences provide useful supplemental information to confirm intracapsular and extracapsular rupture. Breast MRI for evaluation of implant integrity does not require intravenous contrast material. The use of MRI contrast material in patients with breast augmentation is indicated when infection or malignancy is suspected. Radiologists should have a thorough understanding of the different techniques for breast augmentation, normal imaging features, and complications specific to breast augmentation. An invited commentary by Ojeda-Fournier is available online. ©RSNA, 2022.

Radiomics Based on DCE-MRI Improved Diagnostic Performance Compared to BI-RADS Analysis in Identifying Sclerosing Adenosis of the Breast.

PurposeSclerosing adenosis (SA) is a benign lesion that could mimic breast carcinoma and be evaluated as malignancy by Breast Imaging-Reporting and Data System (BI-RADS) analysis. We aimed to construct and validate the performance of radiomic model based on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) compared to BI-RADS analysis to identify SA.MethodsSixty-seven patients with invasive ductal carcinoma (IDC) and 58 patients with SA were included in this retrospective study from two institutions. The 125 patients were divided into a training cohort (n= 88) from institution I and a validation cohort from institution II (n=37). Dynamic contrast-enhanced sequences including one pre-contrast and five dynamic post-contrast series were obtained for all cases with different 3T scanners. Single-phase enhancement, multi-phase enhancement, and dynamic radiomic features were extracted from DCE-MRI. The least absolute shrinkage and selection operator (LASSO) logistic regression and cross-validation was performed to build the radscore of each single-phase enhancement and the final model combined multi-phase and dynamic radiomic features. The diagnostic performance of radiomics was evaluated by receiver operating characteristic (ROC) analysis and compared to the performance of BI-RADS analysis. The classification performance was tested using external validation.ResultsIn the training cohort, the AUCs of BI-RADS analysis were 0.71 (95%CI [0.60, 0.80]), 0.78 (95%CI [0.67, 0.86]), and 0.80 (95%CI [0.70, 0.88]), respectively. In single-phase analysis, the second enhanced phase radiomic signature achieved the highest AUC of 0.88 (95%CI [0.79, 0.94]) in distinguishing SA from IDC. Nine multi-phase radiomic features and two dynamic radiomic features showed the best predictive ability for final model building. The final model improved the AUC to 0.92 (95%CI [0.84, 0.97]), and showed statistically significant differences with BI-RADS analysis (p<0.05 for all). In the validation cohort, the AUC of the final model was 0.90 (95%CI [0.75, 0.97]), which was higher than all BI-RADS analyses and showed statistically significant differences with one of the BI-RADS analysis observers (p = 0.03).ConclusionsRadiomics based on DCE-MRI could show better diagnostic performance compared to BI-RADS analysis in differentiating SA from IDC, which may contribute to clinical diagnosis and treatment.

BI-RADS BERT and Using Section Segmentation to Understand Radiology Reports.

Radiology reports are one of the main forms of communication between radiologists and other clinicians, and contain important information for patient care. In order to use this information for research and automated patient care programs, it is necessary to convert the raw text into structured data suitable for analysis. State-of-the-art natural language processing (NLP) domain-specific contextual word embeddings have been shown to achieve impressive accuracy for these tasks in medicine, but have yet to be utilized for section structure segmentation. In this work, we pre-trained a contextual embedding BERT model using breast radiology reports and developed a classifier that incorporated the embedding with auxiliary global textual features in order to perform section segmentation. This model achieved 98% accuracy in segregating free-text reports, sentence by sentence, into sections of information outlined in the Breast Imaging Reporting and Data System (BI-RADS) lexicon, which is a significant improvement over the classic BERT model without auxiliary information. We then evaluated whether using section segmentation improved the downstream extraction of clinically relevant information such as modality/procedure, previous cancer, menopausal status, purpose of exam, breast density, and breast MRI background parenchymal enhancement. Using the BERT model pre-trained on breast radiology reports, combined with section segmentation, resulted in an overall accuracy of 95.9% in the field extraction tasks. This is a 17% improvement, compared to an overall accuracy of 78.9% for field extraction with models using classic BERT embeddings and not using section segmentation. Our work shows the strength of using BERT in the analysis of radiology reports and the advantages of section segmentation by identifying the key features of patient factors recorded in breast radiology reports.

Underestimation of Breast Cancer Based on Ultrasound Breast Imaging-Reporting and Data System (BI-RADS) Assessment of Ductal Carcinoma In Situ: A Retrospective Study of 296 Lesions

Background: Ultrasound Breast Imaging-Reporting and Data System (BI-RADS) classification may underestimate ductal carcinoma in situ (DCIS). Currently, there is a lack of research on ultrasound BI-RADS underestimating DCIS. Objectives: To improve the diagnosis of DCIS, this study aimed to investigate factors associated with the underestimation of DCIS, based on ultrasound BI-RADS assessments. Patients and Methods: In this cross-sectional study, consecutive patients with breast ultrasound BI-RADS classification and biopsy results were retrospectively examined. DCIS was found in the pathology reports of all patients. DCIS cases classified as BI-RADS 4A or lower were considered as underestimations of DCIS, while DCIS cases classified as BI-RADS 4B or higher were considered as non-underestimation of DCIS. The demographics, clinical manifestations, features of breast images, BI-RADS classification, and pathological results of the two groups were compared to explore possible associated factors. A stepwise logistic regression analysis was also carried out based on the significance of factors associated with the underestimation of DCIS according to the BI-RADS assessment. Results: Between January 2015 and May 2017, a total of 296 breast DCIS lesions were diagnosed in 294 female patients. Overall, 65 lesions (22.0%) were underestimated DCIS, and 231 lesions (78.0%) were non-underestimated DCIS; no significant differences were found between their clinical presentations. The univariate analysis showed that the age of the patients, presence of microinvasions, maximum lesion diameter, shape, margin, orientation, echo pattern, posterior acoustic features, ultrasound pattern, and vascularity of lesions were possibly associated factors, which could lead to the underestimation of DCIS. The logistic regression analysis showed that age above 50 years, maximum lesion diameter &lt; 10 mm, lack of microinvasion, and circumscribed margins were associated with the underestimation of DCIS. Conclusion: In this study, 22% of DCIS lesions was underestimated by the BI-RADS assessment. The patient’s age, maximum lesion diameter, microinvasion, and lesion margin were associated with the underestimation of DCIS.

Comprehensive quantitative malignant risk prediction of pure grouped amorphous calcifications: clinico-mammographic nomogram.

Pure grouped amorphous calcifications are classified as Breast Imaging Reporting and Data System (BI-RADS) category 4B suspicious calcifications and recommended for biopsy. However, the biopsies often reveal benign findings, especially in screening mammograms (92.4-97.2%). Mammograms of 699 pure grouped amorphous calcifications with final pathological results were analyzed in this retrospective study. The maximum span (MS) of the group of calcifications and the MS of the parallel/vertical direction of the mammary duct (MPS/MVS) were measured, and the MPS to MVS ratio was calculated. Based on the MS and ratio, 2 prediction nomograms with other clinic-mammographic features were developed. The discrimination performance of the models was assessed and compared by the area under the receiver operating characteristic curve (AUC). Scatterplots were created to determine the cutoff values with fewer misdiagnoses of malignant calcifications and fewer false positives. Ultimately, 2 prediction models were successfully developed based on the 4 risk factors of age, purpose of the mammogram, whether multiple or single calcifications, and the MS [odds ratio (OR) =1.06, P=0.02]/ratio (OR =6.05, P<0.001). Both models had good discrimination. The ratio model performed better than the MS model in the training cohort (AUC of 0.875 and 0.834, respectively, P=0.003) and validation cohort (AUC 0.908 and 0.867, respectively, P=0.047). For the group with probably benign calcifications (as detected by the ratio nomogram), the malignancy rates were 2.7% [95% confidence interval (CI): 1.00% to 6.53%] and 1.19% (95% CI: 0.06% to 7.37%) in the training and validation cohorts, respectively, and 44.12% and 47.70% of benign biopsies were detected in the training and validation cohorts, respectively. The clinico-mammographic quantitative malignancy risk prediction nomogram showed favorable discrimination and calibration performance. The ratio model showed better diagnostic efficiency than the MS model, and identified >40% of benign biopsies.

Ultrasound high-definition microvasculature imaging with novel quantitative biomarkers improves breast cancer detection accuracy

ObjectivesTo overcome the limitations of power Doppler in imaging angiogenesis, we sought to develop and investigate new quantitative biomarkers of a contrast-free ultrasound microvasculature imaging technique for differentiation of benign from malignant pathologies of breast lesion.MethodsIn this prospective study, a new high-definition microvasculature imaging (HDMI) was tested on 521 patients with 527 ultrasound-identified suspicious breast masses indicated for biopsy. Four new morphological features of tumor microvessels, microvessel fractal dimension (mvFD), Murray’s deviation (MD), bifurcation angle (BA), and spatial vascularity pattern (SVP) as well as initial biomarkers were extracted and analyzed, and the results correlated with pathology. Multivariable logistic regression analysis was used to study the performance of different prediction models, initial biomarkers, new biomarkers, and combined new and initial biomarkers in differentiating benign from malignant lesions.ResultsThe new HDMI biomarkers, mvFD, BA, MD, and SVP, were statistically significantly different in malignant and benign lesions, regardless of tumor size. Sensitivity and specificity of the new biomarkers in lesions > 20 mm were 95.6% and 100%, respectively. Combining the new and initial biomarkers together showed an AUC, sensitivity, and specificity of 97% (95% CI: 95–98%), 93.8%, and 89.2%, respectively, for all lesions regardless of mass size. The classification was further improved by adding the Breast Imaging Reporting and Data System (BI-RADS) score to the prediction model, showing an AUC, sensitivity, and specificity of 97% (95% CI: 95–98%), 93.8%, and 89.2%, respectively.ConclusionThe addition of new quantitative HDMI biomarkers significantly improved the accuracy in breast lesion characterization when used as a complementary imaging tool to the conventional ultrasound.Key Points• Novel quantitative biomarkers extracted from tumor microvessel images increase the sensitivity and specificity in discriminating malignant from benign breast masses.• New HDMI biomarkers Murray’s deviation, bifurcation angles, microvessel fractal dimension, and spatial vascularity pattern outperformed the initial biomarkers.• The addition of BI-RADS scores based on US descriptors to the multivariable analysis using all biomarkers remarkably increased the sensitivity, specificity, and AUC in all size groups.

Multichannel DenseNet Architecture for Classification of Mammographic Breast Density for Breast Cancer Detection.

Percentage mammographic breast density (MBD) is one of the most notable biomarkers. It is assessed visually with the support of radiologists with the four qualitative Breast Imaging Reporting and Data System (BIRADS) categories. It is demanding for radiologists to differentiate between the two variably allocated BIRADS classes, namely, “BIRADS C and BIRADS D.” Recently, convolution neural networks have been found superior in classification tasks due to their ability to extract local features with shared weight architecture and space invariance characteristics. The proposed study intends to examine an artificial intelligence (AI)-based MBD classifier toward developing a latent computer-assisted tool for radiologists to distinguish the BIRADS class in modern clinical progress. This article proposes a multichannel DenseNet architecture for MBD classification. The proposed architecture consists of four-channel DenseNet transfer learning architecture to extract significant features from a single patient's two a mediolateral oblique (MLO) and two craniocaudal (CC) views of digital mammograms. The performance of the proposed classifier is evaluated using 200 cases consisting of 800 digital mammograms of the different BIRADS density classes with validated density ground truth. The classifier's performance is assessed with quantitative metrics such as precision, responsiveness, specificity, and the area under the curve (AUC). The concluding preliminary outcomes reveal that this intended multichannel model has delivered good performance with an accuracy of 96.67% during training and 90.06% during testing and an average AUC of 0.9625. Obtained results are also validated qualitatively with the help of a radiologist expert in the field of MBD. Proposed architecture achieved state-of-the-art results with a fewer number of images and with less computation power.

An Artificial Intelligence-based Mammography Screening Protocol for Breast Cancer: Outcome and Radiologist Workload.

Background Developments in artificial intelligence (AI) systems to assist radiologists in reading mammograms could improve breast cancer screening efficiency. Purpose To investigate whether an AI system could detect normal, moderate-risk, and suspicious mammograms in a screening sample to safely reduce radiologist workload and evaluate across Breast Imaging Reporting and Data System (BI-RADS) densities. Materials and Methods This retrospective simulation study analyzed mammographic examination data consecutively collected from January 2014 to December 2015 in the Danish Capital Region breast cancer screening program. All mammograms were scored from 0 to 10, representing the risk of malignancy, using an AI tool. During simulation, normal mammograms (score < 5) would be excluded from radiologist reading and suspicious mammograms (score > recall threshold [RT]) would be recalled. Two radiologists read the remaining mammograms. The RT was fitted using another independent cohort (same institution) by matching to the radiologist sensitivity. This protocol was further applied to each BI-RADS density. Screening outcomes were measured using the sensitivity, specificity, workload, and false-positive rate. The AI-based screening was tested for noninferiority sensitivity compared with radiologist screening using the Farrington-Manning test. Specificities were compared using the McNemar test. Results The study sample comprised 114 421 screenings for breast cancer in 114 421 women, resulting in 791 screen-detected, 327 interval, and 1473 long-term cancers and 2107 false-positive screenings. The mean age of the women was 59 years ± 6 (SD). The AI-based screening sensitivity was 69.7% (779 of 1118; 95% CI: 66.9, 72.4) and was noninferior (P = .02) to the radiologist screening sensitivity of 70.8% (791 of 1118; 95% CI: 68.0, 73.5). The AI-based screening specificity was 98.6% (111 725 of 113 303; 95% CI: 98.5, 98.7), which was higher (P < .001) than the radiologist specificity of 98.1% (111 196 of 113 303; 95% CI: 98.1, 98.2). The radiologist workload was reduced by 62.6% (71 585 of 114 421), and 25.1% (529 of 2107) of false-positive screenings were avoided. Screening results were consistent across BI-RADS densities, although not significantly so for sensitivity. Conclusion Artificial intelligence (AI)-based screening could detect normal, moderate-risk, and suspicious mammograms in a breast cancer screening program, which may reduce the radiologist workload. AI-based screening performed consistently across breast densities. © RSNA, 2022 Online supplemental material is available for this article.

Added value of ultrafast sequence in abbreviated breast MRI surveillance in women with a personal history of breast cancer: A multireader study

PurposeTo evaluate the added value of ultrafast MRI in abbreviated breast MRI (AB-MRI) surveillance in women with a personal history of breast cancer (PHBC). MethodBetween September 2017 and November 2019, consecutive postoperative surveillance AB-MRIs with ultrafast MRIs (20 images with a 4.0-second temporal resolution using 4D time-resolved angiography with keyhole technique) were retrospectively collected. Four blinded radiologists independently classified the Breast Imaging Reporting and Data System (BI-RADS) category for AB-MRI alone versus the combined protocol (AB-MRI + ultrafast MRI). Readers were recommended to change BI-RADS category according to the time to enhancement cut-off of 12 s in ultrafast MRI. McNemar test and generalized estimation equation model were used to compare the diagnostic performances. ResultsA total of 867 MRI examinations in 867 women (mean age ± standard deviation, 51 years ± 8) were evaluated. The sensitivity of both protocols among all readers was the same, at 90% (9/10). Addition of ultrafast MRI improved the specificity (a mean of 95.3% vs. 88.6 %, p < 0.001 for all readers) and positive predictive value 1 (PPV1) (a mean of 21% vs. 10%, p < 0.001 for all readers) compared to AB-MRI alone. Downgrading BI-RADS category 3 to 2 in four readers in a mean of 6.7% (57 of 857) of negative or benign findings was the main reason for the improved specificity and PPV1. ConclusionAddition of ultrafast MRI to AB-MRI improved the specificity and PPV1 by reducing unnecessary short-term follow-ups without compromising sensitivity in postoperative surveillance.

Color-map virtual touch tissue imaging (CMV) combined with BI-RADS for the diagnosis of breast lesions.

To investigate the importance of color-map virtual touch tissue imaging (CMV) in assisting Breast Imaging Reporting and Data Systems (BI-RADS) in diagnosing malignant breast lesions. A dataset included 134 patients and 146 breast lesions was assembled. All patients underwent biopsy or surgical excision of breast lesions, and pathological results were obtained. All patients with breast lesions also underwent conventional ultrasound (US) and CMV. Each lesion was assigned a CMV score based on the color pattern of the lesion and surrounding breast tissue and a BI-RADS classification rating based on US characteristics. We compared the diagnostic performance of using BI-RADS and CMV separately and their combination. BI-RADS (odds ratio [OR]: 3.665; 95% confidence interval [CI]: 2.147, 6.258) and CMV (OR: 6.616; 95% CI: 2.272, 19.270) were independent predictors of breast malignancy (all P < 0.05). The area under the receiver operating characteristic curves (AUC) for either CMV or BI-RADS alone was inferior to that of the combination (0.877 vs. 0.962; 0.938 vs. 0.962; all P < 0.05). The performance of BI-RADS in diagnosing breast lesions is significantly improved by combining CMV. Therefore, we recommend CMV as an adjunct to BI-RADS.

The Breast Imaging-Reporting and Data System (BI-RADS) Made Easy

: The main advantage of the Breast Imaging-Reporting and Data System (BI-RADS) system is that when correctly applied, the American College of Radiology (ACR) and BI-RADS classifications are the only necessary tools for a practical and scientific communication between radiologists and clinicians regarding the breast findings. The BI-RADS provides a common standardized language of communication by minimizing subjective interpretations. A non-specialist clinician, without an in-depth knowledge of breast imaging, should be assisted on how to manage or refer patients based on standard algorithms. Dedicated breast clinicians, despite their high level of knowledge and experience regarding breast disease and management, are not radiologists; therefore, the BI-RADS can help them organize a follow-up and perform further diagnostic examinations. Besides, radiologists who are specialized in using different modalities communicate better with each other about breast images if the BI-RADS terminology and management guidelines are applied. In this study, we present a concise and simplified description for all sections of the BI-RADS to facilitate its understanding and practical use for all practitioners, not only breast radiologists.