Abstract
This study aimed to assess the utility of second-look ultrasonography (US) in differentiating breast imaging reporting and data system (BI-RADS) 4 calcifications initially detected on mammography (MG). BI-RADS 4 calcifications have a wide range of positive predictive values. We hypothesized that second-look US would help distinguish BI-RADS 4 calcifications without clinical manifestations and other abnormalities on MG. This study included 1622 pure BI-RADS 4 calcifications in 1510 women (112 patients with bilateral calcifications). The cases were randomly divided into training (85%) and testing (15%) datasets. Two nomograms were developed to differentiate BI-RADS 4 calcifications in the training dataset: the MG-US nomogram, based on multifactorial logistic regression and incorporated clinical information, MG, and second-look US characteristics, and the MG nomogram, based on clinical information and mammographic characteristics. Calibration of the MG-US nomogram was performed using calibration curves. The discriminative ability and clinical utility of both nomograms were compared using the area under the receiver operating characteristic curve (AUC) and the decision analysis curve (DCA) in the test dataset. The clinical information and imaging characteristics were comparable between the training and test datasets. The bias-corrected calibration curves of the MG-US nomogram closely approximate the ideal line for both datasets. In the test dataset, the MG-US nomogram exhibited a higher AUC than the MG nomogram (0.899 vs 0.852, P = .01). DCA demonstrated the superiority of the MG-US nomogram over the MG nomogram. Second-look US features, including ultrasonic calcifications, lesions, and moderate or marked color flow, were valuable for distinguishing BI-RADS 4 calcifications without clinical manifestations and other abnormalities on MG.
Keywords: Calcification, Mammography, Nomogram, Prediction, Ultrasonography
1. Introduction
Despite its limited specificity, mammography (MG) remains the primary modality for screening breast calcifications. The 5th edition of the American College of Radiology Breast Imaging Reporting and Data System (5th ACR-BI-RADS) classifies “amorphous,” “coarse heterogeneous,” and “fine pleomorphic” calcifications as BI-RADS 4B, while “fine linear” or “fine-linear branching” calcifications are assigned to BI-RADS 4C.[1] Although BI-RADS 4A calcifications are common in clinical practice, they lack a specific classification system. The distribution and morphology of calcifications are fundamental factors in assessing breast calcifications. Some studies have reported that amorphous calcifications with regional or grouped distributions can be classified as BI-RADS 4A.[2,3]
Ultrasonography (US) has a higher sensitivity for lesions than MG. Additionally, US-guided core needle biopsy is more comfortable, easier, and safer than MG-guided stereotactic vacuum-assisted biopsy. Therefore, clinicians or radiologists often perform a US examination focused on areas where calcifications are revealed by MG, referred to as “second-look” US. Wang et al suggested that second-look US can assist in distinguishing between BI-RADS 4 and 5 calcifications.[4] However, we believe that BI-RADS 4 calcifications need to be analyzed separately due to their wider positive predictive value range compared to BI-RADS 5 calcifications. Moreover, many patients in their study had clinical manifestations such as masses or nipple discharge. Previous studies have reported that clinical manifestations are significantly associated with breast cancer.[5]
Therefore, we sought to investigate whether second-look US can assist in distinguishing BI-RADS 4 calcifications detected on MG without clinical manifestations or other abnormalities.
2. Methods
2.1. Ethical approval
This retrospective study was approved by the institutional review board of our institution. Informed consent was waived due to the retrospective nature of the study.
2.2. Inclusion and exclusion criteria
A total of 2419 BI-RADS 4A, 4 B, and 4C calcifications were selected from our MG database spanning from January 2015 to March 2020 (Figs. 1 and 2). The exclusion criteria comprised cases lacking pathological results from surgical excision or biopsies; follow-up periods shorter than 24 months; clinical manifestations (such as palpable mass or nipple discharge); MG findings of the mass, asymmetry, or architectural distortion; and cases without second-look US images. Ultimately, 1622 pure BI-RADS 4 calcifications in 1510 women (including 112 patients with bilateral calcifications) with a median age of 48 years (range, 24–88 years) were enrolled. The cases were randomly divided into training (85%) and validation (15%) datasets.
Figure 1.
Flow chart of patient screening.
Figure 2.
Left breast of a 52-yr-old woman. (A) Magnified mediolateral mammography shows grouped amorphous microcalcifications (circle), which was classified as BI-RADS 4A. The Final pathological examination revealed an apocrine metaplasia. (B) Second-look ultrasonography shows a mass without microcalcification. (C) Second-look ultrasonography shows moderate color flow in the mass.
2.3. Imaging analysis
All mammograms were obtained using a digital mammography system (Senographe DS; GE Medical Systems, Milwaukee, WI, USA). Magnification mammography was performed to detect calcifications. Two qualified radiologists, each with over ten years of experience in breast imaging, assessed all mammograms using the 5th BI-RADS system. Second-look US was conducted by 2 other experienced radiologists with more than 10 years of experience in breast imaging analysis (Philips Healthcare iU22, Bothell, WA, USA). The extent of calcification, lesion definition, and color flow on the second-look US were rigorously evaluated and categorized. The lesion was defined as a mass or architectural distortion on the second-look US. The extent of calcification was determined as the greatest diameter on either the mediolateral oblique or craniocaudal images. Color flow on second-look US was categorized into 4 levels: 0 – absent; 1 – minimal, with 1 or 2 pixels containing flow (usually <1 mm in diameter); 2 – moderate, with a primary vessel observed in the area and/or several small vessels noted; and 3 – marked, with 4 or more substantial color flow vessels visualized. In cases of discrepancy between the 2 radiologists, a consensus was reached through discussion.
2.4. Clinical data collection
Clinical data, including age, personal history of breast cancer, follow-up duration, biopsy method, and pathological results, were extracted from electronic medical records. If second-look US detects lesions or calcifications, we perform a US-guided core needle biopsy with specimen radiography using a 14-gauge needle. If the second-look US result was negative, we performed an MG-guided stereotactic vacuum-assisted biopsy with specimen radiography. Open surgery was recommended if core biopsy showed high-risk lesions, including atypical papilloma, atypical ductal or lobular hyperplasia, lobular carcinoma in situ, and complex sclerosing lesions. Patients with benign diseases underwent a minimum follow-up of 24 months with US or MG. Follow-up data will be collected until April 1, 2022.
2.5. Statistical Analysis
Statistical analyses were conducted using the SPSS software (IBM Corporation, Armonk, NY, USA) and R (version 4.2.1). The Chi-square or Fisher exact test assessed categorical variables, whereas the Mann–Whitney U test analyzed continuous variables. Multivariate logistic regression identified malignancy risk factors. A statistical P value of less than.05 was considered significant. Cohen kappa (k) statistics evaluated interobserver agreement for the retrospective assessment of medical imaging features. The degree of agreement was classified as slight agreement (k ≤ 0.20), fair agreement (k = 0.21–0.40), moderate agreement (k = 0.41–0.60), substantial agreement (k = 0.61–0.80), and almost perfect agreement (k = 0.81–1.00).
For enhanced interpretation and analysis, continuous variables were transformed into categorical variables based on data distribution and clinical experience. In the training dataset, 2 nomograms were constructed: the MG-US nomogram, developed from multivariate logistic regression results, and the MG nomogram, derived from multivariate logistic regression results excluding US factors. Calibration of the MG-US nomogram was assessed using calibration curves based on 1000 bootstrapped samples in both the training and test datasets. Discrimination and practical utility of the MG-US and MG nomograms were compared using the area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA) curve in the test dataset.
3. Results
3.1. Clinical information and imaging characteristics
In this study, the overall malignancy rate was 37.2% (603 out of 1622 patients). Malignancy rates within the training and test datasets were 37.4% (515 out of 1376) and 35.7% (88 out of 246), respectively. Table 1 compares the clinical and imaging characteristics between patients in the training and test datasets. Among the 1376 patients in the training dataset, 867 underwent 14-gauge US-guided core needle biopsies with specimen radiography, while 509 underwent MG-guided stereotactic vacuum-assisted biopsies with specimen radiography. Surgical excision followed the diagnosis of high-risk lesions based on the sampled specimens in 155 cases. Among these 155 patients, 24 (15.5%) were diagnosed with ductal carcinoma in situ and 12 (7.7%) with invasive ductal carcinoma. No cancers occurred during the follow-up period in patients with benign diseases diagnosed using sampled specimens. The median follow-up duration was 38.4 months, ranging from 24 to 54 months.
Table 1.
Comparisons of clinical and imaging characteristics of training dataset and test dataset.
| Test dataset | Training dataset | P value | |
|---|---|---|---|
| N = 246 | N = 1376 | ||
| Pathology | .67 | ||
| Benign | 158 (64.2%) | 861 (62.6%) | |
| Malignant | 88 (35.8%) | 515 (37.4%) | |
| Age | .27 | ||
| <50 yr | 167 (67.9%) | 880 (64.0%) | |
| ≥50 yr | 79 (32.1%) | 496 (36.0%) | |
| Personal history of breast cancer | .10 | ||
| No | 191 (77.6%) | 1133 (82.3%) | |
| Yes | 55 (22.4%) | 243 (17.7%) | |
| Density | .29 | ||
| Fatty | 66 (26.8%) | 419 (30.5%) | |
| Dense | 180 (73.2%) | 957 (69.5%) | |
| Distribution | .15 | ||
| Diffuse or regional | 72 (29.3%) | 340 (24.7%) | |
| Grouped, linear or segmental | 174 (70.7%) | 1036 (75.3%) | |
| Morphology | .21 | ||
| Punctate or amorphous | 108 (43.9%) | 542 (39.4%) | |
| Coarse heterogeneous or Fine pleomorphic | 138 (56.1%) | 834 (60.6%) | |
| Extent | .52 | ||
| <10mm | 161 (65.4%) | 908 (66.0%) | |
| ≥10mm | 85 (34.6%) | 468 (34.0%) | |
| Ultrasonic calcifications | .29 | ||
| No | 173 (70.3%) | 917 (66.6%) | |
| Yes | 73 (29.7%) | 459 (33.4%) | |
| Ultrasonic lesions | .41 | ||
| No | 183 (74.4%) | 985 (71.6%) | |
| Yes | 63 (25.6%) | 391 (28.4%) | |
| Color flow | .54 | ||
| Absent or Minimal | 177 (72.0%) | 1019 (74.1%) | |
| Moderate or Marked | 69 (28.0%) | 357 (25.9%) |
3.2. Repeatability
Kappa statistics demonstrated excellent interobserver agreement between the 2 readers for assessing mammographic and ultrasonographic features: morphology (k = 0.824, P < .001), distribution (k = 0.836, P < .001), as well as a high level of agreement for ultrasonography: ultrasonic calcifications (k = 0.862, P < .001), ultrasonic lesions (k = 0.884, P < .001), and ultrasonic color flow (k = 0.843, P < .001).
3.3. Risk factors of malignancy
Table 2 compares the clinical and imaging characteristics of the benign and malignant subgroups in the training dataset. The study identified age over 50 years and a personal history of breast cancer as risk factors for malignancy when BI-RADS 4 calcifications were detected on MG. Additionally, mammographic appearances such as dense breasts; grouped, linear, or segmental distribution; coarse heterogeneous or fine pleomorphic morphology; and calcification extent exceeding 10 mm were associated with breast cancer. Ultrasonic calcifications, lesions, and moderate or marked color flow on second-look US were also identified as risk factors. These factors were further validated using multifactorial logistic regression analysis (Table 3).
Table 2.
Comparisons of clinical and imaging characteristics of subgroups of the training dataset.
| Benign | Malignant | P value | |
|---|---|---|---|
| N = 861 | N = 515 | ||
| Age | <.001 | ||
| <50 yr | 598 (69.5%) | 282 (54.8%) | |
| ≥50 yr | 263 (30.5%) | 233 (45.2%) | |
| Personal history of breast cancer | <.001 | ||
| No | 740 (85.9%) | 393 (76.3%) | |
| Yes | 121 (14.1%) | 122 (23.7%) | |
| Density | <.001 | ||
| Fatty | 340 (39.5%) | 79 (15.3%) | |
| Dense | 521 (60.5%) | 436 (84.7%) | |
| Distribution | <.001 | ||
| Diffuse or regional | 285 (33.1%) | 55 (10.7%) | |
| Grouped, linear or segmental | 576 (66.9%) | 460 (89.3%) | |
| Morphology | <.001 | ||
| Punctate or amorphous | 471 (54.7%) | 71 (13.8%) | |
| Coarse heterogeneous or Fine pleomorphic | 390 (45.3%) | 444 (86.2%) | |
| Extent | <.001 | ||
| <10mm | 607 (70.5%) | 301 (58.4%) | |
| ≥10mm | 254 (29.5%) | 214 (41.6%) | |
| Ultrasonic calcifications | <.001 | ||
| No | 630 (73.2%) | 287 (55.7%) | |
| Yes | 231 (26.8%) | 228 (44.3%) | |
| Ultrasonic lesions | <.001 | ||
| No | 712 (82.7%) | 273 (53.0%) | |
| Yes | 149 (17.3%) | 242 (47.0%) | |
| Color flow | <.001 | ||
| Absent or minimal | 695 (80.7%) | 324 (62.9%) | |
| Moderate or marked | 166 (19.3%) | 191 (37.1%) |
Table 3.
Multifactorial logistic regression analysis of clinical and imaging characteristics.
| Crude OR (95%CI) | Adj. OR (95%CI) | P (Wald test) | P (LR-test) | |
|---|---|---|---|---|
| Age: ≥50 yr vs <50 yr | 1.88 (1.5, 2.36) | 1.6 (1.2, 2.12) | <.01 | <.01 |
| Personal history of breast cancer: Yes vs No | 1.9 (1.44, 2.51) | 2.02 (1.41, 2.89) | <.001 | <.001 |
| Density: Dense vs Fatty | 3.6 (2.73, 4.75) | 2.19 (1.57, 3.06) | <.001 | <.001 |
| Distribution: Grouped, linear or segmental vs diffuse or regional | 4.14 (3.02, 5.66) | 3.13 (2.12, 4.6) | <.001 | <.001 |
| Morphology: Coarse heterogeneous or fine pleomorphic vs punctate or amorphous | 7.55 (5.68, 10.03) | 7.96 (5.57, 11.38) | <.001 | <.001 |
| Extent:≥10 mm vs < 10 mm | 1.7 (1.35, 2.13) | 1.85 (1.38, 2.48) | <.001 | <.001 |
| Ultrasonic calcifications: Yes vs No | 2.17 (1.72, 2.73) | 2.42 (1.79, 3.25) | <.001 | <.001 |
| Ultrasonic lesions: Yes vs No | 4.24 (3.31, 5.42) | 6.76 (4.84, 9.43) | <.001 | <.001 |
| Color flow: Moderate or marked vs absent or minimal | 2.47 (1.93, 3.16) | 2.31 (1.69, 3.16) | <.001 | <.001 |
3.4. Comparisons of MG-US nomogram and US nomogram
The Hosmer–Lemeshow Goodness-of-fit testing of the MG-US nomogram (Fig. 3) and MG nomogram yielded P-values of .15 and .25, respectively. Calibration curves of the MG-US nomogram for both the training and test datasets indicated that the bias-corrected curves closely approximated the ideal lines (Fig. 4). The MG-US nomogram demonstrated superior discriminative ability compared with the MG nomogram, as evidenced by a higher AUC (0.899 vs 0.852, P = .008, Fig. 5) and superior performance in DCA (Fig. 6).
Figure 3.
Nomogram for distinguishment of BI-RADS 4 calcifications; morphology A: punctate or amorphous; morphology B: coarse heterogeneous or fine pleomorphic.
Figure 4.
Internal validation of the nomogram using 1000 bootstrapped samples in the training dataset (A) and test dataset (B), the bias-corrected curves were all close to the ideal lines.
Figure 5.
The ROCs of the MG-US nomogram (red) and MG nomogram (blue) in the test dataset.
Figure 6.
Decision curve analysis (DCA) of the MG-US nomogram (red) and MG nomogram (blue) in the test dataset. The DCA demonstrated that utilizing the MG-US nomogram to distinguish BI-RADS 4 calcifications was superior to using MG nomogram.
4. Discussion
This study represents the first large-scale investigation into the utility of second-look US in distinguishing BI-RADS 4 calcifications without clinical manifestations and other abnormalities on MG. Our findings highlight the value of age, personal history of breast cancer, mammographic appearance, and second-look US features, including ultrasonic calcifications, lesions, and color flow, in this distinction.
Previous research has emphasized the predictive role of lesions associated with microcalcifications on second-look US as an independent predictor of malignancy, with calcifications showing negative second-look US findings more likely to be benign.[4,6,7] Wang et al observed BI-RADS 4 and 5 calcifications and found that 67.6% of their group showed clinical manifestations.[4] The malignancy rate in their study was 68.2%, which is much higher than that reported in our study (37.2%). Kang et al analyzed amorphous, indistinct, and fine granular microcalcifications detected using MG without other abnormalities.[6] However, according to the 5th ACR-BI-RADS, these microcalcifications can be categorized as BI-RADS 3, 4, or 5.[1] In addition, the sample size was small (n = 37) and the distribution of calcifications was ignored. Pustahija et al also studied BI-RADS 4 and 5 calcifications without other abnormalities, reporting a 38.4% breast cancer pathology rate, with visible microcalcifications on second-look US more likely to be malignant in the BI-RADS 4 and 5 microcalcification setting (visible vs invisible: 85.4% vs 74%, P = .18).[7] In our study, nearly half of the visible calcifications on second-look US were associated with breast cancer. Given the independent predictive role of clinical manifestation in breast cancer and the strong recommendation for surgery in BI-RADS 5 calcifications,[5] improving the diagnostic accuracy of BI-RADS 4 calcifications without clinical manifestations and other abnormalities on MG could help avoid unnecessary interventional operations.
Previous investigations have not revealed a correlation between color flow on second-look US and suspicious calcifications. Our study, however, revealed that moderate or marked color flow on second-look US was associated with malignancy compared to absent or mild color flow. The mechanisms underlying this phenomenon can be attributed to processes such as angiogenesis and the development of capillary networks stimulated by tumor cells,[8,9] leading to increased blood flow within and around the lesion, contributing to the observed color flow patterns on US.
In our study, masses and architectural distortions detected on second-look US were associated with malignancies, underscoring the significance of second-look US in detecting malignancies in cases where breast cancers are nonpalpable and manifest solely as microcalcifications on MG. Notably, patients often to prefer US-guided biopsy to MG-guided biopsy due to the former’s greater comfort. However, lesions and calcifications are frequently overlooked during US screening. Therefore, second-look US is necessary to help diagnose the nature of BI-RADS 4 calcifications and facilitate the application of US-guided biopsy. We generally do not perform color Doppler in areas where present no lesions or calcifications during screening US, therefore, applying color Doppler of second-look US may reveal occult breast cancers typically often overlooked during screening US.
In terms of clinical predictors of malignant calcifications, age over 50 years and a personal history of breast cancer emerged as independent predictors of malignant calcifications in our study, similar to previous studies on breast cancer.[4,10–13] With regard to mammographic appearances, our study validated the significance of mammographic features, including dense breasts, grouped, linear, or segmental distribution, heterogeneous or fine pleomorphic morphology, and calcification extent exceeding 10 mm in predicting malignant calcifications, which is also consistent with previous research.[2,14,15]
The present study has limitations. First, it was a single-center, retrospective study, which may have introduced some degree of selection bias and limited the generalizability of our findings to a broader population. Second, we did not incorporate cutting-edge techniques such as contrast-enhanced ultrasonography, superb microvascular imaging, or contrast-enhanced mammography. Nevertheless, it is noteworthy that our substantial sample size contributed to the robustness and reliability of our results.
5. Conclusion
Our study demonstrated the significance of second-look US features, such as ultrasonic calcifications, lesions, and moderate or marked color flow, in evaluating BI-RADS 4 calcifications without clinical manifestations and other abnormalities on MG.
Acknowledgments
We would like to thank Editage (www.editage.com) for English language editing.
Author contributions
Data curation: Sheng Cheng, Muzhen He, Yingbin Yu.
Formal analysis: Sheng Cheng, Ning Lin, Hui Zhang.
Investigation: Sheng Cheng, Ning Lin, Muzhen He.
Methodology: Sheng Cheng.
Software: Sheng Cheng.
Validation: Sheng Cheng.
Writing – original draft: Sheng Cheng.
Funding acquisition: Lin Zhu.
Project administration: Lin Zhu, Mengbo Lin, Hui Zhang.
Supervision: Lin Zhu, Hui Zhang.
Writing – review & editing: Lin Zhu.
Resources: Yingbin Yu, Mengbo Lin.
Abbreviations:
- ACR
- American College of Radiology
- AUC
- Area Under the Receiver Operating Characteristic Curve
- BI-RADS
- Breast Imaging Reporting and Data System
- DCA
- Decision Analysis Curve
- MG
- mammography
- SPSS
- Statistical Package for the Social Sciences
- US
- ultrasonography
Natural Science Foundation of Fujian Province of China (grant number 2023J011192); Natural Science Foundation of Fujian Province of China (grant number 2021J01379); Young and Middle-aged Talents Training Project of Fujian Provincial Health Commission (grant number 2021GGA001); Joint Funds for the Innovation of Science and Technology of Fujian Province, China(grant number 2023Y9302).
The authors declare that they have no competing interests.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
How to cite this article: Cheng S, Zhu L, Lin N, He M, Yu Y, Lin M, Zhang H. Utility of Second-look Ultrasonography in Distinguishing BI-RADS 4 Calcifications Detected on Mammography: An observational study. Medicine 2024;103:28(e38841).
Contributor Information
Sheng Cheng, Email: chengsheng0712@163.com.
Ning Lin, Email: mengbolin34@126.com.
Muzhen He, Email: muzhenhee@126.com.
Yingbin Yu, Email: yingbinyuf@126.com.
Mengbo Lin, Email: mengbolin34@126.com.
Hui Zhang, Email: zhanghui35001@126.com.
References
- [1].Spak DA, Plaxco JS, Santiago L, Dryden MJ, Dogan BE. BI-RADS® fifth edition: a summary of changes. Diagn Interv Imaging. 2017;98:179–90. [DOI] [PubMed] [Google Scholar]
- [2].Kim SY, Kim HY, Kim EK, Kim MJ, Moon HJ, Yoon JH. Evaluation of malignancy risk stratification of microcalcifications detected on mammography: a study based on the 5th edition of BI-RADS. Ann Surg Oncol. 2015;22:2895–901. [DOI] [PubMed] [Google Scholar]
- [3].Kim JH, Kim EK, Kim MJ, Moon HJ, Yoon JH. “Category 4A” microcalcifications: how should this subcategory be applied to microcalcifications seen on mammography? Acta Radiol. 2017;59:147–53. [DOI] [PubMed] [Google Scholar]
- [4].Wang HL, Lai JG, Li J, et al. Does establishing a preoperative nomogram including ultrasonographic findings help predict the likelihood of malignancy in patients with microcalcifications? Cancer Imaging. 2019;19:46–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Singareeka Raghavendra A, Alameddine HF, Andersen CR, et al. Influencers of the decision to undergo contralateral prophylactic mastectomy among women with unilateral breast cancer. Cancers (Basel). 2021;13:2050. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Kang SS, Ko E, Han BK, et al. Breast US in patients who had microcalcifications with low concern of malignancy on screening mammography. Eur J Radiol. 2008;67:285–91. [DOI] [PubMed] [Google Scholar]
- [7].Pustahija A, Ivanac G, Brkljacic B. US and MRI in the evaluation of mammographic BI-RADS 4 and 5 microcalcifications. Diagn Interv Radiol. 2018;24:187–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].de Heer EC, Jalving M, Harris AL. HIFs, angiogenesis, and metabolism: elusive enemies in breast cancer. J Clin Invest. 2020;130:5074–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Cao J, Liu X, Yang Y, et al. Decylubiquinone suppresses breast cancer growth and metastasis by inhibiting angiogenesis via the ROS/p53/ BAI1 signaling pathway. Angiogenesis. 2020;23:325–38. [DOI] [PubMed] [Google Scholar]
- [10].Hu Y, Yang YP, Gu R, et al. Does patient age affect the PPV3 of ACR BI-RADS Ultrasound categories 4 and 5 in the diagnostic setting? Eur Radiol. 2018;28:2492–8. [DOI] [PubMed] [Google Scholar]
- [11].Lei SY, Zheng RS, Zhang SW, et al. Breast cancer incidence and mortality in women in China: temporal trends and projections to 2030. Cancer Biol Med. 2021;18:900–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Lewinsohn R, Zheng Y, Rosenberg SM, et al. Fertility preferences and practices among young women with breast cancer: germline genetic carriers versus noncarriers. Clin Breast Cancer. 2023;23:317–23. [DOI] [PubMed] [Google Scholar]
- [13].Cardoso F, van’t Veer LJ, Bogaerts J, et al. MINDACT Investigators. 70-Gene signature as an aid to treatment decisions in early-stage breast cancer. N Engl J Med. 2016;375:717–29. [DOI] [PubMed] [Google Scholar]
- [14].Metaxa L, Healy N, O’Keeffe SA. Breast microcalcifications: the UK RCR 5-point breast imaging system or BI-RADS; which is the better predictor of malignancy? Br J Radiol. 2019;92:20190177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Kaltenbach B, Brandenbusch V, Möbus V, et al. A matrix of morphology and distribution of calcifications in the breast: analysis of 849 vacuum-assisted biopsies. Eur J Radiol. 2017;86:221–6. [DOI] [PubMed] [Google Scholar]