Negative mammograms in symptomatic patients with breast cancer

Deep learning for detection of iso-dense, obscure masses in mammographically dense breasts

OBJECTIVES

To analyze the performance of deep learning in isodense/obscure masses in dense breasts. To build and validate a deep learning (DL) model using core radiology principles and analyze its performance in isodense/obscure masses. To show performance on screening mammography as well as diagnostic mammography distribution.

METHODS

This was a retrospective, single-institution, multi-centre study with external validation. For model building, we took a 3-pronged approach. First, we explicitly taught the network to learn features other than density differences: such as spiculations and architectural distortion. Second, we used the opposite breast to enable the detection of asymmetries. Third, we systematically enhanced each image by piece-wise-linear transformation. We tested the network on a diagnostic mammography dataset (2569 images with 243 cancers, January to June 2018) and a screening mammography dataset (2146 images with 59 cancers, patient recruitment from January to April 2021) from a different centre (external validation).

RESULTS

When trained with our proposed technique (and compared with baseline network), sensitivity for malignancy increased from 82.7 to 84.7% at 0.2 False positives per image (FPI) in the diagnostic mammography dataset, 67.9 to 73.8% in the subset of patients with dense breasts, 74.6 to 85.3 in the subset of patients with isodense/obscure cancers and 84.9 to 88.7 in an external validation test set with a screening mammography distribution. We showed that our sensitivity exceeded currently reported values (0.90 at 0.2 FPI) on a public benchmark dataset (INBreast).

CONCLUSION

Modelling traditional mammographic teaching into a DL framework can help improve cancer detection accuracy in dense breasts.

CLINICAL RELEVANCE STATEMENT

Incorporating medical knowledge into neural network design can help us overcome some limitations associated with specific modalities. In this paper, we show how one such deep neural network can help improve performance on mammographically dense breasts.

KEY POINTS

• Although state-of-the-art deep learning networks achieve good results in cancer detection in mammography in general, isodense, obscure masses and mammographically dense breasts posed a challenge to deep learning networks. • Collaborative network design and incorporation of traditional radiology teaching into the deep learning approach helped mitigate the problem. • The accuracy of deep learning networks may be translatable to different patient distributions. We showed the results of our network on screening as well as diagnostic mammography datasets.

Contrast-Enhanced Mammography: Technique, Indications, and Review of Current Literature

Purpose of Review

To provide an update on contrast-enhanced mammography (CEM) regarding current technique and interpretation, the performance of this modality versus conventional breast imaging modalities (mammography, ultrasound, and MRI), existing clinical applications, potential challenges, and pitfalls.

Recent Findings

Multiple studies have shown that the low-energy, non-contrast-enhanced images obtained when performing CEM are non-inferior to full-field digital mammography with the added benefit of recombined post-contrast images, which have been shown to provide comparable information compared to MRI without sacrificing sensitivity and negative predictive values. While CEMs' usefulness for further diagnostic characterization of indeterminate breast findings is apparent, additional studies have provided strong evidence of potential roles in screening intermediate to high-risk populations, evaluation of disease extent, and monitoring response to therapy, particularly in patients in whom MRI is either unavailable or contraindicated. Others have shown that some patients prefer CEM over MRI given the ease of performance and patient comfort. Additionally, some health systems may find significantly reduced costs compared to MRI. Currently, CEM is hindered by the limited availability of CEM-guided tissue sampling and issues of intravenous contrast administration. However, commercially available CEM-guided biopsy systems are on the horizon, and small changes in practice workflow can be quickly adopted. As of now, MRI remains a mainstay of high-risk screening, evaluation of the extent of disease, and monitoring response to therapy, but smaller studies have suggested that CEM may be equivalent to MRI for these indications, and larger confirmatory studies are needed.

Summary

CEM is an emerging problem-solving breast imaging modality that provides complementary information to conventional imaging modalities and may potentially be used in place of MRI for specific indications and/or patient populations.

Kinetic patterns of benign and malignant breast lesions on contrast enhanced digital mammogram

Purpose

To evaluate the kinetic patterns of benign and malignant breast lesions using contrast-enhanced digital mammogram (CEDM).

Methods

Women with suspicious breast lesions on mammography or ultrasound were enrolled. Single-view mediolateral oblique (MLO) CEDM of an affected breast was acquired at 2, 3, 4, 7, and 10 min after injection of contrast agent. Three readers visually and semi-quantitatively analyzed the enhancement of suspicious lesions. The kinetic pattern of each lesion was classified as persistent, plateau, or washout over two time intervals, 2–4 min and 2–10 min, by comparing the signal intensity at the first time interval with that at the second.

Results

There were 73 malignant and 75 benign lesions in 148 patients (mean age: 52 years). Benign and malignant breast lesions showed the highest signal intensity at 3 min and 2 min, respectively. Average areas under receiver operating characteristic (ROC) curve for diagnostic accuracy based on lesion enhancement at different time points were 0.73 at 2 min, 0.72 at 3 min, 0.69 at 4 min, 0.67 at 7 min, and 0.64 at 10 min. Diagnostic performance was significantly better at 2, 3, and 4 min than at 7 and 10 min (all p < 0.05). A washout kinetic pattern was significantly associated with malignant lesions at 2–4 min and 2–10 min frames according to two of the three readers’ interpretations (all p ≤ 0.001).

Conclusion

Applications of optimal time intervals and kinetic patterns show promise in differentiation of benign and malignant breast lesions on CEDM.

Breast cancer screening with tomosynthesis--initial experiences

Experiences gained so far using tomosynthesis for breast cancer screening will be reported. A short summary of results from preparatory studies will also be presented. The sensitivity and specificity of breast tomosynthesis (BT) will be compared with conventional two-dimensional digital mammography (DM) for breast cancer screening in a population-based study. Over 2000 women have been examined so far with BT and DM. The BT reading is significantly more time-consuming than the DM reading. Preparatory studies have shown that BT has a higher diagnostic precision and higher accuracy of size measurements and stage determination than DM. There is potential to use lower compression force with BT compared with DM, without decreasing the diagnostic accuracy. BT might play an important role in clinical as well as screening mammography. A large-scale population-based study to investigate BT as a screening modality is underway.

X-ray tomosynthesis: a review of its use for breast and chest imaging

Tomosynthesis is a three-dimensional imaging technique based on the reconstruction of several planar radiographs. During the image acquisition in tomosynthesis, the X-ray tube moves around the detector which is often stationary, and a number of projection images are taken from different angles. Individual slices from the reconstructed volume can be studied. With the effective reduction of the visibility of the overlapping normal tissue, the detection of pathological lesions is improved when compared with projection radiography. Up to now, tomosynthesis has mainly been used for breast and chest examinations and, to some extent, also for orthopaedic, angiographic and dental investigations. For chest, tomosynthesis is used as an alternative to computed tomography with significantly lower cost and radiation dose to the patient. Breast tomosynthesis has, in several studies, proved to be an effective tool for improving detection of breast lesions. As tomosynthesis has many properties that make it suitable as a modality for screening, including good diagnostic performance, short examination time and low radiation dose, it is a strong competitor to the current gold standard breast screening modality, i.e. mammography. In this paper, the principles of tomosynthesis will be presented as well as a few clinical studies showing the potential role of tomosynthesis in clinical routine examinations.

Characterization of breast lesions with CE-MR multimodal morphological and kinetic analysis: comparison with conventional mammography and high-resolution ultrasound

OBJECTIVE

Evaluation of the diagnostic value of magnetic resonance mammography and comparison with conventional mammography and ultrasonography in cases of women with suspicious breast lesions.

SUBJECTS AND METHODS

Sixty-nine women (age range 39-68 years) with 78 focal breast lesions were examined with mammography, ultrasonography and dynamic magnetic resonance mammography. The lesions were classified according to the Breast Imaging Reporting and Data System (BI-RADS) lexicon of the American College of Radiology for each diagnostic method. Histological reports were available after biopsy or surgical excision of the lesions.

RESULTS

Pathological examination confirmed that 53 lesions were malignant and 25 benign. Conventional mammography estimated a total of 59/78 lesions as malignant with 44 true positive lesions, ultrasonography estimated a total of 50/78 lesions as malignant with 44 true positive lesions and magnetic resonance mammography estimated a total of 66/78 lesions as malignant with 52 true positive lesions. Sensitivity and specificity of magnetic resonance mammography in the diagnosis of malignancy was 98.1% and 44%, of conventional mammography 83% and 40% and of ultrasonography 83% and 76%. Negative predictive value for magnetic resonance mammography was 91.7%, for ultrasonography 67.9% and for mammography 52.6% for malignancies.

CONCLUSION

Magnetic resonance mammography has the highest negative predictive value compared with mammography and ultrasound in cases of suspicious breast lesions. The combination of morphologic and enhancement criteria can improve the diagnostic capability of magnetic resonance mammography (MRM) in breast lesion characterization.

Des cancers du sein sans signe en mammographie : quand et pourquoi ?

Breast carcinomas with false negative mammogram correspond to carcinomas diagnosed with means other than mammogram when the mammogram, at the time of diagnosis, showed no significant abnormality. The rate of false mammogram is difficult to ascertain because few studies have been published on this subject. The absence of mammographic abnormality is related to histopathological characteristics of the tumor and mammographic features of the patient's breast tissue. The small size of the tumor, the lack of microcalcifications often due to tumor necrosis and the absence of any significant associated desmoplastic reaction are the main histopathological factors encountered, particularly for dense breasts.

Missed breast carcinoma: pitfalls and pearls

Mammography is the standard of reference for the detection of breast carcinoma, yet 10%-30% of breast cancers may be missed at mammography. Possible causes for missed breast cancers include dense parenchyma obscuring a lesion, poor positioning or technique, perception error, incorrect interpretation of a suspect finding, subtle features of malignancy, and slow growth of a lesion. Recent studies have emphasized the use of alternative imaging modalities to detect and diagnose breast carcinoma, including ultrasonography (US), magnetic resonance imaging, and nuclear medicine studies. However, the radiologist can take a number of steps that will significantly enhance the accuracy of image interpretation at mammography and decrease the false-negative rate. These steps include performing diagnostic as well as screening mammography, reviewing clinical data and using US to help assess a palpable or mammographically detected mass, strictly adhering to positioning and technical requirements, being alert to subtle features of breast cancers, comparing recent images with earlier mammograms to look for subtle increases in lesion size, looking for additional lesions when one abnormality is seen, and judging a lesion by its most malignant features.

Use of the American College of Radiology BI-RADS to report on the mammographic evaluation of women with signs and symptoms of breast disease

PURPOSE

To examine whether mammographic assessments and recommendations are linked as expected, based on the Breast Imaging Reporting and Data System (BI-RADS), for the evaluation of women with signs and symptoms of breast disease.

MATERIALS AND METHODS

Eight mammography registries from the Breast Cancer Surveillance Consortium contributed mammographic data from 1996 through 1997 for women 25 years of age or older, with signs or symptoms of breast cancer. The association of assessments and recommendations and the relationship of self-reported symptoms to assessments are described.

RESULTS

A total of 51,673 diagnostic mammograms were included in the analyses and the expected management recommendation was provided 85%-90% of the time for mammograms classified as assessment categories 1, 2, 4, or 5. Category 3 ("probably benign finding") had the most variability in associated management recommendations, with only 40% (2,998 of 7,423) of cases associated with the recommendation for short interval follow-up. Of the 1,648 category 0 mammograms ("needs additional imaging") that did not have a final assessment, 64% were recommended for additional imaging, while another 20% of the cases were recommended for either a consultation or biopsy. The number of women who reported a lump as a symptom decreased with age but was associated with higher BI-RADS assessments.

CONCLUSION

BI-RADS assessment categories were generally used as intended for all categories but 0 and 3. Additional education about the use of these categories may be warranted. The inconsistencies between assessment category and management recommendations may present difficulties in conducting outcome audits.

Diagnostic value of radiological breast imaging in a non-screening population

The aim of this study was to assess the diagnostic performance of breast imaging in the diagnosis of breast cancer in a non-screening population. In a consecutive set of patients referred for mammography in one year, the results of palpation and radiological breast imaging were scored on a 5-point grading scale and linked to pathology as gold standard after a follow up period of one year. The diagnostic performance was studied by logistic regression analysis and ROC-curves. There were 1,944 breast examinations in 1,890 patients and 3,816 breasts. Pathology results reported 118 malignancies in 115 women. With a cut-off point between benign and uncertain benign a sensitivity of 89% and a specificity of 98% was found for radiological imaging. ROC-curves showed a significant increase in diagnostic performance when radiology was added to results of palpation and age (p = 0.007). Radiological imaging tests have a large diagnostic value in the detection of breast cancer in addition to palpation and age. A sensitivity close to 100% could be reached.

False-negative findings of malignant breast lesions on preoperative magnetic resonance mammography

The aim of this study was to evaluate causes and percentages of false negative diagnoses of malignant breast lesions on preoperative dynamic magnetic resonance mammography (MRM). MRM was performed in 223 patients with 234 histopathologically proven malignant breast lesions (193 invasive carcinoma, 41 CIS) which were analyzed prospectively by routine analysis prior to surgery and re-analyzed by specialists, retrospectively. False negative findings were re-evaluated with respect to contrast enhancement, size and shape of lesions, reading errors, and technical problems. Preoperative analysis missed 27 of 234 malignant breast lesions (sensitivity 88.5%) including 15 of 193 invasive cancers (sensitivity 92%) and 12 of 41 CIS (sensitivity 71%). Five of 193 invasive cancers (four invasive lobular, one invasive tubular carcinoma) and five of 41 CIS lesions were missed due to delayed or no contrast enhancement. The remaining 17 false negative diagnoses were due to reading errors (n=8), previous core biopsies (n=3), metal induced artefacts (n=3), localization outside the field of view (n=1), incorrect injection (n=1) or movement artefacts (n=1). Using dynamic MR mammography, there were 4.3% slow contrast enhancing malignant breast lesions and a maximum sensitivity of 95.7% for detection of all malignant breast lesions (97.4% for invasive breast cancer, 87.8% for carcinoma in situ) can be achieved in a preselected preoperative population.

Application of magnetic resonance imaging to early detection of breast cancer

Since its introduction approximately 10 years ago, there has been extensive progress in the application of magnetic resonance imaging (MRI) to the detection and diagnosis of breast cancer. Contrast-enhanced MRI has been shown to have value in the diagnostic work-up of women who present with mammogram or clinical abnormalities. In addition, it has been demonstrated that MRI can detect mammogram occult multifocal cancer in patients who present with unifocal disease. Advances in risk stratification and limitations in mammography have stimulated interest in the use of MRI to screen high-risk women for cancer. Several studies of MRI high-risk screening are ongoing. Preliminary results are encouraging.

Effect of variations in operational definitions on performance estimates for screening mammography

RATIONALE AND OBJECTIVES

The Mammography Quality Standards Act requires practices to measure limited aspects of their performance. The authors conducted this study to calculate the differences in measurements of sensitivity and specificity due only to differences in the definitions used in the analysis. This included definitions for case inclusion.

MATERIALS AND METHODS

Data from the New Mexico Mammography Project for January 1991 to December 1995 on 136,540 women who underwent screening mammography were analyzed. A starting definition was created for each performance measure. The components of the definition were varied, and estimates of sensitivity and specificity for the different definitions were calculated.

RESULTS

Sensitivity was lower and specificity was higher when assessed on the basis of the results of all imaging performed in the screening work-up rather than on the initial screening examination alone. Sensitivity was higher and specificity was lower in women who did not undergo rather than in women who did recently undergo a previous examination. When the definition of a positive examination included cases that were recommended for short-term follow-up, the work-up sensitivity was slightly higher and the work-up specificity was considerably lower. Longer follow-up times for determining the diagnosis of cancer were associated with decreasing sensitivity, particularly when the follow-up period extended beyond 12 months.

CONCLUSION

Variations in the operational definitions for measures of mammographic performance affect these estimates. To facilitate valid comparisons, reports need to be explicit regarding the definitions and methods used.