Automatic detection of various patterns of calcification distribution on mammograms: a graph convolution network approach (Preprint)

UNSTRUCTURED Purpose: To develop a novel artificial intelligence (AI) model algorithm focusing on automatic detection and classification of various patterns of calcification distribution in mammographic images using a unique graph convolution approach. Materials and methods: Images from 200 patients classified as Category 4 or 5 according to the American College of Radiology Breast Imaging Reporting and Database System, which showed calcifications according to the mammographic reports and diagnosed breast cancers. The calcification distributions were classified as either diffuse, segmental, regional, grouped, or linear. Excluded were mammograms with (1) breast cancer as a single or combined characterization such as a mass, asymmetry, or architectural distortion with or without calcifications; (2) hidden calcifications that were difficult to mark; or (3) incomplete medical records. Results: A graph convolutional network-based model was developed. 401 mammographic images from 200 cases of breast cancer were divided based on calcification distribution pattern: diffuse (n = 24), regional (n = 111), group (n = 201), linear (n = 8) or segmental (n = 57). The classification performances were measured using metrics including precision, recall, F1 score, accuracy and multi-class area under receiver operating characteristic curve. The proposed achieved precision of 0.483 ± 0.015, sensitivity of 0.606 (0.030), specificity of 0.862 ± 0.018, F1 score of 0.527 ± 0.035, accuracy of 60.642% ± 3.040% and area under the curve of 0.754 ± 0.019, finding method to be superior compared to all baseline models. The predicted linear and diffuse classifications were highly similar to the ground truth, and the predicted grouped and regional classifications were also superior compared to baseline models. Conclusion: The proposed deep neural network framework is an AI solution to automatically detect and classify calcification distribution patterns on mammographic images highly suspected of showing breast cancers. Further study of the AI model in an actual clinical setting and additional data collection will improve its performance.

Download Full-text

Posterior breast cancer: Mammographic and ultrasonographic features

Vojnosanitetski pregled ◽

10.2298/vsp120726032j ◽

2013 ◽

Vol 70 (11) ◽

pp. 1034-1038

Author(s):

Ana Jankovic ◽

Mirjan Nadrljanski ◽

Vesna Plesinac-Karapandzic ◽

Nebojsa Ivanovic ◽

Zoran Radojicic ◽

...

Keyword(s):

Breast Cancer ◽

Breast Imaging ◽

Color Doppler ◽

Lobular Carcinoma ◽

Architectural Distortion ◽

Breast Cancers ◽

Significant Difference ◽

Diagnostic Mammography ◽

Anatomical Localization ◽

Mass Lesions

Background/Aim. Posterior breast cancers are located in the prepectoral region of the breast. Owing to this distinctive anatomical localization, physical examination and mammographic or ultrasonographic evaluation can be difficult. The purpose of the study was to assess possibilities of diagnostic mammography and breast ultrasonography in detection and differentiation of posterior breast cancers. Methods. The study included 40 women with palpable, histopathological confirmed posterior breast cancer. Mammographic and ultrasonographic features were defined according to Breast Imaging Reporting and Data System (BI-RADS) lexicon. Results. Based on standard two-view mammography 87.5%, of the cases were classified as BI-RADS 4 and 5 categories, while after additional mammographic views all the cases were defined as BIRADS 4 and 5 categories. Among 96 mammographic descriptors, the most frequent were: spiculated mass (24.0%), architectural distortion (16.7%), clustered microcalcifications (12.6%) and focal asymmetric density (12.6%). The differentiation of the spiculated mass was significantly associated with the possibility to visualize the lesion at two-view mammography (p = 0.009), without the association with lesion diameter (p = 0.083) or histopathological type (p = 0.055). Mammographic signs of invasive lobular carcinoma were significantly different from other histopathological types (architectural distortion, p = 0.003; focal asymmetric density, p = 0.019; association of four or five subtle signs of malignancy, p = 0.006). All cancers were detectable by ultrasonography. Mass lesions were found in 82.0% of the cases. Among 153 ultrasonographic descriptors, the most frequent were: irregular mass (15.7%), lobulated mass (7.2%), abnormal color Doppler signals (20.3%), posterior acoustic attenuation (18.3%). Ultrasonographic BI-RADS 4 and 5 categories were defined in 72.5% of the cases, without a significant difference among various histopathological types (p = 0.109). Conclusion. Standard two-view mammography followed by additional mammographic projections is an effective way to demonstrate the spiculated mass and to classify the prepectoral lesion as category BI-RADS 4 or 5. Additional ultrasonography can overcome the mimicry of invasive lobular breast carcinoma at mammography.

Download Full-text

An Antibody-based Blood Test Utilizing a Panel of Biomarkers as a New Method for Improved Breast Cancer Diagnosis

Biomarkers in Cancer ◽

10.4137/bic.s13236 ◽

2013 ◽

Vol 5 ◽

pp. BIC.S13236 ◽

Cited By ~ 9

Author(s):

Galit Yahalom ◽

Daria Weiss ◽

Ilya Novikov ◽

Therese B. Bevers ◽

Laszlo G. Radvanyi ◽

...

Keyword(s):

Breast Cancer ◽

Breast Imaging ◽

Blood Test ◽

Characteristic Curve ◽

Clinical Status ◽

Area Under The Curve ◽

Breast Cancer Diagnosis ◽

Receiver Operator Characteristic Curve ◽

Enzyme Linked Immunosorbent Assay ◽

Breast Cancer Patients

In order to develop a new tool for diagnosis of breast cancer based on autoantibodies against a panel of biomarkers, a clinical trial including blood samples from 507 subjects was conducted. All subjects showed a breast abnormality on exam or breast imaging and final biopsy pathology of either breast cancer patients or healthy controls. Using an enzyme-linked immunosorbent assay, the samples were tested for autoantibodies against a predetermined number of biomarkers in various models that were used to determine a diagnosis, which was compared to the clinical status. Our new assay achieved a sensitivity of 95.2% [CI = 92.8–96.8%] at a fixed specificity of 49.5%. Receiver-operator characteristic curve analysis showed an area under the curve of 80.1% [CI = 72.6–87.6%]. These results suggest that a blood test which is based on models comprising several autoantibodies to specific biomarkers may be a new and novel tool for improving the diagnostic evaluation of breast cancer.

Download Full-text

Fat Composition Measured by Proton Spectroscopy: A Breast Cancer Tumor Marker?

Diagnostics ◽

10.3390/diagnostics11030564 ◽

2021 ◽

Vol 11 (3) ◽

pp. 564

Author(s):

Almir Bitencourt ◽

Varadan Sevilimedu ◽

Elizabeth A. Morris ◽

Katja Pinker ◽

Sunitha B. Thakur

Keyword(s):

Breast Cancer ◽

Saturation Index ◽

Multivariable Analysis ◽

Area Under The Curve ◽

Resonance Spectroscopy ◽

Breast Cancers ◽

Her2 Positive ◽

Tumor Tissues ◽

Cancer Tumor ◽

Fat Composition

Altered metabolism including lipids is an emerging hallmark of breast cancer. The purpose of this study was to investigate if breast cancers exhibit different magnetic resonance spectroscopy (MRS)-based lipid composition than normal fibroglandular tissue (FGT). MRS spectra, using the stimulated echo acquisition mode sequence, were collected with a 3T scanner from patients with suspicious lesions and contralateral normal tissue. Fat peaks at 1.3 + 1.6 ppm (L13 + L16), 2.1 + 2.3 ppm (L21 + L23), 2.8 ppm (L28), 4.1 + 4.3 ppm (L41 + L43), and 5.2 + 5.3 ppm (L52 + L53) were quantified using LCModel software. The saturation index (SI), number of double bods (NBD), mono and polyunsaturated fatty acids (MUFA and PUFA), and mean chain length (MCL) were also computed. Results showed that mean concentrations of all lipid metabolites and PUFA were significantly lower in tumors compared with that of normal FGT (p ≤ 0.002 and 0.04, respectively). The measure best separating normal and tumor tissues after adjusting with multivariable analysis was L21 + L23, which yielded an area under the curve of 0.87 (95% CI: 0.75–0.98). Similar results were obtained between HER2 positive versus HER2 negative tumors. Hence, MRS-based lipid measurements may serve as independent variables in a multivariate approach to increase the specificity of breast cancer characterization.

Download Full-text

Breast cancer detection using deep convolutional neural networks and support vector machines

PeerJ ◽

10.7717/peerj.6201 ◽

2019 ◽

Vol 7 ◽

pp. e6201 ◽

Cited By ~ 48

Author(s):

Dina A. Ragab ◽

Maha Sharkas ◽

Stephen Marshall ◽

Jinchang Ren

Keyword(s):

Breast Cancer ◽

Breast Imaging ◽

Input Data ◽

Data Augmentation ◽

Area Under The Curve ◽

Screening Mammography ◽

Support Vector ◽

Svm Classifier ◽

Deep Convolutional Neural Networks ◽

Cad System

It is important to detect breast cancer as early as possible. In this manuscript, a new methodology for classifying breast cancer using deep learning and some segmentation techniques are introduced. A new computer aided detection (CAD) system is proposed for classifying benign and malignant mass tumors in breast mammography images. In this CAD system, two segmentation approaches are used. The first approach involves determining the region of interest (ROI) manually, while the second approach uses the technique of threshold and region based. The deep convolutional neural network (DCNN) is used for feature extraction. A well-known DCNN architecture named AlexNet is used and is fine-tuned to classify two classes instead of 1,000 classes. The last fully connected (fc) layer is connected to the support vector machine (SVM) classifier to obtain better accuracy. The results are obtained using the following publicly available datasets (1) the digital database for screening mammography (DDSM); and (2) the Curated Breast Imaging Subset of DDSM (CBIS-DDSM). Training on a large number of data gives high accuracy rate. Nevertheless, the biomedical datasets contain a relatively small number of samples due to limited patient volume. Accordingly, data augmentation is a method for increasing the size of the input data by generating new data from the original input data. There are many forms for the data augmentation; the one used here is the rotation. The accuracy of the new-trained DCNN architecture is 71.01% when cropping the ROI manually from the mammogram. The highest area under the curve (AUC) achieved was 0.88 (88%) for the samples obtained from both segmentation techniques. Moreover, when using the samples obtained from the CBIS-DDSM, the accuracy of the DCNN is increased to 73.6%. Consequently, the SVM accuracy becomes 87.2% with an AUC equaling to 0.94 (94%). This is the highest AUC value compared to previous work using the same conditions.

Download Full-text

Breast Imaging During Pregnancy and Lactation

Journal of Breast Imaging ◽

10.1093/jbi/wbz065 ◽

2019 ◽

Vol 1 (4) ◽

pp. 342-351

Author(s):

Lisa Abramson ◽

Lindsey Massaro ◽

J Jaime Alberty-Oller ◽

Amy Melsaether

Keyword(s):

Breast Cancer ◽

Breast Imaging ◽

Treatment Options ◽

Nipple Discharge ◽

Cancer Staging ◽

Imaging Features ◽

Breast Cancers ◽

Bloody Nipple Discharge ◽

Breast Cancer Staging ◽

Pregnancy And Lactation

Abstract Breast imaging during pregnancy and lactation is important in order to avoid delays in the diagnosis and treatment of pregnancy-associated breast cancers. Radiologists have an opportunity to improve breast cancer detection by becoming familiar with appropriate breast imaging and providing recommendations to women and their referring physicians. Importantly, during pregnancy and lactation, both screening and diagnostic breast imaging can be safely performed. Here we describe when and how to screen, how to work up palpable masses, and evaluate bloody nipple discharge. The imaging features of common findings in the breasts of pregnant and lactating women are also reviewed. Finally, we address breast cancer staging and provide a brief primer on treatment options for pregnancy-associated breast cancers.

Download Full-text

Radiological review of prior screening mammograms of screen-detected breast cancer

European Radiology ◽

10.1007/s00330-020-07130-y ◽

2020 ◽

Cited By ~ 1

Author(s):

Tone Hovda ◽

Kaitlyn Tsuruda ◽

Solveig Roth Hoff ◽

Kristine Kleivi Sahlberg ◽

Solveig Hofvind

Keyword(s):

Breast Cancer ◽

Mammographic Density ◽

Breast Imaging ◽

Screening Program ◽

Tumor Characteristics ◽

Breast Cancers ◽

Mammographic Finding ◽

Estrogen Receptor Positivity ◽

Screening Mammograms ◽

Screening And Diagnosis

Abstract Objective To perform a radiological review of mammograms from prior screening and diagnosis of screen-detected breast cancer in BreastScreen Norway, a population-based screening program. Methods We performed a consensus-based informed review of mammograms from prior screening and diagnosis for screen-detected breast cancers. Mammographic density and findings on screening and diagnostic mammograms were classified according to the Breast Imaging-Reporting and Data System®. Cases were classified based on visible findings on prior screening mammograms as true (no findings), missed (obvious findings), minimal signs (minor/non-specific findings), or occult (no findings at diagnosis). Histopathologic tumor characteristics were extracted from the Cancer Registry of Norway. The Bonferroni correction was used to adjust for multiple testing; p < 0.001 was considered statistically significant. Results The study included mammograms for 1225 women with screen-detected breast cancer. Mean age was 62 years ± 5 (SD); 46% (567/1225) were classified as true, 22% (266/1225) as missed, and 32% (392/1225) as minimal signs. No difference in mammographic density was observed between the classification categories. At diagnosis, 59% (336/567) of true and 70% (185/266) of missed cancers were classified as masses (p = 0.004). The percentage of histological grade 3 cancers was higher for true (30% (138/469)) than for missed (14% (33/234)) cancers (p < 0.001). Estrogen receptor positivity was observed in 86% (387/469) of true and 95% (215/234) of missed (p < 0.001) cancers. Conclusions We classified 22% of the screen-detected cancers as missed based on a review of prior screening mammograms with diagnostic images available. One main goal of the study was quality improvement of radiologists’ performance and the program. Visible findings on prior screening mammograms were not necessarily indicative of screening failure. Key Points • After a consensus-based informed review, 46% of screen-detected breast cancers were classified as true, 22% as missed, and 32% as minimal signs. • Less favorable prognostic and predictive tumor characteristics were observed in true screen-detected breast cancer compared with missed. • The most frequent mammographic finding for all classification categories at the time of diagnosis was mass, while the most frequent mammographic finding on prior screening mammograms was a mass for missed cancers and asymmetry for minimal signs.

Download Full-text

Triple-negative invasive breast cancer (TNBC): Mammographic, US, and MR imaging features according to androgen receptor (AR) expression.

Journal of Clinical Oncology ◽

10.1200/jco.2014.32.26_suppl.159 ◽

2014 ◽

Vol 32 (26_suppl) ◽

pp. 159-159

Author(s):

Woo Kyung Moon

Keyword(s):

Breast Cancer ◽

Mr Imaging ◽

Breast Imaging ◽

Gene Signature ◽

Irregular Shape ◽

Consecutive Series ◽

Imaging Features ◽

Breast Cancers ◽

Pathologic Features ◽

Chi Squared

159 Background: A subset of TNBC is characterized by an androgen gene signature and early clinical trials have demonstrated clinical benefit with the use of the AR antagonist, bicalutamide, for the treatment of patients with AR+, estrogen receptor/progesterone receptor- breast cancer. Methods: AR expression was assessed immunohistochemically in 125 patients (median age; 54 years, range; 26-82 years) with TNBC from a consecutive series of 1,086 operable invasive breast cancers. Two experienced breast imaging radiologists (6 and 24 years of experience, respectively) reviewed the mammograms, US, and MR images without knowledge of clinicopathologic findings. The imaging and pathologic features of 33 AR-positive TNBCs were compared with those of 92 AR-negative TNBCs by using the Fisher’s exact or chi-squared tests. Results: AR expression in TNBC is significantly associated with mammographic findings (P < 0.001), lesion type at MR imaging (P < 0.001), and mass shape or margin at ultrasound (P < 0.001; P= 0.002). The highest PPVs for AR-positive cancer were non-mass enhancement on MR imaging (PPV, 1.00; 95% CI: 0.61, 1.00), calcifications only seen on mammography (PPV, 1.00; 95% CI: 0.37, 1.00), and spiculated masses on US (PPV, 1.00; 95% CI: 0.22, 1.00). Conclusions: AR-positive and AR-negative tumors have distinct imaging features in TNBC. The presence of calcifications or focal asymmetries at mammography, the presence of echogenic halo or non-complex hypoechoic masses at US, masses with irregular shape or indistinct margins at mammography and US, and masses with irregular shape or spiculated margins, or non-mass lesions at MR imaging were associated with AR expression in TNBC. These imaging features may be used to predict AR status, which could assist in treatment planning, prediction of response, and assessment of prognosis for patients with TNBC.

Download Full-text

Computer-aided classification of MRI for pathological complete response to neoadjuvant chemotherapy in breast cancer

Future Oncology ◽

10.2217/fon-2021-1212 ◽

2021 ◽

Author(s):

Shaolei Yan ◽

Haiyong Peng ◽

Qiujie Yu ◽

Xiaodan Chen ◽

Yue Liu ◽

...

Keyword(s):

Breast Cancer ◽

Neoadjuvant Chemotherapy ◽

Pathological Complete Response ◽

Characteristic Curve ◽

Area Under The Curve ◽

Complete Response ◽

Automated Classification ◽

General Applicability ◽

Computer Aided ◽

Auc Value

Background: To determine suitable optimal classifiers and examine the general applicability of computer-aided classification to compare the differences between a computer-aided system and radiologists in predicting pathological complete response (pCR) from patients with breast cancer receiving neoadjuvant chemotherapy. Methods: We analyzed a total of 455 masses and used the U-Net network and ResNet to execute MRI segmentation and pCR classification. The diagnostic performance of radiologists, the computer-aided system and a combination of radiologists and computer-aided system were compared using receiver operating characteristic curve analysis. Results: The combination of radiologists and computer-aided system had the best performance for predicting pCR with an area under the curve (AUC) value of 0.899, significantly higher than that of radiologists alone (AUC: 0.700) and computer-aided system alone (AUC: 0.835). Conclusion: An automated classification system is feasible to predict the pCR to neoadjuvant chemotherapy in patients with breast cancer and can complement MRI.

Download Full-text

Pros and Cons for Automated Breast Ultrasound (ABUS): A Narrative Review

Journal of Personalized Medicine ◽

10.3390/jpm11080703 ◽

2021 ◽

Vol 11 (8) ◽

pp. 703

Author(s):

Ioana Boca (Bene) ◽

Anca Ileana Ciurea ◽

Cristiana Augusta Ciortea ◽

Sorin Marian Dudea

Keyword(s):

Breast Cancer ◽

Breast Ultrasound ◽

Architectural Distortion ◽

Breast Cancers ◽

Ultrasound Technique ◽

Advantages And Disadvantages ◽

Increased Risk ◽

Pros And Cons ◽

Low Sensitivity ◽

Automated Breast Ultrasound

Automated breast ultrasound (ABUS) is an ultrasound technique that tends to be increasingly used as a supplementary technique in the evaluation of patients with dense glandular breasts. Patients with dense breasts have an increased risk of developing breast cancer compared to patients with fatty breasts. Furthermore, for this group of patients, mammography has a low sensitivity in detecting breast cancers, especially if it is not associated with architectural distortion or calcifications. ABUS is a standardized examination with many advantages in both screening and diagnostic settings: it increases the detection rate of breast cancer, improves the workflow, and reduces the examination time. On the other hand, like any imaging technique, ABUS has disadvantages and even some limitations. Many disadvantages can be diminished by additional attention and training. Disadvantages regarding image acquisition are the inability to assess the axilla, the vascularization, and the elasticity of a lesion, while concerning the interpretation, the disadvantages are the artifacts due to poor positioning, lack of contact, motion or lesion related. This article reviews and discusses the indications, the advantages, and disadvantages of the method and also the sources of error in the ABUS examination.

Download Full-text

Detecting accurate tumor size across imaging modalities in breast cancer

Breast Disease ◽

10.3233/bd-201021 ◽

2021 ◽

pp. 1-6

Author(s):

Olutayo Sogunro ◽

Constance Cashen ◽

Sami Fakir ◽

Julie Stausmire ◽

Nancy Buderer

Keyword(s):

Breast Cancer ◽

Tumor Size ◽

Breast Imaging ◽

Imaging Modality ◽

Breast Cancer Diagnosis ◽

Imaging Modalities ◽

Clinical Staging ◽

P Value ◽

Pathology Report ◽

Breast Cancers

BACKGROUND: Of the most common imaging modalities for breast cancer diagnosis – mammogram (MAM), ultrasound (US), magnetic resonance imaging (MRI) – it has not been well established which of these most accurately corresponds to the histological tumor size. OBJECTIVE: To determine which imaging modality (MAM, US, MRI) is most accurate for determining the histological tumor size of breast lesions. METHODS: A retrospective study of 76 breast cancers found in 73 female patients who received MAM, US, and/or MRI was performed. 239 charts were reviewed and 73 patients met inclusion criteria. Analysis was performed using signed rank tests comparing the reported tumor size on the imaging modality to the tumor size on pathology report. RESULTS: Mammography and ultrasonography underestimated tumor size by 3.5 mm and 4 mm (p-values < 0.002), respectively. MRI tends to overestimate tumor size by 3 mm (p-value = 0.0570). Mammogram was equivalent to pathological size within 1 mm 24% of the time and within 2 mm 35% of the time. CONCLUSIONS: No one single modality is the most accurate for detecting tumor size. When interpreting the size reported on breast imaging modalities, the amount of underestimation and overestimation in tumor size should be considered for both clinical staging and surgical decision-making.

Download Full-text