Imaging Of Microcalcifications Research Articles

The impact of (simulated) resolution on breast cancer diagnosis based on high-resolution 3D micro-CT microcalcification images.

Breast microcalcifications (MCs) are considered to be a robust marker of breast cancer. A machine learning model can provide breast cancer diagnosis based on properties of individual MCs - if their characteristics are captured at high resolution and in3D. The main purpose of the study was to explore the impact of image resolution (8µm, 16µm, 32µm, 64µm) when diagnosing breast cancer using radiomics features extracted from individual high resolution 3D micro-CT MCimages. Breast MCs extracted from 86 female patients were analyzed at four different spatial resolutions: 8µm (original resolution) and 16µm, 32µm, 64µm (simulated image resolutions). Radiomic features were extracted at each image resolution in an attempt, to find a compact feature signature allowing to distinguish benign and malignant MCs. Machine learning algorithms were used for classifying individual MCs and samples (i.e., patients). For sample diagnosis, a custom-based thresholding approach was used to combine individual MC results into sample results. We conducted classification experiments when using (a) the same MCs visible in 8µm, 16µm, 32µm, and 64µm resolution; (b) the same MCs visible in 8µm, 16µm, and 32µm resolution; (c) the same MCs visible in 8µm and 16µm resolution; (d) all MCs visible in 8µm, 16µm, 32µm, and 64µm resolution. Accuracy, sensitivity, specificity, AUC, and F1 score were computed for eachexperiment. The individual MC results yielded an accuracy of 77.27%, AUC of 83.83%, F1 score of 77.25%, sensitivity of 80.86%, and specificity of 72.2% at 8µm resolution. For the individual MC classifications we report for the F1 scores: a 2.29% drop when using 16µm instead of 8µm, a 4.01% drop when using 32µm instead of 8µm, a 10.69% drop when using 64µm instead of 8µm. The sample results yielded an accuracy and F1 score of 81.4%, sensitivity of 80.43%, and specificity value of 82.5% at 8µm. For the sample classifications we report for F1 score values: a 6.3% drop when using 16µm instead of 8µm, a 4.91% drop when using 32µm instead of 8µm, and a 6.3% drop when using 64µm instead of 8µm. The highest classification results are obtained at the highest resolution (8µm). If breast MCs characteristics could be visualized/captured in 3D at a higher resolution compared to what is used nowadays in digital mammograms (approximately 70µm), breast cancer diagnosis will beimproved.

Persistent Homology-Based Machine Learning Method for Filtering and Classifying Mammographic Microcalcification Images in Early Cancer Detection.

Microcalcifications in mammogram images are primary indicators for detecting the early stages of breast cancer. However, dense tissues and noise in the images make it challenging to classify the microcalcifications. Currently, preprocessing procedures such as noise removal techniques are applied directly on the images, which may produce a blurry effect and loss of image details. Further, most of the features used in classification models focus on local information of the images and are often burdened with details, resulting in data complexity. This research proposed a filtering and feature extraction technique using persistent homology (PH), a powerful mathematical tool used to study the structure of complex datasets and patterns. The filtering process is not performed directly on the image matrix but through the diagrams arising from PH. These diagrams will enable us to distinguish prominent characteristics of the image from noise. The filtered diagrams are then vectorised using PH features. Supervised machine learning models are trained on the MIAS and DDSM datasets to evaluate the extracted features' efficacy in discriminating between benign and malignant classes and to obtain the optimal filtering level. This study reveals that appropriate PH filtering levels and features can improve classification accuracy in early cancer detection.

Application of the mineral-binding protein fetuin-A for the detection of calcified lesions.

Rationale: Calcium plays an essential role in the biology of vertebrates. Calcium content in body fluids is maintained within a narrow physiologic range by feedback control. Phosphate is equally important for metabolism and is likewise controlled, albeit over a wider range. This results in a nearly supersaturated state of calcium phosphate in body liquids driving mineral precipitation in soft tissues, which is actively prevented by calcification inhibitors. The hepatic plasma protein fetuin-A is a circulating mineralization inhibitor regulating calcium phosphate crystal growth and calcified matrix metabolism. Ectopic mineralization is associated with many pathological conditions aggravating their outcome. Current diagnostic methods lack sensitivity towards microcalcifications representing the initial stages of the process. Given the irreversibility of established calcifications, novel diagnostic tools capable of detecting nascent calcium phosphate deposits are highly desirable. Methods: We designed fluorescent fusion proteins consisting of fetuin-A coupled to a green or red fluorescent protein derivate, mEmerald or mRuby3, respectively. The proteins were expressed in mammalian cell lines. Sequence optimization resolved folding issues and increased sensitivity of mineral binding. Chimeric proteins were tested for their ability to detect calcifications in cell cultures and tissue sections retrieved from calcification-prone mice. Results: We employed novel genetically labeled fetuin-A-based fluorescent proteins for the detection of ectopic calcifications. We show that fetuin-A-based imaging agents are non-toxic and suitable for live imaging of microcalcifications beyond the detection limit of conventional staining techniques. The ability of fetuin-A to preferentially bind nascent calcium phosphate crystals allowed the resolution of histopathological detail of early kidney damage that went previously undetected. Endogenous expression of fetuin-A fluorescent fusion proteins allowed extended live imaging of calcifying cells with unprecedented sensitivity and specificity. Conclusion: Ectopic microcalcifications represent a major clinical concern lacking effective diagnostic and treatment options. In this paper, we describe novel highly sensitive fetuin-A-based fluorescent probes for imaging microcalcifications. We show that fusion proteins consisting of a fetuin-A mineral binding moiety and a fluorescent protein are superior to the routine methods for detecting calcifications. They also surpass in continuous live cell imaging the chemically fluorescence labeled fetuin-A, which we established previously.

Microcalcification Discrimination in Mammography Using Deep Convolutional Neural Network: Towards Rapid and Early Breast Cancer Diagnosis.

Breast cancer is among the most common types of cancer in women and under the cases of misdiagnosed, or delayed in treatment, the mortality risk is high. The existence of breast microcalcifications is common in breast cancer patients and they are an effective indicator for early sign of breast cancer. However, microcalcifications are often missed and wrongly classified during screening due to their small sizes and indirect scattering in mammogram images. Motivated by this issue, this project proposes an adaptive transfer learning deep convolutional neural network in segmenting breast mammogram images with calcifications cases for early breast cancer diagnosis and intervention. Mammogram images of breast microcalcifications are utilized to train several deep neural network models and their performance is compared. Image filtering of the region of interest images was conducted to remove possible artifacts and noises to enhance the quality of the images before the training. Different hyperparameters such as epoch, batch size, etc were tuned to obtain the best possible result. In addition, the performance of the proposed fine-tuned hyperparameter of ResNet50 is compared with another state-of-the-art machine learning network such as ResNet34, VGG16, and AlexNet. Confusion matrices were utilized for comparison. The result from this study shows that the proposed ResNet50 achieves the highest accuracy with a value of 97.58%, followed by ResNet34 of 97.35%, VGG16 96.97%, and finally AlexNet of 83.06%.

Improved automated early detection of breast cancer based on high resolution 3D micro-CT microcalcification images

BackgroundThe detection of suspicious microcalcifications on mammography represents one of the earliest signs of a malignant breast tumor. Assessing microcalcifications’ characteristics based on their appearance on 2D breast imaging modalities is in many cases challenging for radiologists. The aims of this study were to: (a) analyse the association of shape and texture properties of breast microcalcifications (extracted by scanning breast tissue with a high resolution 3D scanner) with malignancy, (b) evaluate microcalcifications’ potential to diagnose benign/malignant patients.MethodsBiopsy samples of 94 female patients with suspicious microcalcifications detected during a mammography, were scanned using a micro-CT scanner at a resolution of 9 μm. Several preprocessing techniques were applied on 3504 extracted microcalcifications. A high amount of radiomic features were extracted in an attempt to capture differences among microcalcifications occurring in benign and malignant lesions. Machine learning algorithms were used to diagnose: (a) individual microcalcifications, (b) samples. For the samples, several methodologies to combine individual microcalcification results into sample results were evaluated.ResultsWe could classify individual microcalcifications with 77.32% accuracy, 61.15% sensitivity and 89.76% specificity. At the sample level diagnosis, we achieved an accuracy of 84.04%, sensitivity of 86.27% and specificity of 81.39%.ConclusionsBy studying microcalcifications’ characteristics at a level of details beyond what is currently possible by using conventional breast imaging modalities, our classification results demonstrated a strong association between breast microcalcifications and malignancies. Microcalcification’s texture features extracted in transform domains, have higher discriminating power to classify benign/malignant individual microcalcifications and samples compared to pure shape-features.

Artificial Intelligence Development for Detecting Microcalcification within Mammography

Background: Artificial Intelligence (AI) is the recently advanced technology in machine learning that is increasingly used to help radiologists, especially when working in arduous conditions. Microsoft Corporation offered a free-trial service called Custom Vision to develop AI for images. Objective: To study the possibility of AI development from free-trial service for detecting microcalcification within mammography. Materials and Methods: Radiological images of breast cancer-proven patients who underwent mammography between 2018 and 2019 were used to train AI to detect microcalcification. The training processes were divided into five iterations of 30, 60, 100, 130, and 160 lesion datasets. After each training, the AI was tested as “Performance Per Tag” and clinical performance. There were three types of training, quick, 1-hour, and 2-hour trainings. Results: The present study included 116 microcalcification images with 206 lesions from 56 breast cancer patients. The 160-tag iteration presented the best performance with a precision of 80.0%, a recall of 12.5%, a mean average precision of 30.5%, and a prediction rate of 32.14%. The performance of the 1-hour training was better than the quick training but was not different from the 2-hour training. Conclusion: Health personnel can easily develop AI for the detection of microcalcification in mammography. However, the AI development is further required, and the result should be interpreted along with radiologist. Keywords: AI, Microcalcification, Mammography, Machine learning

18F]-sodium fluoride autoradiography imaging of nephrocalcinosis in donor kidneys and explanted kidney allografts

Nephrocalcinosis is present in up to 43% of kidney allograft biopsies at one-year after transplantation and is associated with inferior graft function and poor graft survival. We studied [18F]-sodium fluoride ([18F]-NaF) imaging of microcalcifications in donor kidneys (n = 7) and explanted kidney allografts (n = 13). Three µm paraffin-embedded serial sections were used for histological evaluation of calcification (Alizarin Red; Von Kossa staining) and ex-vivo [18F]-NaF autoradiography. The images were fused to evaluate if microcalcification areas corresponded with [18F]-NaF uptake areas. Based on histological analyses, tubulointerstitial and glomerular microcalcifications were present in 19/20 and 7/20 samples, respectively. Using autoradiography, [18F]-NaF uptake was found in 19/20 samples, with significantly more tracer activity in kidney allograft compared to deceased donor kidney samples (p = 0.019). Alizarin Red staining of active microcalcifications demonstrated good correlation (Spearman’s rho of 0.81, p < 0.001) and Von Kossa staining of consolidated calcifications demonstrated significant but weak correlation (0.62, p = 0.003) with [18F]-NaF activity. This correlation between ex-vivo [18F]-NaF uptake and histology-proven microcalcifications, is the first step towards an imaging method to identify microcalcifications in active nephrocalcinosis. This may lead to better understanding of the etiology of microcalcifications and its impact on kidney transplant function.

Coherent anti-Stokes Raman scattering imaging of microcalcifications associated with breast cancer.

Chemical imaging of calcifications was demonstrated in the depth of a tissue. Using long wavelength excitation, broadband coherent anti-Stokes Raman scattering and hierarchical cluster analysis, imaging and chemical analysis were performed 2 mm below the skin level in a model system. Applications to breast cancer diagnostics and imaging are discussed together with the methods to further extend the depth and improve the spatial resolution of chemical imaging.

18Fluoride-based molecular imaging of coronary atherosclerosis in HIV infected patients

Background and aimsMolecular imaging with 18Fluorodeoxyglucose (FDG) and 18F-sodium-fluoride (NaF) captures arterial inflammation and micro-calcification and can reveal potentially unstable atherosclerotic plaques. MethodsWe performed FDG and NaF PET/CT imaging in two clinically similar cohorts of patients living with HIV (PLWH) with no symptomatic cardiovascular disease. The prevalence and intensity of coronary artery uptake of each tracer, measured as target-to-background ratio (TBR), were assessed in patients at low and high cardiovascular risk. ResultsNinety-three PLWH were submitted to PET/CT imaging with FDG (N = 43) and NaF (N = 50); 42% were at low and 58% at high cardiovascular risk. The intensity of uptake and multivessel coronary artery uptake were significantly higher with NaF than FDG both in low and high-risk patients. When each 18F-tracer was tested in low and high-risk patients, an equal proportion of subjects showed no vessel, single and multivessel NaF uptake; the same was true for no and single vessel uptake of FDG (no multivessel FDG uptake was noted). Waist circumference, CRP, D-dimer, HIV duration and treatment with nucleoside reverse transcriptase inhibitors were associated with high NaF uptake in univariable analyses; D-dimer remained significant in multivariable analyses (OR = 1.05; p=0.02). There were no significant associations with FDG uptake. ConclusionsThe prevalence of coronary artery uptake was higher with NaF compared to FDG both in high and low risk patients, hence microcalcification imaging may be a more sensitive tool to detect coronary atherosclerosis than inflammation imaging. However, the uptake of each 18Fluoride tracer was similar between low and high-risk subjects, and this underscores the discordance between clinical and imaging based risk assessment. Future investigation should address the prognostic significance of NaF coronary artery uptake.

Compressibility and Density Weighting for Ultrasound Scattering Tomography.

A novel reconstruction technique based on delay-and-sum ultrasound tomography algorithms is introduced. This reconstruction technique enables ultrasound tomographic systems to produce compressibility- and density-weighted images (D-WI) for spherically symmetric ultrasound scatterers. The efficacy of this reconstruction technique was demonstrated with simulation and phantom experiments. Separation between a dense wire target and a compressible thread target within a phantom was quantified, showing an increased signal of the wire target for D-WIs. The suppression of background scatter in D-WIs was also quantified. Resolution was calculated for these reconstruction techniques, exemplifying the half-wavelength diffraction-limited resolution capabilities of this ultrasound scattering tomography system. The proposed technique offers a enhanced minimum-detectable density-contrast sensitivity compared to traditional B-scan imaging for a 100- target. This enhanced detectability is expected to prove advantageous for microcalcification imaging.

Abstract P2-12-09: histopathologic correlation of residual tumor in breast with residual mammographic microcalcification and magnetic resonance imaging after neoadjuvant systemic treatment

Abstract Introduction. In patients with advanced breast cancer, neoadjuvant chemo-therapy is performed to increase breast conserving surgery. Although MRI is known to be accurate in predicting residual cancer after treatment, if calcification remains, the issue of whether to perform surgery based on the residual tumor prediction range in mammography and MRI has not yet been solved. The objective of this study was to estimate the accuracy of residual mammographic (MMG) microcalcification and enhancing lesion in magnetic resonance imaging (MRI) in predicting residual tumor after neoadjuvant systemic treatment (NST). Method. This is a single-center, retrospective study. We included patients with breast cancer who underwent NST and have microcalcifications in the post NST mammogram and had the surgery from January 2, 2013 to December 30, 2014. All the patients had post NST imaging exams of MMG and MRI. Final pathologic tumor size with histopathology and biomarker status was obtained after surgery. Analysis of correlation between image findings and pathology was evaluated. Result. Of 151 patients that were included in this study, 125 patients (82.8%) had residual invasive tumor and 26 patients (17.2%) had pathologic complete response. In overall, MRI correlated better than MMG in predicting tumor size (intraclass correlation coefficient [ICC] = 0.769 vs 0.651), but for HR+/HER2- subtype, MMG had higher correlation than MRI (ICC = 0.747 vs 0.575). Specially in HR- subtype, MRI had strong correlation with pathology (ICC for HR-/HER2+ = 0.939 and TN = 0.75), while MMG tend to overestimate the tumor size (ICC for HR-/HER2+ = 0.543 and TN = 0.479). Discussion. Overall post-NST residual microcalcifications on MMG have lower correlation with residual tumor size than MRI. In other than HR+/HER2- subtype, the extent of calcifications on pre-OP evaluation may not be accurate in evaluating the residual extent of the tumor after NST. Citation Format: Um E. histopathologic correlation of residual tumor in breast with residual mammographic microcalcification and magnetic resonance imaging after neoadjuvant systemic treatment [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P2-12-09.

Rapid Differential Diagnosis of Breast Microcalcification Using Targeted Near-Infrared Fluorophores.

Early detection and differential diagnosis of breast microcalcifications are of significant importance in effective treatment of early breast cancer, because mineral composition of breast calcification is directly associated with different pathological states. However, applying image-based modalities for component identification in breast calcification remains challenging, because no calcification-specific contrast agent is available to distinguish between benign and malignant (type I and type II, respectively) calcifications of breast lesions. In this study, real-time near-infrared (NIR) fluorescence imaging of breast microcalcifications using targeted NIR fluorophores in combination with dual-channel NIR fluorescence imaging system is reported. This strategy can be used to solve major problem in mammography and ultrasonography methods for the differentiation of benign and malignant microcalcifications. Thus, this novel technology shows significant potential for breast cancer diagnosis and image-guided surgery performed with increased precision and efficiency by providing differential diagnosis of breast microcalcifications.

Improved Diagnostics by Assessing the Micromorphology of Breast Calcifications via X-Ray Dark-Field Radiography.

Breast microcalcifications play an essential role in the detection and evaluation of early breast cancer in clinical diagnostics. However, in digital mammography, microcalcifications are merely graded with respect to their global appearance within the mammogram, while their interior microstructure remains spatially unresolved and therefore not considered in cancer risk stratification. In this article, we exploit the sub-pixel resolution sensitivity of X-ray dark-field contrast for clinical microcalcification assessment. We demonstrate that the micromorphology, rather than chemical composition of microcalcification clusters (as hypothesised by recent literature), determines their absorption and small-angle scattering characteristics. We show that a quantitative classification of the inherent microstructure as ultra-fine, fine, pleomorphic and coarse textured is possible. Insights underlying the micromorphological nature of breast calcifications are verified by comprehensive high-resolution micro-CT measurements. We test the determined microtexture of microcalcifications as an indicator for malignancy and demonstrate its potential to improve breast cancer diagnosis, by providing a non-invasive tool for sub-resolution microcalcification assessment. Our results indicate that dark-field imaging of microcalcifications may enhance the diagnostic validity of current microcalcification analysis and reduce the number of invasive procedures.

Technical feasibility proof for high-resolution low-dose photon-counting CT of the breast.

• Photon-counting detector technology allows for spatial resolution better than 100 μm. • pcBCT allows for dose levels in the screening mammography range. • pcBCT provides the highest quality imaging of microcalcifications.

WE-FG-207A-03: Low-Dose Cone-Beam Breast CT: Physics and Technology Development

WE-FG-207A-03: Low-Dose Cone-Beam Breast CT: Physics and Technology Development

WE-FG-207A-02: Why We Need Breast CT? - Clinical Perspective

WE-FG-207A-02: Why We Need Breast CT? - Clinical Perspective

WE-FG-207A-04: Performance Characteristics of Photon-Counting Breast CT

WE-FG-207A-04: Performance Characteristics of Photon-Counting Breast CT

WE-FG-207A-00: Advances in Dedicated Breast CT

WE-FG-207A-00: Advances in Dedicated Breast CT

WE-FG-207A-01: Introduction to Dedicated Breast CT - Early Studies

WE-FG-207A-01: Introduction to Dedicated Breast CT - Early Studies

WE-FG-207A-05: Dedicated Breast CT as a Diagnostic Imaging Tool: Physics and Clinical Feasibility

WE-FG-207A-05: Dedicated Breast CT as a Diagnostic Imaging Tool: Physics and Clinical Feasibility