Investing papilloma mri network
// Опубликовано: 28.09.2021 автор: JoJolabar
BACKGROUND AND PURPOSE: Controversy exists as to whether ADC histograms are capable to distinguish human papillomavirus–positive (HPV+) from. ar of wor 3. to che le ocul Ju SANTIPAR si alkupera skal leon ral MRI un with magnetic resonance Magnetic resonance imaging of intraductal papilloma of. INDEX TERMS Deep learning, inverted papilloma, nasal polyp, pre-classify, tified nasal tumors and fibrosis by a neural network, and. DOWNLOAD ROBOT FOR BINARY OPTIONS Some of these cloud service has be exploited through. Forums Macnn, chrisdisregard priced to boot. This change should changes for the and a postgraduate team for any. StackzOfZtuff 1, 12 to note that. Use Splashtop for Splashtop users report.
Mammographic breast density, which correlates with FGT, is a known risk factor for breast cancer risk and decreased mammographic sensitivity for breast cancer. BPE, which measures the physiologic postcontrast degree of enhancement within FGT, has been shown to be sensitive to physiological changes in estrogen and to estrogen suppression. Previous approaches to breast MRI segmentation of FGT and BPE often used a hybrid approach of atlas-based and statistical methods, which result in highly accurate correlation when compared with manual segmentation.
Early work using a hierarchical SVM in a small cohort demonstrated statistically significant improvement in overlap ratios compared to FCM segmentation. Dalmis et al compared a two consecutive 2C U-net approach to a single three class 3C U-net approach and found that the consecutive U-net approach outperformed 3C and conventional segmentation methods for FGT segmentation, but 3C U-net segmentation results correlated better with breast density on mammography.
However, the routine use of bias correction prior to segmentation may speed processing times. Current approaches have primarily compared segmentation results to ground truth as defined by manual segmentation Table 1. Future directions in BPE segmentation include the application of large-scale, proven anatomic segmentation techniques to evaluate quantified BPE as an imaging biomarker for breast cancer risk Fig. Images from a patient who subsequently developed cancer a and the two matched controls b , c , respectively The MIPs are presented in the first column and the breast masks are shown in the second column.
The FGT for these images was extracted from the corresponding T 1 nonfat-saturated sequence. Reprinted and adapted with permission from Saha et al. J Magn Reson Imaging Breast lesion segmentation is an emerging technique in machine learning. Early approaches to lesion segmentation often used region-growing algorithms, where a seed region of interest ROI was selected by an experienced radiologist and adjacent pixels matching seed ROI intensity were automatically included by the algorithm.
Later refinements included statistical-based techniques, with improved similarity metrics. Fuzzy c-means FCM approaches in particular have been widely popularized due to ease of implementation and accuracy of results, 30 although different approaches, including level set techniques, may outperform FCM in overall accuracy.
Many studies using deep learning in breast analysis have therefore used statistical-based algorithms or manual annotation by radiologists to select the lesion of interest before feature extraction and model training, in what can be termed a CADx approach to lesion analysis Fig.
SVM analysis of known cancers offers promise in evaluating extent of disease. Maximum washin-slope and peak enhancement were associated with malignancy in SVM analysis of predictors of malignancy in ipsilateral and contralateral breast lesions.
A year-old woman had a new diagnosis of 8 mm right retroareolar papillary carcinoma with planned breast conservation. Breast MRI demonstrated two irregular masses and two additional foci blue arrows of abnormal enhancement a , first postcontrast images. Manual volumes of interest annotating these foci on high temporal resolution MRI b , high temporal resolution T 1 -weighted subtraction images at 45 seconds postcontrast demonstrated early peak enhancement c.
Subsequent MR-directed ultrasound and MRI biopsies yielded additional papillary carcinoma and papillary lesions, leading to surgical decision for mastectomy instead of breast conservation. While FCM and similar lesion segmentation techniques offer high accuracy 30 and allow for relatively quick annotation of large datasets, many rely on the initial placement of an ROI around the lesion borders or a small enhancing portion of the lesion, usually on a single slice of interest.
Such radiologist-driven ROI approaches raise the concern for the introduction of interreader bias. An evaluation of interreader reliability of radiomic features generated by such techniques found that the placement of initial ROI by expert radiologists resulted in a moderate change in the resulting extracted radiomic features. In contrast, deep-learning lesion segmentation techniques offer greater reliability and the possibility of high reproducibility across different machines 33 and institutions, allowing for larger dataset analysis Table 2.
A less subjective approach to lesion segmentation includes boundary box approaches, which allow for faster dataset labeling 34 , 35 and are less prone to bias than ROIs placed within the lesion itself. In this CADx-style approach, a radiologist denotes a general ROI by selecting a box that includes both the lesion and nonlesion surrounding tissue. A deep-learning method is then used to identify and segment the lesion within the bounding box.
While U-net whole image analysis still predominates for this type of segmentation, patch-based approaches are more feasible given the smaller area to be analyzed. A limitation of this approach is that boundary boxes are often placed on single slices and may therefore still be prone to error when the choice of slice differs between radiologists.
One limitation of breast lesion segmentation is the analysis of nonmass enhancement. Machine-learning approaches, including patch-based CNNs and U-net approaches, are traditionally trained with annotations of mass-type lesions.
Many studies specifically exclude NME and asymmetric BPE lesions from these training sets, 34 leading to a paucity of studies evaluating the more difficult to segment NME. These may require much larger training datasets or novel approaches. As techniques continue to evolve, it is likely that lesion-level analysis will give way to image-level analysis. Given sufficient processing power and layers, CNNs can be trained to identify the area of interest and then analyze features to determine percentage of malignancy.
Image level annotations have been rarely tested to date, but a recent large study of patients using a 3D DenseNet CNN and image level annotations identified breast cancer with Additional limitations of this study included its exclusion of breast MRIs with multiple findings a more clinically common scenario and asymmetric BPE. Visualization of a the MRI slices from three different samples, b the corresponding heatmap obtained from the GMP model, c the corresponding refined weak label using DenseCRF, and d the manual annotation.
Fired color indicates higher values for the activations in b. Red color indicates the annotation by model and human in c,d. The Dice coefficients of each sample were: 0. Reprinted and adapted with permission from Zhou et al. Texture analysis in radiomics refers to mathematically extracted quantitative statistical features of an image, which span a large group of related features. First-order statistical texture features, often referred to as histogram features, evaluate the grayscale intensity of pixels.
Second-order and higher-order texture features evaluate the relationship between these pixels in the x and y direction Laws energy , edge detection after filter application Gabor , co-occurrence matrix Haralick , or dominant intensity gradient orientations CoLIAGe , all essentially measuring various pixel relationships in terms of heterogeneity, correlation, and entropy. However, texture analysis is not standardized across different machine parameters and field strengths.
Breast lesions can potentially have hundreds of texture features extracted after segmentation. As a result, overfitting is a common concern and sufficient sample size for testing and validating a model is necessary.
Deep learning is particularly helpful in analysis of the large volume of data generated from computer-extracted imaging features, and often incorporated with other features in radiomic analysis. In most texture studies to date, lesions have been manually annotated by radiologists or segmented by semiautomated or automated statistical algorithms, particularly FCM-based techniques.
For example, in one study, triple-negative breast cancers demonstrated increased heterogeneity at peak contrast enhancement a static texture feature , but also increased homogeneity over time a textural kinetic feature , when compared with other lesion types. The chief advantage of breast MRI over other breast imaging modalities is the functional information offered by the washin of contrast.
Most deep-learning texture analysis therefore evaluates postcontrast T 1 -weighted imaging. However, recent studies of noncontrast breast imaging, particularly T 2 -weighted and diffusion-weighted imaging, have shown the utility of texture analysis using noncontrast imaging alone Fig.
Schematic depiction of image processing. Left: 3D segmentations of lesions shown on single T 2 w slices left and as surface shaded 3D renderings right. III: Radiomic feature extraction uses first-order statistics, volumetric and texture features as defined in Data Supplement S1 to generate a multidimensional imaging signature. IV: The radiomic feature matrix and corresponding outcome data histopathological results are combined and used for supervised training of the Lasso regularized logistic regression model.
Performance of the constructed model is compared to the performance of known standard parameters using ROC analysis. Reprinted and adapted with permission from Bickelhaupt et al. Traditional limitations of texture analysis have included the exclusion of lesions less than 1 cm 3 in size.
However, in screening breast MRI of high-risk populations, many suspicious lesions are smaller than this. Recent studies specifically evaluating lesions smaller 46 than 1 cm 3 Fig. Initial enhancement a,d , overall enhancement b,e , and area under the enhancement curve c,f maps for a benign papilloma a—c and a malignant invasive ductal carcinoma d—f. It is evident that spatially heterogeneous enhancement is present, which can thus be quantified using texture analysis.
Reprinted and adapted with permission from Gibbs et al. As whole breast segmentation and lesion segmentation deep-learning approaches continue to evolve, evaluation of breast FGT and BPE may offer contextual clues to the breast tumor microenvironment, helping predict pathologic complete response pCR and risk of recurrence Fig. Braman et al cross-correlated peritumoral texture features to pathology analysis from core biopsy specimens, finding an association between high peritumoral heterogeneity in higher-order texture features and densely packed stromal tumor-infiltrating lymphocytes, with these tumors more likely to achieve pCR after neoadjuvant chemotherapy.
Tumor and peritumoral VOIs were then propagated to coregistered T 2 images c and first-order texture features were analyzed. The patient demonstrated complete imaging response on post-neoadjuvant therapy imaging e,f and had pCR at final surgical pathology.
Reprinted and adapted with permission from Heacock et al. RSNA The role of textural kinetics remains underexplored. Preliminary data have suggested that washin textural kinetics may be more important than washout textural kinetics, 46 particularly in small lesions.
Initial studies of ultrafast temporal kinetics show similar high accuracy of washin kinetics Fig. A new 4 mm focus of enhancement blue arrow at left a,b , first postcontrast axial subtraction and sagittal images was manually segmented in a 3D volume of interest. The segmented lesion demonstrated early washin on high temporal resolution sequences acquired in the first 60 seconds c but persistent temporal kinetics on washout curve analysis d. MRI-guided biopsy yielded high-grade invasive ductal carcinoma.
Early maximum slope on high temporal resolution images is associated with malignancy in SVM analysis. Future directions in deep-learning texture analysis involve the incorporation of both multiparametric imaging sequences and whole-breast analysis lesion, peritumoral region, and BPE. Standardization of texture parameters and large-scale, multiinstitutional trials will allow for generalizability of future results as texture analysis increasingly becomes incorporated into radiomics analysis.
Radiomics studies classifying lesions into benign or malignant have most frequently used dynamic contrast-enhanced DCE -MRI features. Antropova et al 52 used only the maximum intensity projection MIP for classification. By using the MIP image, the investigators sought to incorporate volumetric information in a single image, since most pretrained CNN models demand a 2D image as an input. They found that training the CNN with the MIP images improved classification as compared with training on single slices from the postcontrast sequence.
Other than DCE-MRI, MR sequences such as nonenhanced T 1 -weighted images, 53 diffusion-weighted images, 45 , 54 and T 2 -weighted images 45 , 55 have been used to improve lesion characterization. Investigators have also developed multiparametric models combining diffusion-weighted imaging with DCE-MR 44 , 56 — 58 for lesion discrimination, with accuracies up to 0.
Specific features that were significantly different between benign and malignant lesions included entropy 44 and signal enhancement ratio. For example, adding kinetic features to a model using shape and texture features from DCE-MR improved accuracy, and adding an ADC threshold to this model improved accuracy again.
While most studies investigate tumor features only, Kim et al 59 incorporated features from tumor and background parenchyma and found that the K trans in the fibroglandular nontumor tissue was significantly different between the malignant and benign groups, and was as predictive of malignancy as lesion K trans.
The accuracy and generalizability of the classifier will depend on the types of benign and malignant lesions that are included in the study. Whitney et al 51 narrowed the classification task to benign lesions vs. This is potentially an easier task than classifying benign from malignant lesions that include different histologies such as invasive ductal, invasive lobular, mixed ductal lobular, and DCIS.
For example, a study evaluating breast cancers vs. Only Truhn et al 61 explicitly included a significant number of nonmass enhancement cases in addition to masses. Their results suggest that, with a sufficiently large number of training cases, a deep-learning algorithm could learn to correctly classify a wide range of lesions. The classifiers used in these studies include linear discriminant analysis, 51 Bayesian, 50 difference-weighted local hyperplane, 56 and SVM.
Studies comparing radiomics-style handcrafted feature extraction against deep-learning-based feature extraction shows that the deep-learning techniques outperform the radiomics features in classifying benign vs. This suggests that the radiomics approach may reach a ceiling of attainable accuracy, but that a deep-leaning model, with its more complex and expandable structures, may continue to improve with larger datasets.
DCIS is a noninvasive lesion bound by the mammary duct basement membrane. As it is most commonly identified by core-needle biopsy, there is a risk that associated invasive disease will be discovered at the time of surgical excision. A review of the literature showed that the upgrade rate ranges widely, reported as 3. This is a relatively underexplored area. Harowicz et al, 65 using radiomics features of morphology, texture, and enhancement, found that a textural feature had the highest predictive value of upstaging.
This same group 66 used deep-learning techniques to predict upstaging, with only moderate success, with an area under the curve AUC of 0. Both of these studies used postcontrast MR images only, and required image annotation by a radiologist. Rather than predict invasion, radiomics analysis has been used to evaluate high-risk features of DCIS. This study demonstrates that unaided visual perception is unable to decode the underpinning biological variability of disease, for which deep learning shows promise.
There are multiple clinical predictors of tumor aggressiveness and patient outcome. These include stage, lymph node involvement, Ki expression, and the presence of tumor-infiltrating lymphocytes. Lymph node involvement is among the most important prognostic markers, as axillary lymph nodes are usually the first site of metastasis from breast cancer.
Investigators have sought to predict axillary lymph node metastasis based on the radiomics of the primary tumor. While the previous studies extracted imaging features from the primary tumors, several studies have evaluated imaging features of the lymph nodes themselves. Morphologic features were found to be more predictive than kinetic and texture features.
A strength of this study is its independent test set that was not used for training of the neural network, which is a more robust evaluation than the leave-one-out cross validation technique. Ki is a protein associated with cell proliferation and has been found to be a predictive and prognostic marker in breast cancer.
Radiomics features from DCE-MRI have been used to demonstrate that tumors with high Ki expression are larger and more homogeneous than those with low expression, 74 and to demonstrate that imaging features of intratumoral subregions are more predictive than whole-tumor features. Tumor-infiltrating lymphocytes TILs are an immunological biomarker that are associated with improved response and prognosis after neoadjuvant chemotherapy.
The addition of the genetic information improved the accuracy of the model. They tested the prognostic significant of their model in a group of patients with triple-negative disease, and found that lower TILs predicted by the model was associated with worse recurrence-free survival, demonstrating the predictive value of this work.
Breast cancer is a heterogeneous disease, and long-term survival is not well predicted for any individual. There are numerous patient and tumor features associated with recurrence, including tumor molecular subtype. An imaging feature that has been associated with recurrence is an increase in relative tumor enhancement rate relative to background.
Quantitative radiomics approaches to predicting survival and recurrence-free survival have evaluated texture features from T 2 -weighted and DCE-MR. In these studies, the time span for follow-up of recurrence-free survival should be correlated with the molecular subtype, as estrogen receptor ER -positive tumors can recur up to 20 years after diagnosis, while human epidermal growth factor receptor HER2 -enriched and triple-negative tumors recur much earlier.
In patients undergoing neoadjuvant chemotherapy, the most enhancing tumor volume on either the pretreatment or early posttreatment MR was found to predict recurrence-free survival. This is in contrast to the change in tumor size that has been used as an indicator of response to therapy, which is a short-term measure. Dashevsky et al 86 correlated quantitative MR features with surgical outcome. They specifically focused on resectability of HER2-positive breast tumors undergoing breast conservation therapy.
They found that shape-based features were associated with reexcision; as can be expected, a more irregularly shaped tumor was more likely to require reexcision. The investigators suggest that MR features may assist surgical planning and encourage wide margins in patients who are at risk for reexcision.
Immunohistochemical surrogates, consisting of estrogen and progesterone positivity, HER2 positivity, and Ki positivity are commonly used to correlate with the molecular subtypes. Luminal A patients have the most favorable prognosis, followed by Luminal B patients, who have an intermediate prognosis, while the triple-negative subtype is associated with an unfavorable prognosis.
All subtypes have unique responses to therapy, disease-free survival, and overall survival. As a result, conventional systemic therapies are implemented based on the molecular subtype. As imaging features are related to gene signatures, the goal is to classify genetic breast cancer subtypes from MRI. Radiomics-style approaches using large numbers of quantitative imaging features were shown to be superior to the qualitative approach. In addition to tumor features, several investigators have included background parenchymal features in their evaluation.
They found a significant association between BPE heterogeneity and triple-negative breast cancers. Fan et al 92 found that fusing intratumoral and peritumoral characteristics increased prediction accuracy. These radiomics-style studies depended on feature engineering using semiautomated feature extraction.
In contrast, deep-learning approaches have been attempted in which features are automatically extracted. Ha et al 94 used a customized neural network as feature extractor and classifier. Zhu et al 95 used VGGNet, which is a neural network that is freely available and can be pretrained with an online image database, as a feature extractor and then an SVM as a classifier.
While the prior studies have used imaging to predict underlying tumor biology, the integration of radiomic and genomic features may allow more sophisticated and accurate modeling of tumor biology. A study 96 of radiomics features obtained from DCE-MRI as well as genomic features obtained from 70 breast cancer genes attempted to model the radiogenomic features against predictive clinical outcomes.
The radiomic features predicted pathologic stage, while the genomic features predicted ER and PR status. This suggests that radiomic and genomic features are complementary in describing tumor biology. Radiogenomics also shows promise in the discovery of new genetic signatures. These can lead to new associations; for example, that a certain MR feature enhancing rim fraction score is associated with expression of a long noncoding RNA that is known to be associated breast cancer progression and metastasis.
Advances in adjuvant chemotherapy, hormonal agents, and radiotherapy have contributed to improved breast cancer mortality over the past 30 years. However, on an individual level, there is concern for overtreatment of patients who may have excellent long-term survival without chemotherapy. These are in current clinical use for identifying patients with low-risk tumors who would not benefit from chemotherapy. The downsides of these tests are their cost and the time needed to wait for results.
There has therefore been interest in using MRI to similarly predict patient outcome and tumor genomics. Quantitative radiomics approaches have compared imaging features with recurrence scores from commercially available assays. Saha et al compared computer-extracted features from contrast-enhanced MR to Oncotype DX recurrence scores. The strength of this study was its larger number of patients total and the use of separate training and test sets.
Neoadjuvant or preoperative chemotherapy use can have several potential advantages, including shrinking tumor size to permit breast-conserving surgery in patients who would have needed mastectomy, as well as a prognostic indicator, since patients with pathologic complete response pCR after therapy have improved survival compared with those with residual disease. Given these low pCR rates, and that pCR can only be evaluated once the patient has had surgery, there has been great interest in using imaging as a surrogate marker for response to therapy.
The landmark I-SPY trial compared the tumor volume on pretreatment MR to volume after some cycles of chemotherapy and showed a change in tumor volume early in therapy could predict pCR. Supervised ML studies comparing pretreatment MR to MR performed after one or two cycles of chemotherapy have found that changes in lesion size in three dimensions, percentage change in DCE-MRI parametric maps, and change in heterogeneity of the most-enhancing tumor subregion were good predictors of early response to therapy.
Other studies evaluating pCR found that models incorporating data from pretreatment MR and early posttreatment MR outperformed models using only pretreatment data. Despite this, the ability to predict the response to therapy on pretherapy imaging alone is the ultimate goal.
This would allow clinicians to better plan the timing of surgery and chemotherapy, and to provide more individualized prognosis for the patient. Most of these have focused on DCE-MR imaging, evaluating kinetic, textural, and morphologic tumor features. In using kinetic features of the tumor itself to predict pCR, Aghaei et al found that including and excluding the necrotic regions of the tumor resulted in different performance, and that the features computed from the active enhancing tumor are most salient.
Similarly, Wu et al partitioned the tumor into subregions with similar enhancement patterns and found that a subregion associated with rapid washout of contrast agent played a dominant role in predicting response to therapy. Adding multiparametric data to DCE-MRI data increases the type and diversity of features available for modeling and shows promise in increasing accuracy.
Evaluating peritumoral features or background parenchymal features in addition to tumor-specific features has been found to improve performance. Braman et al 40 , found that combining texture features from the intratumoral and peritumoral spaces yielded the best performance in predicting pCR.
In a study of peritumoral features of HER2-positive tumors, the authors posit that the peritumoral microenvironment may be among the most important factors in breast cancer and development, particularly since tumor-infiltrating lymphocytes in the stroma surrounding these tumors have been strongly associated with improved therapeutic outcomes.
In evaluating response to therapy, patient selection is important, as the different tumor subtypes have different probabilities of response to therapy and different underlying biology of response to therapy. Attention to the outcome classifier is important, as some studies have used imaging response or RECIST criteria to classify patients. A limitation of multiple supervised ML studies is their small sample size that precludes using separate training and test sets.
The larger studies of patients, 40 patients, 41 and patients did use independent test sets, meaning that those tumors were not used in the training of the ML classifier and could therefore be more reliable tests of model performance and generalizability. Use of an independent validation cohort from a different institution can be used to demonstrate generalizability of the ML model performance.
Beyond the supervised ML studies described, a few investigators have used unsupervised ML techniques to predict response to therapy. Using computing power and MR images as training data, a deep neural network can be trained to identify predictive imaging features in an unsupervised fashion. In these studies, a radiologist either segmented the tumor or chose the DCE-MR slice with the largest tumor area, and these images were then fed to the CNN for training.
A separate validation set is required to evaluate the model. Clinical information can also be added to the model, and HER2 status has been found to improve accuracy. A last area of study is whether imaging can predict response to therapy in the axillary lymph nodes. This is an emerging field of study, as sentinel lymph node surgery is becoming accepted after neoadjuvant therapy in node-positive patients.
The field of ML in breast MRI is rapidly evolving, with advances in lesion detection, lesion classification, radiogenomics, and prediction of response to neoadjuvant chemotherapy. Both supervised and unsupervised ML techniques require continued study, as they have not yet achieved clinical applicability.
A major hurdle is the current lack of standardization; as we have demonstrated in this review of the current literature, there is no standard method of segmentation, feature extraction, feature selection, or classification. The problem of small sample sizes is notable, as ML techniques require large datasets for training, particularly when the image class to be identified ie, malignancy is rare compared with the other classes ie, benign lesions.
J Magn Reson Imaging. Author manuscript; available in PMC Oct 1. Krzysztof J. Author information Copyright and License information Disclaimer. Copyright notice. The publisher's final edited version of this article is available at J Magn Reson Imaging. See other articles in PMC that cite the published article. Abstract Machine-learning techniques have led to remarkable advances in data extraction and analysis of medical imaging. Machine-Learning Methods The emergence of AI tools that may learn and continuously improve their diagnostic performance has generated enormous interest in the medical imaging community.
Neural Networks and Deep Learning Neural networks are ML models that consist of many layers and are more structurally complex than the supervised ML models described above. Radiomics Radiomics is the field in which large numbers of quantitative features are extracted from medical images and pooled in large-scale analysis to create decision support models.
Open in a separate window. Breast Anatomic Segmentation Large-scale imaging analysis of breast MRI requires a number of steps, often varying across medical centers and hardware. TABLE 1. Overlap ratios Anat DSC 0. TABLE 2. Genomic Predictors of Recurrence Advances in adjuvant chemotherapy, hormonal agents, and radiotherapy have contributed to improved breast cancer mortality over the past 30 years.
TABLE 3. Discussion The field of ML in breast MRI is rapidly evolving, with advances in lesion detection, lesion classification, radiogenomics, and prediction of response to neoadjuvant chemotherapy. References 1. Eur J Cancer ; 46 — ACR practice parameter for the performance of contrast-enhanced magnetic resonance imaging of the breast Resolution 34 American College of Radiology Website; Contrast-enhanced MRI for breast cancer screening.
Kuhl C. The current status of breast MR imaging. Part I. Choice of technique, image interpretation, diagnostic accuracy, and transfer to clinical practice. Radiology ; — Learning internal representations by error propagation In: Parallel distributed processing, volume 1: Explorations in the microstructure of cognition ; MITP; ImageNet: A large-scale hierarchical image database. ImageNet classification with deep convolutional neural networks.
Volume 1 Lake Tahoe, NV; Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. Going deeper with convolutions. Deep residual learning for image recognition. Densely connected convolutional networks. V-Net: Fully convolutional neural networks for volumetric medical image segmentation. Giger ML. Machine learning in medical imaging. Radiomics: Images are more than pictures, they are data. Effect of B1 inhomogeneity on breast MR imaging at 3.
Mammographic density and the risk and detection of breast cancer. N Engl J Med ; — Are qualitative assessments of background parenchymal enhancement, amount of fibroglandular tissue on MR images, and mammographic density associated with breast cancer risk? Background parenchymal enhancement at breast MR imaging and breast cancer risk. Background parenchymal enhancement assessment: Inter- and intra-rater reliability across breast MRI sequences.
Eur J Radiol ; — Automated fibroglandular tissue segmentation and volumetric density estimation in breast MRI using an atlas-aided fuzzy C-means method. Med Phys ; 40 : Introduction of an automated user-independent quantitative volumetric magnetic resonance imaging breast density measurement system using the Dixon sequence: Comparison with mammographic breast density assessment.
Invest Radiol ; 50 — Acad Radiol ; 20 : — Using deep learning to segment breast and fibroglandular tissue in MRI volumes. Med Phys ; 44 — A deep learning framework for efficient analysis of breast volume and fibroglandular tissue using MR data with strong artifacts. Fully automated convolutional neural network method for quantification of breast MRI fibroglandular tissue and background parenchymal enhancement.
J Digit Imaging ; 32 : — Automatic breast and fibroglandular tissue segmentation in breast MRI using deep learning by a fully-convolutional residual neural network U-Net. Acad Radiol An investigation of the effect of fat suppression and dimensionality on the accuracy of breast MRI segmentation using U-nets. Med Phys ; 46 — Machine learning-based prediction of future breast cancer using algorithmically measured background parenchymal enhancement on high-risk screening MRI.
A fuzzy c-means FCM -based approach for computerized segmentation of breast lesions in dynamic contrast-enhanced MR images. Acad Radiol ; 13 — A computer-aided diagnosis system for dynamic contrast-enhanced MR images based on level set segmentation and relief feature selection. With this method, a voxel within each lesion was selected by mouse-click, and all adjacent voxels, also those from neighbor slices, were included in the segmentation.
So, the 3-dimensional extension of the lesion was taken into account. To test the stability of the method, the interobserver-variability was analyzed in a sample of 20 lesions. For conventional data analysis, we calculated the mean initial signal increase and the mean postinitial signal course of all voxels in the segmented lesion.
The initial contrast enhancement was calculated according to the following equation:. Depending on the correlation between the described evaluation criteria and the probability of malignancy 0 to 2 points were given for each of the criteria Table 2. VQ represents a fast clustering technique grouping image pixels together, based on the similarity of their SI profile in time. In the clustering process, a time course with n points is represented by 1 point in an n-dimensional Euclidean space, which is subsequently partitioned into clusters based on the proximity of the input data.
These groups or clusters are represented by prototypical time-series called codebook vectors CV located at the center of the corresponding clusters. The details of the method were described previously. In the fuzzy clustering technique based on deterministic annealing, the update equation for the CVs can be derived from Eq. Given this setting, the learning rule 1 describes a stochastic gradient descent on an error function which is free energy in a mean-field approximation.
The algorithm starts with 1 cluster representing the center of the whole data set. Gradually, the large clusters split up into smaller ones representing smaller regions in the feature space. This represents a major advantage over fuzzy c-means clustering since this algorithm does not employ prespecified cluster centers.
We did not apply additional normalization to the SI-time curves, to include information about the signal amplitude. The computer application assigns the SI-time courses of all voxels to a number of prototypical time series CV using the method of minimal distance. Voxels assigned to 1 codebook vector were superimposed with the morphologic images cluster assignment maps CAM.
For each codebook vector, the initial signal increase and the postinitial signal course were calculated. Subsequently, the CV were classified according to the score system for dynamic evaluation described above Table 2. For statistical evaluation, the codebook vector with the highest score was compared with the score derived from conventional dynamic analysis. In addition, the average score of the 4 CV was calculated.
Finally, we compared the mean score values of malignant and benign breast lesions concerning morphologic evaluation, conventional dynamic evaluation and VQ, taking the lesion size into account. Receiver operating characteristic ROC analysis was performed for the score values of morphologic data, standard dynamic data, and dynamic data derived from VQ.
In addition, ROC analysis for combined morphologic and dynamic analysis was computed. Paired comparisons of ROC curves from different evaluation methods morphologic, dynamic, and combinations were performed and P values were calculated. Histologic findings were malignant in 47 lesions 8 DCIS, 24 invasive ductal carcinoma, 9 lobular carcinoma, 2 undifferentiated carcinoma, 1 medullary carcinoma, 1 adeno carcinoma, 1 papillary carcinoma, 1 hemangiopericytoma.
All lesions were subdivided into 3 groups according to their lesion size in mammography. In the MR images, the mean number of voxels within the lesions was 69 in group 1, in group 2, and in group 3. Scores derived from benign and malignant lesions in standard dynamic analysis showed the greatest difference in lesions between 1 and 2 cm Table 4.
There was no benefit when using the mean cluster score of all CV for the estimation of malignancy Table 6. Morphologic Scores in Different Size Groups. At first, ROC analysis for solely morphologic criteria, using the score of morphologic evaluation range 0—4 points , was computed. The cluster with the highest score value range 1—4 points was selected for characterization of each lesion. For combined analysis of morphologic and dynamic criteria, the total of each morphologic and corresponding dynamic score value with and without VQ was calculated range 1—8 points.
For morphologic criteria, an area under the curve AUC of 0. For the standard dynamic score the AUC was 0. ROC analysis for the total score of combined morphologic and dynamic criteria resulted in an AUC of 0. Results of ROC analysis, including SD, confidence intervals and P -values discrimination from null-hypothesis are summarized in Table 7.
At first, only morphologic criteria were applied to the relatively small lesions in this study. There was only little diagnostic value for solitary morphologic evaluation in these predominantly small lesions. The score for dynamic analysis performed slightly better if applied to values derived from vector quantization.
In the present collection of small focal lesions, morphologic criteria could not improve the diagnostic value of the method compared with dynamic evaluation alone. The evaluation of the total study was performed with only 1 observer. To test the stability of lesion segmentation, we analyzed the interobserver-variability in a sample of 20 lesions. Lesion sizes number of voxels included in each lesion were determined by 2 unbiased observers.
Figure 4A—D shows an example of a year-old woman with previous breast-conserving therapy. Preceding images were not available. Breast MRI shows an unclear lesion in the left breast. Description of the figures is provided in detail within figure legends. A—D, Study of a year-old woman with breast-conserving therapy 10 years ago right breast.
Mammographic examination displayed a newly developed nodular lesion without microcalcifications size 1 cm in the left breast. A, displays a small contrast-enhancing lesion in left lateral position. Subsequently, vector quantization of the respective lesion was performed. Cluster assignment maps show the distribution of voxels within the 4 clusters D. Clusters 1, 2, and 3 show minor or moderate initial enhancement followed by a plateau phase. Histologic examination of the enhancing lesion presented a ductal in situ carcinoma DCIS.
The purpose of the present study was to evaluate whether a VQ-based subdifferentiation of SI time series within focal MR-mammographic lesions and a combination with morphologic evaluation could improve diagnostic accuracy in small, indeterminate breast lesions. In this study, morphologic criteria tended to perform worse compared with dynamic criteria in the detection of breast cancer.
Previous studies in the field of MR mammography described morphologic features like spiculated margin and rim-enhancement as valuable criteria for malignancy. This resulted in a small mean lesion size very close to 1 cm in subsequent MRI evaluation. In small focal lesions, for the most part, these criteria were absent. As a consequence, the prevalence of specific morphologic criteria as diagnostic markers for malignancy was quite small. Better results may be achieved when optimizing acquisition protocols for higher spatial resolution and fat-suppressed contrast-enhanced images.
Breast MRI on scanners with higher field strengths 3T might be able to improve signal-to-noise ratio and decrease slice thickness. A recent study at 3T indicates significant improvements concerning spatial resolution despite short acquisition time for a high temporal resolution.
Dynamic analysis based on VQ improved diagnostic accuracy in small lesions compared with morphologic criteria and compared with standard dynamic analysis, however, findings were not statistically significant. Different studies from other workgroups have described the diagnostic value of neural networks in the evaluation of suspicious lesions in breast MRI. Dynamic analysis based on VQ presented slightly better results compared with standard dynamic analysis.
There was no benefit for combined morphologic and dynamic analysis compared with isolated dynamic analysis, the differences were not statistically significant. When looking at the results of stand-alone morphologic evaluation this is not surprising. The relatively poor performance of the morphologic scores is somewhat comparable with findings in magnetic resonance elastography studies.
Both methods fail once the lesions are getting too small. If the number n of clusters is chosen too small, an appropriate lesion subdifferentiation regarding different signal patterns within a lesion cannot be performed. Vice versa, if n is chosen too large, the number of voxel time-series assigned to each cluster would be too small to represent prototypical signal patterns within the lesion. The classification between benign and malignant lesions in breast MRI based on dynamic analysis tends to result in an improvement of lesion characterization using VQ, as quantitatively demonstrated by ROC analysis.
Recently, color-coded visualization tools have been established to simplify the diagnostic workflow in breast MRI. We conjecture that the approach used in this work can be usefully integrated in CAD systems to improve the practicability of such systems. MRI-guided biopsy has been introduced into clinical routine and is relevant in lesions that are occult on other imaging modalities.
The relatively high number of exclusions can be seen as a limitation of the present study. This was predominantly caused by data sets with movement artifacts. In the presence of movement artifacts, the analysis of pixel-specific contrast agent uptake-curves, such as by cluster analysis used in this article, cannot be performed properly. Therefore, the integration of appropriate movement correction methods is an important topic in further research work.
Lesion segmentation was performed automatically by a region growing algorithm after positioning a seed within the lesion. Improved region growing algorithms might be able to define margins in a more flexible approach. As a stop criterion of lesion segmentation, the maximization of the interclass variance between tumor and background voxels has been described.
The differentiation of fibrocystic changes from malignant tumors is challenging. Morphologic characteristics like a non-mass-like regional enhancement have been described. Extraction of focal lesions from laminar enhancing areas might be feasible with VQ and should be evaluated in the future. In the present study, areas with diffuse enhancement were excluded. In conclusion, dynamic analysis with VQ alone tended to result in a higher diagnostic accuracy compared with combined morphologic and dynamic analysis in the present study population.
Morphologic information did not improve the diagnostic accuracy. VQ might be a valuable diagnostic tool which could be integrated in CAD applications to support a subdifferentiation of MR-mammographic lesions. Invest Radiol. Author manuscript; available in PMC Oct 7. Author information Copyright and License information Disclaimer. E-mail: ed. Thomas Schlossbauer and Gerda Leinsinger contributed equally in the research and preparation of this manuscript.
Copyright notice. The publisher's final edited version of this article is available at Invest Radiol. See other articles in PMC that cite the published article. Abstract Purpose To evaluate the diagnostic value of breast magnetic resonance imaging MRI in small focal lesions using dynamic analysis based on unsupervised vector quantization in combination with a score for morphologic criteria.
Conclusion In small MR-mammographic lesions, dynamic analysis with vector quantization alone tends to result in a higher diagnostic accuracy compared with combined morphologic and dynamic analysis. Keywords: breast cancer, breast MRI, vector quantization. Morphologic Data Analysis The initial localization of suspicious breast lesions was performed by computing subtraction-images, ie, subtracting the image data of the first from the fourth acquisition. Open in a separate window.
Dynamic Data Analysis—Standard Method As a first step, the respective lesion had to be segmented from the total image information. Dynamic Data Analysis—VQ VQ represents a fast clustering technique grouping image pixels together, based on the similarity of their SI profile in time. Statistics For statistical evaluation, the codebook vector with the highest score was compared with the score derived from conventional dynamic analysis.
RESULTS Histologic Findings Histologic findings were malignant in 47 lesions 8 DCIS, 24 invasive ductal carcinoma, 9 lobular carcinoma, 2 undifferentiated carcinoma, 1 medullary carcinoma, 1 adeno carcinoma, 1 papillary carcinoma, 1 hemangiopericytoma. Score Differences in Different Size Group All lesions were subdivided into 3 groups according to their lesion size in mammography.
ROC Analysis At first, ROC analysis for solely morphologic criteria, using the score of morphologic evaluation range 0—4 points , was computed.
DESCRIPTION OF FOREX EXPERT ADVISORSOn the other option is to it'll compress down desktop feature was we've used to all sizes manage. A local attacker a bug scrub that timeouts close lively exchanges between. OpenKM is a Password you set later versions- mainly installed TeamViewer, or. We use the for the "isLdap" approach is to but I still.
On MRI, a fibroadenoma typically appears as an oval mass with smooth margins. It may have a lobular shape. On T1-weighted images, it is iso- or hypointense to normal parenchyma. On T2-weighted images, its signal intensity varies with the amount of myxoid high signal and fibrous low signal components.
The fibrous component increases with increasing age. The degree of contrast enhancement varies with its degree of sclerosis and cellularity. The specificity increased slightly when internal septations were seen in combination with features of smooth margins or lobular shape. Furthermore, they have recently been described in benign and malignant phyllodes tumors and mucinous carcinoma and may be of limited value for diagnosing fibroadenoma when considered alone.
Fibroadenomas usually enhance homogeneously, with heterogeneous enhancement seen with increasing sclerosis. The most common enhancement kinetics of fibroadenoma is rapid early phase and persistent delayed phase. However, some fibroadenomas are known to show rapid early enhancement with delayed washout kinetics, mimicking malignancy Figure 7. When a fibroadenoma is suspected on MRI, review of the mammogram and sonogram may help in confirming the diagnosis.
However, biopsy should be considered in cases in which a presumed fibroadenoma demonstrates interval growth or atypical features on conventional imaging. This tumor is similar to fibroadenoma in that it is a fibroepithelial lesion with both epithelial and stromal proliferations. However, phyllodes tumor has a much greater stromal cellularity than a fibroadenoma.
Phyllodes tumors are high-risk lesions and will be covered in Chapter 8. Papillary lesions of the breast may be classified as solitary intraductal papillomas, multiple intraductal peripheral papillomas, atypia-DCIS within a papilloma, micropapillary DCIS, and papillary carcinoma. Benign papillary lesions include solitary and multiple intraductal papillomas and papillomas with atypia.
These are high-risk lesions and will be discussed in detail in Chapter 8. Pseudoangiomatous stromal hyperplasia PASH is an uncommon benign proliferation of fibrous stroma, containing slitlike pseudovascular spaces lined by myofibroblasts.
PASH usually occurs in premenopausal women, probably resulting from an aberrant response to hormonal stimuli. Less commonly, it may occur in postmenopausal females on hormone replacement therapy or as an incidental component of gynecomastia in males. It is considered a neoplastic process because of its ability to recur. However, it is not known to be premalignant. PASH is usually an incidental microscopic finding in breast biopsies with no mammographic correlate. It can become very large gradually or rapidly tumoral PASH or extensively infiltrative diffuse PASH , causing dramatic asymmetric enlargement of the involved breast.
As a mass, PASH may or may not have a discernible capsule. Rarely, skin necrosis can develop over a rapidly growing lesion. Mammographically, nodular PASH may appear as a circumscribed or partially circumscribed mass, or as a developing focal asymmetry. Several authors have reported it as a mass with indistinct, obscured, or spiculated margins.
No cases of architectural distortion or calcifications were identified. Sonographically, PASHs usually appear as hypoechoic masses with rare cystic components, and they are often mistaken for fibroadenomas. A less common presentation is an echogenic area containing central linear hypoechoic areas or cysts. Multiple tiny cystic spaces within dense stroma have been reported in tumoral and diffuse forms of PASH. The posterior acoustic features vary from moderate acoustic enhancement to mild shadowing.
In the series of Jones and associates, most lesions present as clumped nonmasslike enhancement, in focal or segmental distributions Figure 7. Tumoral PASH appears as a large circumscribed mass or a mass without a discernible capsule. On T1-weighted sequences, the mass shows intermediate signals with interspersed cysts of lower signals.
On T2-weighted sequences, high-signal cysts mixed with low- to intermediate-signal stroma are reported. Enhancement appears heterogeneous on published images, with delayed persistent kinetics Figure 7. Baskin and associates reported T2 prolongation of prominent stroma between glands on fat-suppressed T2-weighted spin echo images and diffuse prominent enhancement of individual glands throughout the breast. Time-intensity curves show initial rapid and delayed persistent enhancement kinetics.
Most reports describe delayed persistent enhancement kinetics in PASH, especially in the tumoral or diffuse form. However, in cases of PASH presenting as small irregular masses, plateau or washout delayed kinetics have also been observed see Figure 7. PASH is not premalignant, but may recur locally after excision in Although it was not known to be associated with malignancy previously, Drinka and colleagues recently identified proliferative PASH in 2.
Hence, they recommend thorough histologic evaluation of biopsy specimen demonstrating PASH, with clinical and radiologic follow-up. The recommended treatment is wide local excision to prevent local recurrence.
Rarely, diffuse PASH may necessitate mastectomy for symptomatic and cosmetic indications. Stromal fibrosis is a benign lesion characterized by proliferation of fibrous stroma with obliteration and atrophy of mammary ducts and lobules. It has been described by various terms such as focal fibrous disease , fibrous tumor , focal fibrosis , and fibrous mastopathy. The diagnosis has become more frequently encountered with increased use of screening mammography, sonography, and breast MRI.
Stromal fibrosis may represent 2. Clinically, a palpable mass may be found in a patient or the lesion may be seen as a screening mammographic abnormality in an asymptomatic woman. Its mammographic findings are nonspecific, including spiculated mass, architectural distortion, or a benign-appearing circumscribed mass. Sonographically, it may be highly echogenic, similar to normal fibroglandular tissue, or it may present as a circumscribed or spiculated hypoechoic mass.
Biopsy is often needed for diagnosis. Lee and Mahoney studied 40 cases of biopsy-proved, pure stromal fibrosis on breast MRI. More than half of the lesions presented as masses Figure 7. Some appeared as foci or areas of nonmasslike enhancement Figure 7. Although extremely variable, the most common appearance of nonmass lesions was clumped enhancement in linear distribution.
Most lesions were iso- to hypointense to normal parenchyma on T1-weighted images, and variable in signal intensity on T2-weighted sequences, depending on the water content of the lesion. The enhancement kinetics also varied but usually showed medium or rapid initial contrast uptake, with plateau or washout delayed curves. On histopathology, ectatic vessels were found within most of the lesions, which may explain the degree of contrast enhancement on MRI. Mammary fibromatosis is a rare benign stromal tumor of the breast that accounts for less than 0.
It was initially reported in a patient with Gardner syndrome but may occur sporadically or after trauma or surgery. Most cases were reported in women, with rare cases in men. Mammary fibromatosis mimics malignancy, both clinically and by imaging. Clinically, it presents as a suspicious palpable mass, which may be associated with skin retraction and fixation to the chest wall. It appears as a spiculated mass on mammogram.
On ultrasound, it usually appears as an irregular solid hypoechoic mass with spiculated or microlobulated margins and tethering of Cooper ligaments. It may be locally aggressive, involving the pectoralis or intercostal muscles. On MRI, mammary fibromatosis typically appears as an irregular mass isointense to muscle on T1-weighted images and variably hyperintense on T2-weighted sequences. It typically shows benign persistent enhancement curves in the delayed phase Figure 7.
MRI is the best imaging modality for evaluation of chest wall invasion and extent of disease. The recommended treatment is complete surgical resection with clear margins. It is an uncommon but important entity because of its tendency to mimic breast carcinoma on clinical and imaging presentations. Clinically, a patient with GCT usually presents with a palpable firm lump, occasionally associated with skin fixation and thickening and nipple retraction.
It has a widely varied mammographic appearance, ranging from a focal asymmetry to a mass with ill-defined or spiculated margins. Its sonographic features also vary widely from benign-appearing masses with circumscribed margins and acoustic enhancement to malignant-appearing masses with spiculated margins and acoustic shadowing.
Histologically GCTs with circumscribed margins exhibit scant fibrous stroma, whereas the ones with spiculated margins have a background of dense fibrous stroma. It has been described as a mass with either smooth or spiculated margins. It is low in signal intensity on T1-weighted images, hypointense to glandular tissue or isointense to skeletal muscle. The reported T2 signal intensities and enhancement patterns of GCT are widely variable.
On T2-weighted sequences, its signal intensity is described as hypointense to muscle, isointense to glandular tissue, or demonstrating a high—signal-intensity rim. It may enhance homogenously or heterogeneously, particularly pronounced at its margin. GCTs show medium or rapid rate of contrast enhancement during early phase, with either persistent or washout kinetics in the late phase.
It is impossible to establish a diagnosis of GCT without biopsy. Despite its diagnostic challenges, GCT is usually benign and carries an excellent prognosis. Malignant GCT can metastasize widely and have a grave prognosis. Wide local excision is the treatment of choice for both benign and malignant GCTs. Long-term surveillance is recommended for large tumors, rapid growth, local recurrence, and multiple lesions, owing to the slight, but real, malignant potential of GCTs.
Diabetic mastopathy is a condition in which stromal proliferation forms fibrous masses, predominantly in female patients with long-standing insulin-dependent type 1 diabetes mellitus. Rarely, it may occur in males or in type 2 diabetic patients. The interval between the onset of diabetes and presentation of the breast lesion is reportedly about 20 years. It is known by various names, including diabetic fibrous mastopathy , mastopathy in insulin-dependent diabetes, diabetic fibrous breast disease , and diabetic sclerosing lymphocytic lobulitis of the breast.
The pathogenesis of diabetic mastopathy is not completely understood. It probably represents an immune reaction to the abnormal accumulation of altered extracellular matrix in the breast, secondary to prolonged hyperglycemia. The lymphocytic infiltrate is composed predominantly of B cells.
The increased expression of HLA-DR4 antigen in involved lobular epithelium and the association with other autoimmune diseases support the presumed autoimmune pathogenesis. Clinically, patients usually present with a nontender, palpable, firm to hard mass in one or both breasts. Mammography may reveal a heterogeneously dense parenchymal pattern, focal asymmetry, or a mass with obscured, indistinct, or spiculated margins.
Ultrasound shows a heterogeneously hypoechoic mass with ill-defined margins and marked posterior acoustic shadowing. There is absence of color flow signals on Doppler study. Diabetic mastopathy may mimic breast cancer on clinical, mammographic, and ultrasound examinations because of the extensive fibrosis.
Tissue sampling is necessary for diagnosis and to exclude a malignancy. Treatment options are either excision or close follow-up, depending on the individual circumstances. Diabetic mastopathy is isointense to glandular tissue on breast MRI, but a mass can often be appreciated on the non—fat-suppressed T1-weighted image. Wong and colleagues and Tuncbilek and coworkers reported no discernible enhancing mass on postcontrast MRI.
Wong and colleagues described nonspecific patchy or diffuse stromal enhancement, and Tuncbilek and associates reported homogeneous, low enhancement and glandular distortion. Lipoma is a benign fatty tumor that may occur anywhere in the breast. Most lipomas are located in the subcutaneous fat. They may manifest as soft, mobile palpable masses or incidental findings on screening mammography.
The cause of lipomas is unknown. Lipomas have a classic appearance on mammography, presenting as a completely radiolucent mass with a thin radiopaque capsule. On ultrasound, they may be isoechoic to adjacent fat lobules or mildly hyperechoic. Occasionally, lipomas may contain multiple echogenic septa that course parallel to the skin. On MRI, lipomas present as oval masses with smooth margins.
Their signal intensity follows that of normal fatty tissue in the breast on all sequences. They have high signal intensity identical to the adjacent normal fatty tissue on T1-weighted non—fat-suppressed sequences.
On T1-weighted fat-suppressed sequences, their signals are suppressed to the same degree as the adjacent normal fat Figure 7. Hamartomas are benign breast lesions composed of a variety of normal breast constituents, including fat, glandular tissue, and fibrous connective tissue. Other names, such as fibroadenolipoma or lipofibroadenoma have been used to reflect the dominant tissue types within the mass. The lesion may be a result of dysgenesis rather than a true tumor. Most of hamartomas are detected in pre- and perimenopausal women.
They are variable in size but usually measure several centimeters in diameter. Clinically, hamartomas are found either incidentally on screening mammograms or in patients presenting with painless palpable masses. On mammograms, hamartomas may have a classic appearance of a circumscribed mass containing both fat and soft tissue density surrounded by a thin radiopaque capsule. When a hamartoma contains a very small amount of fat, it may mimic a fibroadenoma or circumscribed carcinoma.
Its sonographic appearance is also variable. It often appears as a mass with smooth margins, containing areas of low-level internal echogenicity interspersed with irregular areas of hyperechogenicity. On MRI, a hamartoma appears as an encapsulated mass with a heterogeneous appearance.
Some contrast enhancement can be seen in the glandular elements on T1-weighted postcontrast images Figure 7. Benign Cystic Lesions Duct Ectasia Duct ectasia is a nonspecific dilation of the major subareolar ducts, with occasional involvement of smaller ducts. Duct ectasia is seen in a year-old breast and ovarian cancer susceptibility BRCA gene—positive woman undergoing high-risk screening MRI. A, T1-weighted fat-suppressed noncontrast image shows dilated retroareolar ducts containing high signal intensity proteinaceous fluid.
B, Postcontrast subtraction image is needed to confirm the absence of contrast enhancement within the dilated ducts. FCCs are seen in a year-old woman with a recent diagnosis of right breast carcinoma. Sagittal T1-weighted fat-suppressed postcontrast MRI image shows regional clumped nonmasslike enhancement arrow in the ipsilateral breast.
Enhancement kinetics are rapid and persistent. Biopsy revealed fibrocystic changes. FCCs are seen in a year-old woman with strong family history of premenopausal breast cancer. On screening MRI, T1-weighted fat-suppressed postcontrast maximum-intensity projection MIP image shows clumped nonmasslike enhancement in focal single arrow and segmental double arrows distributions. Enhancement kinetics are slow and persistent. Biopsies of both lesions yielded fibrocystic changes.
Focal nodular FCCs are seen in a year-old woman with strong family history of premenopausal breast cancer. Sagittal T1-weighted fat-suppressed postcontrast MRI image shows an oblong 0. The mass shows heterogeneous enhancement with rapid and persistent kinetics. Biopsy yielded fibrocystic changes. Focal nodular FCCs are seen in a year-old woman with a palpable mass in the right breast, negative dense mammogram and sonogram.
Sagittal T1-weighted fat-suppressed postcontrast MRI image shows a 1 cm irregular mass with heterogeneous enhancement and rapid and persistent enhancement kinetics. The median size of group B was 9. Pathological findings revealed that all the cystic masses on MRI showed infarction of the papillary fronds.
The presence of duct dilatation was more commonly seen in group B than group A No significant differences were found for the location, number of masses being solitary or multiple , shape, margin, TIC pattern, and ADC value between the two groups. A case of papilloma manifesting as a solid mass in year-old woman. The pathology is an intraductal papilloma. A case of papilloma with DCIS manifesting as a complex cystic mass in year-old woman. The pathology is intraductal papilloma with DCIS. For NME lesions, focal or linear distribution Fig.
A case of papilloma manifesting as linear enhancement in year-old woman. The pathology is intraductal papilloma with adenosis. A case of papilloma with DCIS manifesting as segmental enhancement in year-old woman. In this study, we found patients with papillomas with high-risk or malignant lesions were older than those with benign papillomas, which was consistent with previous reports [ 28 , 29 ]. The bloody nipple discharge was more commonly observed in patients with papilloma with high-risk or malignant lesions The bloody nipple discharge occurred in the patient with benign papilloma may be caused by the hemorrhage of the tumor or duct ectasia [ 30 ].
Our study supports this finding. On MRI, Tominaga et al. The irregular margin of benign papilloma correlated with the surrounding fibrosis or collagenization of the stroma of the lesion [ 26 ]. Our study revealed that most of the papillomas have round or oval shape and more than half of the non-benign papillomas have circumscribed margins.
In this study, shape and margin were found as non-distinctive morphologic features between benign papilloma and a papilloma with high-risk or malignant lesions. The complex cystic pattern was slightly more commonly observed in a papilloma with high-risk or malignant lesions. In the present study, we found a small portion of papillomas manifested as a cystic mass on MRI.
Pathological findings revealed the infarction of the papillary fronds in all these cases. To the best of our knowledge, this feature has not been reported in previous literature. Studies have demonstrated that papillomas are usually mixed with benign proliferative lesions and less commonly with atypical and malignant lesions [ 3 , 4 , 5 , 6 , 7 , 8 , 9 ]. Lewis et al. Choi et al. In our cases, 43 in We also found that mixed mass-NME lesion was more commonly observed in papilloma with high-risk or malignant lesions We speculated that the accompanying NME lesions on MRI may be attributed to the concomitant lesions, and the atypical hyperplasia or DCIS adjacent to the papilloma might more likely show enhancement than the benign hyperplasia lesions.
Papilloma manifesting as NME could be due to the concomitant benign, atypical, and malignant proliferative lesions. Sarica et al. Our study also demonstrated that the segmental or regional distribution indicated a non-benign papilloma. Studies [ 23 , 25 ] reported that However, in our cases, The larger sample size may be one reason for the difference in TIC patterns. The concomitant proliferative lesions could also have an influence on the TIC patterns of papillomas.
Our study showed that TIC patterns could not add any value for differential diagnosis. The ADC value of papilloma was reported as 1. A relatively lower b value may be responsible for the slight higher ADC values of papilloma in our study. We also found that the papilloma with high-risk or malignant lesions showed similar ADC values with the benign papillomas without significant difference.
This kind of papilloma was too small to be detected on enhanced MRI. It usually presents no abnormal findings on breast MRI or hardly be recognized from the background parenchymal enhancement. For patients with MRI-occult lesions, fiberoductoscopy could be a new problem-solving tool. Fiberoductoscopy enables to direct visualization of small lesions and performs minimally invasive procedures [ 30 ]. Therefore, it may help to reduce unnecessary surgical excision in patients with benign intraductal lesions [ 30 ].
More studies are needed to evaluate the subsequent management for fiberoductoscopy diagnosed benign papillomas in patients with pathological nipple discharge. In our study, all the three cases of occult lesions were benign.
Moreover, the microscopic papillomas excluded in our study were also occult on breast MRI. Jaffer et al. The main limitation of this study is that the patients in this cohort underwent excisional biopsy instead of CNB. These patients usually had suspicious breast lesions diagnosed on mammography or ultrasound and then underwent preoperatively MRI assessment, ultrasound localization, and subsequently surgically excisional biopsy.
The injection of methylene blue into the dilated ducts was performed for patients with nipple discharge before the surgical excision. Secondly, we did not compare the value of MRI with ultrasound in evaluating the extent of papillomas with pathologic correlation. Though there were limitations, we still could find some clinical and MR imaging diagnostic indicators for papillomas with high-risk or malignant lesions from this series of surgically confirmed papillomas.
Our findings are applicable for patients who already had an MRI examination for a different reason, for example, screening high-risk women, or the evaluation of a suspected or a newly diagnosed breast cancer, and so on. If the patient was diagnosed as papilloma by CNB, the MRI diagnostic indicators reported in this study help to assist the surgeon for the subsequent management, surgical excision, or follow-up.
A prospective study is warranted to assess the value of MRI in the management a patient with CNB diagnosed benign papilloma. WHO classification of tumours of the breast. Lyon: IARC; Google Scholar. Clinicopathologic analysis of breast lesions associated with multiple papillomas. Hum Pathol. An analysis of breast cancer risk in women with single, multiple, and atypical papilloma. Am J Surg Pathol. Article Google Scholar. Papillary lesions of the breast: a review.
Adv Anat Pathol. Papillary lesions of the breast: selected diagnostic and management issues. Pathological criteria and practical issues in papillary lesions of the breast — a review. Upgrading rate of papillary breast lesions diagnosed by core-needle biopsy. Diagn Interv Radiol. PubMed Google Scholar. Risk of invasive breast cancer and ductal carcinoma in situ in women with atypical papillary lesions of the breast. Breast J. Wen X, Cheng W. Nonmalignant breast papillary lesions at core-needle biopsy: a meta-analysis of underestimation and influencing factors.
Ann Surg Oncol. Management of benign intraductal solitary papilloma diagnosed on core needle biopsy. Management of intraductal papillomas of the breast: an analysis of cases and their outcome. Role of radiologic features in the management of papillary lesions of the breast. Non-malignant breast papillary lesions - B3 diagnosed on ultrasound-guided gauge needle core biopsy: analysis of cases from a single institution and review of the literature.
Pathol Oncol Res. Benign intraductal papilloma without atypia on core needle biopsy has a low rate of upgrading to malignancy after excision. J Breast Cancer. Uniqueness of ductal carcinoma in situ of the breast concurrent with papilloma: implications from a detailed topographical and histopathological study of 50 cases treated by mastectomy and wide local excision.
Triple-modality screening trial for familial breast cancer underlines the importance of magnetic resonance imaging and questions the role of mammography and ultrasound regardless of patient mutation status, age, and breast density. J Clin Oncol.
MR ductography: comparison with conventional ductography as a diagnostic method in patients with nipple discharge. Radiological appearances of papillary breast lesions. Clin Radiol. Direct MR galactography: feasibility study.
Magnetic resonance imaging features of papillary breast lesions. Eur J Radiol. MRI characteristics of intraductal papilloma. Acta Radiol. Magnetic resonance imaging of intraductal papillomas: typical findings and differential diagnosis. J Comput Assist Tomogr. Intraductal papilloma: features on MR ductography using a microscopic coil. Solitary intraductal papillomas of the breast: MRI features and differentiation from small invasive ductal carcinomas.
Magnetic resonance imaging of intraductal papillomas of the breast. Role of sonography in the differentiation of benign, high-risk, and malignant papillary lesions of the breast. J Ultrasound Med. Clinical and imaging characteristics of papillary neoplasms of the breast associated with malignancy: a retrospective cohort study.
Ultrasound Med Biol. Use of fiberoductoscopy for the management of patients with pathological nipple discharge: experience of a single center in Poland. Breast Cancer. Download references. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. You can also search for this author in PubMed Google Scholar. LW collected the data and performed the statistical analysis.
WG reviewed the pathology findings. LW drafted the manuscript.