Mammographic Microcalcifications Research Articles

Size-adapted microcalcification segmentation in mammography utilizing scale-space signatures

The purpose of this study is size-adapted segmentation of individual microcalcifications in mammography, based on microcalcification scale-space signature estimation, enabling robust scale selection for initialization of multiscale active contours. Segmentation accuracy was evaluated by the area overlap measure, by comparing the proposed method and two recently proposed ones to expert manual delineations. The method achieved area overlap of 0.61 ± 0.15 outperforming statistically ( p < 0.001) the other two methods (0.53 ± 0.18, 0.42 ± 0.16). Only the proposed method performed equally for both small (<460 μm) and large (≥460 μm) microcalcifications. Results indicate an accurate method, which could be utilized in computer-aided diagnosis schemes of microcalcification clusters.

Breast magnetic resonance imaging as a problem-solving modality in mammographic BI-RADS 3 lesions

The probability of a mammographic Breast Imaging Reporting and Data System (BI-RADS) 3 lesion being cancer is considered to be less than 2%. Therefore, the work-up of a mammographic BI-RADS 3 lesion should be biopsy or follow-up mammography after 6 months. However, most patients referred for biopsy have benign disease. Although the negative predictive value (NPV) of magnetic resonance imaging (MRI) is highest of all imaging techniques, it is not yet common practice to use breast MRI as a problem-solving modality to exclude patients for further diagnostic work-up. Therefore, in this meta-analysis the usefulness of breast MRI as a problem-solving modality in mammographic BI-RADS 3 lesions is investigated. After a systematic search only 5 out of 61 studies met the inclusion criteria. The NPV in 2 of those studies was reported to be 100%. It was concluded that MRI can be used as an adjunctive tool to mammographic BI-RADS 3 findings to exclude patients for further diagnostic work-up. The other 3 studies assessed the accuracy of MRI in mammographic BI-RADS 3 microcalcifications. These studies reported an NPV of MRI between 76% and 97%. Therefore, MRI cannot be implemented as a diagnostic tool to evaluate mammographic microcalcifications at this time. The first solid data indicate that breast MRI might be useful as a problem-solving modality to exclude patients with non-calcified mammographic BI-RADS 3 lesions for further diagnostic work-up. However, further research is needed to verify these results.

0620: Breast Ultrasound of DCIS. Correlation with Mammography, MRI and Pathology

Ultrasound(US) performed with a 10-13-MHz transducer and optimal technique can be used as a complement to mammography in detecting and evaluating DCIS of the breast. There are several circumstances in which US may be of benefit in the evaluation of patients with DCIS. First, US can be used to visualize large-size (over 10 mm) clusters of microcalcifications with a high suspicion of malignancy (estimated likelihood of malignancy 75% or higher, using mammographic assessment criteria). The main benefit of identifying a sonographic abnormality in women with mammographically detected microcalcifications is to allow the use of US to guide interventional procedures, such as needle biopsy and needle localization. US-guided procedures are less expensive and faster than stereotactically-guided procedures. In addition, for those institutions that do not have stereotactic equipment, the use of US in selected cases would extend the role of precutaneous biopsy at these sites. US may also be used as a guide to biopsy the invasive component of tumors in patients presented with mammographic microcalcifications. Second, US might be used to increase the specificity of mammography and helps to reduce the number of surgical or core biopsies in women with microcalcifications. Third, US can be used to reveal occult DCIS in patients with dense breasts. A microlobulated mass with mild hypoechogenicity, ductal extension, and normal acoustic transmission was the most common US finding in noncalcified DCIS. Fourth, US can be used to evaluate patients with a nipple discharge when ductography is not possible or indeterminate.

Flache epitheliale Atypie

According to the WHO, flat epithelial atypia (FEA) is defined as a neoplastic epithelial proliferation of ductal type in either a single or in multiple terminal duct lobular unit(s) limited to the periphery of the ductules in a clinging growth pattern. The atypical cells may form between one and several layers of epithelial cells that show low grade cytologic atypia. FEA most often presents as mammographic microcalcifications, which are typically round (secretory type and psammomatous calcification in an eosinophilic matrix, so-called ossifying calcifications). Clinical relevance is dependent on whether the lesion appears in isolation or whether it is an excision biopsy or a minimally invasive biopsy. Currently available data suggest that the risk of subsequent breast carcinoma in the ipsilateral breast is very low following the diagnosis of FEA. The differential diagnosis should include atypical ductal hyperplasia, low-grade clinging ductal carcinoma in situ, blunt duct adenosis and apocrine metaplasia.

Lobular neoplasia on core biopsy and risk factors to predict upstaging at surgical excision.

Abstract Abstract #5009 Introduction: Lobular neoplasia (including atypical lobular hyperplasia and lobular carcinoma in-situ) identified on core biopsy (CB) is often recommended for surgical excision because the risk of upstaging to invasive carcinoma is documented to be 0-50%. We sought to identify risk factors to predict upstaging at surgical excisional biopsy.&amp;#x2028; Methods: Retrospective chart review was used to identify women with CB revealing lobular neoplasia (LN) as highest risk pathological diagnosis. LN was defined as atypical lobular hyperplasia (ALH) and/or lobular carcinoma in situ (LCIS). Radiologic findings were correlated with pathology and excision results were recorded when available. Follow-up radiologic data was also recorded when available.&amp;#x2028; Results: From 1997-2008, 45 women (age ranging from 33 to 81) underwent CB revealing LN. Ten had LCIS only, 34 had ALH only and 1 had both. Twenty five (56%) underwent surgical excision and 20 (44%) were followed clinically and mammographically. Core biopsy was recommended for mammographic microcalcifications (three with associated mass or focal asymmetry) in 29 pts (64%), while 14 CB (31%) were recommended for either mammographic mass or focal asymmetry only. Two CB (4%) were recommended for MRI abnormalities. For patients who underwent surgical excision, 8 (32%) were upstaged to invasive carcinoma or DCIS. Two of these 8 (25%) had LCIS on core biopsy and 6 (75%) had ALH only. Of the 6 pts with ALH only, the original core biopsy was done for focal asymmetry or mass in 3, abnormal MRI in 1 and microcalcifications in the remaining 2 (one with synchronous contralateral carcinoma). Patients not upstaged at surgical excision had original CB done for LCIS only in 5 (31%) and ALH only in 11 (69%). Of these 11 pts with ALH who were not upstaged, 8 underwent original CB for mass, focal asymmetry or MRI finding. Patients with LN on core biopsy done exclusively for mammographic microcalcifications underwent excision only 46% (13/28) of the time. Of those excised 4 were upstaged at surgical excision ( 3 with DCIS and 1 tubular carcinoma). Follow-up mammographic data was available for 65% (13/20) of patients who did not have surgical excision. At median follow-up of 16 months no new biopsies have been recommended and no new lesions have been identified for these patients.&amp;#x2028; Conclusion :Of all patients who underwent surgical excision for LN found on core biopsy, 32% had a final diagnosis of either DCIS or invasive cancer. Of the 25 patients who underwent surgical excision, 24 were associated with one of the following risk factors: mass or focal asymmetry on mammogram, LCIS, synchronous contralateral carcinoma and non-concordant core biopsy. We did not identify any significant risk factors to predict upstaging at surgical excision, however, this may be due to selection bias. Further studies with larger number of patients are needed to identify risk factors to predict upstaging at surgical excision. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 5009.

A New Resolution Enhancement Technology Using the Independent Sub-Pixel Driving for the Medical Liquid Crystal Displays

A new resolution enhancement technology using the independent sub-pixel driving technology was developed for the medical monochrome liquid crystal displays (LCDs). Each pixel of monochrome LCDs, which employ the color liquid crystal panels that its color filters are removed, consists of three sub-pixels. In the new LCD system implemented with this new technology, the sub-pixel intensities were modulated according to the detailed image information, and consequently resolution was enhanced three times. Thus the new technology realized a 15 mega-pixels (MP) super high-resolution LCD (SHR-LCD) out of a conventional 5 MP LCD and a 9 MP SHR-LCD out of a conventional 3 MP LCD. Physical measurements and perceptual evaluations performed in this study proved that the achieved 15 MP (through our new technology) was appropriate and efficient to depict the finer anatomical structures such as the micro calcifications in mammography.

Malignant-Appearing Microcalcifications at the Lumpectomy Site with the Use of FloSeal Hemostatic Sealant

The aim of this article is to describe the benign mammographic calcifications that occur at the lumpectomy site after the use of a topical hemostatic sealant (FloSeal Matrix Hemostatic Sealant). These calcifications can have an appearance similar to that of recurrent carcinoma. Application of FloSeal hemostatic sealant in the lumpectomy cavity results in benign mammographic microcalcifications that could be misinterpreted as malignant.

B‐spline active rays segmentation of microcalcifications in mammography

Accurate segmentation of microcalcifications in mammography is crucial for the quantification of morphologic properties by features incorporated in computer-aided diagnosis schemes. A novel segmentation method is proposed implementing active rays (polar-transformed active contours) on B-spline wavelet representation to identify microcalcification contour point estimates in a coarse-to-fine strategy at two levels of analysis. An iterative region growing method is used to delineate the final microcalcification contour curve, with pixel aggregation constrained by the microcalcification contour point estimates. A radial gradient-based method was also implemented for comparative purposes. The methods were tested on a dataset consisting of 149 mainly pleomorphic microcalcification clusters originating from 130 mammograms of the DDSM database. Segmentation accuracy of both methods was evaluated by three radiologists, based on a five-point rating scale. The radiologists' average accuracy ratings were 3.96 +/- 0.77, 3.97 +/- 0.80, and 3.83 +/- 0.89 for the proposed method, and 2.91 +/- 0.86, 2.10 +/- 0.94, and 2.56 +/- 0.76 for the radial gradient-based method, respectively, while the differences in accuracy ratings between the two segmentation methods were statistically significant (Wilcoxon signed-ranks test, p < 0.05). The effect of the two segmentation methods in the classification of benign from malignant microcalcification clusters was also investigated. A least square minimum distance classifier was employed based on cluster features reflecting three morphological properties of individual microcalcifications (area, length, and relative contrast). Classification performance was evaluated by means of the area under ROC curve (Az). The area and length morphologic features demonstrated a statistically significant (Mann-Whitney U-test, p < 0.05) higher patient-based classification performance when extracted from microcalcifications segmented by the proposed method (0.82 +/- 0.06 and 0.86 +/- .05, respectively), as compared to segmentation by the radial gradient-based method (0.71 +/- 0.08 and 0.75 +/- 0.08). The proposed method demonstrates improved segmentation accuracy, fulfilling human visual criteria, and enhances the ability of morphologic features to characterize microcalcification clusters.

Prediction of the presence of invasive disease from the measurement of extent of malignant microcalcification on mammography and ductal carcinoma in situ grade at core biopsy

To determine whether the extent of microcalcification and ductal carcinoma in situ (DCIS) grade can be used to accurately predict the presence and size of invasive cancer in cases of malignant microcalcification. Over a 10-year period, 402 cases of malignant microcalcification from an NHS screening programme were analysed. For each case, measurement of mammographic microcalcification extent, DCIS grade, and the presence and size of invasive carcinoma from the excised surgical specimen were recorded. The final histological diagnosis was DCIS only in 71% (284/402) and DCIS with a focus of invasive disease in 29% (118/402). The risk of invasive disease increased with increasing size of microcalcification from 20% (27/136) for cluster size less than 11mm, to 45% (18/40) for cluster size more than 60mm. The risk of invasive disease also increased with increasing histological grade of DCIS from 13% (4/31) with low-grade DCIS to 36% (86/239) with high-grade DCIS. There were significant associations with the presence of invasive disease for cluster size (p=0.0001) and DCIS grade (p=0.003), and when using univariate analysis with simple [cluster size (p=0.01) and grade (p=0.01)] and multiple [cluster size (p=0.02) and grade (p=0.02)] logistic regression, respectively. The Hosmer-Lemeshow goodness-of-fit test suggests that the multiple logistic regression model has a good fit (p=0.99). The multidisciplinary team can use these data in individual cases to estimate the risk of invasive cancer and decide whether to carry out an axillary staging procedure.

Triple Test Score for the Evaluation of Invasive Ductal and Lobular Breast Cancer

The aim of our study was to compare the preoperative sum score diagnostics of invasive ductal and lobular cancers using three or four diagnostic methods. The novelty of this study is the examination of this phenomenon based on sum score, no such papers can be found in the literature. Ductal cancers have higher score values indicating easier diagnostics, but the difference in distribution of the scores was significant (p = 0.0086) only in case of the triple-test. The score values give appropriate opportunity to create their order of diagnostic power which was the same by both histologic types and in their subgroups with low sum-score: the strongest was cytology, followed by mammography, ultrasound and physical examination. No significant difference was found between the two histologic group in their mammographic appearances-stellate, circumscribed, assymmetric distortion or microcalcification-(p = 0.0694). In low score subgroup besides the occult forms, structural distortion and indeterminate microcalcifications overweighed the stellate and circumscribed lesions typical for the whole groups. In symptomless cases of both histologic groups only one strongly malignant diagnostic test result warrants the right diagnosis. Summarizing the score distribution of the results in case of four diagnostic tools the higher scores-indicating malignancy-were more frequent in the ductal group compared to the lobular ones. Extra attention has to be paid to rare radiomorphologic appearances and to the most deterministic examination, namely cytology.

Predictors of Surgical Margin Status in Breast-Conserving Surgery Within a Breast Screening Program

Breast-conserving surgery (BCS) requires clear surgical margins to minimize local recurrence. We sought to identify groups of patients at higher risk of involved margins who might benefit from preoperative counselling and/or more generous excision at the first operation. We reviewed demographic, clinical, radiological and pathological records of all women diagnosed with ductal carcinoma in situ (DCIS) or invasive cancer (IC) through a population-based breast screening program in Melbourne, Australia between 1994 and 2005. A total of 2,160 women were diagnosed with DCIS or IC. We excluded 199 who had mastectomy (TM) as initial procedure or had missing data. Three hundred and thirteen had a diagnostic biopsy. Of 1,648 women who had BCS after a preoperative diagnosis of DCIS or IC, 13.5% had involved margins, 16.6% had close (</=1 mm), and 69.8% clear (>1 mm) margins. Of the patients, 281/1,648 (17.1%) underwent re-excision, of whom 93 (33.1%) had residual disease identified. Mammographic microcalcifications (P < 0.0001), absence of a mammographic mass (P = 0.002), presence of DCIS (P < 0.0001), high tumour grade (P < 0.0001), large size (P < 0.0001), multifocal disease (P < 0.0001) and lobular histology (P = 0.005) were associated with involved margins. Microcalcifications (odds ratio [OR] 1.97), large size (OR 4.22) and multifocal disease (OR 2.85) were independently associated with involved margins. Residual disease was associated with involved margins (P < 0.0001), presence of DCIS (P = 0.05) and large tumour size (P = 0.01). After BCS, patients with mammographic microcalcifications, larger tumour size and multifocal tumours are more likely to have involved margins. Patients with involved margins, large tumour size and/or a DCIS component are more likely to have residual disease on re-excision.

Mammographic Microcalcification in an Autogenously Reconstructed Breast Simulating Recurrent Carcinoma

Breast cancer is a common cancer among women. The transverse rectus abdominis myocutaneous (TRAM) flap is a popular option because not only does it provide a breast with satisfactory bulk composed of autogenous tissue but it also provides an abdominal dermolipectomy to the patient. Fat necrosis remains a common problem following TRAM flap reconstruction, occurring in 10% to 36% of patients undergoing the procedure. A 44-year-old woman underwent a modified radical mastectomy followed by pedicled TRAM flap reconstruction after 5 months. Follow-up mammography 27 months after TRAM flap reconstruction showed a cluster of microcalcifications in the deep retroareolar area and recurrent breast carcinoma was highly suspected. Physical examination did not detect any abnormality of the reconstructed breast. Stereotactic hook localization was performed and an excisional biopsy was successfully done. The histological features of the resected specimens corresponded to fat necrosis change. Only with the awareness of the existence of such entity and careful follow-up can the occurrence of fat necrosis in TRAM flap reconstructed breasts be accurately detected and appropriately treated.

Predicting the presence of invasive carcinoma from the extent of mammographic malignant microcalcification and grade of ductal carcinoma in situon core biopsy: results from a National Health Service Screening Programme

Four hundred and two cases of malignant microcalcification presenting through a National Health Service screening programme over a 10-year period were analysed. The final histological diagnosis was ductal carcinoma in situ (DCIS) only in 71% (284/402 cases) and DCIS with a focus of invasive disease in 29% (118/402 cases). There were significant associations with the presence of invasive disease for the histological grade of DCIS (χ2 = 9.06, P = 0.003) and for the cluster size (χ2 = 12.18, P = 0.0001). The proportion of cases with invasive tumour increased with increasing DCIS grade, from 13% (4/31) for low-grade DCIS to 36% (86/239) for high-grade DCIS; increasing cluster size, from 20% (27/136) 60 mm; and with increasing DCIS grade within each microcalcification cluster size category: 60 mm from 25% to 55% for low-grade to high-grade DCIS, respectively. The Hosmer–Lemeshow goodness-of-fit test supports this model (χ2 = 0.59, P = 0.96). The multidisciplinary team can use these data to estimate the risk of invasive cancer and to formulate appropriate management to minimise the number of patients requiring more than one operation, currently 49% for those with DCIS on core biopsy in a screening population [1].

Can texture of tissue surrounding microcalcifications in mammography be used for breast cancer diagnosis?

This study investigates whether texture properties of the tissue surrounding microcalcifications (MCs) can contribute to breast cancer diagnosis. A case sample of 100 MC clusters (46 benign, 54 malignant) from 85 dense mammographic images included in the Digital Database for Screening Mammography, is analyzed. Regions of interest containing clusters are processed using wavelet-based enhancement and individual MCs are segmented by local thresholding. The segmented MCs are removed from original image data and the surrounding tissue area is subjected to texture analysis. The feasibility of four texture feature sets (first-order statistics, gray level co-occurrence matrices, gray level run length matrices and Laws’ texture energy measures) in discriminating malignant from benign tissue was investigated using a k-nearest neighbor classifier. Laws’ texture energy measures achieved the best classification accuracy 89% (sensitivity 90.74% and specificity 86.96%).

Mammographic bi-dimensional product: a powerful predictor of successful excision of ductal carcinoma in situ

The aim of this analysis was to ascertain whether uni-dimensional measurement of mammographic microcalcification, the product of bi-dimensional measurement, calcification morphology, and pathological grade are helpful in predicting successful single therapeutic wide local excision (WLE) of ductal carcinoma in situ (DCIS). The study group comprised 505 patients whose mammograms showed the DCIS as calcification, and in whom a non-operative diagnosis had been obtained and WLE attempted. The extents of mammographic calcifications was measured in two planes at 90 degrees on the oblique view, the appearances classified as comedo, granular, or punctate. DCIS was graded using cyto-nuclear characteristics. Three hundred and forty-two patients had a successful first WLE and 163 patients had further surgery. A uni-dimensional measurement of <35 mm and a bi-dimensional product of <800 mm(2) were associated with successful excision (69 versus 54%, p=0.02 and 70 versus 27%, p=0.0001, respectively). Mammographic calcification morphology and histological grade did not influence the likelihood of a successful first WLE. For high-grade DCIS, the upper limit of the bi-dimensional product associated with successful WLE was 800 mm(2) (69 versus 24%, p=0.0003). In contrast, for non-high-grade DCIS, the cut-off was 400 mm(2) (73 versus 33%, p=0.01). Analyses based on mammographic calcification morphology gave similar findings. The mammographic bi-dimensional product is a powerful predictor of successful WLE of DCIS when combined with histological grade and/or calcification morphology.

Tamper Detection and Recovery for Medical Images Using Near-lossless Information Hiding Technique

Digital medical images are very easy to be modified for illegal purposes. For example, microcalcification in mammography is an important diagnostic clue, and it can be wiped off intentionally for insurance purposes or added intentionally into a normal mammography. In this paper, we proposed two methods to tamper detection and recovery for a medical image. A 1024 x 1024 x-ray mammogram was chosen to test the ability of tamper detection and recovery. At first, a medical image is divided into several blocks. For each block, an adaptive robust digital watermarking method combined with the modulo operation is used to hide both the authentication message and the recovery information. In the first method, each block is embedded with the authentication message and the recovery information of other blocks. Because the recovered block is too small and excessively compressed, the concept of region of interest (ROI) is introduced into the second method. If there are no tampered blocks, the original image can be obtained with only the stego image. When the ROI, such as microcalcification in mammography, is tampered with, an approximate image will be obtained from other blocks. From the experimental results, the proposed near-lossless method is proven to effectively detect a tampered medical image and recover the original ROI image. In this study, an adaptive robust digital watermarking method combined with the operation of modulo 256 was chosen to achieve information hiding and image authentication. With the proposal method, any random changes on the stego image will be detected in high probability.

Contrast-enhanced MR imaging in patients with BI-RADS 3-5 microcalcifications

The aim of this study was to evaluate the role of magnetic resonance imaging (MRI) in patients with microcalcifications classed as Breast Imaging Reporting and Data Systems (BI-RADS) 3-5. Fifty-five patients with mammographic microcalcifications classified as BI-RADS categories 3, 4 or 5 underwent MRI and biopsy with stereotactic vacuum-assisted biopsy (VAB). Our gold standard was microhistology in all cases and histology with histological grading in patients who underwent surgery. Patients with a microhistological diagnosis of benign lesions underwent mammographic follow-up for at least 12 months. MRI was performed with a 1.5-Tesla (T) unit, and T1 coronal three-dimensional (3D) fast low-angle shot sequences were acquired before and after injection of paramagnetic contrast agent (0.1 mmol/kg). MRI findings, according to the Fisher score, were classified into BI-RADS classes. In patients with cancer who underwent surgery, we retrospectively compared the extension of the mammographic and MRI findings with histological extension. Histology revealed 26 ductal in situ cancers (DCIS) and ductal microinvasive cancers (DCmic), three atypical ductal hyperplasias (ADH) and 26 benign conditions. Histological grading of the 26 patients with cancer revealed four cases of G1, 11 cases of G2 and 11 cases of G3. If we consider mammographic BI-RADS category 3 as benign and BI-RADS 4 and 5 as malignant, mammography had 77% sensitivity, 59% specificity, 63% positive predictive value (PPV), 74% negative predictive value (NPV) and 67.2% diagnostic accuracy. If we consider MRI BI-RADS categories 1, 2 and 3 as benign and 4 and 5 as malignant, MRI had 73% sensitivity, 76% specificity, 73% PPV, 76% NPV and 74.5% diagnostic accuracy. As regards disease extension, mammography had 45% sensitivity and MRI had 84.6% sensitivity. Mammography and stereotactic biopsy still remain the only techniques for characterising microcalcifications. MRI cannot be considered a diagnostic tool for evaluating microcalcifications. It is, however, useful for identifying DCIS with more aggressive histological grades. An important application of MRI in patients with DCIS associated with suspicious microcalcifications could be to evaluate disease extension after a microhistological diagnosis of malignancy, as it allows a more accurate presurgical planning.

Breast MRI More Sensitive Than Mammography for Ductal Carcinoma In Situ

Breast magnetic resonance imaging is generally considered less sensitive than mammography for detection of ductal carcinoma in situ (DCIS), but this conclusion comes from studies in which MRI failed to detect DCIS that was already identified by mammographic microcalcification. Might MRI detect DCIS that is not apparent on mammography? In this …

Bayesian Network to Predict Breast Cancer Risk of Mammographic Microcalcifications and Reduce Number of Benign Biopsy Results: Initial Experience

To retrospectively determine whether a Bayesian network (BN) computer model can accurately predict the probability of breast cancer on the basis of risk factors and mammographic appearance of microcalcifications, to improve the positive predictive value (PPV) of biopsy, with pathologic examination and follow-up as reference standards. The institutional review board approved this HIPAA-compliant study; informed consent was not required. Results of 111 consecutive image-guided breast biopsies performed for microcalcifications deemed suspicious by radiologists were analyzed. Mammograms obtained before biopsy were analyzed in a blinded manner by a breast imager who recorded Breast Imaging Reporting and Data System (BI-RADS) descriptors and provided a probability of malignancy. The BN uses probabilistic relationships between breast disease and mammography findings to estimate the risk of malignancy. Probability estimates from the radiologist and the BN were used to create receiver operating characteristic (ROC) curves, and area under the ROC curve (A(z)) values were compared. PPV of biopsy was also evaluated on the basis of these probability estimates. The BN and the radiologist achieved A(z) values of 0.919 and 0.916, respectively, which were not significantly different. If the 34 patients estimated by the BN to have less than a 10% probability of malignancy had not undergone biopsy, the PPV of biopsy would have increased from 21.6% to 31.2% without missing a breast cancer (P < .001). At this level, the radiologist's probability estimation improved the PPV to 30.0% (P < .001). A probabilistic model that includes BI-RADS descriptors for microcalcifications can distinguish between benign and malignant abnormalities at mammography as well as a breast imaging specialist can and may be able to improve the PPV of image-guided breast biopsy.

Interpretation of mammographic breast microcalcification: interobserver variability between radiologists and mammographers and analysis of best mammographic predictors of histopathological outcomes

Interpretation of mammographic breast microcalcification: interobserver variability between radiologists and mammographers and analysis of best mammographic predictors of histopathological outcomes