Diagnostic accuracy of automated ACR BI-RADS breast density classification using deep convolutional neural networks.

OBJECTIVES: High breast density is a well-known risk factor for breast cancer. This study aimed to develop and adapt two (MLO, CC) deep convolutional neural networks (DCNN) for automatic breast density classification on synthetic 2D tomosynthesis reconstructions. METHODS: In total, 4605 synthetic 2D images (1665 patients, age: 57 ± 37 years) were labeled according to the ACR (American College of Radiology) density (A-D). Two DCNNs with 11 convolutional layers and 3 fully connected layers each, were trained with 70% of the data, whereas 20% was used for validation. The remaining 10% were used as a separate test dataset with 460 images (380 patients). All mammograms in the test dataset were read blinded by two radiologists (reader 1 with two and reader 2 with 11 years of dedicated mammographic experience in breast imaging), and the consensus was formed as the reference standard. The inter- and intra-reader reliabilities were assessed by calculating Cohen's kappa coefficients, and diagnostic accuracy measures of automated classification were evaluated. RESULTS: The two models for MLO and CC projections had a mean sensitivity of 80.4% (95%-CI 72.2-86.9), a specificity of 89.3% (95%-CI 85.4-92.3), and an accuracy of 89.6% (95%-CI 88.1-90.9) in the differentiation between ACR A/B and ACR C/D. DCNN versus human and inter-reader agreement were both "substantial" (Cohen's kappa: 0.61 versus 0.63). CONCLUSION: The DCNN allows accurate, standardized, and observer-independent classification of breast density based on the ACR BI-RADS system. KEY POINTS: • A DCNN performs on par with human experts in breast density assessment for synthetic 2D tomosynthesis reconstructions. • The proposed technique may be useful for accurate, standardized, and observer-independent breast density evaluation of tomosynthesis.

Tall cell carcinoma of the breast with reversed polarity (TCCRP) with mutations in the IDH2 and PIK3CA genes: a case report.

Tall cell carcinoma with reversed polarity (TCCRP) is a rare breast carcinoma with low malignant potential, initially named "breast tumor resembling the tall cell variant of papillary thyroid carcinoma", which has recently been recognized as a separate entity in the 5th edition of the WHO (World Health Organization) classification of breast tumors. Since the first report of this entity in 2003, more than 40 cases have been reported in the literature. Here, we report another case of this rare tumor in a 60-year-old woman. We performed immunohistochemical analyses and next-generation-sequencing (NGS) using the Oncomine™ Comprehensive DNA Panel (Thermo Fisher Scientific). The tumor showed the typical morphological features of TCCRP and a "triple-negative" phenotype. Moreover, we identified pathogenic mutations in the IDH2 (p.R172G) and PIK3CA (p.H1047R) genes. We report a case of TCCRP of the breast showing the characteristic morphologic, immunohistochemical and molecular features of this entity. There is still a limited number of cases with comprehensive molecular analyses reported in the literature. Therefore, we herewith contribute to a better understanding of the morphological and molecular characteristics as well as the clinical behavior of this rare entity.

Correction to: Second International Consensus Conference on lesions of uncertain malignant potential in the breast (B3 lesions).

The article Second International Consensus Conference on lesions of uncertain malignant potential in the breast (B3 lesions), written by Christoph J Rageth, Elizabeth AM O'Flynn, Katja Pinker, Rahel A Kubik-Huch, Alexander Mundinger, Thomas Decker, Christoph Tausch, Florian Dammann, Pascal A. Baltzer, Eva Maria Fallenberg, Maria P Foschini, Sophie Dellas, Michael Knauer, Caroline Malhaire, Martin Sonnenschein, Andreas Boos, Elisabeth Morris, Zsuzsanna Varga, was originally published electronically on the publisher's internet portal (currently SpringerLink) on November 30, 2018 without open access.

Second International Consensus Conference on lesions of uncertain malignant potential in the breast (B3 lesions).

PURPOSE: The second International Consensus Conference on B3 lesions was held in Zurich, Switzerland, in March 2018, organized by the International Breast Ultrasound School to re-evaluate the consensus recommendations. METHODS: This study (1) evaluated how management recommendations of the first Zurich Consensus Conference of 2016 on B3 lesions had influenced daily practice and (2) reviewed current literature towards recommendations to biopsy. RESULTS: In 2018, the consensus recommendations for management of B3 lesions remained almost unchanged: For flat epithelial atypia (FEA), classical lobular neoplasia (LN), papillary lesions (PL) and radial scars (RS) diagnosed on core-needle biopsy (CNB) or vacuum-assisted biopsy (VAB), excision by VAB in preference to open surgery, and for atypical ductal hyperplasia (ADH) and phyllodes tumors (PT) diagnosed at VAB or CNB, first-line open surgical excision (OE) with follow-up surveillance imaging for 5 years. Analyzing the Database of the Swiss Minimally Invasive Breast Biopsies (MIBB) with more than 30,000 procedures recorded, there was a significant increase in recommending more frequent surveillance of LN [65% in 2018 vs. 51% in 2016 (p = 0.004)], FEA (72% in 2018 vs. 62% in 2016 (p = 0.005)), and PL [(76% in 2018 vs. 70% in 2016 (p = 0.04)] diagnosed on VAB. A trend to more frequent surveillance was also noted also for RS [77% in 2018 vs. 67% in 2016 (p = 0.07)]. CONCLUSIONS: Minimally invasive management of B3 lesions (except ADH and PT) with VAB continues to be appropriate as an alternative to first-line OE in most cases, but with more frequent surveillance, especially for LN.

Combined visual and motor evoked potentials predict multiple sclerosis disability after 20 years.

BACKGROUND: The development of predictors of multiple sclerosis (MS) disability is difficult due to the complex interplay of pathophysiological and adaptive processes. OBJECTIVE: The purpose of this study was to investigate whether combined evoked potential (EP)-measures allow prediction of MS disability after 20 years. METHODS: We examined 28 patients with clinically definite MS according to Poser's criteria with Expanded Disability Status Scale (EDSS) scores, combined visual and motor EPs at entry (T0), 6 (T1), 12 (T2) and 24 (T3) months, and a cranial magnetic resonance imaging (MRI) scan at T0 and T2. EDSS testing was repeated at year 14 (T4) and year 20 (T5). Spearman rank correlation was used. We performed a multivariable regression analysis to examine predictive relationships of the sum of z-transformed EP latencies (s-EPT0) and other baseline variables with EDSST5. RESULTS: We found that s-EPT0 correlated with EDSST5 (rho=0.72, p<0.0001) and ΔEDSST5-T0 (rho=0.50, p=0.006). Backward selection resulted in the prediction model: E (EDSST5)=3.91-2.22×therapy+0.079×age+0.057×s-EPT0 (Model 1, R (2)=0.58) with therapy as binary variable (1=any disease-modifying therapy between T3 and T5, 0=no therapy). Neither EDSST0 nor T2-lesion or gadolinium (Gd)-enhancing lesion quantities at T0 improved prediction of EDSST5. The area under the receiver operating characteristic (ROC) curve was 0.89 for model 1. CONCLUSIONS: These results further support a role for combined EP-measures as predictors of long-term disability in MS.

Mammotome: less invasive than ABBI with similar accuracy for early breast cancer detection.

We performed a prospective analysis of two consecutive biopsy cohorts investigated by the same team to compare the Mammotome system with the ABBI procedure. From April 1997 to August 2003 a series of 413 nonpalpable mammographic lesions in 387 women (median age 56 years, range 30-84 years) were stereotactically biopsied in the University Hospital of Basel, Switzerland. Until October 1999 the ABBI system was applied exclusively, it was subsequently superseded by the Mammotome device in our clinic. Main outcome measures were accuracy, technical demand, and morbidity. Sensitivity (97.3%/96.8%), negative predictive value (99.2%/98.7%), and diagnostic accuracy (99.4%/99.1%) regarding the detection of malignancy were excellent for both techniques (ABBI/ Mammotome). The Mammotome procedure was faster and less invasive, thus causing significantly less morbidity. The larger specimen obtained by the ABBI procedure resulted in more detailed histology. In conclusion, recommend the Mammotome system as the method of choice for detecting nonpalpable early breast cancer.