Deep transfer learning for detection of breast arterial calcifications on mammograms: a comparative study.

INTRODUCTION: Breast arterial calcifications (BAC) are common incidental findings on routine mammograms, which have been suggested as a sex-specific biomarker of cardiovascular disease (CVD) risk. Previous work showed the efficacy of a pretrained convolutional network (CNN), VCG16, for automatic BAC detection. In this study, we further tested the method by a comparative analysis with other ten CNNs. MATERIAL AND METHODS: Four-view standard mammography exams from 1,493 women were included in this retrospective study and labeled as BAC or non-BAC by experts. The comparative study was conducted using eleven pretrained convolutional networks (CNNs) with varying depths from five architectures including Xception, VGG, ResNetV2, MobileNet, and DenseNet, fine-tuned for the binary BAC classification task. Performance evaluation involved area under the receiver operating characteristics curve (AUC-ROC) analysis, F1-score (harmonic mean of precision and recall), and generalized gradient-weighted class activation mapping (Grad-CAM++) for visual explanations. RESULTS: The dataset exhibited a BAC prevalence of 194/1,493 women (13.0%) and 581/5,972 images (9.7%). Among the retrained models, VGG, MobileNet, and DenseNet demonstrated the most promising results, achieving AUC-ROCs > 0.70 in both training and independent testing subsets. In terms of testing F1-score, VGG16 ranked first, higher than MobileNet (0.51) and VGG19 (0.46). Qualitative analysis showed that the Grad-CAM++ heatmaps generated by VGG16 consistently outperformed those produced by others, offering a finer-grained and discriminative localization of calcified regions within images. CONCLUSION: Deep transfer learning showed promise in automated BAC detection on mammograms, where relatively shallow networks demonstrated superior performances requiring shorter training times and reduced resources. RELEVANCE STATEMENT: Deep transfer learning is a promising approach to enhance reporting BAC on mammograms and facilitate developing efficient tools for cardiovascular risk stratification in women, leveraging large-scale mammographic screening programs. KEY POINTS: • We tested different pretrained convolutional networks (CNNs) for BAC detection on mammograms. • VGG and MobileNet demonstrated promising performances, outperforming their deeper, more complex counterparts. • Visual explanations using Grad-CAM++ highlighted VGG16's superior performance in localizing BAC.

MRI-derived extracellular volume as a biomarker of cancer therapy cardiotoxicity: systematic review and meta-analysis.

OBJECTIVES: MRI-derived extracellular volume (ECV) allows characterization of myocardial changes before the onset of overt pathology, which may be caused by cancer therapy cardiotoxicity. Our purpose was to review studies exploring the role of MRI-derived ECV as an early cardiotoxicity biomarker to guide timely intervention. MATERIALS AND METHODS: In April 2022, we performed a systematic search on EMBASE and PubMed for articles on MRI-derived ECV as a biomarker of cancer therapy cardiotoxicity. Two blinded researchers screened the retrieved articles, including those reporting ECV values at least 3 months from cardiotoxic treatment. Data extraction was performed for each article, including clinical and technical data, and ECV values. Pooled ECV was calculated using the random effects model and compared among different treatment regimens and among those who did or did not experience overt cardiac dysfunction. Meta-regression analyses were conducted to appraise which clinical or technical variables yielded a significant impact on ECV. RESULTS: Overall, 19 studies were included. Study populations ranged from 9 to 236 patients, for a total of 1123 individuals, with an average age ranging from 12.5 to 74 years. Most studies included patients with breast or esophageal cancer, treated with anthracyclines and chest radiotherapy. Pooled ECV was 28.44% (95% confidence interval, CI, 26.85-30.03%) among subjects who had undergone cardiotoxic cancer therapy, versus 25.23% (95%CI 23.31-27.14%) among those who had not (p = .003). CONCLUSION: A higher ECV in patients who underwent cardiotoxic treatment could imply subclinical changes in the myocardium, present even before overt cardiac pathology is detectable. CLINICAL RELEVANCE STATEMENT: The ability to detect subclinical changes in the myocardium displayed by ECV suggests its use as an early biomarker of cancer therapy-related cardiotoxicity. KEY POINTS: • Cardiotoxicity is a common adverse effect of cancer therapy; therefore, its prompt detection could improve patient outcomes. • Pooled MRI-derived myocardial extracellular volume was higher in patients who underwent cardiotoxic cancer therapy than in those who did not (28.44% versus 25.23%, p = .003). • MRI-derived myocardial extracellular volume represents a potential early biomarker of cancer therapy cardiotoxicity.

Detection and quantification of breast arterial calcifications on mammograms: a deep learning approach.

OBJECTIVE: Breast arterial calcifications (BAC) are a sex-specific cardiovascular disease biomarker that might improve cardiovascular risk stratification in women. We implemented a deep convolutional neural network for automatic BAC detection and quantification. METHODS: In this retrospective study, four readers labelled four-view mammograms as BAC positive (BAC+) or BAC negative (BAC-) at image level. Starting from a pretrained VGG16 model, we trained a convolutional neural network to discriminate BAC+ and BAC- mammograms. Accuracy, F1 score, and area under the receiver operating characteristic curve (AUC-ROC) were used to assess the diagnostic performance. Predictions of calcified areas were generated using the generalized gradient-weighted class activation mapping (Grad-CAM++) method, and their correlation with manual measurement of BAC length in a subset of cases was assessed using Spearman ρ. RESULTS: A total 1493 women (198 BAC+) with a median age of 59 years (interquartile range 52-68) were included and partitioned in a training set of 410 cases (1640 views, 398 BAC+), validation set of 222 cases (888 views, 89 BAC+), and test set of 229 cases (916 views, 94 BAC+). The accuracy, F1 score, and AUC-ROC were 0.94, 0.86, and 0.98 in the training set; 0.96, 0.74, and 0.96 in the validation set; and 0.97, 0.80, and 0.95 in the test set, respectively. In 112 analyzed views, the Grad-CAM++ predictions displayed a strong correlation with BAC measured length (ρ = 0.88, p < 0.001). CONCLUSION: Our model showed promising performances in BAC detection and in quantification of BAC burden, showing a strong correlation with manual measurements. CLINICAL RELEVANCE STATEMENT: Integrating our model to clinical practice could improve BAC reporting without increasing clinical workload, facilitating large-scale studies on the impact of BAC as a biomarker of cardiovascular risk, raising awareness on women's cardiovascular health, and leveraging mammographic screening. KEY POINTS: • We implemented a deep convolutional neural network (CNN) for BAC detection and quantification. • Our CNN had an area under the receiving operator curve of 0.95 for BAC detection in the test set composed of 916 views, 94 of which were BAC+ . • Furthermore, our CNN showed a strong correlation with manual BAC measurements (ρ = 0.88) in a set of 112 views.

The predictive role of right ventricular late gadolinium enhancement in patients with tetralogy of Fallot undergoing pulmonary valve replacement.

BACKGROUND: Our purpose was to evaluate the correlations between right ventricular (RV) late gadolinium enhancement (LGE) at cardiac magnetic resonance (CMR) in patients with tetralogy of Fallot (ToF) scheduled for pulmonary valve replacement (PVR) and post-PVR functional data. METHODS: We retrospectively reviewed ToF patients scheduled for PVR who underwent two CMR examinations at our institution, one before the procedure (CMR-0), including contrast-enhanced sequences, and one after the procedure (CMR-1). Functional left and RV data were obtained by segmenting short-axis stacks on both CMR examinations, and normalised variations were calculated by dividing differences between CMR-1 and CMR-0 by the intercurring time interval, whereas the RV scar burden was assessed on CMR-0 LGE sequences both semiquantitatively and quantitatively. Data were reported as median and interquartile range, differences were appraised with the Mann-Whitney U test, while correlations were assessed with Spearman's ρ. RESULTS: Fifteen patients with a median age of 25 years (16-29), including 9 (60%) males, with a median time interval between CMR-0 and CMR-1 of 17 months (12-23), were retrospectively reviewed. The semiquantitative LGE score at CMR-0 was 7 (6-9), and LGE volume was 4.49 mL (3.70-5.78), covering 5.63% (4.92-7.00) of the RV. RV LGE score showed a moderate positive correlation with the normalised variation of RV stroke volume (ρ = 0.662, p = 0.007) and a borderline moderate positive correlation with the normalised variation of RV end-diastolic indexed volume (ρ = 0.513, p = 0.050). CONCLUSIONS: The assessment of RV LGE before PVR may provide insights on post-PVR functional data, potentially facilitating a patient-tailored treatment pathway.

Mammography biomarkers of cardiovascular and musculoskeletal health: A review.

Breast density (BD) and breast arterial calcifications (BAC) can expand the role of mammography. In premenopause, BD is related to body fat composition: breast adipose tissue and total volume are potential indicators of fat storage in visceral depots, associated with higher risk of cardiovascular disease (CVD). Women with fatty breast have an increased likelihood of hypercholesterolemia. Women without cardiometabolic diseases with higher BD have a lower risk of diabetes mellitus, hypertension, chest pain, and peripheral vascular disease, while those with lower BD are at increased risk of cardiometabolic diseases. BAC, the expression of Monckeberg sclerosis, are associated with CVD risk. Their prevalence, 13 % overall, rises after menopause and is reduced in women aged over 65 receiving hormonal replacement therapy. Due to their distinct pathogenesis, BAC are associated with hypertension but not with other cardiovascular risk factors. Women with BAC have an increased risk of acute myocardial infarction, ischemic stroke, and CVD death; furthermore, moderate to severe BAC load is associated with coronary artery disease. The clinical use of BAC assessment is limited by their time-consuming manual/visual quantification, an issue possibly solved by artificial intelligence-based approaches addressing BAC complex topology as well as their large spectrum of extent and x-ray attenuations. A link between BD, BAC, and osteoporosis has been reported, but data are still inconclusive. Systematic, standardised reporting of BD and BAC should be encouraged.

Psoas Cross-Sectional Measurements Using Manual CT Segmentation before and after Endovascular Aortic Repair (EVAR).

Sarcopenia has been associated with an increased incidence of adverse outcomes, including higher mortality, after endovascular aortic repair (EVAR). We aim to use computed tomography (CT) to quantify changes in total psoas muscles area (PMA) and psoas muscle density (PMD) after EVAR, and to evaluate the reproducibility of both measurements. PMA and PMD were assessed via manual segmentation of the psoas muscle on pre- and post-operative CT scans belonging to consecutive patients who underwent EVAR. Wilcoxon test was used to compare PMA and PMD before and after EVAR, and inter- and intra-reader agreements of both methods were evaluated through Bland−Altman analysis. A total of 50 patients, 42 of them males (84%), were included in the study. PMA changes from 1243 mm2 (1006−1445 mm2) to 1102 mm2 (IQR 937−1331 mm2), after EVAR (p < 0.001). PMD did not vary between pre-EVAR (33 HU, IQR 26.5−38.7 HU) and post-EVAR (32 HU, IQR 26−37 HU, p = 0.630). At inter-reader Bland−Altman analysis, PMA showed a bias of 64.0 mm2 and a coefficient of repeatability (CoR) of 359.2 mm2, whereas PMD showed a bias of −2.43 HU and a CoR of 6.19 HU. At intra-reader Bland−Altman analysis, PMA showed a bias of −81.1 mm2 and a CoR of 394.6 mm2, whereas PMD showed a bias of 1.41 HU and a CoR of 6.36 HU. In conclusion, PMA decreases after EVAR. A good intra and inter-reader reproducibility was observed for both PMA and PMD. We thus propose to use PMA during the follow-up of patients who underwent EVAR to monitor muscle depletion after surgery.

Quantification of epicardial adipose tissue in obese patients using an open-bore MR scanner.

BACKGROUND: Our aim was to evaluate the reproducibility of epicardial adipose tissue (EAT) volume, measured on scans performed using an open-bore magnetic resonance scanner. METHODS: Consecutive patients referred for bariatric surgery, aged between 18 and 65 years who agreed to undergo cardiac imaging (MRI), were prospectively enrolled. All those with cardiac pathology or contraindications to MRI were excluded. MRI was performed on a 1.0-T open-bore scanner, and EAT was segmented on all scans at both systolic and diastolic phase by two independent readers (R1 with four years of experience and R2 with one year). Data were reported as median and interquartile range; agreement and differences were appraised with Bland-Altman analyses and Wilcoxon tests, respectively. RESULTS: Fourteen patients, 11 females (79%) aged 44 (41-50) years, underwent cardiac MRI. For the first and second readings, respectively, EAT volume was 86 (78-95) cm3 and 85 (79-91) cm3 at systole and 82 (74-95) cm3 and 81 (75-94) cm3 at diastole for R1, and 89 (79-99) cm3 and 93 (84-98) cm3 at systole and 92 (85-103) cm3 and 93 (82-94) cm3 at diastole for R2. R1 had the best reproducibility at diastole (bias 0.3 cm3, standard deviation of the differences (SD) 3.3 cm3). R2 had the worst reproducibility at diastole (bias 3.9 cm3, SD 12.1 cm3). The only significant difference between systole and diastole was at the first reading by R1 (p = 0.016). The greatest bias was that of inter-reader reproducibility at diastole (-9.4 cm3). CONCLUSIONS: Reproducibility was within clinically acceptable limits in most instances.

Development and Validation of an AI-driven Mammographic Breast Density Classification Tool Based on Radiologist Consensus.

Mammographic breast density (BD) is commonly visually assessed using the Breast Imaging Reporting and Data System (BI-RADS) four-category scale. To overcome inter- and intraobserver variability of visual assessment, the authors retrospectively developed and externally validated a software for BD classification based on convolutional neural networks from mammograms obtained between 2017 and 2020. The tool was trained using the majority BD category determined by seven board-certified radiologists who independently visually assessed 760 mediolateral oblique (MLO) images in 380 women (mean age, 57 years ± 6 [SD]) from center 1; this process mimicked training from a consensus of several human readers. External validation of the model was performed by the three radiologists whose BD assessment was closest to the majority (consensus) of the initial seven on a dataset of 384 MLO images in 197 women (mean age, 56 years ± 13) obtained from center 2. The model achieved an accuracy of 89.3% in distinguishing BI-RADS a or b (nondense breasts) versus c or d (dense breasts) categories, with an agreement of 90.4% (178 of 197 mammograms) and a reliability of 0.807 (Cohen κ) compared with the mode of the three readers. This study demonstrates accuracy and reliability of a fully automated software for BD classification. Keywords: Mammography, Breast, Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning Algorithms Supplemental material is available for this article. © RSNA, 2022.

Potential role of epicardial adipose tissue as a biomarker of anthracycline cardiotoxicity.

BACKGROUND: We investigated the radiodensity of epicardial (EAT), subcutaneous (SAT), and visceral adipose tissue (VAT) before and after treatment with anthracyclines in a population of breast cancer (BC) patients, and in controls not treated with anthracyclines, to detect a potential role of EAT density as a biomarker of changes related to chemotherapy cardiotoxicity. METHODS: We reviewed BC patients treated with anthracyclines who underwent CT before (CT-t0) and after (CT-t1) chemotherapy, and age- and sex-matched controls who underwent two CT examinations at comparable intervals. On non-contrast scans, EAT was segmented contouring the pericardium and thresholding between -190 and -30 Hounsfield units (HU), and SAT and VAT were segmented with two 15-mm diameter regions of interest thresholded between -195 and -45 HU. RESULTS: Thirty-two female patients and 32 controls were included. There were no differences in age (p = 0.439) and follow-up duration (p = 0.162) between patients and controls. Between CT-t0 and CT-t1, EAT density decreased in BC patients (-66 HU, interquartile range [IQR] -71 to -63 HU, to -71 HU, IQR -75 to -66 HU, p = 0.003), while it did not vary in controls (p = 0.955). SAT density increased from CT-t0 to CT-t1 in BC patients (-107 HU, IQR -111 to -105 HU, to -105 HU, IQR -110 to -100 HU, p = 0.014), whereas it did not change in controls (p = 0.477). VAT density did not vary in either BC patients (p = 0.911) or controls (p = 0.627). CONCLUSIONS: EAT density appears to be influenced by anthracycline treatment for BC, well known for its cardiotoxicity, shifting towards lower values indicative of a less active metabolism.

Semiquantitative score of breast arterial calcifications on mammography (BAC-SS): intra- and inter-reader reproducibility.

BACKGROUND: Breast arterial calcifications (BAC), representing Mönckeberg's sclerosis of the tunica media of breast arteries, are an imaging biomarker for cardiovascular risk stratification in the female population. Our aim was to estimate the intra- and inter-reader reproducibility of a semiquantitative score for BAC assessment (BAC-SS). METHODS: Consecutive women who underwent screening mammography at our center from January 1st to January 31st, 2018 were retrieved and included according to BAC presence. Two readers (R1 and R2) independently applied the BAC-SS to medio-lateral oblique views, obtaining a BAC score by summing: (I) number of calcified vessels (from 0 to n); (II) vessel opacification, i.e., the degree of artery coverage by calcium bright pixels (0 or 1); and (III) length class of calcified vessels (from 0 to 4). R1 repeated the assessment 2 weeks later. Scoring time was recorded. Cohen's κ statistics and Bland-Altman analysis were used. RESULTS: Among 408 women, 57 (14%) had BAC; 114 medio-lateral oblique views were assessed. Median BAC score was 4 [interquartile range (IQR): 3-6] for R1 and 4 (IQR: 2-6) for R2 (P=0.417) while median scoring time was 156 s (IQR: 99-314 s) for R1 and 191 s (IQR: 137-292 s) for R2 (P=0.743). Bland-Altman analysis showed a 77% intra-reader reproducibility [bias: 0.193, coefficient of repeatability (CoR): 0.955] and a 64% inter-reader reproducibility (bias: 0.211, CoR: 1.516). Cohen's κ for BAC presence was 0.968 for intra-reader agreement and 0.937 for inter-reader agreement. CONCLUSIONS: Our BAC-SS has a good intra- and inter-reader reproducibility, within acceptable scoring times. A large-scale study is warranted to test its ability to stratify cardiovascular risk in women.

Breast arterial calcifications as a biomarker of cardiovascular risk: radiologists' awareness, reporting, and action. A survey among the EUSOBI members.

OBJECTIVES: To investigate the knowledge of radiologists on breast arterial calcifications (BAC) and attitude about BAC reporting, communication to women, and subsequent action. METHODS: An online survey was offered to EUSOBI members, with 17 questions focused on demographics, level of experience, clinical setting, awareness of BAC association with cardiovascular risk, mammographic reporting, modality of BAC assessment, and action habits. Descriptive statistics were used. RESULTS: Among 1084 EUSOBI members, 378 (34.9%) responded to the survey, 361/378 (95.5%) radiologists, 263 females (69.6%), 112 males (29.6%), and 3 (0.8%) who did not specify their gender. Of 378 respondents, 305 (80.7%) declared to be aware of BAC meaning in terms of cardiovascular risk and 234 (61.9%) to routinely include BAC in mammogram reports, when detected. Excluding one inconsistent answer, simple annotation of BAC presence was declared by 151/233 (64.8%), distinction between low versus extensive BAC burden by 59/233 (25.3%), and usage of an ordinal scale by 22/233 (9.5%) and of a cardinal scale by 1/233 (0.4%). Among these 233 radiologists reporting BAC, 106 (45.5%) declared to orally inform the woman and, in case of severe BAC burden, 103 (44.2%) to investigate cardiovascular history, and 92 (39.5%) to refer the woman to a cardiologist. CONCLUSION: Among EUSOBI respondents, over 80% declared to be aware of BAC cardiovascular meaning and over 60% to include BAC in the report. Qualitative BAC assessment predominates. About 40% of respondents who report on BAC, in the case of severe BAC burden, investigate cardiovascular history and/or refer the woman to a cardiologist. KEY POINTS: • Of 1084 EUSOBI members, 378 (35%) participated: 81% of respondents are aware of breast arterial calcification (BAC) cardiovascular meaning and 62% include BAC in the mammogram report. • Of those reporting BAC, description of presence was declared by 65%, low versus extensive burden distinction by 25%, usage of an ordinal scale by 10%, and of a cardinal scale by 0.4%; 46% inform the woman and, in case of severe BAC burden, 44% examine cardiovascular history, and 40% refer her to a cardiologist. • European breast radiologists may be ready for large-scale studies to ascertain the role of BAC assessment in the comprehensive framework of female cardiovascular disease prevention.

Computed tomography-derived myocardial extracellular volume: an early biomarker of cardiotoxicity in esophageal cancer patients undergoing radiation therapy.

OBJECTIVES: We aimed to assess extracellular volume (ECV) through non-gated, contrast-enhanced computed tomography (CT) before and after radiation therapy (RT) in patients with esophageal cancer (EC). MATERIALS AND METHODS: EC patients who had undergone CT before and after RT were retrospectively assessed. Patients with preexisting cardiovascular disease or with heavily artifacted CT were excluded. ECV was calculated using density values for the myocardial septum and blood pool. Data were reported as mean and standard deviation or median and interquartile range according to their distribution; t test or Wilcoxon and Pearson r or Spearman ρ were subsequently used. RESULTS: Twenty-one patients with stage ≥ IB EC, aged 64 ± 18 years, were included. Mean and maximum RT doses were 21.2 Gy (16.9-24.1) and 42.5 Gy (41.8-49.2), respectively. At baseline (n = 21), hematocrit was 39% ± 4%, ECV 27.9% ± 3.5%; 35 days (30-38) after RT (n = 20), hematocrit was 36% ± 4%, lower than at baseline (p = 0.002), ECV 30.3% ± 8.3%, higher than at baseline (p = 0.081); at follow-up 420 days (244-624) after RT (n = 13), hematocrit was 36% ± 5%, lower than at baseline (p = 0.030), ECV 31.4% ± 4.5%, higher than at baseline (p = 0.011). No patients showed signs of overt cardiotoxicity. ECV early after RT was moderately positively correlated with maximum RT dose (ρ = 0.50, p = 0.036). CONCLUSIONS: In EC patients, CT-derived myocardial ECV was increased after RT and may thus appear as a potential early biomarker of cardiotoxicity.