Citations

Breast cancer is the second-leading cause of mortality among women around the world. Ultrasound (US) is one of the noninvasive imaging modalities used to diagnose breast lesions and monitor the prognosis of cancer patients. It has the highest sensitivity for diagnosing breast masses, but it shows increased false negativity due to its high operator dependency. Underserved areas do not have sufficient US expertise to diagnose breast lesions, resulting in delayed management of breast lesions. Deep learning neural networks may have the potential to facilitate early decision-making by physicians by rapidly yet accurately diagnosing and monitoring their prognosis. This article reviews the recent research trends on neural networks for breast mass ultrasound, including and beyond diagnosis. We discussed original research recently conducted to analyze which modes of ultrasound and which models have been used for which purposes, and where they show the best performance. Our analysis reveals that lesion classification showed the highest performance compared to those used for other purposes. We also found that fewer studies were performed for prognosis than diagnosis. We also discussed the limitations and future directions of ongoing research on neural networks for breast ultrasound.

Abstract Two inherited autosomal dominant cancer predisposition conditions - BRCA related hereditary breast and ovarian cancer (HBOC) and Lynch syndrome (LS) - are termed Centers for Disease Control and Prevention Tier 1 (CDCT1) genetic conditions, for which early identification and intervention have a positive impact on public health. Selection of individuals for genetic testing for these conditions is based on personal and family history. The goal of this study was to evaluate whether screening in a tertiary medical center using exome sequencing could efficiently identify carriers of HBOC and LS and to determine the frequency of incremental carriers identified outside of traditional clinical practice guidelines. Participants from three geographically diverse Mayo Clinic practices in the USA (Rochester MN, Phoenix AZ, Jacksonville FL) consented to clinical “Exome+” sequencing (Helix, San Mateo, CA; Mayo GeneGuide, Rochester, MN) in the TAPESTRY study (NCT05212428), which links sequencing data with electronic health records and the return of CDCT1 genetic findings. For this study we focused on two inherited cancer predisposition conditions: BRCA-related HBOC (BRCA1 and BRCA2) and LS (MLH1, MSH2, MSH6, PMS2 and EPCAM). Detailed chart review to collect demographic information, personal and family history, and assessment of clinical practice guidelines for genetic evaluation (National Cancer Control Network 2021). To date, 44,306 patients have enrolled and sequenced in TAPESTRY. Annotation and interpretation of all variants in 7 genes for HBOC and LS resulted in identification of 550 carriers (prevalence 1.24%) which included 387 with HBOC (27.2% BRCA1, 42.8% BRCA2) and 163 with LS (12.3% MSH6, 8.8% PMS2, 4.5% MLH1, 3.8% MSH2 and 0.2% EPCAM). Demographics of the cohort included: 62.7% female, mean age 55.2 years, non-white race 9.6% and 3.8% Hispanic/Latino ethnicity. A personal history of cancer was present in 46.4% of test positive patients, including 22.5% of HBOC patients with breast or ovarian cancer and 20.9% of LS patients with colorectal or uterine cancer. More than half of the patients (52.1%) were newly diagnosed with HBOC and LS based on the results of this study. Overall, 39.2% of HBOC/LS carriers identified through Exome+ sequencing did not satisfy NCCN criteria for genetic evaluation, this was higher for LS (56.2%) compared to HBOC (32%). Amongst those newly diagnosed with HBOC/LS (n=286), NCCN criteria were not satisfied in 60% of cases (78% for LS and 51% for HBOC). Of the pathogenic germline variant carriers who met NCCN guidelines for testing, 34.2% were not aware of their diagnosis prior to participation in this study. Our results emphasize the need for wide genomic screening for CDCT1 cancer predisposition syndromes. Such screening could identify at-risk carriers of HBOC and LS, who would not otherwise be identified through clinical practice guidelines. Citation Format: Emily Gay, Niloy Jewel Samadder, Michelle L. Bublitz, Melanie M. Peterson, Tammy A. Wilson, Lorelei A. Bandel, Sebastian M. Armasu, Robert A. Vierkant, Matthew J. Ferber, Eric W. Klee, Nicholas B. Larson, Teresa M. Kruisselbrink, Timothy B. Curry, Jan B. Egan, Jennifer L. Kemppainen, Jessa S. Bidwell, Jennifer L. Anderson, Tammy M. McAllister, T'Nita S. Walker, Katie L. Kunze, Vanda Lindpere, Michael A. Golafshar, Margaret Klint, Richard J. Presutti, William V. Bobo, Aleksander Sekulic, Jolene M. Summer Bolster, Cheryl L. Willman, Konstantinos N. Lazaridis. Genetic screening in a tertiary medical center identifies carriers of cancer predisposition diseases that would be missed by clinical guidelines [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5768.

Axillary lymph nodes (ALNs) are generally among the first to be impacted by the metastatic spread of breast tumors. Consequently, evaluating their condition is of great importance as the removal of affected ALNs can potentially curtail the further spread of cancer cells to other parts of the body. Conventional ultrasound imaging suffers from low sensitivity and specificity and does not provide a comprehensive picture of the status of the node. In this work, we utilize an ultrasound microvessel imaging technique to visualize vascular structures within ALNs and, through several image processing steps, characterize the morphology of these structures. Several morphological metrics are used as biomarkers to differentiate between reactive and metastatic lymph nodes of patients with suspicious ALNs. A group of 68 patients is included in this study. Pathology examination results are used as the gold-standard reference for data labeling. A support vector machine (SVM) model is trained, and its performance is evaluated. Statistical analysis of the discriminative potential of individual biomarkers and receiver operating characteristic (ROC) curve analysis of the trained model are presented. Preliminary results suggest that a sensitivity of 0.88, specificity of 0.88, and area under the curve of 0.94 are achievable. Work supported by NIH-R01CA239548.

Observation of thyroid nodules in ultrasound examinations is a common occurrence. However, accurate diagnosis of the malignancy status of these nodules is a challenging task relying only on conventional ultrasound imaging. Therefore, increasing the sensitivity and specificity of non-invasive screening methods can be helpful in reducing the probability of missed diagnosis and decreasing the rate of unnecessary biopsies. In this feasibility study, a quantitative ultrasound microvasculature imaging technique is applied to ultrasound data collected from a cohort of 92 patients. This technique consists of first enabling the visualization of vascular networks inside the scanned nodules, followed by a morphometric analysis framework with a number of biomarkers to distinguish between benign and malignant thyroid nodules. Statistical tests are performed to find significant differences between the distribution of the values of the biomarkers. Subsequently, these biomarkers are utilized to train classification models that solely benefit from morphological biomarkers and those trained with additional clinical features. The receiver operating characteristic (ROC) curve analysis of these models, which helps us evaluate their predictive power, is presented. In our preliminary results, the model trained with combined morphological and clinical features achieves an AUC of 0.90, a sensitivity of 0.81, and a specificity of 0.79. [Work supported by R01CA239548.]

Low specificity in current ultrasound modalities for thyroid cancer detection necessitates the development of new imaging modalities for optimal characterization of thyroid nodules. Herein, the quantitative biomarkers of a new high-definition microvessel imaging (HDMI) were evaluated for discrimination of benign from malignant thyroid nodules. Without the help of contrast agents, this new ultrasound-based quantitative technique utilizes processing methods including clutter filtering, denoising, vessel enhancement filtering, morphological filtering, and vessel segmentation to resolve tumor microvessels at size scales of a few hundred microns and enables the extraction of vessel morphological features as new tumor biomarkers. We evaluated quantitative HDMI on 92 patients with 92 thyroid nodules identified in ultrasound. A total of 12 biomarkers derived from vessel morphological parameters were associated with pathology results. Using the Wilcoxon rank-sum test, six of the twelve biomarkers were significantly different in distribution between the malignant and benign nodules (all p < 0.01). A support vector machine (SVM)-based classification model was trained on these six biomarkers, and the receiver operating characteristic curve (ROC) showed an area under the curve (AUC) of 0.9005 (95% CI: [0.8279,0.9732]) with sensitivity, specificity, and accuracy of 0.7778, 0.9474, and 0.8929, respectively. When additional clinical data, namely TI-RADS, age, and nodule size were added to the features, model performance reached an AUC of 0.9044 (95% CI: [0.8331,0.9757]) with sensitivity, specificity, and accuracy of 0.8750, 0.8235, and 0.8400, respectively. Our findings suggest that tumor vessel morphological features may improve the characterization of thyroid nodules.

Introduction: Over 6 million people in the U.S. have heart failure (HF), less than half are expected to survive beyond 5 years. The need to better stratify mortality risk in HF is recognized. Frailty is associated with mortality in HF but not routinely measured clinically. As frailty is linked to inflammation and malnutrition, we hypothesized that the Metabolic Vulnerability Index (MVX) a multimarker score of systemic inflammation (small HDL particles, GlycA) and malnutrition (leucine, valine, isoleucine, citrate), could serve as a biomarker of frailty to predict mortality risk. Methods: Clinical data and plasma were collected from 1,389 patients from a HF community cohort between 2003-2012. We measured frailty using the Rockwood Index as the proportion of deficits present out of 32 physical limitations and comorbidities. MVX was calculated from the nuclear magnetic resonance LipoProfile® test. Patients were categorized by frailty (0-0.15; 0.16-0.27; 0.28-0.78) and MVX (33.4-50, 50-60, 60-70, 70-85.8) cutpoints. Cox models estimated the association of frailty and MVX assignment with mortality, adjusted for Meta-Analysis Global Group in Chronic HF (MAGGIC) score, a validated clinical risk score for HF mortality. Results: Frailty and MVX scores were available in 985 patients (median age 77, IQR: 67-84; 48% women). Higher frailty was associated with higher MVX (p-trend < 0.001). The highest frailty and MVX groups experienced large increases in risk of death, after adjustment for MAGGIC score (HR=3.3, 95% CI=2.6-4.2) and (HR=2.7, 95% CI=2.1-3.5), respectively. When adjusted for one another and MAGGIC score, MVX and frailty associations with death were only minimally attenuated: frailty (HR=3.2, 95% CI=2.5-4.0) and MVX (HR=2.4, 95% CI=1.9-3.2) (Figure 1). Conclusion: In this community cohort of patients with HF, frailty and MVX are positively associated with one another. However, both indicators are independently associated with an increased risk of death and can contribute to risk stratification.

Introduction: Heart Failure (HF) is a syndrome of high comorbidity and mortality. The impact of comorbidities on life expectancy in HF is well studied, but guideline-recommended disease-centric risk scores such as the Meta-analysis Global Group in Chronic Heart Failure (MAGGIC) score may not fully account for the effect of comorbidities. The Charlson Comorbidity Index (Charlson) is a validated measure that quantifies the prognostic impact of comorbidities. Objective: To examine the incremental value of Charlson over MAGGIC in predicting death in HF. Methods: In a population-based HF cohort, we computed MAGGIC and Charlson scores using electronically linked medical records. Patients were divided based on ejection fraction (EF) into reduced [EF <50%, HFrEF] or preserved [EF ≥50%, HFpEF]. Cox regression was performed to assess the association between the risk scores and mortality. The incremental value of Charlson over MAGGIC was tested by calculating the area under the ROC [receiver operating characteristic] curves (AUCs) with 95% confidence interval (CI). Results: We studied 1,388 patients with HF validated using the Framingham criteria (mean age, 75±13 years, 48% female, mean EF, 49±17%). The median Charlson was 7.0 (interquartile range 5.0-9.0) and was higher in HFpEF than HFrEF (7.0 [5.0-9.0] vs. 6.0 [4.0-8.3] respectively. Over a median follow-up of 13.9 years (13.3-14.2), mortality was 17.4% (95% CI, 15.4 -19.3%) and 36.8% (95% CI, 34.2-39.3%) at 1, and 3 years respectively. In regression analysis, increasing MAGGIC and Charlson were independently associated with death (HR 1.09, 95% CI 1.08-1.11 and HR,1.22, 95% CI 1.20-1.24, both p <0.001) irrespective of EF. Discrimination ability was improved by adding Charlson over MAGGIC (Table), reaching statistical significance at 3 years. Conclusion: In this HF cohort, the Charlson index provided an added value over MAGGIC in predicting the risk of death at 3 years. These findings emphasize the importance of accounting for comorbidities in mid-term risk stratification in HF.

Background: Heart failure (HF) is associated with metabolic alterations, including ketogenesis. However, determinants of ketogenesis and risk of mortality in HF is not defined. Total ketone bodies (KB) include β-hydroxybutyrate, acetoacetate, and acetone and can be measured in plasma by nuclear magnetic resonance (NMR). The aim of this study is to determine the relationship between KB and clinical characteristics in a community HF cohort and to assess the association between KB and all-cause mortality. Methods: A population-based cohort of 1,389 HF patients was prospectively enrolled between 2003 and 2012. Plasma KB was measured by LP4 NMR LipoProfile ® assay/test on the Vantera® NMR analyzer platform. A conditional inference tree method (ctree R Package) was used to determine optimal KB group cut points. Associations between clinical characteristics and KB were measured with Wilcoxon rank sum test and Pearson’s Chi-squared test. Kaplan-Meier method estimated survival. Cox regression analyses were used to estimate associations between KB concentrations and mortality. Results: Among the 1,382 HF patients with KB measurements, the median age was 78 years (IQR 68-84) and 52% were men. Median KB was 180 μM (IQR 134-308). Patients were divided into two groups with lower KB (≤471.5 μM) and higher KB (>471.5 μM). Patients with higher KB (N=210) had lower BMI, higher BNP, and were more likely to be in the New York Heart Association class III-IV; however, these patients were less likely to have hyperlipidemia, coronary disease, or diabetes mellitus (P < 0.05). Age, sex, creatinine, ejection fraction, or Meta-Analysis Global Group in Chronic HF (MAGGIC) score did not differ by KB group. Higher KB was associated with worse survival (figure). After adjustment for the MAGGIC score, higher KB was associated with increased risk of mortality (HR 1.3; 95% CI, 1.08-1.48). Conclusions: In this community HF cohort, higher KB was associated with increased mortality, independent of the MAGGIC score.

Introduction: Over six million U.S. adults have heart failure (HF). Evaluation of kidney function is critical to the care of patients with HF but may not be fully captured by traditional measures, such as estimated glomerular filtration rate (eGFR) or the Meta-Analysis Global Group in Chronic HF (MAGGIC) risk score. Identifying patients at highest risk of progressive chronic renal insufficiency (PCRI) may improve risk stratification in HF, beyond these traditional measures. High-throughput proteomics has allowed for the development of PCRI risk scores based on novel kidney function biomarkers. However, the prognostic value of a proteomic-based PCRI risk score in HF has not been explored. Hypothesis: Proteomic-based PCRI risk scores will improve risk stratification in HF. Methods: Clinical data and plasma were collected from 1,389 patients in a HF community cohort from Southeastern Minnesota (2003-2012). Results from the aptamer-based technology SomaScan® were used to derive PCRI risk scores using the SomaSignal™ Kidney Prognosis test, a protein-based algorithm developed to predict risk of PCRI within four years. Cox proportional hazard models were used to estimate the association between quintiles of PCRI and mortality, after adjustment for the MAGGIC score. Results: PCRI risk scores were available for 1,349 patients who were on average 75±13 years of age, 48% female, and had a median eGFR of 57 mL/min/1.73m 2 . There was a positive association with mortality across PCRI risk quintiles, after adjustment for the MAGGIC score. (Figure). Overall, the highest quintile was associated over a two-fold higher risk of mortality compared with the lowest quintile (HR 2.3, 95% CI 1.8,2.8). The higher risk of mortality remained in analyses stratified by patients with (HR 1.5, 95% CI 1.1,2.1) and without (HR 2.2, 95% CI 1.4,3.5) an eGFR <60 mL/min/1.73 m 2 . Conclusions: In this community HF cohort, proteomic-based PCRI risk scores improve risk stratification, independent of traditional measures of kidney function.