A5079669314- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- Artificial Intelligence in Healthcare and Education
- Machine Learning in Healthcare
- Nutrition and Health in Aging
- Advanced X-ray and CT Imaging
- Domain Adaptation and Few-Shot Learning
- Fetal and Pediatric Neurological Disorders
- MRI in cancer diagnosis
- Cerebrovascular and Carotid Artery Diseases
- Radiology practices and education
- Advanced Neuroimaging Techniques and Applications
- Medical Imaging and Analysis
- Advanced Neural Network Applications
- Acute Ischemic Stroke Management
- Body Composition Measurement Techniques
- Glioma Diagnosis and Treatment
- Cardiac Imaging and Diagnostics
- Frailty in Older Adults
- Lung Cancer Diagnosis and Treatment
- COVID-19 diagnosis using AI
- Brain Tumor Detection and Classification
- Brain Metastases and Treatment
- Cardiac and Coronary Surgery Techniques
- Advanced MRI Techniques and Applications
Athinoula A. Martinos Center for Biomedical Imaging
2021-2025
Harvard University
2022-2025
Massachusetts General Hospital
2020-2025
Mass General Brigham
2019-2025
Office of Science
2022-2025
Deutsches Herzzentrum München
2024
Charité - Universitätsmedizin Berlin
2024
Humboldt-Universität zu Berlin
2024
Freie Universität Berlin
2024
University of Alberta
2024
Background Although CT-based body composition (BC) metrics may inform disease risk and outcomes, obtaining these has been too resource intensive for large-scale use. Thus, population-wide distributions of BC remain uncertain. Purpose To demonstrate the validity fully automated, deep learning analysis from abdominal CT examinations, to define demographically adjusted reference curves, illustrate advantage use curves compared with standard methods, along their biologic significance in...
Abstract The integration of artificial intelligence into clinical workflows requires reliable and robust models. Repeatability is a key attribute model robustness. Ideal repeatable models output predictions without variation during independent tests carried out under similar conditions. However, slight variations, though not ideal, may be unavoidable acceptable in practice. During development evaluation, much attention given to classification performance while repeatability rarely assessed,...
Body composition on chest CT scans encompasses a set of important imaging biomarkers. This study developed and validated fully automated analysis pipeline for multi-vertebral level assessment muscle adipose tissue routine scans. retrospectively trained two convolutional neural networks 629 from patients (55% women; mean age, 67 years ± 10 [standard deviation]) obtained between 2014 2017 prior to lobectomy primary lung cancer at three institutions. A slice-selection network was identify an...
Non-contrast head CT (NCCT) is extremely insensitive for early (< 3-6 h) acute infarct identification. We developed a deep learning model that detects and delineates suspected infarcts on NCCT, using diffusion MRI as ground truth (3566 NCCT/MRI training patient pairs). The substantially outperformed 3 expert neuroradiologists test set of 150 scans patients who were potential candidates thrombectomy (60 stroke-negative, 90 stroke-positive middle cerebral artery territory only infarcts), with...
Confirmation of pregnancy viability (presence fetal cardiac activity) and diagnosis presentation (head or buttock in the maternal pelvis) are first essential components ultrasound assessment obstetrics. The former is useful assessing presence an on-going latter for labour management. We propose automated framework detection heartbeat from a predefined free-hand sweep abdomen. Our method exploits key anatomical sonographic image patterns carefully designed scanning protocols to develop, time,...
Interpretation of ultrasound videos the fetal heart is crucial for antenatal diagnosis congenital disease (CHD). We believe that automated image analysis techniques could make an important contribution towards improving CHD detection rates. However, to our knowledge, no previous work has been done in this area. With goal mind, paper presents a framework tracking key variables describe content each frame freehand 2D scanning healthy heart. This represents first step developing tools can...
Background Adrenal masses are common, but radiology reporting and recommendations for management can be variable. Purpose To create a machine learning algorithm to segment adrenal glands on contrast-enhanced CT images classify as normal or mass-containing assess performance. Materials Methods This retrospective study included two groups of abdominal examinations (development data set secondary test set). in the development were manually segmented by radiologists. Images both classified...
The remarkable advances of artificial intelligence (AI) technology are revolutionizing established approaches to the acquisition, interpretation, and analysis biomedical imaging data. Development, validation, continuous refinement AI tools requires easy access large high-quality annotated datasets, which both representative diverse. National Cancer Institute (NCI) Imaging Data Commons (IDC) hosts diverse publicly available cancer image data collections. By harmonizing all based on industry...
Automatic analysis of fetal echocardiography screening images could aid in the identification congenital heart diseases. The first step towards automatic is locating an image and identifying viewing (imaging) plane. This highly challenging since small with relatively indistinct anatomical structural appearance. further compounded by presence artefacts ultrasound images. Herein we provide a state-of-art solution for detecting classifying each individual frame as belonging to one standard...
. Ultrasound is extensively utilized as a convenient and cost-effective method in emergency situations. Unfortunately, the limited availability of skilled clinicians hinders wider adoption point-of-care ultrasound. To overcome this challenge, paper aims to aid less experienced healthcare providers lung ultrasound scans.
Abstract Public imaging datasets are critical for the development and evaluation of automated tools in cancer imaging. Unfortunately, many do not include annotations or image-derived features, complicating downstream analysis. Artificial intelligence-based annotation have been shown to achieve acceptable performance can be used automatically annotate large datasets. As part effort enrich public data available within NCI Imaging Data Commons (IDC), here we introduce AI-generated two...
Privacy-preserving open-source large language models show potential as an alternative to OpenAI’s GPT-4 for accurate chest radiograph report classification.
Cancer patients with spinal metastases may undergo surgery without clear assessments of prognosis, thereby impacting the optimal palliative strategy. Because morbidity adversely impact recovery and initiation adjuvant therapies, evaluation risk factors associated mortality complications is critical. Evaluation body composition cancer as a surrogate for frailty an emerging area study improving preoperative stratification.
BACKGROUND. CT-based body composition (BC) measurements have historically been too resource intensive to analyze for widespread use and lacked robust comparison with traditional weight metrics predicting cardiovascular risk. OBJECTIVE. The aim of this study was determine whether BC obtained from routine CT scans by a fully automated deep learning algorithm could predict subsequent events independently weight, BMI, additional risk factors. METHODS. This retrospective included 9752 outpatients...
BACKGROUND: Obesity is a risk factor for diverticulitis. However, it remains unclear whether visceral fat area, more precise measurement of abdominal fat, associated with the OBJECTIVE: To estimate incident and recurrent diverticulitis according to area. DESIGN: A retrospective cohort study SETTINGS: The Mass General Brigham Biobank PATIENTS: 6,654 patients who underwent CT clinical indications had no diagnosis diverticulitis, inflammatory bowel disease, or cancer before scan. MAIN OUTCOME...