A5049617352- Radiomics and Machine Learning in Medical Imaging
- Brain Tumor Detection and Classification
- Glioma Diagnosis and Treatment
- AI in cancer detection
- Advanced Neural Network Applications
- MRI in cancer diagnosis
- Artificial Intelligence in Healthcare and Education
- Medical Imaging Techniques and Applications
- Cancer Genomics and Diagnostics
- Advanced X-ray and CT Imaging
- Medical Image Segmentation Techniques
- Medical Imaging and Analysis
- Personal Information Management and User Behavior
- IoT-based Smart Home Systems
- COVID-19 diagnosis using AI
- Autonomous Vehicle Technology and Safety
- Human-Automation Interaction and Safety
- Computational Drug Discovery Methods
- Cell Image Analysis Techniques
- Colorectal Cancer Screening and Detection
- Statistical Methods in Clinical Trials
- Advanced MRI Techniques and Applications
- Machine Learning in Bioinformatics
- IoT and GPS-based Vehicle Safety Systems
- Metabolomics and Mass Spectrometry Studies
Indiana University School of Medicine
2024-2025
Indiana University – Purdue University Indianapolis
2024-2025
University of Indianapolis
2025
University of Pennsylvania
2019-2023
Brain extraction, or skull-stripping, is an essential pre-processing step in neuro-imaging that has a direct impact on the quality of all subsequent processing and analyses steps. It also key requirement multi-institutional collaborations to comply with privacy-preserving regulations. Existing automated methods, including Deep Learning (DL) based methods have obtained state-of-the-art results recent years, primarily targeted brain extraction without considering pathologically-affected...
Deep Learning (DL) has the potential to optimize machine learning in both scientific and clinical communities. However, greater expertise is required develop DL algorithms, variability of implementations hinders their reproducibility, translation, deployment. Here we present community-driven Generally Nuanced Framework (GaNDLF), with goal lowering these barriers. GaNDLF makes mechanism development, training, inference more stable, reproducible, interpretable, scalable, without requiring an...
Convolutional neural network (CNN) models perform state of the art performance on image classification, localization, and segmentation tasks. Limitations in computer hardware, most notably small memory size deep learning accelerator cards, prevent relatively large images, such as those from medical satellite imaging, being processed a whole their original resolution. A fully convolutional topology, U-Net, is typically trained down-sampled images inference resolution, by simply dividing...
Abstract Objective. De-centralized data analysis becomes an increasingly preferred option in the healthcare domain, as it alleviates need for sharing primary patient across collaborating institutions. This highlights consistent harmonized curation, pre-processing, and identification of regions interest based on uniform criteria. Approach. Towards this end, manuscript describes Fe derated T umor S egmentation (FeTS) tool, terms software architecture functionality. Main results. The aim FeTS...
Testicular cancer is one of the most common malignancies in males aged 20-44, with its two major subtypes being Seminoma and Non-Seminomatous Germ Cell Tumors (NSGCT). Accurate classification essential for patient management treatment, after initial histologic diagnosis on radical orchiectomy specimens or biopsies metastases. Histopathological diagnosis, though effective, time-consuming, costly, subject to inter-observer variability. Digital pathology combined artificial intelligence (AI)...
Abstract Breast cancer is one of the most pervasive forms and its inherent intra- inter-tumor heterogeneity contributes towards poor prognosis. Multiple studies have reported results from either private institutional data or publicly available datasets. However, current public datasets are limited in terms having consistency in: a) quality, b) quality expert annotation pathology, c) availability baseline computational algorithms. To address these limitations, here we propose enhancement...
We evaluate the performance of federated learning (FL) in developing deep models for analysis digitized tissue sections. A classification application was considered as example use case, on quantifiying distribution tumor infiltrating lymphocytes within whole slide images (WSIs). model trained using 50*50 square micron patches extracted from WSIs. simulated a FL environment which dataset, generated WSIs cancer numerous anatomical sites available by The Cancer Genome Atlas repository, is...
Multiple Sclerosis (MS) is a demyelinating disease of the central nervous system that affects nearly 1 million adults in United States. Magnetic Resonance Imaging (MRI) plays vital role diagnosis and treatment monitoring MS patients. In particular, follow-up MRI with T2-FLAIR images brain, depicting white matter lesions, mainstay for activity making decisions. this article, we present computational approach has been deployed integrated into real-world routine clinical workflow, focusing on...
Abstract PURPOSE Robustness and generalizability of artificial intelligent (AI) methods is reliant on the training data size diversity, which are currently hindered in multi-institutional healthcare collaborations by ownership legal concerns. To address these, we introduce Federated Tumor Segmentation (FeTS) Initiative, as an international consortium using federated learning (FL) for data-private collaborations, where AI models leverage at participating institutions, without sharing between...
Glioblastoma is the most common primary adult brain tumor, with a grim prognosis - median survival of 12-18 months following treatment, and 4 otherwise. widely infiltrative in cerebral hemispheres well-defined by heterogeneous molecular micro-environmental histopathologic profiles, which pose major obstacle treatment. Correctly diagnosing these tumors assessing their heterogeneity crucial for choosing precise treatment potentially enhancing patient rates. In gold-standard...
Time to biochemical recurrence in prostate cancer is essential for prognostic monitoring of the progression patients after prostatectomy, which assesses efficacy surgery. In this work, we proposed leverage multiple instance learning through a two-stage ``thinking fast \& slow'' strategy time (TTR) prediction. The first (``thinking fast'') stage finds most relevant WSI area and second slow'') leverages higher resolution patches predict TTR. Our approach reveals mean C-index ($Ci$) 0.733...
Abstract Glioblastoma, the most common malignant primary adult brain tumor, poses significant diagnostic and treatment challenges due to its heterogeneous molecular micro-environmental profiles. To this end, we organize BraTS-Path challenge provide a public benchmarking environment comprehensive dataset develop validate AI models for identifying distinct histopathologic glioblastoma sub-regions in H&E-stained digitized tissue sections. We identified 188 multi-institutional slides of...
Abstract BACKGROUND Skull-stripping describes essential pre-processing in neuro-imaging, directly impacting subsequent analyses. Existing skull-stripping algorithms are typically developed and validated only on T1-weighted MRI scans without apparent gliomas, hence may fail when applied neuro-oncology scans. Furthermore, most have large computational footprint lack generalization to different acquisition protocols, limiting their clinical use. We sought identify a practical, generalizable,...
Abstract BACKGROUND Glioblastomas are arguably the most aggressive, infiltrative, and heterogeneous adult brain tumor. Biophysical modeling of glioblastoma growth has shown its predictive value towards clinical endpoints, enabling more informed decision-making. However, mathematically rigorous formulations biophysical come with a large computational footprint, hindering their application to studies. METHODS We present deep learning (DL)-based logistical regression model, estimate in seconds...
Abstract BACKGROUND Diffuse astrocytic glioma are common and aggressive malignant primary brain tumors with grim prognosis. Artificial intelligence (AI) has shown promise across predictive, prognostic, diagnostic neuro-oncology applications, towards improving patient management. However, clinical translation deployment hampered by AI models’ requirements for explicit acceleration cards, which not typically considered in environments. Here, we seek the execution of models such...