Our Research Goals
Our laboratory leads in the advancement of new computational methods for digital pathology, medical imaging, clinical text mining, and multi-modal AI. Our team strives to unlock valuable insights from large, complex, and heterogeneous biomedical datasets. We transform this intricate data into actionable knowledge by providing researchers and clinicians with intelligent tools that enable a deeper understanding of this data and enhance clinical decision-making processes.
We have close partnerships with numerous clinical departments at Dartmouth Health, in particular the Departments of Pathology and Radiology. These multifaceted collaborations enable us to build novel computational methods, advance the field of digital pathology and radiomics, and transform how clinicians review and interpret medical images. Our team is dedicated to expanding the role of AI in medical research and clinical practice, while upholding a deep commitment to diversity, equity, and inclusion in the field of precision health.
Research Highlights
Precision Cancer Diagnosis and Care with AI
The incorporation of AI into cancer care holds tremendous potential for improving patient outcomes. Our laboratory is at the forefront of integrating deep learning methods into cancer diagnosis, prognosis, and precision treatment. We build AI models to identify indicative patterns in clinical data and provide novel insights to aid medical professionals in their decision-making processes. Our research spans a broad spectrum of diseases, including breast, colorectal, lung, and renal cancers. Our work has not only improved the accuracy of diagnoses but also enabled the identification of novel biomarkers and the prediction of prognosis and treatment responses.
Innovations in Histopathology Image Analysis Using Deep Learning
Our laboratory is focused on building innovative methods for histopathology image analysis using deep learning technologies. We have pioneered the development of new methods for histology image characterization and the identification of digital biomarkers based on histological features, utilizing novel computer vision techniques. These innovative methods have enhanced the precision and efficiency of histopathology image analysis, resulting in improved diagnostic accuracy and more informed treatment plans.
Pioneering Developments in Foundational AI Research
Our laboratory is heavily invested in advancing the foundational methods and techniques that underpin new advancements in AI. This foundational research forms the backbone of our more specialized work on biomedical applications and is critical in driving broader impact and innovation in AI. We have developed new algorithms, models, and methodologies in machine learning, natural language processing (NLP), and computer vision that can be leveraged across a wide range of AI applications.
Broadening AI's Reach: Real-World Healthcare Applications
The research conducted in our laboratory transcends the boundaries of specific diseases, exploring pragmatic applications of AI in healthcare. For example, we have devised methods for detecting substance use through social media, offering a broadly applicable tool for screening and outreach. In addition, our work has enhanced communication in primary care settings by automating the identification of medication references in visit conversations using NLP. We have also contributed to acute care by detecting pneumoperitoneum on chest radiographs with deep learning models. These diverse applications aim at expanding the use of AI in healthcare to benefit various clinical disciplines and improve patient care.