BETH ISRAEL DEACONESS MEDICAL CENTER

AI in healthcare includes using AI to analyze scientific literature, training robots for surgery, FDA developing AI regulation frameworks, ACR launching an AI quality registry, TrialGPT matching patients with clinical trials, combining LLMs improving diagnosis accuracy, AI bias affecting clinical outcomes, Mount Sinai opening an AI and human health center, Abridge partnering with Beth Israel to evaluate AI summaries, and Nvidia increasing AI-enabled robots in health systems.

Researchers developed an algorithm to assess how well Alzheimer's disease models mimic patient brain function and gene expression, identifying shared pathways crucial for drug discovery. The 'Alzheimer's in a dish' model, which accelerates drug testing, was confirmed to accurately reflect human brain changes, enabling rapid and scalable drug testing.

HIV research has transformed the infection from a death sentence into a manageable chronic condition, affecting 1.2 million people in the U.S. and nearly 40 million worldwide. Key advancements include antiretroviral therapy (ART) and pre-exposure prophylaxis (PrEP), with ongoing efforts focused on developing a vaccine. Insights from HIV research have also significantly influenced the development of COVID-19 vaccines and treatments for other diseases.

A new algorithm confirms that 3D models using mature brain cells accurately mimic human brain biology, enabling rapid and accurate drug testing for Alzheimer’s disease, identifying 83 dysregulated pathways and highlighting potential drug targets.

Despite access to obesity drugs like Wegovy, some patients, like Danielle Griffin, find they don't lose weight. Experts estimate 20% of patients may not respond well to these medications due to genetic, hormonal, and brain regulation variability. Alternative treatments and lifestyle changes are suggested for non-responders.

Used MIMIC-III and eICU datasets for sepsis treatment model training, with 80% of MIMIC-III data for training and 20% for testing. Implemented Double-Dueling-Deep-Q-Network (DDDQN) with Highlight coefficient and over/under estimation monitoring for offline RL training.

IVBM's Boston event in 2024 focused on 'Elevating Value in Cancer Care,' discussing immunotherapy challenges in NSCLC, HER2 targeting in breast cancer, predictive biomarkers for sarcomas, and pharmacy decision-making. Panelists highlighted the need for optimized dosing, comprehensive biomarker testing, and equitable access to immunotherapy, emphasizing a more thoughtful approach to advancing cancer care.