🚨 Pulse of Innovation 🚨

First Human Patient Received a Neuralink Implant called “Telepathy”

• A quadriplegic patient was implanted with Telepathy, a coin-sized implant around the area of the brain that controls motions and will wirelessly transmit data from the brain to an app where the signals will be decoded, eventually developing a machine learning model.

• Down the line, Telepathy, the Brain Computer Interface, will be used to control the brain function that helps the patient improve motion, benefitting patients such as those suffering from ALS. A registry is set up to increase patient participation.

Google’s AI Learning to Diagnose Patients

• Real-world data from over 100,000 physician-patient dialogues, ICU medical notes, USMLE clinical scenarios, and simulated clinical scenarios were used to create an Articulate Medical Intelligence Explorer (AMIE) to help diagnose.

• The primary purpose of the AMIE is not to replace human physicians but to aid in diagnoses, particularly during wellness visits, and also to help patients better understand their symptoms and associated diagnoses.

🧑🏼‍🔬 Bench to Bedside👨🏽‍🔬

AI can detect factors associated with blood loss in cholecystectomy

• A subset of 243 surgical videos was used to train an AI-powered computer vision (CV) model, which was then used to detect various surgical action triplets of instruments, motions, and anatomic structure

• The trained CV model was then used on all 243 videos and identified factors associated with low v. moderate blood loss, which were longer use of irrigator, location of irrigation, and surgeon experience

AI-based risk prediction for early detection of pancreatic cancer

• From 55 US hospitals, clinical and demographic data of over 1.5 million patients, 35,387 of which had pancreatic cancer, was used to develop AI models

• Patients received periodic risk assessments by AI models and were followed for up to 18 months after their last risk assessment

• Among the pancreatic cancer patients, the algorithm identified 35% to be high risk 6 to 18 months before their actual diagnosis. The current screening tools only identify 10% of pancreatic cancers.

🩺 Start-Up Stethoscope 💵 

Synchron’s Brain Computer Interface (BCI) could be the first to receive commercial approval.

• While Neuralink is beginning its clinical trials, Synchron has significant headway with the implantation of its device in 10 patients, 4 in Australia and 6 in the United States, and is now in talks with the FDA to seek commercialization of its product.

• Unlike Neuralink, the Synchron device called Stentrode is implanted via the Jugular vein. It works by wireless transmission of the brain signals to its app and focuses on paraplegic patients such as ALS. Given the head start it could become one of the first FDA-approved commercially available product

Vital Interaction, specializing in patient retention, secures $15 million in series A funding

• Vital Interaction’s proprietary product scans practice management system for data such as clinical and billing information to strategize customizable targeted patient outreach campaigns for reactivation and retention of patients

• Currently, over 13,000 providers use Vital Interaction with annual outreach to over 5 million patients

🤖Patients First, AI Second🤖

Biden administration on use of AI for improvement of health outcomes

• On the Biden administration’s initiative, 28 leading healthcare providers and payers announced voluntary commitments to the safe, secure, and trustworthy purchase and use of AI in healthcare.

• It reiterates the administration’s emphasis on patient privacy, risk identification, and mitigation, as well as the development of methods for the responsible and safe use of AI for improving health outcomes

👩‍⚕️AI in the Clinic 🏥

Community Health System (CHS) implements generative AI through Google Cloud integration.

• CHS, one of the larger health systems in the country, spanning over 15 states, has migrated its data to the Fast Healthcare Interoperability Resources 1 (FHIR)-based clinical data platform of the Google Cloud

• This migration will enable better use of generative AI, improving clinical documentation, administrative tasks, and patient information distribution.

Stanford uses AI to navigate health information sharing with teens and their families.

• A natural language processing AI model identified inappropriate sensitive content in the patient’s medical records. Another AI model was used to detect inappropriate guardian access to teen’s accounts to view such sensitive information.

• The combination of these two AI monitoring systems was designed to provide a comprehensive solution to optimally support teen and family engagement in their healthcare.