OncoPower

OncoAI, our AI-based engagement platform is rebranded as WittyAI, given we have a large number of patients with non-cancer indications using it. It is also integrated with Open Evidence to fetch data supported analysis to patient's questions Try it today: https://wittyai.org/signup.

New updated streaming version of https://oncoai.io is live. Leverage your medical records to build a personalized OncoAI Health Agent. You can upload your reports and scans to get guidance for discussing with your doctor. Try it FREE today at https://oncoai.io/signup #GPT4 #gemini #oncopower, #healthcare #diagnosis, #Chatbots #chatbot #LLM #LLMs, #ArtificialIntelligence

Excellent work. CORAL: Expert-Curated Oncology Reports to Advance Language Model Inference https://lnkd.in/g2y7yJHc

Our OncoAI is a signatory of pledge to Build AI for a Better Future Read the pledge here https://lnkd.in/eb9i-xQ3

COMING SOON: 100% evidence-based physician-grade guidance coming on our https://oncoai.io, a personalized Digital Health Agent platform. Try it FREE today at https://oncoai.io/signup and join thousands of consumer health members. #gpt #chatgpt4 Glass Health, Gemini, OpenAI, OpenEvidence

Here is an interesting article. It argues that current regulations for AI in healthcare focus too much on processes, rather than patient outcomes. This could lead to delays in approving new AI tools that could benefit patients. The authors propose a new approach that would require companies to demonstrate that their AI tools improve patient outcomes before being used in healthcare using rigorous clinical studies, such as randomized controlled trials. The problem is currently randomized clinical trials are very hard to complete because of fewer patients participation. Finding the right balance between regulation and innovation is essential for ensuring that AI is used safely and effectively in healthcare. AI in healthcare has the potential to revolutionize patient care, but overly cautious regulations could stifle innovation. https://lnkd.in/gFjFSUbB #AI #healthcare #regulation #innovation

Data from 100 patients with strokes was given to GPT-4 & an experts: 1) GPT-4s treatment suggestion agreed with the expert in 80 cases, in 5 it was right & expert wrong, in 5 expert was right & AI wrong, 10 were judgement. 2) GPT-4 predicted mortality better than specialist AIs., 3) GPT-4 mortality predictions separated patients for mortality a year forward. Of 25 with highest risk, there were 13 deaths, compared to 6/75 in the rest #llm #llms #ai #healthcare #stroke Evaluating GPT-4 as a Clinical Decision Support Tool in Ischemic Stroke Management: https://lnkd.in/espvYy4F

More progress is needed with vision models. After we added vision model capabilities to OncoPower's AI app https://about.oncoai.io, we have seen a lot of patients upload scans and reports to get AI-related guidance. But it is incredibly hard to meet patients' expectations vision model's responses compared to a traditional chat like GPT-4. We yet to see a medical LLM (hopefully Truveta) that can match GPT-4 which is still the best. #LLMs #GPT #GPT4 #healthcare

Truveta Language Model (TLM) (from Terry Myerson and his team) is the most unique and fascinating project among all medical LLM initiatives that are underway. Its exclusive focus on EHR source data, size of its data set (reportedly 100 million+ patient journeys) and the involvement some of large health systems in US provides significant competitive advantages over others including Med-PaLM M (MPM) and Hippocratic AI. Several AI initiatives in healthcare including our own (https://about.oncoai.io/) are eagerly waiting for their GTM plans. #llm #healthcare #patients. #ai https://lnkd.in/ggMK6QgB

This application of LLM for informed consent is interesting and provides a template for using LLM in other patient facing use cases (Ex- e-Consent using LLM generated video). They have addressed the bias, and ownership issues pretty well. #llms #chatgpt #gpt4 #healthcare #oncology #patientcare ?https://lnkd.in/eVV8xwvu