What's New In Surgical AI: 10/28 Edition
What's New In Surgical AI: 10/28 Edition
Shakeup in AI documentation, and a Recap of Digital Neurosurgery
Welcome back! If you’re new to ctrl-alt-operate, we do the work of keeping up with AI, so you don’t have to. We’re grounded in our clinical-first context, so you can be a discerning consumer and developer. We’ll help you decide when you’re ready to bring A.I. into the clinic, hospital or O.R.
Table of Contents
📰 The News: The News: Some New Tools, Words of Warning, and Wearables
🤿 Deep Dive: Report from Digital Neurosurgery 2023
📰 The News: Some New Tools, Words of Warning, and Wearables
The clinical documentation ecosystem is getting frothy…
As nicely pointed out by @rikrenard on X, there seems to be a looming shut down of Orthopedics AI documentation company, Robin Healthcare. Yet Simultaneously, Kaiser is doubling down on their AI investments- they announced a partnership with Nabla while simultaneously investing in AI startup Abridge.
Interesting play by Kaiser - perhaps a thesis that there will be many players in this world, it’s tough to predict who will come out on top, so best to hedge.
(As a brief aside, we built dotphrase.ai, yet another AI clinical documentation tool, to be super easy to use and wildly cost effective. But we are biased 😉)
How many hours of your life have you spent editing surgical video for a talk or journal submission? What if AI could give you hours back in the future. Descript's AI video tool looks too good to be true. They’ve already changed how people create video - automatically scrubbing any “uh” or “umms” in the audio. But now they allow you to edit the words you’ve said (via text), and it automatically changes what is said in the video. Next level.
Always a good read, Eric Topol with some words of caution for physicians: shape AI or someone else will. This is a belief Dan and I have held for years. AI is coming - just look at this survey of hospital CIOs/CTOs. Who should decide how AI is used in clinical care settings? If we don’t ensure providers (and patients) get a say, they won’t.
Two new AI wearables to talk about. The first is Rewind Pendant, which would (selectively) record and transcribe your life, for you to reference or chat with in the future
.
The second is Humane’s AI Pin which premiered at the Paris fashion week.
Is healthcare the perfect domain for wearables? We frequently have both our hands in use, and especially in the inpatient setting are often away from a computer but needing clinical information. Something to think about…
A bit more on the amusing front, Meta launched generative AI stickers, allowing users to create their own little emoji’s to attach to Messenger conversations. Unfortunately the screening on these wasn’t great, and led to some questionable AI generated stickers.
Lots to go over this week. AI moves fast. We help you keep up.
Deep Dive: Report from Digital Neurosurgery 2023
If you’re wondering why the newsletter has been sparse over the past three months, let me explain. We recently undertook a 10-month-long process to found a new meeting, which we called “Digital Neurosurgery” (www.digitalnsgy.com). This meeting, with over 150 attendees and speakers, was a resounding success and created a series of impactful opportunities to reimagine how neurosurgery, and surgery writ large, interacts with the broader AI/ML community.
We will post some of the talks shortly - from luminaries such as Craig Mundie and Sebastien Bubeck $MSFT, leaders at GOOG 0.00%↑, medical strategics, MDT 0.00%↑ and ISRG 0.00%↑,
I want to a take a few minutes to share the inside story of why we thought yet-another surgical meeting was needed, how we went about creating this meeting, and what we hoped to achieve. Lastly, I’ll outline what will follow from the meeting.
Why create a new meeting?
Our mutual interest in all things surgery, data, and computer vision should be no surprise to readers of this newsletter. Yet, even after dozens of abstracts, papers, awards, and so on, we felt that publishing another paper in a surgical meeting wouldn’t move the needle. There is a great deal of phenomenal work happening in surgical data science, “surgical AI”, and many other “slashies,” but translating that work into impact remains challenging. We wanted to shake things up.
A few focused meetings bring clinicians into discussion with computational advances, such as MICCAI (and CLINICCAI), while remaining rooted in academic discourse. In order to impact the direction of our field, we have to look beyond existing structures. So, we sought to aim bigger: including venture, big tech companies, medical device manufacturers, and others who are typically left out of the conversation. Furthermore, we recognized that simply enlarging our view within medicine wouldn’t do justice to the vibrant community of computational innovation that has already brought products to markets in selected use cases. So, we invited voices from outside of medicine: sports, self-driving cares, and military/defense. These areas are farther along the innovation curve - we wanted to solicit and internalize their experiences.
How did we do it?
Well, the simplest answer is that we created a team - including one of the foremost members of our field, Dr. Sam Browd, an accomplished neurosurgeon scientist innovator and entrepreneur who has been walking the walk for two decades. We brought in other neurosurgeons with defined expertise, interest, and reach. We formed a small planning committee with a few kind souls who donated their time and effort to make sure every detail went off without a hitch (Caroline, Rick, and Amanda - much love to y’all). Oh, and we got some financial sponsors, without whom the event wouldn’t have been possible. Then we beat the bushes for attendees, speakers, interesting folks from all walks of the world. We did the logistics and legwork (many thanks to Stanford Neurosurgery, and many others for supporting us along the way). Speakers cancelled at the last minute, venues changed pricing, and all the usual joys. We used a couple technology elements that I would recommend (eventbrite, though others may be cheaper), and an app (Whova) that was excellent. We worked our tails off for the last four months to get everything into shape and used every resource and connection we had. It truly took a village
How did it go?
In one word, transformative. Having leaders from Neurosurgery (department chairs, AANS/CNS/Washington Committee leadership, etc.), engineering/industry, and business/venture in one room was powerful. The fact everyone was discussing the future and the ability for AI to change how we treat our patients was even more powerful. But don’t take our word for it. Here is some feedback:
“I have attended these types of meetings for 20+ years. This has been the best yet”
“I came here as a blank slate just ready to learn, and I have learned more in 3 days than I ever thought possible”
“You must do this again. It would be a tragedy not to”
“Everyone raves about the meeting. It must continue.”
Collaborations were started and new colleagues introduced. Even a year ago, when Ctrl-Alt-Operate started, people didn’t really care about A.I. How things have changed!
Rest assured, when tickets for 2024 open up, Ctrl-Alt-Operate readers will be the first to find out.
Now what?
Time to capitalize on the momentum from the meeting, both inside neurosurgery and beyond it. When we wrote our prior editorial we laid out our plan for this year, including the post meeting period. We want to achieve several aims with the consensus built around the meeting:
1. Catalyze “Organized Surgery” around the topic of AI. We are discussing the creation of a joint section / task force on AI to facilitate data standards and best practices and speak with a united voice. We aim to connect these efforts across surgical societies and other organizations. Lastly, we aim to bring hospitals and payers into the dialogue.
2. Accelerate multi-institutional research structures. We have founded one of these, the Surgical Data Science Collective (SDSC, www.surgicalvideo.io), oriented around surgical video data, but additional efforts are needed. We firmly believe that research organizations that can operate outside of the identity and data management space of a single hospital or academic institution will be the engines of progress. We demonstrated the progress of SDSC at the meeting, in the hopes of encouraging les autres.
3. Sharpen future conferences to accelerate cross-disciplinary innovation in ML and computer vision for surgery. We wrote that “new meetings can galvanize progress and create challenges by using a [Common Task Framework] to focus the attention of the field on clinically impactful use cases within neurosurgery… Neurosurgeons should partner with scientists to define [Frameworks] that are clinically impactful… and operationalize innovations in our field.” Expect this to be a major theme of our work in the next twelve months.
4. Increase awareness, adoption, and analysis of surgical video. In essence, we aim to persuade surgeons that they *should* be recording and managing their precious video data, as well as other preoperative data sources. One opportunity that we haven’t yet tapped into is the partnership with patients, who may be our best advocates for surgical quality and care improvement.
Feeling inspired? Drop us a line and let us know what you liked.
Like all surgeons, we are always looking to get better. Send us your M&M style roastings or favorable Press-Gainey ratings by email at ctrl.alt.operate@gmail.com