S3 E11: "Block Box 4: Our convo with Kwesi Kwofie and Kris Schroeder!"

Okay. Buckle up. Next stop, Orlando. Okay. But I get to pick what we listen to. 

No way. Driver's choice. That's the rule. Okay. How's this? 

No. I'm not in the mood for that. Okay. Okay. How about this? 

I do love it, but I've just heard it too much. Did what? Listeners, he's trying to say Oahu. Wait. What's that? 

Oh, yeah. I heard about this. It's that podcast with those two guys. Oh, yeah. And it's about regional anesthesia and POKUS. 

Right? What's the name of it? Block it like it's hot. What's up, loyal listeners? You're in for a treat today. 

Two fantastic guests that we interviewed in the block box at the AZRA Spring meeting. So many nuggets to take home. First up is Kwasey Kofi from Dalhousie University who brought his dad joke a game and left us with a lot to think about with respect to needle nerve contact and nerve injury. And now when you start talking, you should get hear you. You can and you can hear yourself okay? 

I can hear myself, and I can hear you. Do I really sound like this? Yeah. You you sound great. Oh my goodness. 

You sound great. Yeah. Amit, we are so privileged now to have with us Kwesi Kofi. Of the Dalhousie University in Halifax, Canada. Oh my god, guys. 

I am so honored. I cannot believe it. We have been talking about Kwesi and the work that he's been putting out from some of the early episodes on the podcast. Right? So Oh, absolutely. 

And we're both privileged to have known Kwesi for a number of years. So to have you on the podcast is like a double whammy, friend and someone who we've referenced their clinical work. In fact, we I remember we got called out one day because we were giving him and his group so many shout outs. So, like, this is just like the you're promoing these guys only. Well, I love our I love our our team so much. 

Everyone works is working so hard, and we appreciate the the shout outs. Yeah. Well, you've got great productivity of your group over there, and you've had some real legends in in regional anesthesia that have come through and worked through that institution. So it's amazing. For us, it's lovely to have you call us tiny but mighty. 

Yeah. Right. You really you really are. I mean, you you have built that division, that that regional anesthesia group into a powerhouse. I wouldn't say that I built it, but I but No. 

You wouldn't because you're Because you're super modest. No. But I think we have a we just have a really great team of of staff and support and and and amazing fellows over the years. Uh-huh. We've just been blessed with people who really have just an absolute passion Yeah. 

For for regional anesthesia, taking care of patients, and and really the magic that regional anesthesia is and I often I often tell our fellows even after all these years, regional anesthesia to me really feels like like when I put on my my surgical cap, but I go and change in the morning like I'm going to, like, like, wizard school. Yeah. And you're doing these, like, magic tricks. You know? You you're you're you're doing a procedure and someone's arm goes numb, or you're doing a procedure and someone leg goes numb, and the patients go, wow. 

Yeah. And I I would say, you know, like, David Blaine or Harry Potter. He He can't do those kind of those kind of those kind of magic. So it really is like a it's like wizard school. I think yeah. 

I think it's like I I love it so much. In 22 gauge. Yeah. Yeah. You know, I've I've never heard anyone describe it like that, but some days you feel just like that. 

Right? It's magic. I still get that emotion when I'm injecting a small amount of medication in an area. I'm like, is it gonna work? And and then it it doesn't it it is amazing, but I've never thought about it. 

And the patients are amazed. Of course. It's not just about us. Right? Yeah. 

They're they're just they're just they're just bewildered just like they got stopped on, a, you know, a a magic trick on the street. Yeah. Absolutely. It's I I love the way you describe that, and I have I I have to admit, I have copied a bit of that. Like, I heard you say that once. 

I know it was about a couple years Oh, really? Oh, yeah. No. I remember you Oh, I believe it. I remember you saying that, and I I thought, that is exactly how I feel. 

And I when residents come through or medical students, and they're like, what do you they say to me, what do you like so much about this stuff? That is what I go to. I said, look. I've been doing this for fifteen year. You know? 

Twenty five years. Yeah. Yeah. I did. Was just counting. 

Who's counting? Beep. Years. And and I still when I get up in the morning, I look forward to coming to work. It's not boring, and yet I'm doing the same thing every day because it is it is magical. 

But I prefer this analogy. I've heard other people when they're talking about surgical specialists using the superhero analogy, which I don't I don't I don't this is not this doesn't that doesn't resonate with me, but this does. A 100%. It is magic sometimes. And I I don't like it. 

I love it. Oh, wow. Yeah. Yeah. Well, that passion that passion comes through in I mean, I've seen you work clinically, and I I I I loved watching you work when you were we were together in New York, but also in your research and and your educational talents. 

I I have you ever seen Kwesi teach your workshop? So I have, but I've also seen him lecture, and I've seen him present. Yeah. And I love his slides. So, you know, when you when you come to conferences when I come to conferences, you're at a stage now where, of course, you wanna come and get the new information, but also you wanna learn about teaching style and teaching method because we all become better when you embrace techniques from other people. 

I remember sitting at one of the Azure meetings, just watch your slides again. I love them because they were clean and they were bold. So, yeah, we've bigged you up enough. So now we're gonna now we're gonna knock you down by Don't worry. Don't worry. 

Now we now we need to hear a joke. Okay. See A joke? Yeah. I I I often I often tease my my daughter once once in a while and and tease her with little little dad jokes. 

One that one that really was a groaner for her the other day was what do you what do you give an astronaut that's grumpy? Because this is really good. I could tell with some of you. What do you give an astronaut that is grumpy? I don't think. 

Give him some space. Oh, that's so good. Okay. I'm loving that. But my my little piece of space. 

Did the what did the the drummer name his daughters? Oh, Anna one Anna You got it. Yeah. Oh, that's the first time I've ever seen you actually get the joke. Yeah. 

That's right. Yes. Was the punch line. Yeah. Was was impressive. 

Anna one Anna two. Yeah. I was gonna say, but I'm spinning at that side. One, Anna 2. I don't know. 

Dental. No. No. Oh, very good. Okay. 

That's awesome. And I love the fact you you you had one straight off another say, listen. Let's talk about regional anesthesia. So one of the early papers of yours that kind of got us thinking was talking about nerve injury or doing a particular procedure, and you did some early cadaveric work with stains and actually made us realize that sometimes we thought we were distant from the nerve, but actually we're not that far. Actually, some of our needle placements may be more deleterious than we realized. 

How did you get interest in this work? And and tell me how it's changed your practice. Well, really, I was I was I was inspired. I our group was inspired initially by, by the the work of Brian Seitz and and his group who were looking at periplexus and intra plexus approach blocks, which you guys have talked about before. And one of my my mentors, doctor Jennifer Serb, really was like a bull in a China shop Uh-huh. 

With this idea. And she she really started this process of getting these together, and we collaborated very closely on this over the years. So, you know, it's really the inspiration from doctor Seitz and and Jennifer's amazing, amazing, recently retired. Okay. So for those people who haven't had a chance to read some of your early papers, tell us a little bit about what you found. 

So Some early work. So so we were using small volume injections of India ink. When I mean small, I'm I'm talking about point one mils. These aren't clinical volumes. Uh-huh. 

They're small volumes. And the reason why you have to use small volumes is because what we're trying to do is we're trying to find out what is the interaction between the needle and the and the nerve itself. Because it's very difficult to be able to discern that. So we use the injection to try to tell us exactly where the needle was at the time of injection. And we know that the injectate actually doesn't pass through the epineurium and the perineurium. 

So it's not something that's diffused. Not oh, okay. It does not diffuse. And and and we investigated that in some some early studies. But we the the first cadaver study that we did after demonstrating that the the ink didn't pass was the study looking at interplexus and periplexus interscalene blocks. 

Because we know that whether you actually get in between the c five and c six nerve root or adjacent to the the envelope that eventually becomes, I think, the the space around the superior trunk. Whether you're into that or not, the block actually works clinically the same. Uh-huh. So if we know that those would both be clinically effective, well, maybe we can see if there's a difference between the likelihood of getting into the the nerves whether you're doing those injections or not. So we did cadaver a cadaver study, and we randomized either doing an injection by a periplexus approach or an intraplexus approach. 

And then we took the histology samples and tried to see where the actual dye end up. And what we found is if you did a periplexus approach, which is clinically about the same effect, that there was no cases where the ink was deep to the epineurium. Mhmm. In a significant number of cases, the in the intra plexus group, the ink was deep to the epineurium. But more importantly, in twelve percent of the cases, the ink was actually deep to the perineurium. 

So it was in the fascicle, and you could actually see Wow. The ink disrupting the that that dense connected space between the axons. And the only way you could have got that stain there was by having the needle placed in that location. 100%. What would and There's something more interesting, though. 

Okay. There's something the the the even more interesting thing is that when you looked at where the ink went after you saw the disruption in the fascicle, what you'd see is that the ink in a single nerve, only a 100 microliters of ink would actually be deep to the perineurium in every single fascicle in the nerve Oh my. Without being in between the fascicles. There was no ink between the fascicles. That's what I was gonna say. 

That was that was the most impactful thing, and you have a amazing micrograph in the publication showing ink, you know, inside the shell of the epineurium on this one. But oh, wait. Also here. Here. Here. 

Here. Here. Here. And it's not like a wash or stain all over the place. It's quite clearly and it is an it was an moment for a lot of us when we realized, oh, these things are interconnecting. 

Yeah. And we just don't don't see it, don't appreciate it. But when you put a very small amount of ink in there, it gets to all these fascicle. Now and we since we know local anesthetics are neurotoxic Mhmm. The job of the perineurium is to protect those axons. 

That's not a good thing. Yeah. Absolutely. And I think the one of the ways the other thing that that that I I think was an moment for me is not only that these these these axons are they're kind of changing dance partners as they as they sort of go through. Right? 

God, I love that. That that that is But it's happening over such a short a short distance, right, that only a 100 microliters of paint can bathe all of those. So, you know, you can imagine the axon that's going going along, and all of a sudden, it's like it's like taking, an off ramp to another fascicle and then taking another off ramp to another fascicle. And they're constantly changing dance partners as they're as they're driving along the highway. Like, I think it's I think it's I think it's very Yeah. 

Yeah. Yeah. It was it was a real moment when when when I came to understood that. It was actually well described by there was a night paper bike. Yes. 

And Quinn Hogan had a great I I use his his from, like, 2008 or something because the best picture I've ever seen that shows a little cartoon that shows these interconnected webs. Yeah. And I use that slide all the time. Me too. Yeah. 

Well Now, Chris, I had a question for you. Got a clinical scenario to pose to you. K? I got I got a colleague. We'll call him Kyle. 

Alright. Kyle says to me, I got a busy hand list today. I can't wait around for this block to take thirty minutes. So I'm gonna do my super quick. I get a block, and I'm and I'm gonna put the. 

He's goading me. I know he's goading me. Middle into the middle of the bag of marbles. Okay. Okay. 

Yeah. Let Yeah. Because I'm a I know you wanna blow them up. Because I know you wanna blow up those fascicles. Because I wanna because I wanna get home. 

You dirty dog. Let's answer the question because I'm not gonna I'm not gonna say anything until you deliver the question. Let's hear the answer, and I'll tell you the real story. And I say, but, Kyle, you you can put low kleinocytic outside the brachial plexus sheath, and a block still works. And he goes, yeah. 

But I I I I really need this block to work fat. What's your what do you say to Kyle? So Am I Kyle, by the way? It's a a leading leading question. So so so this has been, like, well described. 

And it it it in the literature, there are a lot of papers that that are actually advocating these types of blocks. So the so the classic one that I that I like to talk about is the is the targeted intracluster approach, the Blake brachial plexus. We call it a cluster of some other Yeah. That's a cluster something. So so But that sounds it sounds bad. 

I there's a There's a handful of different studies that have looked at this, and they've compared to, like, every other brachial plexus block you could imagine. Uh-huh. And what they see is, you know, about five minutes faster onset Uh-huh. Compared to, you know, every other brachial plexus block that you would imagine. But but we looked at and when they do injections, they're doing somewhere between in their studies, they did somewhere mean of somewhere between five and seven injections. 

Right. But they did up to 12 individual injections, cluster injections That sounds like a in some patients. Lot of injections. So so we did another cadaver study. Uh-huh. 

And we did a single injection Yeah. Supraclavicular cluster injection. This is an intentional injection deep to the perineum. That's our goal is to inject deep to the perineum. So we're we're trying to do exactly what they're describing, but only once. 

And, again, we looked at histology just like we did. We only injected two hundred microliters of ink, so 0.2 mils, small volume. And we saw the exact same findings that we did except even higher. So those were unintentional injections, the interscaling. We were trying to stay out of the nerves, and we saw twelve percent intravascular injections. 

Here, we're trying to get in. And, actually, you know, my teaching when I was early on told me that, oh, you know, these nerves have, you know, all this connective tissue and fascicles. They're gonna bob and weave out of the way. Yeah. Don't worry. 

You could put your needle there because you won't go into them. Running back. They're kinda, like, bobbing and leaving out of the way. I hope you this is video podcast. I love the movie. 

He's good on the dance floor for sure. I've seen him dance. It's awesome. Bubbly pops first. So what we found was twenty four percent incidence of of intravascular injection, subperineural injection. 

And and the images look the same. You can see this disruption of the of the axons in the fascicle, and that same subperineural injectate throughout the entire nerve with only point two microliters of of ink. The only way that you can get that is buying is is by have your having your needle deep to the perineurium. And you can see on multiples on on multiple of our histological slides that you can see the fascicles disrupted. And you can even see in some where, you know, you could it looks like the perineurium's ripping. 

It's making me feel scared just hearing this. That's not a good that's not a good idea, like, ripping. I mean, the the the good news is that that, you know, nerves are very resilient. Yeah. And the likelihood of nerve injury actually is very low. 

You know? Probably even pretty low if you're doing bad things. I would call those bad things. But you don't need to do that. Okay. 

And if it was my nerve Yeah. I I I think that's a a bit invasive. You know? I I think if you're putting your nerve you're putting your needle into a position that's potentially injurious like that, I would say, you know, you've got some nerve. Oh. 

Even when we weren't expecting it, so he drops it. Okay. He was made for this podcast. He was. I need to ask a question. 

Do you do supraclavicular brachial plexus blocks now? Yes. They're they're I would say they're not my preferred technique. I prefer an infraclavicular block for a bunch of reasons. One related to the studies that we're talking about. 

The other because, you know, infraclavicular brachial plexus block, there's a small subsection of blocks. I call them one and done. Mhmm. Yeah. You you put the needle in one position. 

You you put your local, and you're done. You know? Like Drop the mic. You know, an inner inner scaling block is one and done. Right? 

You don't need multiple places. Interscaling, one and done. No video gaming. Yeah. Yes. 

Yes. Yeah. Infraclavicular, one and done. You know, pop sciatic, one and done. You know? 

So I have to tell the fellows that you love the one and done. Very efficient. Very effective. So and supraclavicular brachial plexus is not necessarily a one and done or tell me. Well, I mean, I don't I don't know of anyone who's really described a one and done, in the supraclavicular brachial plexus. 

Well, I know one guy said or he not. But he didn't he denies it now. I remember Jeff came to a meeting in London, and he put a video. This is right around the time we would we first met. We were thinking about this thing. 

He showed a video of his needle going in the eight ball corner pocket. Inject the stuff that lifted up beautifully. He's like, see? How that was amazing. You don't need to go into the plexus. 

And I was like, so hold on. So you did that. That was your only injection, and that worked completely for surgical anesthesia. He's like, yeah. Yeah. 

And when I asked him about it afterwards, he denies that he actually said that. But where do you place your needle around the nerves when you're doing a supraclavicular brachial plexus block? I I find if you're if you're if you've if you only are placing and so I used to start just by trying to place a needle into the corner pocket Yeah. And then and then above and sort of having that kind of sandwich effect. One below, one above. 

Goddamn. Sand. Clinically, I find that those often don't didn't set well. And I and I sort of became very disillusioned Correct. With supraclavicular So how did you make it work? 

I will I will say the other thing that I worry about answer the question very well. Is that is that corner pocket because and people often describe the supraclavicular space as as, like, trunks or or divisions, but they're but they're not. They're roots, trunks, and divisions depending on where you're looking. And particularly in that corner pocket, that's the t one nerve root that's coming right out there. You can you can see it. 

And there's a couple great papers that describe that. So it's a combination of both. So you can get right into that t one nerve root in that corner pocket, which you often can't see wells, which I think is a I worry about a little bit. So my my goal is to do a corner pocket injection watching for that t one nerve root. And I'm usually doing an injection. 

I try to follow the superior trunk down, inject just below the superior trunk into that into that plexus. Between the superior trunk and the middle trunk. Correct. Sometimes we're not there's an arch that can sometimes run through there. For sure. 

For sure. Dorsal scapular artery. And I and I I always image for it. And I so I I wanna make sure that I that I see it. But how can you be sure that that injection that you're taking between the superior trunk and the middle trunk isn't crossing any any areas or any any epineural components? 

I mean, how do you know? So that's the so that's the that's the ultrasound part. And then the other part, I I think, comes back to multimodal monitoring. Yeah. Okay. 

Yeah. But Sounds like more. But but but sometimes patients don't have, like, a very sometimes you go, you know, c five, c six, and then there's a very small distance, and then you're suddenly in the suprachoroid complex. You don't always get that beautiful suprachoroid trunk. So you can often follow it. 

Sometimes you can't. The other thing that you can do is is is I'm a big fan of of kind of dynamic scanning as you're as you're moving your needle, you can often see so there's nice paper that that talked about sort of direct needle approaches and tangential needle approaches, and I often talk about about this. But as you're advancing your needle, you can see whether or not you're actually disrupting the neural structures or whether they're kind of rolling around you. I know exactly. Kind of dynamic scanning, I think it's sometimes hard to explain that to to to residents. 

But I spent a lot of time talking with the fellows and really, you know, trying to get them to understand those subtleties of of what you're seeing of the tissues as you're disrupting them. And I I often talk about using your needle as a as a tool to be able to sort of the side of the needle to be able to push things to the side out of the way to be able to get into the fascial plane where the nerve lives. Because, you know, in in in the end, that's what we're trying to do. Yeah. You know, delicately deposit the local anesthetic into the fascial plane where the nerve lives without damaging the essential architecture of the nerve. 

I gotta tell you. Listening to this conversation, if I was a first year anesthesia resident, I'd be like, why on earth is supraclavicular even a thing? Mhmm. Wait. If I've got one and done with infra, why are we taking the chance up here? 

Yeah. If if, a, it's hazardous, b, doesn't always work that well because Yeah. The plexus is spread out so far and, like, you gotta do multiple injections to make sure you're getting everything. I don't know. But you know what? 

What do do? I I really tend to tend to avoid them, and I usually use them only when I'm pushing the corner. But you you will get patients who have you know, like, maybe they have, like, a porta cat that's in the way. Sure. What have you but I but I agree with you. 

Is the super clap dead? That's that's what I I am so happy you asked that question because, look, somebody's written on our glass bot block box there. Supraclavicular greater than greater than greater than infra. So one of our one of our people that listed has said that that so I think it's a great question. When I was taught how to do supraclicular brachial plexus blocks, that was back in the day when ultrasound was had less resolution, and we were happy if we saw an artery pulsating at like a mush next to it. 

And we were told to stick the needle inside the mush and inject. So it went, and then that was great. And then sometimes the patient said, ow. And sometimes they didn't, but the blocks always worked. I was at a a meeting where professor Kijin Chin was speaking, and he was talking about this the the right way to do supraclavicular brachial plexus box. 

And the next day, I was meant to be doing a live demo from theater of what I normally did. And professor, Kijin Chin was saying, well, actually, I used to do this, but that's really bad. And, actually, we should be doing that. I was thinking, oh my god. I've been doing it wrong. 

How am gonna change? I started to get nervous from that moment about the technique I was using. And I went through a path thinking maybe I could do a little bit of inject local and say a hydrodissection, the nerves will move out of way. But that that doesn't really happen. Right? 

Yeah. So I think you're right. I find it very difficult to infuse about the supraclavicular brachial plexus plot now. In fact, I don't enjoy doing it. I've said it now. 

I don't enjoy doing the block because I think the only way to really get it to work is to be dirty and be intraclass, and that's bad. You kinda do. You kinda do. I mean, they they they He's they he's they they he's acknowledged that. To make it work nicely, you need to be naughty. 

It is. Yeah. And I and I don't like being naughty. No. What It's it's you have to you have to you have to take a a step towards towards risk. 

I mean, the the more you're you're you're getting close to the to the the epineurium when you don't need to, the the more you're potentially increasing risk. But So I okay. So should we should we just stop doing supraclavicular brachial plexus block? Are we gonna make a pact? Is is it a thing we're not gonna do them anymore? 

You're saying sometimes you might have to. Right? I I think we need we have a lot of tools in the toolbox. Yeah. But but, I mean, for me, Superclav is at the bottom of the toolbox. 

And people make it seem so e oh, it's so easy. It's it's the easiest one to see. Like, can see the right thing. Of the arm was the term that people use. It's the spinal of the arm because you're mashing up all those nerves, man. 

That's I think it's I think it's not easy. I think to do it safely Yeah. Yeah. And and effectively, you need to have to be very precise, conscious of where your needle tip is Yeah. Where the There's a lot of gnarly complex anatomy Yeah. 

I love the word. In that in that supraclavicular space. I think people do respect it enough. Right? I think that's true. 

But but but, Wendy, we also wanna we all wanna make these techniques sort of widely available so so so people can view that. And and even if you haven't done a deep dive into the into all of the the sort of complex anatomy within those those bundles that, you know, maybe you can still do it safely. And I just I struggle on a on a way where I would want someone to do that to me without that knowledge. So it just falls down to the bottom of the toolbox for me. Well, I think that's really interesting is thinking how comfortable would you be with one of your residents doing that technique on yourself? 

And I can say supraclavicular is not one I would be I would enjoy receiving. Let me pose a different question. Popatile sciatic. Mhmm. How how comfortable are you with that same question? 

Because I'll say that that is like the supraclav of the lower limb in the sense and I love a popatile. I think it's it when it when that local disperses within the perineural sheath and they get the owl eyes of both the nerves, like, warms the cockles of my heart. But Sure. But, man, it is tricky to get your needle safely into that sheath sometimes without contacting anything. Any thoughts on thoughts on that? 

Would you agree? Or is it I I don't know. I totally I totally agree. And and, you know, there's a lot of of risk. When you're using a lateral approach of of securing the common peroneal nerve, that I mean, that's that's the access problem. 

You know, it's sometimes the the nerve is sort of, you know, I call it kind of the snowman where there's where it sort of opens up the space for you in between. But most of the time, it's not. They're sort of side by side. And your needle's going straight for that common peroneal nerve. I often use a technique. 

Again, as I talked about before, using your needle as a tool where I'll come, and I describe this to my fellows, particularly when we're doing catheters, to come under the nerve, lift the nerve up, and pull the needle back until the tibial nerve drops off. And it creates a space in between. So I I use that to be able to to be able to get in. It works especially well when you have a stiff needle like a Tui. We do a lot of I do the same thing. 

And I'll tell I'll I'll say, okay. Now your tip is safely past Peyronia. You can't hurt Peyronia anymore. So Lift. Lift. 

Lift. Lift. Lift. Lift. Lift. 

Lift. And then finally, get it so that it's Yeah. It's now when you pop through the sheath because the needle is over the heavy. So you're underneath the common perineal. Correct. 

And then you're you're pulling to a point, and then you're you're pushing lifting the needle up, so then you aim to go over the tibial. Is that what you're saying? Well, because you've got common perineal sitting on top of your 100%. Needle, and then you're going above tibial and pop through at that point. There's one trick to help you to be able to do that effectively. 

So when you're coming in with the needle, if you come in shallow, almost like you're aiming just for kind of the top of the nerve. And just as you get to the nerve, then you steepen out at the end because what happened? Do the same thing. Do you? Oh my god. 

I called out the Gadsden maneuver. Okay. I wanna make sure I'm understood this. Describe exactly what you do then. So so so what you do is you come in with a with a shallow approach. 

So you're coming in I mean, you could do it in either direction, technically, but but I I would start. So I'm posterior to the nerve, meaning that you're closer to the skin. Okay. Okay. Yes? 

Yeah. And you're pointing to the area that's that's that's shallower. Okay. Yeah. And and aiming towards the the sort of top of the nerve, the shallow part of the nerve. 

And then when you get to the nerve, you steepen out, and then you come under the nerve and come. So the effect that happens when you pull back is the muscle that you're into is lifting the needle up for you. So you don't need to torque the needle. The muscle is doing the lifting, the heavy lifting for you. Yeah. 

Because you know how, like, you've got you told training, okay. Go this way. Go this way. Yeah. Yeah. 

Yeah. And now you if you're through the biceps muscle, you it's very hard to now Correct. Make a big redirection. But if you aim deliberately for this part and then the last minute dip down and then you let go of the hub, it's gonna wanna sort of Yeah. Yeah. 

Spring back that way. Oh, I think this is amazing. I I've I've I've taken something great from this, and I I think we need to see a video. Well, I I think we should co brand it now. Yeah. 

Yeah. Yeah. Yeah. The coffee gas. TK. 

Yeah. TK. TK? Yeah. Yeah. 

Yeah. Okay. Don. Yeah. Yeah. 

Don. I wanna I wanna give a a shout out to my my dad. My dad turned 80 this year. Wow. Happy birthday. 

Yeah. Amazing. And he he looks so good, and he's so fresh. I I I hope I look as good as he does when he's he's 80. My dad was a dentist, actually. 

Right. And I don't know if he wanted me to be an anesthesiologist, but but I I think maybe dentists use lots of local. Maybe that's what got me going. But he retired after thirty five years of practice. Wow. 

And all he got was a little plaque. Oh my god. He finished on a check. That is amazing. Right? 

And I didn't see it coming. You did a Ganson trick. That's, like, lead me along the guarded path. Boom. No. 

A dad joke about my dad. Yeah. Oh my god. That's next level. Meta. 

Wow. It's it's gonna sink in for ages. That was unbelievable. Wow. Thank you so much. 

That was amazing. Oh, buddy. That was Thank you. We'll be right back after this word from our sponsor. So, Jeff, I don't really know how to say this. 

Not again. Don't tell me you need a refresher on how to use the nerve stimulator. Very funny. As stimulating as that ancient relic is, you know I left that back in 2009 with my pager? True. 

Okay. So what's going on? So I was gonna say, I miss you, man. Recording the podcast in person was epic. Same here. 

Remember how much fun we had at the Asra spring meeting with our friends at GE Healthcare? Well, buckle up, buddy, because BlockBox is back. No way. You mean GE Healthcare is teaming up with BlockIt like it's hot again? Yep. 

This time at the forty second Ezra meeting this September in Oslo. Well, you heard it here first, listeners. Amit and I will be back in the interactive glass black box right in the exhibition area. This time with all new guests, interactive games, good vibes, and a GE venue system in the box so you can try out some of their slickest tools. Come hang out with us, ask questions, join the fun, and maybe even hop on the podcast. 

And knowing GE, I've got a hunch there may be something new in the pipeline. You're such a tease. If you wanna find out more, meet us at the BlockBox at Ezra this September. Can't wait. We'll see you in Oslo. 

And now back to our show. Next up, we chatted with Chris Schroeder from the University of Wisconsin about cheese curds and the use of artificial intelligence in medical practice and some of the ethical questions that are being raised. Okay. Amit, we are back, and joining us now is Chris Schroeder from Madison, Wisconsin, University of Wisconsin. Chris, thank you so much for joining us. 

Thank you, Chris. Yep. Thank you for having me. Listen. We're we're really excited because I was doing a bit of reading up, and there's lots of things that we can get into talking about your research interest, your clinical interest, but I'm sure you've read the sign on the door before you walked in. 

There is a prerequisite for being on the podcast. We need a joke from you, Chris. Yes. Okay. So I am very fortunate in that at the University of Wisconsin for a number of years was able to work with doctor George Arndt. 

So he's the inventor of the Arndt endobronchic blocker Oh, wow. Arndt airway exchange catheter kit. And so from him, you know, I fortunately was able to receive jokes nearly every day. Oh, he he was known Not just an innovator and inventor, but Oh, yeah. Oh, youngster. 

No. He's fantastic, but Uh-huh. I cannot tell you 99.9 of his jokes. Oh, okay. It's not suitable for N s y w. 

Yeah. No. Not now. But he has one, and it's it's terrible. Alright. 

So you're pre warned. But so what is a cat's favorite salad dressing? What is a cat's favorite salad dressing? I don't know. No. 

I've done anything. I apologize in advance, but it is rich. Thing is that it takes a lot of commitment. Yeah. Once you start that, Jake, know where you gotta go. 

You do. You do. And it's not funny. It isn't at all. Nope. 

But the made it funny. But I would've I would've messed up. I think you delivered that. Is ranch a big thing in The UK? Not massively. 

No. It is. I I've heard of ranch dressing, but because I'm so, you know, widely traveled. Where Chris is from in Wisconsin. Like, okay. 

So cheese is a huge thing in Wisconsin. Did you know that? Is a very American thing, but particularly in the Midwest, like, seems to be drizzled on all kinds of stuff. Wow. What's the what's the weirdest thing you've seen ranch dressing put on? 

I mean, for us, you know, I don't get that exotic, but the cheese curds, for sure. You gotta put them in the ranch. Cheese curds. I heard about cheese. I never tried them. 

What's the story with cheese curds? Story with them? I guess I don't know what's the story. What are what are cheese curds? So basic cheese that's curdled and and you put in ranch? 

No. No. No curdlings involved. You just you take your cheese and you fry it to make it more healthy. Okay. 

And then, you know, to cap off the health factor, you then dip it in ranch. Wow. Yeah. Yeah. I mean, ranch is delicious. 

I'm not gonna lie. I'm not gonna I'm not gonna sit here and pretend I'm not a ranch fan. No. Right. Okay. 

So it's a funny story. I visited you last year. Right? I think I came Correct. When we we did some stuff with your fellows and trainees. 

And I I went to there was a grocery store near my hotel. Went and bought some cheese curds and brought them home. Kids went crazy for them. And now every time I go somewhere, not not just to go to, like, Wisconsin, but every time I go, like, honestly, on this trip, Gigi was like, hey. Can you buy some cheese curds? 

And I'm like, honey, I I wish I could. I'm going to Orlando, not Wisconsin. But It is true that the curds come in two different varieties. Right? So you get the fried curds if you go to a restaurant, state fair, festival of some sort. 

But, yes, from the grocery store, you would get just like the regular curd. And then the important thing with those is that they're squeaky. Yeah. Yeah. When you bite into sounds like Indian cheese. 

Paneer is kind of like a a Oh, yeah. I think paneer is like a a a curd of sorts. Okay. Yeah. I'm struggling again with the mic here. 

This is one of the oh, now I'm back in the game. So cheese curds. I need to try cheese curds. We should we need to find some. Yeah. 

Okay. Come up to Wisconsin. Okay. I I don't know about Orlando. May not be the ideal place to find cheese curds. 

I wouldn't touch an Orlando curd. Yeah. Now, I wanna talk about your research interests. So Sure. Something tells me that you're interested in artificial intelligence. 

Is that something you've done some research into? Yeah. I think part of it is that I have been very fortunate over the years to work with some young people, so some young colleagues, at the time a medical student, and then now one of our residents who presented to our group with an extensive background and history and knowledge of artificial intelligence and how to use it. Right. Because I'm not specifically trained in computer science or and, you know, so I don't know how to do these things by myself. 

You know, what I was able to do working with these guys is kinda connect some of the dots between the computer experts Yeah. And, you know, some potential clinical scenarios where you might be able to use it. So what is your specific interest in artificial intelligence? Is it in anesthesia in general or specifically in regional anesthesia? So, you know, where we've used it thus far, you know, I think one of the first big things that we did was working with Josh Mazzewski. 

He is currently an intern at UCSF. But we used sentiment analysis, which basically is using AI to kind of evaluate the sentiment or tone of the words that exist. And so there was a publication that came out in RAPM, that was an evaluation of regional anesthesia for joint arthroplasty. They did all the work that goes into a meta analysis, collecting this group of people, pouring over the papers. But what we did is we used sentiment analysis of the abstracts that were included in that meta analysis. 

And in a period of seconds, we were able to reach the same positive conclusions related to the use of regional anesthesia. Wow. Really? Wow. That's neat. 

And then, you know, the next project then we did was with Nate Hurley, where basically what we wanted to do is see how large language models could help us with evaluating clinical scenarios. And I think the advantage to using the large language models is that certainly there's a lot of different AI platforms that you could use, but the LLMs are accessible. You know? Anyone can, you know, use ChatGPT or Yeah. You know, that's what I think makes that powerful. 

So, basically, we created a variety of clinical vignettes and then related to the use of anticoagulation prior to interaxial procedures. And queried chat GBT about, you know, hey. Do you think it's safe under these conditions to perform an epidural, spinal anesthetic, things like that? And, you know, we tried to even kinda tip the scales a little bit in favor of the large language model by prompting it, in some cases, to consider ASRA's anticoagulation guidelines. And even despite that, it wasn't really all that great. 

Really? More like It was more it was all over the place, or it was more conservative or more liberal? Or Kinda like flipping a coin. Like, really missed some important things, especially, you know, where the guideline asks you to consider multiple components. So for Warfarin in particular, know, there's a a timing perspective, you know, that you wanted to stay off of the Warfarin for a certain amount of time, but also have a normalized INR. 

That was one where it really, in particular, seemed to struggle. And where was it where was it taking its reference material? Where was that coming from? I mean, that's the thing with some of these large language models is you don't know necessarily. And even if you ask it for references, there's a chance that it would just sort of hallucinate where the reference material came from. 

So you don't you don't know. Yeah. That's interesting. Do you do you see that becoming more and more precise and powerful, though, as it has more data to look at? And that's likely the case. 

So we did another thing. This isn't as much related to regional anesthesia, but, you know, same sort of idea, creating these clinical vignettes and then related to perioperative transfusion guidelines. And so in that scenario, it actually performed much better and the performance improved with subsequent LLM iterations. So as the models improved, as they grew, it seems like it was able to better answer those questions. The other thing with that transfusion paper that we did is we had to take an examination that would be kind of at the level of a graduating physician related to questions surrounding transfusion. 

Oh, really? It did really well. In most cases, especially the more recent models would have easily passed, and in most cases, ex exceeded the scores that most of our trainees would get. Oh, that's interesting. So in an ideal world, as a practicing anesthesiologist, how do you think we should be using AI? 

And you where what we should be doing right now, and where do you see the future? Besides using it to write all my reference letters. He's joking. He's joking. Or maybe not. 

I don't know. Okay. Well, you heard it here first. If you have not tried using these large language models to, like, write letters of reference, I might suggest it. It's incredible. 

Honestly, I I I I I had to confess. I only started doing this recently. Right. Mhmm. So if you got a reference letter from me last year, it was bonafide me sitting down sweating over a keyboard thinking of things to say nice things to say about you. 

But I tried it recently, and I I uploaded this person's entire CV. Mhmm. And said this is what they're applying for. Woah. Go. 

And it it was better than I could ever have done myself. Yep. Now how much did you have to edit? You know, did you go back and you're like, that doesn't really A little bit. Yeah. 

Because it it just doesn't some tonal things. Mhmm. It didn't sound like me. It but what I'm learning about artificial intelligence is the output is very much related to the prompts that you give it. Yes. 

So presumably, as I've heard people to to to train the the model by saying, ask me, 10 questions about myself or ask me to write certain phrases so you understand my natural writing style and then use that natural writing style to do x, y, or zed. And because I I don't think I use high level prompts when I'm doing that. I will normally say, what do you think about this? Or can you give me some idea of how to respond to this? Or I'll write an angry email and upload it and say, can you make this sound but less angry? 

I would like to not be fired. Please. Please yeah. Mhmm. So so do you think do you see a role for you know, because we need to be trained how to give it the right prompt. 

Right? What what are what are your thoughts on that? Yeah. So a lot of times, what I will do is upload an example of a letter that I've written in the past. And then, you know, ideally, then you'd have the applicant CV, a personal statement even makes it better because then they're able to pick out and highlight those things that they think are important to discuss. 

Because a lot of times, you know, especially early career trainees will have things on their CV that just are totally irrelevant. You know, they'll be like, oh, well, they went to high school and, you know, whatever. It's like I was a life I was a lifeguard in high school. Right. Yes. 

Yes. Yeah. So, you know, ideally then, you know, between, you know, incorporating my style, their CV, a personal statement, I have been incredibly amazed by what it's able to produce. Okay. So now I've got two, again, slightly controversial thought processes. 

Number one Shocker. Mhmm. You upload that candidate CV to this the large language. And now it's out there. Who owns it? 

That's true. Who owns it? And and and what happens to that information? And and do we have an obligation to share the fact that we've uploaded that information? It's it's it's some consideration. 

Right? So who owns that information once it sends it out there? Yeah. So likely the purveyors of whatever large language model exists. However, I would contend that in most cases, there's not really any information that exists on a CV that is really all that private. 

You know? Yeah. It's probably out there in some way, shape, or form. Yeah. Yeah. 

Presumably, please please come through. We we're very lucky. Carolina is the big boss over here. Oh my gosh. Thank you so much, Greg. 

Reminded us some refreshing. Thank you very much. So okay. And I think that's probably fair because you might choose thank you, Jeff. So you might you might choose to put the information out on LinkedIn Mhmm. 

Or whatever else it may be. Okay. Now patient data, that's totally different. I mean Yeah. And you'd have to be incredibly cautious and, just thoughtful about how you would use it for any sort of patient related issue. 

Okay. What about academic pursuits? You might write something yourself in your own prose. Are we allowed to put that stuff into these models and say, can you make this sound different? Because I know there's a declaration. 

Many faces will say, gotta say I've used it. Is there a point when they'll say, actually, we're not gonna accept this publication because it's been used? What what are your thoughts on that? I mean, my thoughts are that there are some people who are just not gifted writers. Right? 

So if they've gone through, if they've completed the science portion, if they've done the evaluation, if they've given the draft an honest go, if they use these large language models to improve what's written Yeah. I think for many people, that's great. Certainly, it has to be edited. Certainly, scientific content needs to be something that's provided by the researchers themselves. But if you're using the large language model, I see this is no different necessarily than the AI that already exists in Microsoft Word that helps you improve your spelling. 

Yes. Yes. Yes. Yes. It's a good point, actually, Chris. 

I think I think it's just a tool to get you across the finish line when you've done all the scientific work. And why should someone who's maybe English isn't their first language or it's not a strong suit. Right? I mean, some people are just really good writers. But now that makes that study, which might be rejected otherwise, get out there to the masses. 

We've all read those manuscripts. You've been asked to review some. You think the science is good, but I can't understand what they're saying Right. Or the way they've written it. And it feels like they're being disadvantaged because English isn't the first language. 

That's one example. Right? The other thing is you say people may not have a natural flow with describing their stuff or getting the ideas across. But is that cheating? And actually, if you declare it, is there a way of telling exactly what it is that has been altered from original document? 

Well, did you see the did you see the post? I think Tanya Selak posted this about an AI a manuscript that went through to print. Yeah. And then the bottom, it said, like, part of the prompt was very very obviously, like, part of the prompt. Oh, I didn't see that place where I've heard about this concept. 

Or maybe not print, but it was it got to the edit. It it got through a fairly far way down the editorial process before someone went, hey. That's that's quite clearly Right. Someone has made yeah. Yes. 

I mean, I said I I think you kind of made us think about a few things here, so I I'm interested in that. Do you have any experience of artificial intelligence for regional anesthesia? So, I mean, I think there are so as far as image interpretation or for planning the regional anesthesia techniques? E e either of those. I wanna hear about both of them. 

Yeah. Okay. So, you know, I think there are a lot of different platforms out there. And, you know, I think the radiology folks are ahead of us to a certain degree and using artificial intelligence for image interpretation. In particular, if they use for detecting breast cancer, for mammograms, or things like that, it's been demonstrated that it's incredibly sensitive and functions at a level higher than a radiologist. 

You know, And I don't necessarily doubt that similar sorts of applications might be useful in the world of regional anesthesia. You know, I think you have to be cautious again about it maybe misidentifying something. Right. Because especially in real time, you know, the implications of us, you know, targeting something that AI has identified as a nerve root, but in fact, as the transfer cervical artery, say. You know? 

I mean, the implications of that are quite catastrophic. So I think you still need that baseline knowledge and understanding. But what it might be able to do is democratize regional anesthesia even further and make it so that, you know, maybe our colleagues who trained in an era prior to ultrasound guidance can, you know, now start, you know, participating in this. You know, maybe the emergency medicine physicians who are kinda just starting to bring this into their world can, you know, more effectively utilize ultrasound guided regional anesthesia. And then for just training, teaching, and, you know, having a better understanding of some of the other anatomical structures that are within the field of view Yeah. 

That a lot of times we don't, I think, have kind of moved away from to a certain extent. We're just like, oh, we just need the snowman. One, two, three. You know? And Yeah. 

There's all this other stuff that's potentially relevant to what you're doing that you either want to note or avoid. So, yeah, I think for that, it could be very useful. Do you know what I love about this podcast sometimes is having someone who's so smart and well spoken. Uh-huh. Are you talking about me? 

Are they on next? And you seem more as, like, democratized. Yeah. I did I did don't know. It just it makes it elevates our Yeah. 

For sure. Yeah. For sure. And the other use you were talking about, so other than image interpretation, the other use of of AI potentially in regional anesthesia? Yeah. 

I mean, I think a lot of times it can be an idea or differential generator. So if you have a patient coming in, again, being very thoughtful and cautious about entering anything that would be identifiable, but in, you know, entering, like, a case scenario prompt that would, you know, you know, talk about the the pet the patient's, you know, maybe comorbid medical conditions, the surgical procedure Yeah. And then prompting it to ask, okay. In this setting, what sort of regional anesthesia approaches might be appropriate? You know? 

And Yeah. Have you tried that? Does it work? It does. Yes. 

Yeah. Yeah. You know, we did another study that is not quite published. It was discussed at the Society for Airway Management annual meeting where basically what we did is we created clinical scenarios, clinical vignettes again, and then prompted the large language model to describe how it would conduct airway management in that scenario. Uh-huh. 

We then went through and had the authors of the difficult airway algorithm and then members of the SAM board go through and evaluate the recommendations that it made. And the thing is is that it performed reasonably well in some tasks. So there were some things that it did really well. Like I said, okay. You should try a multitude of techniques or things like that. 

They have their key concepts that they want you to keep track of. But, like, then it kind of failed with Right. Oh, you know, keep being aware of the passage of time or, you know, trying, you know, those kinds of things. So it it it kind of depended on where it did well and where it did not. For regional anesthesia, you know, I think you can kinda make it like when we're baking cookies where, okay, it recommends all these things or for, like, normally, you would do an inner scaling block, but for whatever reason, you can't in this person. 

Uh-huh. And, you know, like, when we bake cookies, it'll give us a recipe and, you know, include brown sugar. It'd be like, oh, well, that's really interesting, but, like, modify the recipe so that it no longer includes brown sugar. And it usually does a pretty good job. And I think, like, the large language model then, you know, could go through and say, okay. 

Well, you always do inner inner scaling blocks, but, like, also, could you consider doing, you know, whatever, superior drunk or supraclavicular or, you know, some of the more distal blocks that maybe aren't part of your practice, aren't part of your normal thought process, but things that aren't necessarily wrong and could be effective. Right. I'm I'm I I'm excited about what the future holds, but I'm also slightly nervous. And I think, you know, we can't we can't rely on these things entirely, but certainly to get you thinking, give you suggestions, and and a direction of travel. I mean, you know what I use it for recently, and this is, you know, we've been really high brow, and then I'm gonna bring the tone down again as normal. 

I was off. I was on leave, and I wanted to make my wife a cocktail. And so I I entered in a whole load of ingredients that we had in the house, and I said, can you recommend some cocktails for us to bake? And it was amazing. Oh, really? 

Yeah. And it and it came up with some stuff that was really good. Yeah. So why don't you make a, you know, an orange and dragon fruit gin fizz or something? And so I I think if we if we are sensible with the prompts, if we respect data confidentiality and we apply some some logic behind what we're doing, it's a tool that we can certainly use. 

Are we gonna just skip over the fact that you have dragon fruit in your house as No. It's yeah. All the time? Because that but this is we just happened to buy this dragon fruit juice thing. But, yeah, it was Yeah. 

I I should have known you to pick up on that. But, Chris, I can't thank you enough for getting us to think about things Yeah. Differently. And we're very we have been very thinking about AI for regional anesthesia in this way, but stepping back and looking at the whole picture and, you know, sentiment analysis and yeah. They've given us lots to think about. 

So it's a it's added another facet to the concept that we put out on it. So thank you so much. Thank you so much for being part of the show. With me. Really, really lovely to meet you. 

Thanks, buddy. Well, that's it for this episode. Thanks for listening, and we'll catch you next time. Oh, and, hey, do us a favor. Would you? 

When you get a chance, remember to Block it like it's hot.