Interventional pulmonologist Diana H. Yu, MD – who performed the first robotic bronchoscopy at UCSF – presents the case for more aggressive lung screening in California and throughout the country, then offers an exciting look at the latest ultrasound and robotic techniques for detecting lesions, assessing risk of malignancy and performing biopsies. She reviews the evidence from recent studies and describes how patients with nodules and effusion can benefit from advanced therapeutic approaches.
Thanks guys for having me. Um, I, my name is Diana Yu. I am one of the 3 interventional pulmonologists at UCSF. Um, I actually joined the group about a year and I'm sorry, there's a little background noise. Um, so I, I joined UCSF about a year and a half ago, but I was faculty at Stanford and USC prior to joining, uh, UCSF. So, um. I wanted to kind of briefly, I've discussed today about what interventional pulmonology um offers. So we'll kinda go through those quickly and, you know, I have a lot of slides, but we'll kind of cut through some of the slides, um, and, uh, with the details. Um, so just a quick disclosure. So, you know, objective for today is to just review, uh, um, do a little bit of introduction of what interventional pulmonology actually means and what services that we provide. We do a variety of diagnostic and therapeutic procedures. I'll go through them pretty briefly so you'll have an idea as to what we, what we do and what we offer. Um, but just a brief background on interventional pulmonary. It's an exciting time to be in the field of IP. In the past decade, there's been remarkable growth in the IP technology and equipment, and, and more importantly, there's been increased clinical and translational research effort to focus on patient-centered outcomes. So, and recently, we've established a uniform accreditation standard for IP fellowship programs in the US which is an important step in Continued evolution of our subspecialty within pulmonary medicine, so we're excited about that and we're transitioning over to a fully ACGME accredited fellowship. So, fellowships have been around for about 20 years, um, at this point. Some of the advanced diagnostic and therapeutic bronchoscopic and plural procedures that we offer surround, I think what's more per uh pertinent to Our group here today is some of the lung nodules that you see. So you'll come across patients with incidental lung nodules or even screen a low-dose CT scan lung nodules. I think that's sort of, uh, majority of the cases that you'll be referring cases to us. Also pleural diseases, so new onset of pleural effusion or pleural diseases that you have questions about or diagnostic, uh, procedures that we can perform, mediastinal and hilar adenopathy. Sometimes you find these on CT scans. So these are also some things that we also diagnose. Uh, and these are sort of our procedures that we offer flexible bronch, linear radial Ebus, um, navigational bronchoscopy. We offer robotic bronchoscopy, thoracentesis, picile chest replacement for diagnostic pur purposes, therapeutic procedures, we offer rigid bronchoscopy with tumor debulking. We do all kinds of Breavement, I'll go over that briefly. Um, balloon dilation, stent placement, uh, and some of, uh, and more pertinent to primary care would be, you know, perhaps long-term management of these pleural effusions, especially malignant diffusions we place in dwelling pleural catheter placements. So that's uh something else that we offer. So, you know, you see a lot of lung nodules, and as you know, lung cancer is by far the leading cause of cancer deaths still among both men and women, making up to about 25% of all cancer deaths. And this is a game-changing NLST, the National Lung Screening Trial, 10 years ago showing low-dose CT scan saves lives with 20% reduction in lung cancer deaths. So significant um change there. The Nelson study published in the New England Journal of Medicine. Now, 3 years ago already, further validated this older NLST study that early detection of lung nodules will improve mortality and volume-based low-dose CT, CT screening, uh, will have an integral part of nodule management, which, um, I believe for the primary care phy. Physicians are, uh, play a critical role. Um, this study shows significant, uh, significantly lower lung cancer mortality of 2.5 deaths per 1000 person years, um, versus a 3.3 for unscreen control groups, so significant change. Um, updated lung cancer screening recommendations by American College of Chest Physicians, the US Preventive. Uh, Services Task Force, American Cancer Society. They all recommend annual low-dose CT scan, uh, in asymptomatic patients from age 50 to 80 with 20+ pack year smoking history or quit within 15 years. So this is an updated guideline. It used to be 55 to 75, 75. With 30 pack year history, so that's, they've extended to catch earlier and also to continue screening for later, um, older patients as well. So that's a new recommendation. Um, estimated 1.5, I'd say 1.6 million new pulmonary nodules are detected with CT scans annually. Um, in addition, you know, because we're doing more screening, um, we see more screen-detected nodules, uh, projected to increase, uh, lung cancer screening. As the lung cancer screening expands, we are expected to see more lung nodules. Um, believe it or not, even Even for California, um, as aggressive as we are about lung cancer screening, and we're talking a handful of percentages actually, uh, patients who meet criteria are getting screened. So, uh, there's a shortage all across the country. Um, we're doing poorly with lung cancer screening. Um, So how do we determine the best course of action for some of these lung nodules? You know, we look at the whole patient as a whole, we look at clinical history, look at the old films, and etc. This is kind of where we get consulted for some of these incidental lung nodules versus um screen lung cancer screening, uh, um, nodules that we detect and You know, I think what's easy is if it, if it's likely benign, we just follow, right? If it's too small, like 23 millimeter long nodules, we're gonna follow these, right? These are not something that you've been are recommended to follow. Um, anything that's huge, that looks like a cancer, you take it out. So, um, the intermediate-risk patients are the ones that really is where the question lies. What do we do with these patients? So these are cases that we see on a daily basis, uh, referrals from primary and, and also even oncology. So, um, you know, we look at the probability of cancer when it comes to lung nodules and you have to realize that it, it's really highly dependent on the size. So, 0 to 1% for nodules less than 5 millimeters, 6 to 28% for nodules between 5 to 10 millimeters, 64 to 82% in nodules greater than 2 centimeters. Um, obviously, we look at the location, you know, characteristics like speculation, the growth rate, these are all important in predicting malignancy. Uh, and there are a lot of prediction models. There's, this is one of them, X-ray.com, and there are different models, male model, VA model, different predictors that look at patient, uh, how do we, how do we risk stratify patients, right? So you look at the smoking history, the age, the localization of the lesion, the size, speculated, non-spiculated, all these things kinda go into a specific calculator that we use to predict whether or not something is cancerous or malignant. And there are multiple ways to approach biopsying lung nodules and what's the best modality to sample a peripheral lung nodule, and the answer is Depends on risks and benefits, right? So we talk about Do we cut these out all the time, you know, do we send them to thoracic surgery for every single lung nodule? The answer is no. I mean, they're great for peripheral lung nodules. What about the deeper ones, um, morbidity, mortality, you know, this is a, this is kind of a combination of older data, but You know, our, our surgeons are great these days with uh minimally invasive robotic fats, right? So, so these are smaller. Resections that we perform now with lower hospitalization stay and complications, but, um, remember that the conversion rate from that to an open, uh, uh, open thoracotomy is about 6.7%. Mortality, um, 1.5 at 30 days, 3 to 6%, uh, at, uh, 60 days. It's not a benign thing, so we need to figure out what it is before we subject everybody to a surgical biopsy or removal of the lung nodule. There's always transthoracic needle biopsy, so our interventional radiology colleagues are excellent here. They do a good job with uh transthoracic needle approach. Um, their diagnostic yield range is between 74 and 96% depending on the size and the location. However, you have to realize the pneumothorax rate ranges between 15 to 44%. It's a huge range. I'd say our colleagues are great. It's about 15% here, um, but it's a huge range of pneumothorax. So, bleeding, pneumothorax is much higher. Um, and yes, getting to the lesion is easy, but what about staging the mediastinum and the hylum, right? So, they're not able to stage the patient fully when they get the biopsy transthoracically. So, um, where we come in is we do a sort of bronchoscopic biopsy approach. You know, I don't wanna burden you with some of these old data. This is an old conventional bronchoscopy data that looks at, you know, how poorly we are diagnosing, um, anything less than 2 centimeters, you know, 33%. diagnostic yield over 2 centimeters, 62, and if it's in the outer third, it's less diagnostic versus inner third. But this is actually data from, this is the old data, right? Things have, things have improved. Um, radioebus is an ultrasound that looks at Uh, it's a radial ultrasound that spins inside of the airway, gives you sort of a shadow of where the lesion is, and you can see imaging down here on the left lower corner is a concentric image of a, a shadow of a lesion, eccentric image on the right, um. You can kind of tell that if it's concentric and you're in the center of it, you have a higher chance of getting it versus something that's on the side, right? So, diagnostic yield for concentric is 75 to 84%, eccentric is 46, 48%. Um, these are multi-center prospective RCT conducted in 5 centers in the US looking at conventional bronchoscopy, fluoroscopy versus thin bronchoscope with radio Ebus. So there's a lot of jargon I'm, I'm, I'm using sort of to talk about, um, specific, uh, technologies that we use, but just realize that some of these older technology hasn't really been great at diagnosing. Anywhere between like 40% to 50%. 60, 70% at best. That's sort of um uh what's been published in the past. The various technologic advances also developed um over the past couple of decades with the intention of improving some of the diagnostic yield of bronchoscopy to biopsy peripheral lesion. Which includes navigational system, thin, ultra thin bronchoscope, radio Evas, um, virtual bronchoscopy, cone beam CT, all of these things, and despite this, accuracy remained, remained close to 70%, far below 90 to 95% yield of a transthoracic biopsy of our IR colleagues. Um, and this is another sort of, uh, study, it's a 15 trial, 1000 nodules looking at about A medium sized 25 millimeter long nodule that was close to the coral surface, um, about 1 centimeter. And successful navigation to the lesion was 97%, but the diagnostic, uh, definitive diagnostic, uh, diagnosis was obtained in only about 65% of the time. So this is also an older study. So, you know, what's the problem with the bronchoscopic biopsy? Can't we do better? And some of the limitations that I'll admit even as an interventional pulmonologist is that Um, it's hard to push our scope to the periphery or the edges of the lung because of size limitations. Um, and aside from that, also, you know, the bronchi all, they all branch at different angles and makes it difficult to kind of navigate, um, using conventional bronchoscopy. So this is sort of where we are now. Fast forward, we have now this new kid on the block, this robotic bronchoscopy that came out, that's available. Um, I've actually performed the first robotic bronchoscopic biopsy at UCSF back in July. We finally acquired ion intuitive, uh, robotic bronchoscopy. So, Just a brief background in that robotic surgery has been performed across many platforms that include urologic, gynecologic, thoracic surgery, and offers potential advantages of improved dexterity, visualization while you're maintaining sort of minimally invasive approach. There's a lot of promise, um, of robotic bronchoscopy that's designed to allow in the bronchial navigation into the lung periphery, so you can navigate all the way to the distance while maintaining catheter stability and, and to also to maximize precision and sampling. So there's a lot of promise that's, that's out there and Um, uh, the 3 robotic-assisted bronchoscopy platform as of March of this year were developed and received, uh, FDA clearance. And I'll go through this briefly. There's 3 different platforms that we have. One of them is Intuitive, uh, Ion. This is, uh, intuitive company down the street in Sunnyvale, they, um, also are the creators of da Vinci that all our our thoracic surgeons use. Um, it's got a planning station and a monitor, and we look at the pre-procedure CT scan, 1 millimeter cut CT scans. We, um, uh, we allow virtual planning using that CT scan through the, uh, semi-automatic pathway creation. It creates a 3-day, 3D airway structure for us to kinda navigate to the closest branch to get to the closest, closest to the lesion to biopsy. So this is one of the, one of the, um, And the highlight of this particular robot is that it has a shape sensing technology, so it gives you sort of positional feedback. So, virtual imaging and what, what you're doing in real life, it kind of adjusts as you go through the airway tree. So it's shape sensing technology that's quite unique. This is a uh different robot called AAS by Monarch. Looks like we're almost playing a video game. It's not as fun as playing a video game, but it's, uh, it's got two robot arms under continuous direct visual control to manipulate the bronchoscope. And so, they're different features. I won't get into it too much. This is the newest kid on the block, really not enough human data at all published by Galaxy. Um, it's integrated tomosynthesis and augmented fluoroscopy. So, um, that's their key feature that they sort of publicize. So, What's the data in it? We're still looking for data. I mean, the initial very first human trial was done in 2019. So, you know, over the past couple of years, um, and multiple versions of the software upgrade, um. The data is pretty promising. I think that I've convinced all my colleagues here to convert to doing robotic bronchoscopy. So, um, even the ones that we were, you know, waiting for more data, um, to convince us that this is a way to go, I think that, um, anecdotally, we've had really great, um, uh, uh, outcome with a robotic bronchoscopy for diagnosis. So this is a study that was published in BMC Pulmonary 2 years ago. It's a lead-ing, um, Uh, stage of multi-center single-arm prospective evaluation of this particular, uh, robot ion. Um, uh, and it's, it's a study out of investigators, um, at the Mayo First Health, MassGen Data-Farber, Dana-Farber BI, Henry Ford and MD Anderson, and they enrolled patients with CT evidence of one or more solid or semi-solid nodules. Subjects were followed at 10 and 30 days post-procedure. Median size of lesions was about 18 millimeters. Distance to the pleura or the fissure was about 4. So these are really just far out there. They're like essentially touching the rib, essentially, these nodules. Um, and procedure duration was about an hour. And biopsy completed in 96.7% of the patients, pneumothorax and bleeding was zero. I think overall, um, pneumothorax and bleeding rate, all comers, uh, you know, with, with the most updated study now, it's about 1% or less, so it's a very safe procedure. This is another study published in 2021. It's a prospective single-center study to deter determine sensitivity for malignancy and overall accuracy of the robotic bronchoscopy using the com beam CT as a secondary confirmation. 52 patients enrolled uh with 7 patients with 2 nodules biopsied. They were able to reach, um, Uh, with just minor adjustments in 9 patients out of the 52 after using cone beam CT and tissue diagnosis obtained in 83% of the biopsies. Um, and the diagnostic yield was 86%. Um, Which Pretty remarkable for, um, this was an early study back in 2021. Um, jumping to the next study, this is a published in Chess 2022. It's a study conducted at Memorial Sloan Kettering Cancer Center. Uh, where all cases referred to robotic bronchoscopy between 2019 and 2020, October of 2019 to July of 2020, captured prospectively, analyzed retrospectively. Medium size, 18 millimeters. Bronchus sign just means that there's an airway directly heading towards the lesion. Um, there was bronchus sign in 63% of the cases, so more bronchus signed, the better and easier for us to actually navigate out to and biopsy. Pneumothorax rate was 1.5%, which is great. Um, this study actually provided post-marketing experience of basically a heterogeneous group of IP and thoracic surgeons and along with a range of uh prior, um, experts in, uh, guided bronchoscopy. The study showed that uh successful navigation was achieved in 98.7% of the time with overall diagnostic yield of 82%. Um, obviously higher on a bigger lesion, so greater than 2 centimeters, diagnostic yield was 96%. Um, So, bigger the better. And this actually is the first study that looked at collaboration between IP and thoracic surgery and operating uh a robotic bronchoscopy in a very high volume academic center. Um, uh, it's a single-center study in memorial. Um, it also suggested that the shape sensing technology or the ion robot may offer a significant advancement in guided biopsy. So it's a pretty good deal there. This is a retrospective data collection of all patients who underwent, um, uh, robotic bronchoscopy, the ion uh biopsy with concurrent use of radio Ebus or Coron CT, and, um, The main takeaway point here is that um, The locations of the lesions vary, right? So more peripheral, difficult for us to get to, more central, easier for us to get to. Um. This study actually was, um, they were able to actually use uh Robotic bronchoscopy to navigate to majority of the lesions, no problem. Um, in fact, even the peripheral lesions, they were able to really, uh, get to and diagnose and get tissue diagnosis. And, um, all cases except 3 cases, um, which needed radio Ebus, uh, Rose, Rose is actually, uh, rapid on-site, uh, cytopathology, um, and combeam CT, um, adjustments. Aside from the three. The robot was able to navigate to the lesion with precision and accuracy. That's the big takeaway is that. The diagnostic yield and the precision to get to the lesion was very high in the um ion robot. I'll skip this here. Um, and just conclude that it does offer a promising future for us to actually, um, enhance our ability to diagnose lung nodules accurately. Um, in fact, just two weeks ago, I, the smallest lesion that I actually even offered to biopsy was 7 millimeters. It's not a standard, but we were able to get tissue diagnosis for a 7 millimeter lung nodule, which was a lung cancer. So, um, I think there's a lot of promise there. Obviously, we need more data. So, um, some of the therapeutic options for central airway obstruction, I don't wanna go into this too much, but we also deal with therapeutic modalities that deal with airway tumors. Um, this is less of an issue for primary, um, care physicians, but we have different technologies to ablate and debulk tumors in the airway. So, for whatever reason, you run into uh patients that are advanced, and they come to you and they can't breathe and they get a CT scan and there's airway obstruction. Um, just know that we're available to actually, um, Uh, provide therapeutic bronchoscopic management of these. So, you know, this is a one of the therapeutic modality is the argon plasma coagulation therapy, imaging 8 here. Figure 8 shows in a bronchial tumor with neovascularization and severe stenosis, B showing the tip of the catheter that burns and destroys the tissue. And we follow it with by um balloon dilation and D and then the final result of a patent airway E. So, something could be very stenotic and we can open it and destroy tumors. Um, this is also a figure on the left are two different types of CRE balloon, uh, dilation, uh, catheter guns that we use, um, and we have, you know, various ways to dilate, destroy tumors. I'll forward also here. This is uh another therapeutic modality which is freezing, um, or cryotherapy that we use, um, that we can remove uh large blood clots. Um, and also debulk tumors with this cryotherapy. This is just a quick sample of what, uh, what used to be an endotracheal tumor that we debulked and placed a stent in, um, while the patient ended up getting chemo radiation to shrink the tumor and we went back and took out the stent. This is sort of the picture on the left showing the extrinsic compression. So tumor was a mediastinal tumor that was pushing into the posterior wall of the trachea causing obstruction, but there's no intrinsic or endotracheal tumor. So we just simply put a stent in here. Pushed everything out of the way. This is an example of metastatic breast cancer in the lung and you can see all this nasty looking tumor inside of the airway on the left. This is pre-procedure and post on the right. Um, lastly, I wanna kind of briefly talk about pleural effusions and malignant diffusion. So, you know, approximately 1.5 million people develop pleural fluid in the US each year, and this is an old study published by Richard Light. Most common exudative effusions are associated with underlying pneumonia. You can see here pneumonia, up to 300,000. A malignancy take up about 200,000. So in the US it's less of an issue, but TB is very prevalent in other uh different parts of the world. You have to remember that this distribution is not the same for um other countries. This is specifically for US. So there's a lot of reasons for having uh pleural effusion, and we typically manage malignant diffusion. I can do a separate study, uh, talk altogether on different types of effusions and how we manage it. Um, but the main point here is that we deal with a lot of cancer-related difffusions, and unfortunately, a lot of these patients come through the clinic and get drained repeatedly before they, they are recommended more of a long-term solution, right? So, what's frustrating for us to see is that people just get referred for ir guided thora, thora, thora. We have a lot of patients who get admitted and then they do the same thing. It's like, Repeat thoras, and on the 10th thora, they're finally like, well, let's, let's get someone involved who can manage this long term. So, a lot of, we have, we see different kinds of cancers that end up metastasizing to the lung. Obviously, primarily lung cancers, um, you know, the data out there, it varies, right? So, I would say about a third of the cases of the lung cancers end up um uh having sequelae of malignant diffusion. Breast cancer, it's a huge range in the teens all the way up to 25%, lymphomas and leukemia, 17%, ovarian, 5%, 6% unknown, 40% other. This is an example of a trapped lung. So you have a fusion that keeps getting drained. Let's say this is cancer-related effusion. What are your options? You have someone who doesn't have a trap lung and they have a fully re-expandable lung. The recommendation is to, if the patient's young and they're active, and they don't want any catheters hanging out, option would be to do two different things. We can do, the surgeons could do that with chemical pleurodesis and, um, and mechanical pleurodesis even and Uh, seal that space with leaving no catheters behind. Um, that's only for re-expandable lung. For, for trapped lungs, such as, you know, you see it here that this already formed little, like a rind around it, and the lung is not coming back out or coming back up. This is kind of trapped in that position. These are not patients that are candidates for chemical pleurodesis or any uh surgical intervention. These are patients that end up coming to us for indwelling pleural catheter placement for long-term management of trap lung. And so, um, I'll show you kind of a vid, a quick, um, Picture of what these look like, but this is a long-term pleural catheter. Uh, patients manage it as an outpatient. We see them in clinic. Uh, it's a procedure that we can do it either in clinic or if they need some anesthesia because they're anxious or they have a lot of pain. We can do this, um, in half an hour or so with just moderate sedation or local anesthesia. And these are catheters that, that we place, and this is ideal for trapped lung physiology and, um, The data is that approximately 50% of the patients actually achieve spontaneous pneumotho pleurodesis because it's a foreign body. So it achieves um spontaneous pleurodesis in about half the case. This is kind of what the catheter looks like. OK. And then we go through all the education. If you get an early referral for us to take a look, we get patients and their families educated on the catheter, what that entails, what does it mean to um maintain a clean catheter. And we also, um, uh, arrange for home health for patients that need long-term management. And this is a quick data um of spontaneous pleurodesis, uh, a study looking at 250 procedures in 233 patients, spontaneous. Pleurodesis was achieved in about 43%. Uh, median time is about 59 days. So I tell patients you have kind of a coin flip chance of achieving auto pleurodesis within two months if you keep the catheter in place and you drain it, um, to keep that space, uh, dry. Um, and it's a safe procedure. The risk of infection cited about 4%, and even in immunocompromised patients. So I'll stop there and um open up for questions. I went through a lot, so I kinda wanted to Um, Leave the rest of the discussion for any questions.
