So, thank you for joining us for this session on the management of sub-solid lung nodules. My name is Anne Gonzalez. I'm based at the McGill University Health Center in Montreal, Canada, and hoping that we are able to keep time for discussion and questions at the end of the session. So, without further ado, I'll introduce Dr. Adam Gutentag, who is an associate professor of radiology here at Vanderbilt, and will be discussing the radiographic features that can help us manage sub-solid lung nodules. Thank you. All right. Thanks, Anne. Okay. Well, let's get started here. As I said, I'm Adam Gutentag from Vanderbilt, just up the road here, and welcome, y'all. I have no disclosures here. So, we're here to discuss the sub-solid pulmonary nodule, which is a problem. That's why it's a big, packed room here. We know that a lot of sub-solid nodules are malignant, and they have a greater risk of being malignant than a solid nodule of the same size, but we also know that these nodules tend to grow quite slowly, and some of them will never grow, and some will never become invasive tumors, and unfortunately, complicating things, some of them, some of these patients have multiple sub-solid nodules, which really complicates our management. As a radiologist, as part of the team managing these patients, I have several goals, some of which are a little bit contradictory. I want to try to identify tumors, lung cancer, as early as possible and as accurately as possible, but I want to avoid sending patients for invasive procedures or treatment who have benign or non-invasive disease. I also don't want to delay diagnosis so long that the patient's tumor stage advances, but I want to minimize the number of scans we do. So, you can see these are kind of contradictory goals. Just to set the vocabulary here, I'm going to be referring to SSNs, or sub-solid nodules, which is an umbrella term encompassing GGNs, or ground glass nodules, and PSNs, which are part-solid nodules, which have some solid component to them. Our understanding of adenocarcinoma is kind of based around this classification that came out over a decade ago now from the pathologists who described this spectrum of disease ranging from adenomatous hyperplasia at the benign end through invasive cancer and passing through adenocarcinoma in situ and minimally invasive cancers. Radiologically, we understand that there is a pretty good but not perfect concordance between the radiographic features and the pathologic features. So the most benign end of the spectrum, we see very small, low-density ground glass nodules, and at the upper end of the spectrum, we see mostly solid nodules. And in between, we have varying gradations of heterogeneous ground glass density up through small solid components up to larger and larger solid components. So what I've been asked to talk about is which radiographic features can help guide the treatment of sub-solid nodules, and the way I interpret that is how can we best distinguish invasive from non-invasive lesions, because the invasive lesions are the ones that we are the most concerned about. And the real question there is what is it that, to me, makes a nodule worrisome enough to hand it off and say, hey, this nodule needs either biopsy or direct treatment? Okay, and that may be PET CT as the next step, it may be biopsy, or it could be ablation or even resection. We understand that a lot of these sub-solid nodules that we see are not going to be cancer. About half of screening-detected sub-solid nodules or part-solid nodules will resolve on short-term follow-up. So here's a pure ground glass nodule that resolved in nine months, and here's a part-solid nodule that resolved in four months, and were presumably inflammatory. And we know these nodules can stay stable. These three lower lobe nodules, pure ground glass, seven years later, they look exactly the same, right? So we don't want to refer these patients to the surgeon, sorry. The real fear we have is waiting too long, right? That's the opposite of being a little too conservative. The problem with being too conservative is you can wait too long. So look at this little ground glass nodule, and we followed it yearly, and it didn't grow very much, but it got more and more solid. And at what point do you pull the trigger and refer this patient for resection or biopsy? If you look at the soft tissue windows on the same patient, there's a little right paratracheal lymph node, and on follow-up, stable on the first year, a little bit bigger on the second year, and then a lot bigger on the third year. And by the time we biopsied this patient, he was already stage 3A because he had those paratracheal lymph nodes were involved by tumor. So we waited too long on that patient. So getting to the radiology part of it, what are the basics? If you're going to be following these nodules, you've got to have thin sections, and that's typically one millimeter thick contiguous sections. It's critical to assess the size and the density of these nodules accurately. It's important to look in multiple planes. You can't just look at the axial plane. You have to look at the reformatted images in coronal and sagittal planes because sometimes the growth or change is going to be more obvious on the orthogonal planes. Critically, is it changing? That's the real thing we're looking for, change over time. And to do that, we have to have old studies. If there are old studies, we've got to get them. Even if they're from another institution, let's make the effort and get them because that can really change the timing of what we're going to do. If we don't have anything old, if we have a new subsolid nodule, the workup begins with short-term follow-up. Typically, that's a three-month follow-up because we know that if the nodule is still there and looks the same at three months, about three-quarters of those are going to have at least some element of malignancy in them. That's a very high number. If at that follow-up there's a bigger than eight millimeter solid component, that's really super suspicious and we usually recommend action, either biopsy or PET CT or something at that time. If we don't see a change and we decide to follow that nodule, the surveillance is going to be at longer intervals for purely ground glass nodules and it's going to be at shorter intervals for the part-solid nodules. We know that in contradistinction to solid nodules, which we usually say two years of stability is evidence of benignity, for these part-solid nodules, we really have to follow them for about five years to be sure that they're benign. So asking, has it grown, raises the question of what counts as growth. These nodules can be very difficult to measure precisely and if you have a measure of some of these nodules twice in the same day, I may come up with different answers. I like showing this slide that shows the cursor measurement at 4.06 millimeters, which is absurdly precise. How big is that nodule really? Is it three? Is it four? I don't know. So we usually say, if we're going to say a nodule has grown, we want to see it growing by two millimeters and not one millimeter, is the way I think about it, because one millimeter is definitely within the range of measurement error. It's important to always compare the current study to the baseline study. If you only are looking at the last study that we did, you can miss slow growth. So always keep that baseline measurement up to compare. So here's an example. Here's a study from 2015. I'm giving you a tenth of a millimeter measurement here and you can see, two-year follow-up, not more than a millimeter change. Each time, it doesn't grow more than a millimeter. But if I look at the first and the last, it's obviously grown. And it's important not to just look at the axial plane. Here's an example of a ground glass nodule that has not changed at all on eight-month follow-up in the axial plane. But in the coronal plane there, you can see, I've marked a nearby vessel with a red arrow, and you can see that the ground glass is starting to encroach upon that vessel. So it's growing in the coronal plane, but not really in the axial plane, not measurably, but in the coronal plane, it was growing measurably. So you have to look at all planes. So what am I looking for? What is it that gets my antenna up and say, hey, this is possibly or probably a cancer? And the ones that I'm going to refer for more aggressive follow-up. So this kind of a thing. A ground glass nodule, not really a measurable solid component, but it's very heterogeneous. It's not homogeneous ground glass density, but heterogeneous. Even though it's a purely ground glass nodule, but it's definitely growing. That's worrying me. A solid component that appears, look at these, you see the first three images look pretty much the same, all ground glass, and then, boom, in the fourth year, there's a solid component. And in the fifth year, that solid component has grown a little bit. Okay, this is almost certainly an invasive lung cancer. The presence of an air bronchogram that persists in a persistent nodule is really characteristic of adenocarcinoma. So that's always worrisome. The presence of a central bubbly lucency, these little faint bubbly nodules, you can see this ground glass nodule has a little solid component and a little central bubbly lucency on follow-up. The solid component doesn't grow, but the ground glass grows, and the bubbly lucency is there, and this is also an invasive adenocarcinoma. When a nodule develops a cystic space, that raises our antenna. Okay, you can see a year out late, that little ground glass nodule now has a cyst associated with it. And then the cyst grows on follow-up. Still no solid component, but then on the fourth study, now it's developed a solid component along its top, okay, best seen on the coronal images. And finally, outside the nodule, is it having some architectural distortion that suggests that there's some fibrosis that's commonly associated with adenocarcinoma? So you can see the adjacent fissure there is being bent towards the tumor, and you can see on the study on the right there that the lateral pleura is being pulled in. So here's a case that I just saw this month, which I thought was really illustrative. This is a ground glass nodule with a very tiny solid component, only a few millimeters, and we did a three-month follow-up, and it looked exactly the same. And then we brought the patient back after another, well, it turned out to be seven months. So on a 10-month follow-up, comparing to baseline, we see several things that have changed. It's two millimeters larger, so it's now measurably larger. The ground glass part has now gotten a little bit denser. It's developing a rim of soft tissue density, right? It's getting a thick rim, and it's tethering the pleura. So all of those things are telling me this is almost certainly a cancer, and this got referred for workup. But I don't want to finish without mentioning radiomics. Some of you folks may be familiar with radiomics, which is basically a process of automatically using artificial intelligence to extract a large number of features from the 3D dataset. That's things that I can see, like location, size, volume, those sorts of things, but it's also tons of things that I can't see with my eyes, sort of statistically and mathematically derived information about the tumor that are invisible to the human eye. And the idea here is that we're going to improve our diagnostic ability to make the diagnosis of cancer more accurately compared to what we usually use now, which are either our eyeballs or using some of the calculators that are available to estimate probability of malignancy. And we're also hoping that this can help with risk stratification and trying to figure out which tumors are invasive and which are non-invasive. And there's even some hope that we can use these radiomic features to predict certain genetic subtypes, like EGFR mutation, that may help us predict response to certain therapies. So I want to finish by making it simple, okay? We'll get away from the computers. We have a subsolid nodule. This is my decision tree, right? We wait three months. If it goes away or shrinks a lot, then it was inflammatory and benign, and we don't have to worry about it anymore. If it's stable, then it's indeterminate. And we have to say, well, does it have frankly suspicious features that we now should push to biopsy, or can we still follow this? And if it's showing changes that are suspicious, did it become larger, is it denser, is it newer and enlarging, like the thing I just showed you, then that's suspicious and that should be pushed either straight to biopsy or even for, in some patients, you might elect to just go ahead and treat it. So that's what the rest of this talk is about. So hopefully I've teed those subjects up for everybody. There's a few references for you, and I really appreciate your attention. So my talk, which I will keep relatively brief, is on the role of PET, if any, and biopsy in the management of subsolid lung nodules. I have no specific disclosures related to this talk. These are my other disclosures. So as per the title, the two questions posed are the role of PET in the evaluation of these subsolid lung nodules and the role of biopsy, transthoracic or bronchoscopic, in their evaluation. And so to follow the precepts of Stephen Covey and the seven habits of highly effective people, I thought we would begin with the end in mind and seek the answers in a recent narrative review from Dr. Dieter Beck. So the answer to the first question is that PET has no role in the management of ground glass nodules. And I quote, using faint uptake to make subjective judgments is unfounded and capricious in the face of extensive evidence regarding CT characteristics of ground glass nodules. That's Frank Dieter Beck at his best, right? And bronchoscopic or CT-guided biopsy of subsolid nodules has substantial false negative rate and the adenocarcinoma subtype with degree of invasiveness really can't be determined from small biopsy specimens. So this had the potential to be a really short talk. So I'll go over a little bit of the data to back up these statements. In fact, maybe the first question is why would you even consider doing a PET scan for these subsolid nodules? And you might say because it's in the conclusion of my radiologist's report. And in fact, our own guidelines recommend PET. So the 2013 guidelines, which are being updated, but recommended PET for part-solid nodules over 8 millimeters that persisted on the follow-up CT that was just mentioned, part-solid nodules that were larger, although there was a mention that really PET didn't have a role if the solid component was under 8 millimeters. In the Fleischner guideline, PET, PET-CT, biopsy, or resection are all recommended for subsolid nodules with particularly suspicious morphology. And some of these features related to contour or bubble lucencies have already been mentioned. They also mentioned possible role of PET for a growing solid component or, again, a solid component that's larger over 8 millimeters. Meanwhile, the British guidelines sort of just state that PET-CT may be suboptimal at characterizing subsolid nodules as either benign or malignant using the conventional criteria that we're used to for solid nodules. So some of the data that we have relating to the role of PET scan for subsolid nodules, a lot of these series are coming from Asia. So one study from Japan published in 2007 looked at PET in relation to the Noguchi classification that antedates the 2011 ISLAC-ATS-ERS classification that we now use since 2011, but essentially really established the poor sensitivity of PET scan for subsolid lesions, particularly in the presence of pure ground glass nodules. Another series from Korea looked at 89 patients with 134 ground glass predominant nodules. They noted that a higher SUV could be noted with larger lesion size, around that 15 millimeters that was mentioned, but it was negatively correlated with the percentage of ground glass. So really the more the ground glass, the less useful the PET scan. And really importantly, they found no true positive correlates in the series for either nodal METs or distant METs and concluded that really there's no advantage of using PET for staging these types of subsolid nodules, particularly the predominant ground glass nodules. This was a multi-center series from Japan by Okada and colleagues. They examined 502 patients with clinical stage 1A lung cancer. All subsolid nodules, they were stage 1A lung cancer, sorry, but separated the subsolid nodules according to the percentage of ground glass. Interestingly, in this series, even in nodules that had over 50% ground glass, overall they still had 6% rate of nodal METs. And they found that this went down to below 1% when they had an SUV less than 1.5. So these authors really concluded that there may be a use for PET, particularly the part solid subsolid nodules. And the authors here concluded that it could be useful in terms of decisions about treatment, which we're going to hear about from Dr. Dieterbeck, but also this paper has been quoted to support this idea of using PET maybe in those who have that larger solid component. So up for discussion. And certainly as far as PET in the workup of screen-detected subsolid nodules, this is data from Giulia Veronese and colleagues in Milan. They examined the role of PET in the evaluation of nodules detected as part of the Italian Cosmos study. This is a single center, single arm screening study and concluded that PET was useful for indeterminate nodules, particularly solid nodules, but really confirmed the very low sensitivity of PET scan for the subsolid nodules, particularly the smaller subsolid nodules less than 15 millimeters. And important to remember that really a negative PET scan is not going to exclude cancer in these patients with subsolid nodules. So in fact, I've put this up to say that there may be a role for PET that we're still trying to figure out, but this is in fact a narrative review that I think was part of your references, Dr. Guttentag, but really the only place where PET's mentioned is these part solid nodules with that solid component greater than 10 millimeters here, greater than eight millimeters in other guidelines. Moving on to biopsies, you've already heard a little bit about this classification of adenocarcinoma that was introduced in 2011 and that is really based on resection specimens. In this document, there was terminology for use with small biopsies that was proposed with a recommendation to add a comment when a biopsy is reported that really if pure lipidic growth is observed in the small biopsy sample, an invasive component can't be excluded. And so really it's important I think for us to do biopsies and review the results of biopsies to remember that the diagnosis of adenocarcinoma in situ or minimally invasive adenocarcinoma really can't be established without having a resection specimen with histologic evaluation of the whole tumor. In terms of some of the data that's been reported looking at biopsy for subsolid nodules, this was a study from Korea looking at diagnostic accuracy of CT-guided core biopsy in 50 patients with predominantly ground glass nodules. The author reported a diagnostic accuracy of 91% overall, although in fact of the 50 patients, there was seven excluded, five with non-diagnostic biopsies, so that's significant out of 50, and two with nonspecific benign diagnosis that had been lost to follow-up. And in fact, follow-up comparison of histology on the biopsy versus the resection specimen did show some of that discordance that we can expect in terms of the specific histologic diagnosis for malignant and premalignant lesions. The series by Kim and colleagues was from Memorial Sloan Kettering looking at CT-guided biopsy in patients who had resected lung adenocarcinomas, and here the overall concordance of biopsy samples with, in terms of the predominant subtype in that classification of adenocarcinoma was 77%, but the sensitivity of TTNB for the aggressive tumor subtypes was only a little under 50%. And in another series, this one from Mass General, 86 patients underwent TTNA or B for subsolid nodules with a high technical success rate and a high sensitivity of malignancy, but in fact, a non-diagnostic biopsy rate that was also significant and varied with the percentage of ground glass opacity and an incidence rate of pneumothorax of around 20% with those TTNAs. So that's what I have for TTNA and B. In terms of bronchoscopy, in fact, really, we know that conventional bronchoscopy has an extremely limited role for evaluation of peripheral lung lesions, especially for lesions below two sonometers where the yield is really around 30%, and there's really no specific studies that I was able to identify that has looked at bronchoscopy for subsolid nodules, and that's including discussing with colleagues. If we look at big series that have been published of advanced diagnostic bronchoscopy for peripheral nodules, in the NAVIGATE trial, which was a multi-center trial of navigation of bronchoscopy, electromagnetic navigation, it was only around 6% of lesions that were ground glass and diagnostic yield, the definition of which has all kinds of issues that were discussed at a session earlier today on bronchoscopy, but it wasn't specifically reported in this subgroup. And in the ACQUIRE multi-center registry, similarly, it was less than 5% of patients for whom data was reported that had these types of subsolid lesions sampled. So we really don't have data on bronchoscopy. So in conclusion, as per our initial answers, really, PET has a limited role in the management of our subsolid nodules. There's really no role for the pure ground glass lesions. There may be a role in lesions with a solid component that's eight to 10 millimeters. That's certainly something that we discuss in our multidisciplinary tumor boards. But always important to remember that a negative PET does not rule out malignancy in these subsolid nodules. And I think TTNA, with aspiration or biopsy, also has a limited role. We have to remember that the subtype of adenocarcinoma can't be reliably determined on these small biopsy samples, that a negative TTNA or TTNB will not rule out malignancy in these subsolid nodules. And one thing I perhaps want to discuss at the end of the session is whether there could be a role for biopsy when patients are being referred for SBRT. So I will stop here. Please don't forget to evaluate the session, as you're being reminded by other chairs. And I will introduce Dr. Doug Ehrenberg, who is a professor of medicine at University of Michigan, and is also our associate editor for the thoracic oncology section of our chest journal. Thank you. And I'll try to... Thank you. All right. Thank you. All right. Can you hear me all right? Thank you. All right. We're good. So I don't usually spend a lot of time on my title slide, but there are two words in this title that I think are doing a lot of heavy lifting. So I'm going to define them up front. So, conservative is code for not necessarily do nothing, but don't put a needle in it and don't biopsy it. So when I talk about conservative management, it's just leaving the darn thing alone, really. And then subsolid, which I think is something that Dr. Gutentag alluded to, is that if you're going to call a lesion ground glass, you better make sure that the CT scan was formatted appropriately. And I think a lot of the confusion about what we're going to find, what we're going to do is we're going to look at the literature and try to figure out what it tells us about doing nothing in ground glass lesions. And where there is controversy, I think it can almost always be traced back to the fact that not all scans are created equal. So if you're going to call something a ground glass lesion, make sure it's thin slice, make sure it has overlap, make sure you look at non-axial sections in addition to axial sections. So with those caveats in mind, let's take a tour through this topic. And I don't have any financial relationships with anybody who cares about this. So I kind of threw out some questions as I thought about this topic. The first is, how do we assess progression? For most nodules, we think growth is concerning. If something grows, it's bad. I think you have to recalibrate that when you're thinking about ground glass lesions. Growth is not the monster in the closet. In fact, I'll show you examples, and I think you had examples of nodules that didn't grow at all, but clearly became much more concerning with time. Growth is not the concern. Too much focus, and one of the problems with the literature that we'll look at is so many studies that have evaluated natural history of these lesions have looked at natural history in terms of growth. And again, I don't think that's the end point that we should be concerned about. You'll see examples of lesions. You probably have seen in your own practice examples of ground glass lesions which didn't grow but which became more solid and may in fact have shrunk. And that's not reflected in most of the studies that we're going to look at. So there's a lot of diverse entities that are thrown into these studies of ground glass lesions that may not all be created equal. And you saw some great examples. Never ever follow, by the way. I have a rule of thumb that I can avoid is never give a talk after a radiologist. They have way better pictures than we do. So the duration of follow-up varies with a lot of these studies. And again, I caution you against defining something as a ground glass lesion on a CT scan photographed with five millimeter slice collimation. You're probably not catching a tiny solid component that could confer more risk. And so the other way I like to reframe all of this is when is the treatment, if there is such a thing, when is the treatment worse than the disease? So here's an example of very different looking lesions which might be lumped under the same umbrella in any given study. So I think you have to be very careful. And this is one of those things where if you're reviewing the literature, you've got to look at the nitty gritty. How did they define a ground glass lesion? And I think all of you can look at these and say they aren't the same. These are not the same. If you have two patients in front of you, I'm going to be much more concerned about the two patients on the right than the two on the left. And I can't tell you why. I don't have radiomics in my head. But I've seen enough of these to know that I'm not too worried about those ones on the left than the ones on the right. I think it's someone who needs to have action sooner rather than later. And here's an example of an anecdote that I have from many years ago of a patient who had a CT scan. Nothing was ever done. Lucky for him, he fell off a ladder and broke his arm and had an x-ray in 2010. And on that x-ray, they had a lung nodule. And sure enough there, that thing is, and if anything, it may be a millimeter and a half smaller, but that's a lot more concerning. So size is really not the concern here. And with that caveat in mind, so many of the papers, particularly in the early 2010s, that asked the question, what is the natural history of these lesions, reported the natural history in terms of growth. And with that limitation, here's a study that shows red dots, green dots, and blue dots. Red dots are nodules that grow. Blue dots are nodules that didn't grow. and the bottom line, the horizontal line, is time, if you look at that, we're talking one, two, three, five, 10 years of follow-up. Now, I think most of us were raised with the two-year rule in mind. These are not two-year lung nodules. These are nodules that you're gonna follow for much, much longer than two years, and for this reason, that the ones that do grow may not grow in a year or two, but after two to three or four or five years, the ones that are going to grow, generally, I think you're going to see. Now, does that mean that I stop following every single patient that I see after I've followed them for five years? No, it doesn't. I try to take into account what I think their life expectancy is. If I see a 65-year-old who's playing golf three days a week, I'm not gonna stop following that nodule that's been stable for five years, because I'm not comfortable with that, and there's no scientific basis in the literature with which to do that, but enough anecdote piles up, and you say to this patient, I don't know what the natural history of this thing is, but I think right now, we're better off not just letting you sail off into the sunset with that 18-millimeter ground glass lesion in your right lower lobe. On the other hand, that same person who's 85 years old with some comorbid cardiovascular disease, I've followed him for five years, I may be comfortable enough saying to them, you probably don't need to have any follow-up. So not all nodules are created equal, and not all nodules should be treated the same simply based on their radiographic characteristics. I think we can probably all agree on that. Another study looked at this amongst some 1,200 nodules in about 800 patients. So as you've noted, Adam, a lot of these patients have more than one, and I think that becomes very problematic when you've got people with two, three, and four ground glass lesions knowing what to do about them. In my own practice, I just look at the worst one, but that's not always the one that becomes a problem, but I think you have to be more careful when there's more than one. They followed these patients for a median of four years, of 1,000 nodules, 69 of them are heterogeneous or part-solid, so that left just under 1,000 ground glass nodules. And over time, they just simply reported what did they do to these patients. Only 35 of nearly 1,000 patients had surgery. And what that tells you is that these investigators are very comfortable following these. They reported really on histology of those that had surgery. What I would really like to know was what happened to the other 937 patients who had nothing done? And I would suspect most of them were either lost to follow-up or were doing fine. We don't know that, but that's, in the mid-2010s, that's the data that we had. And we got some other data from these. And what I'll show you is a lot here. These are very small and hard to read. But you got red lines, green lines, and blue lines here. Red lines are pure ground glass opacity. Green lines are part ground glass. And then you had what was called heterogeneous ground glass. I think that would probably be those more complicated lesions where you can't identify or measure a solid component, but something about it doesn't look right. And then you had these nodules that were described as part solid. So an identifiable, measurable solid component. These were divided into those three categories. Pure ground glass lesions very rarely grew. And when they grew, they very rarely developed a solid component. And that was over a nearly 12-year duration of follow-up. So what this tells you is that the risk of doing nothing is probably very, very low for pure ground glass lesions with all the caveats we talked about earlier. And those red lines are basically the tiny rate of growth observed in those lesions that remain pure ground glass. Another study, again, with some suspicion, when you report a 100% survival rate at 10 years, you have to invoke the possibility that maybe those people didn't need to have surgery in the first place. And so that's what I took away from this study that reported a very large number of people who had pure ground glass lesions. Now, this is back when we still used the term bronchovular carcinoma. Prior to about 2016, that was code for ground glass opacity with all the caveats attendant with slice thickness and everything. People that have 100% 12-year survival probably didn't have an invasive neoplasm to begin with. I think that's an indirect takeaway. The data on these doesn't get very good until we were able to go back and look at people who were enrolled in screening studies. So the last few slides I'm gonna share with you are ground glass nodules that were discovered in the context of a screening study. So, very busy study. I'm just gonna show you, there are about 857 people. About half of them, 491 had follow-up for more than two years. And of those, 89 had persistent ground glass lesions. So we're dealing with 89 people here. It's a small number of patients. Over time, over a sufficient period of follow-up, most of those didn't grow up. 77, that's the bottom left square. 77, nothing happened. And what you see when you look at these studies is a number that keeps coming up is about 15%. About 15% of these progressed and about 15% of them develop a solid component. And that is a recurring theme. It's roughly approximate for studies that define these prospectively as ground glass lesions without a solid component. And these authors actually looked to see what were the predictors of that 15%, rough 15% of people. So it was the initial size, so bigger ground glass lesions are more concerning than smaller ground glass lesions, but size doesn't matter except insofar as it might predict a probability progression. Here's the way I look at it. If I've got a two and a half centimeter ground glass lesion, there's a lot more potential cells in there that can go bad than there are in a seven millimeter ground glass lesion. So from a probabilistic standpoint, the more volume of abnormality you have in there, the more likely, it's like having more lottery tickets, you're more likely to hit the lottery. It's just a bad lottery that you prefer not to hit. But notice also that the mean doubling time of these is over 700 days. And that's a doubling time that we often associate with overdiagnosis, which is also consistent with the 100% 12 year survival we saw earlier. These were nodules, this is an interesting study. So this took a group of subjects who had nodules which had been stable for that five year followup, which is currently the outer end of what's recommended by Fleishner. It says, well what happens if you keep following those people? So these are all people that had nodules that were stable for five years and then followed for another five years. And that's why I like this study. It's a lot of people and it's a lot of followup. And what you see is of these 5,000 or so people, there are 351 that had a ground glass lesion that was nodular in character. Most of them were stable for at least five years and so those 208 people went on to additional followup. 27 of those grew and most did not. And so again, with the caveat that growth isn't necessarily the concerning outcome, progression is defined as growth, is not common. And finally, this is a study from the NLST which I love this study. There's 2,500 subjects out of the 26,000 and the CT arm of all of our favorite study, the NLST, we've milked this study for so much and it's good because it's a great study. About 10% of people had a persistent ground glass opacity. Now of those 2,500, 48 died from lung cancer. Okay, so far so good. That's about 5% of people. Now of those 48, half of them died of a cancer that wasn't in the same lobe as the ground glass opacity. So further good news. So it's probably not likely to evolve into a cancer that can be dangerous and concerning and metastatic. 15 of those had cell types that were not adenocarcinoma. So probably not consistent with origin from a ground glass lesion. So there are 12 out of 2,500 people, half a percent that had a ground glass lesion that may ultimately have been their cause of death. And here's a similar snippet of data from an Italian study, the MILD trial, which looked at these large bars over here. This is over an eight year follow up. How many people with a ground glass lesion died from their cancer? It's a zero, there's none. None of them died from their ground glass lesion. Those who did die from lung cancer, died from a lung cancer that did not arise from that ground glass lesion. So very similar to the NLST data. There were a lot of people that died from an extra pulmonary cancer. I don't know what to make of that. I don't know that ground glass lesions by themselves confer anything other than some susceptibility to being able to develop a neoplasm. And maybe that is more concerning in extra pulmonary neoplasms. But a lot of these people died of something other than cancer. This low bar here, high bar here, that's the definition of over diagnosis. Right, so when we talk about over diagnosis, one of the only ways you can identify that is in a randomized controlled screening study. And the reassuring aspect of this is a ground glass lesion is not risk for death from that lesion itself. It may be an indication of risk for cancer. And what we really need are some of the prospective studies that are ongoing now, and I've identified three of them. One of them is this study being done by the Thoracic Surgery Oncology Group enrolling people with two or more ground glass lesions. I wish they had just done it with one, but they're two or more. The primary endpoint in this is lung cancer specific survival. I didn't put that in there, but there's some important secondary endpoints. I think a really interesting study is being done in China. Again, you see the inclusion criteria, ages of all sorts, persistent ground glass opacities between six and 30 millimeters, less than five millimeter solid component. So you're allowed to have a little teeny tiny solid component. I'm not sure I like the term minimally invasive. I've always likened it to being minimally pregnant. You either are or aren't pregnant, but it's there, and people smarter than me have come up with this designation, so I defer to them. And the final study, which I think is something that's very interesting from the Jaycock Group, is gonna look at a primary outcome of 10-year overall survival. And what this really is is, so these are smaller than two centimeters with what's called a consolidation to tumor ratio of less than 0.25. That means a ground, a solid component less than five millimeters. So really, all these, I think, are going to incorporate subjects similar to what we're concerned about in this session, which is what do you do with these ground glass lesions which look like they're behaving very indolent. I will summarize my own approach is I follow them, and I follow them as long as I think the patient has a reasonable life expectancy, and I try my darndest to talk people out of invasive procedures, because I don't want to create a situation where the treatment or the diagnostic procedure is worse than the condition itself. So to summarize real quickly, the risk, I think, of progression is very low, low enough that I advocate as strongly as I can against invasive procedures for most of these people, with the exception being someone who's developed a complex or cystic or bubbly or solid component that's growing. Progression, in my mind, is defined not by the size or extent of the ground glass opacity. It's defined by whether there is more complex internal structure to it, a solid soft tissue component that looks concerning. I wish I had the radiomic vocabulary to describe what I'm saying, but I think if you close your eyes, you can picture it. Can we predict progression? No, I think that's one of the real challenges in this field is to see whose ground glass opacity has the potential to progress and whose does not, because the idea would be to do as little harm as possible to these people. And if you had a biomarker that predicted a lack of risk, that would be great. And again, I think there are people with pure ground glass lesions that should not be treated. Those with very limited life expectancy, but you could argue the same thing about solid lesions in that case. And I think the most important question to ask is when is the treatment potentially worse than the condition itself? So I guess if I could leave you with one thing is to use that thought to evaluate your patients with sub-solid lesions. And I'll leave it at that. Thank you. And our last speaker is Dr. Frank Diederbeck, who needs no introduction. He's a professor of surgery and director of thoracic surgery at Yale University. Thank you. All right. Well, I was asked to talk about limited resection versus SBRT. I don't have any disclosures. So I can't really do this without talking a little bit about echoing, I think, what Doug was just saying, what other speakers have said. So this is an example here. This is a case report in 2010. So this was a slow progression over 10 years, thought to be worthwhile of a case report. Now we're inundated with these. And I would just draw your attention to this cavitation here. I think the cystic and bubbly stuff is something that we have to pay attention to and is poorly understood. So, what's going on here? Am I, yeah. Okay, so I'll start over again. Okay, we've seen this slide before. So certainly not all ground glass lesions progress or grow. And I kind of think about it this way. So there's certainly some lung cancers, small cell lung cancer that it's gonna grow really rapidly. I think the traditional lung cancer, solid speculated, is somewhere in this range. But we clearly see these less aggressive cancers. And the term I use talking to patients, I'll say, you have a well-behaved cancer. You have a lazy cancer. I don't say, oh, it's not a cancer. I don't even care about this invasive, non-invasive. The whole thing is how is it gonna behave? Is it gonna just sit there and do nothing? I draw an analogy to prostate cancer. So a lot of people know there are forms of prostate cancer where the advice is don't do anything. In 20 years, it's not gonna have hurt you. And when I talk about it that way, most people have no problem talking them off the fence. Okay, so I think there's a spectrum. As you have a spectrum, you have to think a lot more about what are the comorbidities and the age, what's the life expectancy with this patient versus the ground glass lesion. And I think there's a lot of data that I don't have time to go into, but I think not all GGNs are the same. And there's evidence that there's some genetic differences and so forth in the ones that grow and that don't grow and so forth. Not well enough study to take it to the bank. But, you know, and I think to me, I still look a lot at mediastinal windows. I know the radiologists have totally gone away from that and they call something on lung windows solid. But when I flip the mediastinal windows and it disappears, I can tell you it's not as solid as when it stays on mediastinal windows. To me, that's an important distinction, but I know that I'm sort of tilting against windmills here because everybody has kind of gone away from that. So this is a study that was already brought up, I think, the long-term study. Patients follow up for 10 to 15 years, 87% pure GGOs, never progressed. They had some solid component, more progressed, but the point of this is, of those that were eventually resected, of those, 99.2% were stage 1A, were a few that were stage 1B because of visceral pleural involvement, which has unclear prognostic value. Nobody was stage 4, nobody had nodal involvement. This is another study that, you know, was already alluded to. I'm gonna show it to you a little bit differently. So of those pure GGNs here in the green, small number progressed to having a heterogeneous, in other words, a solid component on lung windows, or a part solid component on mediastinal windows. So not too many. You know, more of those that had a heterogeneous component to start with progressed to a solid component. But the bottom line is, of all of these, 7.4% were resected, 1% were invasive adeno. Another 6% were MIA or AIS. Should we be resecting that? I don't really know, but the main point is, again, 98% stage 1A, a few stage 1B, nobody had nodal involvement, no recurrences. So it is safe to wait until there is a change, until there is progression. And so when I talk to patients, I always say it's a well-behaved lung cancer. I don't think it's gonna do anything. It's safe to wait. We'll cross that bridge when we get to it. I often give them a prediction, you know, based on what I see and what I see over time. I say, listen, I think the chance we're gonna intervene on this in five years is 10%, or maybe it's getting up to 50%. But it's safe to wait. Let's deal with it when we get there. So this is kind of my algorithm, pure GGN. I get a CT in 12 months, heterogeneous in six months, part solid in three months. And really what I'm looking for is I'm looking for a new solid area. I prefer the mediastinal windows, at least two millimeters. If you're gonna do lung windows, I would say at least six millimeters. Now, the thing that I would emphasize here is that it has to be a thin-slice CT, and I continue to be amazed, despite Akron and others having guidance that you should be doing thin-slice CTs on these, how many sites are still doing thick-slice CTs? I wish I could change that. And the other thing I would say is that this sort of a marker doesn't take into account this business of how fast is it changing. In a 90-year-old that has slow change over 10 years, I'm gonna react very differently than in a 60-year-old that has a more rapid change. But these are the criteria I use. So I'll get to what I was asked to talk about, but I couldn't help but echo some of what people said, which is don't overdo the intervention. The chance that you're gonna help somebody by excising a pure GGN is a lot lower than the risk you're gonna subject them to no matter what treatment you give them. Okay, so let's talk about direct evidence comparing surgery versus SBRT. Don't wait for anything to show up on this slide. There is nothing. So I'll talk a little bit about some general data for slow-growing lung cancers. I'll talk about some general data for SBRT versus surgery. So these are non-randomized comparisons that have some adjustment for confounders. Now, we've kind of done a systematic review. We've got seven domains of confounding, and green is great, and red is bad, and dark red is really bad, and you can see that all of these adjusted studies didn't adjust well for everything. And so the confidence you have that the treatment, that the effect you're seeing is really due to the treatment as opposed to confounders is not great. But the point of this is just, if you look at the survival, it's really good, you know? So the point is, excellent outcomes, and there's no real data here. This is comparing segment or wedge versus lobe that one is better than the other, but they're both great. Major prospective studies that are ongoing. I'm gonna focus on these studies here. So JCOG 08-02 was just reported, but not really as germane to this study. So the JCOG 12-11 study, I'm waiting for results. They keep saying they should be out, but I thought they were gonna be out last year, you know, still not out yet. A year later, the JCOG 08-04 was a phase two study, you know, wedge for a, you know, almost pure ground class opacity up to a consolidation, so that's lung window, rate of 25%. Five year disease free survival, 99.7%. Hard to imagine you could do any better than that. No local recurrence. So, you know, excellent, you know, excellent results. However, I just, you know, wanna put little bugs in your ear as a part of this talk. The Yoshida trial, so this was a follow-up study to a earlier study, Yoshida trial. Same patients, you know, done. So this was Yoshida trial, a little bit complicated. Five year results, no recurrences. However, they followed the patients for 10 years, and at 10 years, 20% had staple line recurrences. And they were genetically looked at very carefully to see if they were the same, and the conclusion was they were the same. So, I don't know, maybe we need to follow them longer, maybe we need to be a little bit more careful about some of our sub-low bar resections in these. On the other hand, you could argue maybe if it's that slow to progress, maybe it doesn't matter that much. Okay, surgery versus SBRT, 90-day mortality. This is data that I can come up with of, you know, larger series databases that have reported this. So 90-day mortality for SBRT, 1.3%. I'm actually surprised by that. I think most radiation oncologists are surprised by that, but, you know, that really is kind of what is out there. NCDB studies and other things know, yes, we are referring some of our worst patients to SBRT, but it's not like there's nothing that can happen. So there are a number of randomized studies that have been done, and you see here, you know, SBRT versus various types of surgical intervention, size of lesions that were included, and whether these were low-risk patients or sick patients that were high-risk patients. So there was the Rizal study, there was a STAR study, identical studies, really closed after minimal accrual. These were written up as a, you know, meta-analysis of, you know, less than 60 patients. Provocative results for SBRT, really provocative, but, you know, with that few patients, I don't know what, you know, we can really take away from that. ACASOC did a study, another Sabertooth study, closed early in high-risk patients. So the bottom line of these studies is that they only accrued less than 3% of their accrual target. I don't know that we're ever really gonna get much out of that. Other ongoing studies, there's the VALOR study in low-risk patients. There's a StableMake study in high-risk patients. Those are ongoing, I think they are accruing well. I do expect we'll have good results from those. It will take another couple of years. There's another study in China that I'm a little bit skeptical about how it's accruing. So I doubt that we're actually gonna hear anything from that. So these are non-randomized comparisons. Again, you see these, you know, various areas of confounding that are not adjusted for that well. But there are a number where we say, you know, they're adjusted reasonably well, where we, you know, would score them as having a high confidence that what we're seeing is probably related to the treatment. And here where it's green, it means that it was statistically significant in multivariate analysis. So the bottom line is hazard ratio in these non-randomized comparisons, favored surgery over SBRT quite consistently. That's a segmentectomy or wedge over SBRT. And the differences are large enough to be clinically relevant. I'll show you this on a graphic format. Hazard ratios all pointing towards the, or almost all pointing towards the low bar side. And if you look at the differences that are there, for the most part, they're really enough that you say, well, that gets my attention and, you know, seems to be in favor of it. This was one of the best studies, really propensity matched for a lot of factors. They took healthy patients, clinical stage one, healthy patients that had no comorbidities, matched for all these factors. They did a subset analysis here in patients that not only were healthy, but were told you should get a lobectomy, but said, sorry, I'm not doing that. And, you know, fair difference there. So I think that that's some general data we have to look at. And I think most of us kind of have this feeling that probably surgery is enough better, that we probably should do that in most patients unless there's some contraindications. Quality of life data after SBRT. So yellow means no difference, red is bad, orange is maybe a little bit bad. But the bottom line is, quality of life really isn't affected by SBRT, it's excellent. This is the quality of life after surgery. You know, there's a lot of red down here. No question that, you know, that quality of life is better after SBRT. There's maybe a little bit of a signal here for VATS resection as opposed to open resections down here that is largely resolved after six or 12 months. But I think there's a definite quality of life advantage. Another thing that kind of surprised me a little bit, but we did a study a number of years ago kind of looking at new chest morbidity in patients that had gotten surgery or SBRT over time. And with surgery, you paid your price up front. Whereas with SBRT, there was this constant sort of increased rise of new chest morbidity over 24 months. And at two years, it really wasn't all that different. And when I've spoken to radiation oncologists, they have told me, yeah, I think this actually does fit with what we see. So if you're comparing at one month out, surgery versus SBRT, man, SBRT is gonna look a hell of a lot better. If you're comparing complications and some issues two years out, it's not quite as clear how big of a difference there is. I was less aware of this. Change is in FEV1. So on average, there's not much change. On the other hand, you can look at it just saying how many patients had a relative decrease of 10% or 25% or an increase of 10% or 25%. And there's some patients that had an increase, but there's actually a substantial number of patients that had a decrease. The absolute, the average, not too different. But there's a subset of patients that have a decrease, and this rate of decrease is a heck of a lot more than what you could expect from just COPD and progression of COPD, even in the face of ongoing smoking. Another thing I'd point out here is that this is FEV1. So look at the baseline FEV1. I mean, it's really not that bad. These are the high-risk, bad patients. Boy, I don't know, I operate it on people all the time with an FEV1 of 60%, pretty normal. Okay, so I'll just finish with this. This is kind of my comparison of SBRT versus surgery. I think if you have major competing causes of death, to me, it doesn't really factor in much to this decision of SBRT versus surgery. I think it's mostly a decision of do you not treat at all, depending on what their rate of decline of their health is versus the rate of progression of the tumor. I think if you have somebody where you're worried about potential perioperative morbidity or mortality, that certainly does motivate you to think more about SBRT Definitely, if you think there's a chance that you're not gonna complete this by VATS, by a minimally invasive approach, or by RATS, and you might have to open, boy, you better think carefully about a non-operative approach, I think. You know, if VATS is feasible, maybe reasonable to do surgery, but again, if there's a high risk here, I think that's where you need to think about SBRT as an approach. Another factor is, you know, how probable is a satisfying surgical result? That is, you've got a good margin. If you have one of these where you're gonna, ah, I'm gonna struggle, I'm not sure, I'm gonna look at that PATH report, and I'm gonna say, damn, I didn't do as well as I thought, maybe you should have not done surgery in the first place, and you should have done SBRT. You know, patients where you're thinking they have a lot of emphysema, prolonged air leak, um, difficult hyaluronid dissection, all of these things, I think, are the things that would sway me towards saying, I better send this patient to SBRT as opposed to take him to surgery. So, that's it, that's my final slide. Thank you.

sub-solid lung nodules

ground glass opacities

CT scans

benign nodules

invasive tumors

radiologists

thin-section CT scans

growth assessment

solid components

patient management