Thank you very much, Nicole, and thanks, everyone, for filling this room. It's great to see so many people interested in lung cancer screening. I was asked to provide an update on lung cancer screening guidelines. We have updated the chest guidelines recently, and that will be the basis for a lot of the presentation, but I also wanted to provide you with a little bit of a spectrum of the other U.S.-based guidelines and what they're recommending. These are my disclosures. They're related to funding to my institution for research related to early lung cancer detection biomarkers. So each of the guidelines starts with the same evidence base, and the evidence base is really what are the benefits of lung cancer screening and what are the potential harms and what is the balance of those. So I'm going to start by reviewing the evidence base that all of the guidelines are formed from, and then I'll talk to you about recommendations and add a little bit of evidence that helps us to interpret those benefits and harms. It used to be that the recommendations were about do we screen or not, and then they evolved to be who do we screen, and now they also involve recommendations about how we screen, and so I'll share some information from each of the guidelines on both of those topics. When we look at the evidence and when you consider how the evidence is interpreted, I think it's most important to remember that what we're trying to establish here is an optimal balance of benefit and harms. What's good in lung cancer screening is that fewer people die from lung cancer because they're screened, and the harms are related to performing the test or to managing the findings from the test itself. It's important to remember that we're screening. We're not doing diagnostic testing, and that means you're taking somebody who's well and testing them as opposed to somebody who's unwell. In that light, only a slim minority of everybody who's screened ever will see the benefit, but all are exposed to the potential harms. Also, the magnitude of benefit, someone doesn't die, is different than the magnitude of a lot of the harms we'll talk about. You find a small lung nodule, maybe you get a little concerned, and so on, and so you have this fulcrum that can shift based on a patient's values, how much they're concerned about those harms, how much they value that potential benefit, how healthy they are, how skilled the screening facilities are, is smoking cessation part of the program, comorbidities that might be identified on the scan as well. All of that can shift this fulcrum of benefit and harm and leads to slightly different interpretations of this evidence. So the evidence of benefit now comes from at least two studies, two studies that were powered to detect a lung cancer mortality benefit, the National Lung Screening Trial and the Nelson study. In here, it's just important to note that these studies, though studying the same question, did have differences in their methodology, and those differences included the age ranges of individuals that were included, the smoking histories of individuals that were included, the number of scans and their cadence, the control arms, how you detect and determine a positive finding, a positive lung nodule, and ultimately the lung cancer mortality reduction that was proven. There are several other studies that controlled trials that were performed to address this question. None of the others were powered, really, to detect lung cancer mortality reduction. However, when you combine them in a meta-analysis, you still get the positive result of lung cancer mortality reduction that we were hoping for on that benefit side of the balance. When you look at different groups, some of these challenged by the number of studies in them and when put into a meta-analysis, but low-dose CT versus usual care or chest X-ray favors low-dose CT. Annual screening, other screening protocols favors screening. Beginning at age 50, screening until at least age 75, both men and women, altogether, there's benefit from lung cancer screening with low-dose CT. On the other side of that balance are the harms, and many of the harms come from finding lung nodules. So there's a lot of data now about lung nodule identification and management for our screening programs. A lot of it varies, and it depends a lot on what size nodule you say is a positive result, how thin are your slices, what part of the country or world are you in. In the National Lung Screening Trial, which has still perhaps the most robust data, at the end of the baseline and then two annual scans, 39.1% of individuals had at least one lung nodule, four millimeters in size or larger. How you manage the nodules, in particular, how many of these healthy individuals you send for a procedure also varies in the studies, how long was the study, how was it designed. The mean across all studies was 3% of individuals screened end up going for a procedure. When you look at the robust data from the National Lung Screening Trial, again, baseline and two annual scans, 4.2% underwent an invasive procedure, 0.9% a procedure-related complication, and 0.3% a serious complication. 22% of all surgeries performed were performed as part of management of what turned out to be a benign screen-detected nodule, and 37% of all non-surgical procedures were for benign disease. The other things that contribute to the harm side of the balance but have been much more difficult to define are over-diagnosis, more importantly, perhaps over-treatment. An estimate suggests you're going to over-diagnose and treat one lung cancer for every lung cancer death averted. This means you've either found a very indolent cancer that didn't need to be found, or maybe more often in our population, you screen somebody who's got a lot of comorbidities, they're older, bad COPD, and they die of something else before that cancer would have affected them. And then the impact of radiation exposure from the low-dose CT, also very difficult to estimate, but one radiation-induced lung cancer death for every 108 lung cancers detected over 10 years is one estimate. If we put these numbers back on the balance, and this balance is drawn just for the National Lung Screening Trial results, again, the three annual low-dose CTs, they had age 55 to 74, 30-plus PAC years, smoked in the past 15. It took 320 individuals being screened to avert one lung cancer death, and of those 320, 125 would have been found to have a nodule, eight a procedure, 0.1 a serious complication from the procedure, 1.38 were over-diagnosed or one over-diagnosis, small radiation risk, and acceptable cost-effectiveness. Again, patient values and other influences in the screening program can shift that fulcrum. Put another way, put into a diagram, if you had 1,000 individuals screened per the NLST protocol, three fewer deaths, 18 still die of lung cancer, 365 nodules, 25 procedures, three serious complications, and four get over-diagnosed, over-treated. And this can illustrate why implementation becomes so important. If you detect more nodules in your population, if you're more aggressive with procedures or perform lower quality of procedures, if you don't have a system to help communicate the importance of those nodules to your patients, if you're screening a lot of unhealthy individuals, then the design of your program can shift that balance in a direction you don't want to see it go. So with that as our evidence base of all the guidelines, really, evidence base, we're going to spend most of the time on the recommendations for who to screen, considering this new evidence, and then clarify a little bit of guidance on other implementation thoughts. So who to screen? Some of the new evidence between our guidelines highlighted that our screening guidelines identified people at high risk who could benefit, however, also missed a whole lot of people who also have high risk. This particular study says anyone who's smoking more than 20 pack years, if they've quit up to 25 years of quitting, they still have a risk of developing lung cancer that's four times that of someone who's never smoked. So still a potentially high risk group that isn't eligible for lung cancer screening. Who's being screened? Unfortunately, uptake's not been as robust as we'd like, but it is more robust in individuals who have more comorbidities and more functional impairments. So we may not be identifying those most likely to benefit. The age and smoking history-based criteria on which we were basing eligibility can lead to disparities in identifying individuals eligible. The sensitivity of our older recommendations was higher for a white population than an African-American population. The newer guidelines from ourselves and the USPSDF help to equilibrate that, and the use of risk calculators can also help to account for those differences. There's been more evidence on the use of risk calculators. Risk calculators for some time have been known to be able to identify more individuals with lung cancer by screening a fewer, a smaller number of individuals. However, a lot of the factors that are included in those risk calculators can influence that balance of benefit and harm in ways that aren't in the right direction. Risk calculators include age, they include how much you've smoked, they include comorbidities like COPD and a history of prior cancers that can limit your ability to benefit or add to the harms. So for example, a population identified by risk calculators ends up having poor overall survival after treatment for early stage cancers. Potentially different lung cancer phenotypes. Most of the benefit is in adenocarcinoma, your COPD population has a greater portion of squamous cell cancers than the non-COPD population, and competing mortality that may increase out of proportion to the increased lung cancer risk. Also more likely to find a nodule, more risk from an invasive procedure, and more complications from resection for early stage cancer. So another way to look at things would be instead of a risk calculator, how about a benefit calculator? Calculator that tells you who's most likely to benefit from being screened. Risk-based calculators increase the percentage of eligible individuals as you get older. It's part of the risk calculation. Whereas a life gained-based calculator gets a little bit lower as you get older. Individuals identified through a benefit calculator, as opposed to a risk-based calculator, are younger, have fewer comorbidities, and increased quality of life. By adding life expectancy to the risk-based strategies, you're able to more adequately mirror what you get out of a benefit calculator. So with that additional evidence, this is what the CHESS guidelines evolved to in its most recent iteration, for who do you screen? Asymptomatic individuals age 55 to 77 who've smoked 30 pack years, continue to smoke or quit in the last 15 years, we recommend annual screening. This language hasn't changed. It's now a stronger recommendation. Second, asymptomatic individuals who don't meet criteria number one, they're age 50 to 80, smoke 20 pack years, continue to smoke or quit in the last 15, we suggest annual screening. These, as we'll talk about, match the USPSTF guidelines. For individuals who don't meet number one or number two, but you project to have a high net benefit from lung cancer screening, based on the results of validated risk prediction calculations and life expectancy estimates, or based on life-year gain calculations, we suggest annual screening. This will lead to greater equity across race and gender. We provided some examples of thresholds that could be used from the literature, both for benefit and risk calculators connected to life expectancy. And then we go on to recognize that these are not always easy to implement. They're more burdensome to apply to practice, and at the end of the day, this is guideline. It's not insurance or Medicare policy. So health insurers may not pay for patients who meet this recommendation alone. The USPSTF updated their guidelines recently. They performed the same sort of evidence review, and then they add to that with collaborative modeling studies through CISNET. They have five different centers model the ideal population to select for screening. Their modeling included various subgroups, the number and frequency of scans, the harms associated with performing the low-dose CT, and the use of various nodule management strategies. They also assessed risk prediction models to see how they would influence the balance of benefits and harms. And an interesting change occurred, not just to the eligibility criteria. Their modeling is put into this line of efficiency, all of the different age and smoking history-based categories that they looked at. The line of efficiency is benefit to harm. With benefit in this graph I'm pointing at now, number of deaths averted, lung cancer deaths averted, and the harms being the number of scans. And you can see the criteria they selected wasn't as efficient as others. However, when they shift the benefit definition to life years gained, the criteria they selected is one of the criteria along that line of efficiency. So their recommendations are now shifting a little bit more to a benefit of life years gained than it is to simply finding cancers and treating them. So their recommendation, age 50 to 80, 20-pack year history, screen them every year, stop screening when they haven't smoked for 15 years or they have another health problem that limits their life expectancy. They go on to explain a little bit further that this moderate net benefit is influenced, like we talked about in that fulcrum, by identifying people who are at high risk, by having accurate image interpretation, by resolving most false positives with serial imaging. And they looked at the risk models but decided that there was at this time insufficient evidence that using a risk prediction model would improve outcomes. This change leads to an increase in lung cancer mortality reduction from 9.8% to 13% compared to their older criteria. But also an increase in how many people will be screened, about 87% overall. A greater percentage increase in those who are currently not receiving equitable access to screening. And more harm, 2.2 versus 1.9 false positive results per person who is screened. The NCCNs also updated their guidelines using the same evidence base and expert discussion. Their only differences are that they don't have an upper age limit. They say stop screening when someone's not healthy enough. And they don't have a limit to when you quit. So you could quit more than 15 years ago. The American Cancer Society's last screening guideline was in 2018. They've taken it down and endorsed the USPSTF. The CHESS guidelines, the AFP, they have updated their guideline and it will likely be published by the end of this calendar year. The American Academy of Family Practitioners, great news, updated their guidelines to support lung cancer screening based on the USPSTF criteria while still acknowledging that the harms from screening may not be well documented and there are still barriers to screening in community settings that need to be addressed. Other things that are included in all of the guidelines are related to implementation now and I think they're extremely important. Shared decision making, talking to your patients about this balance and letting them make value based decisions are included in all the guidelines. This particular meta-analysis of decision aids used in lung cancer screening shared decision making shows improvement in patient's knowledge, low decision conflict, and high acceptability of the tools. So the CHESS guidelines, we suggest low-dose CT screening programs develop strategies to provide effective counseling and shared decision making visits prior to the performance of the low-dose CT exam. We list the components that should be there and we highlight that the way this looks may be different in centralized and decentralized programs. Decentralized programs need support to do this well. Lung nodule management. Lung RADS is used by most for small nodules and there are guidelines that help you for the higher risk nodules including risk calculators. This slide just reflects some variability in the guidelines that are available on the left for solid nodules. Simply consistent recommendations across guidelines and on the right for subsolid nodules, much more variable guidelines about how often you follow nodules. The Lung RADS criteria update should be published in the next few months as well. So our suggestions, we weren't a lung nodule guideline so we simply suggested that CT screening programs develop a comprehensive approach to lung nodule management, access to multidisciplinary expertise, algorithms for the management of small solid, larger solid, and subsolid nodules. We also suggested that screening programs develop strategies to minimize over-treatment of potentially indolent cancers, particularly those that are ground glass on imaging. Compliance with testing has become a very recognized major issue even at high-quality large programs, compliance is in the 40% range for annual screening. This particular study suggests that centralized programs tend to do a little bit better than decentralized programs. So our guidelines suggest that programs develop strategies to maximize compliance with annual screening exams and evaluation of screen-detected findings and we list some of the strategies in the document. Smoking cessation, nothing specific to screening other than suggesting that it's an opportunity to talk to your patients and that you should use otherwise recommended smoking cessation guidance. Non-lung nodule findings, other things you'll see, thyroid nodules, coronary artery calcification, guidelines exist for their management. We suggest that programs develop strategies to guide the management of these lung nodule, non-lung nodule findings. In particular, strategies to assist the ordering provider in non-centralized programs. USPSTF has implementation recommendations, they do recommend shared decision making, discussing potential benefits, limitations, and harms, and suggest that screening occurs at centers with experience and expertise. Smoking cessation interventions and lung RADS use are also recommended. The NCCN suggests shared decision making and focuses specific attention in three populations, African-American persons who may not meet standard eligibility criteria but have added risk, elderly patients who may not benefit as much by being screened, and then those with risk factors that aren't included in the age and smoking history-based eligibility criteria. They also provide algorithms for lung nodule management, and they provide a table of guidance on how to perform the low-dose CT scan. So a whirlwind tour of some of the guidelines, hopefully help to understand the evidence base on which they were all developed, and provide you with some thoughts about who you should be screening and how you should be running your programs. Thanks very much for your time. I'd like to thank the organizers for inviting me to speak today about eligibility criteria for lung cancer screening and the impact on disparities. I have no disclosures. While there has been substantial progress in cancer prevention, screening, diagnosis, and treatment over the past several decades, addressing cancer, excuse me, addressing cancer health disparities in certain populations is an area in which progress has not kept pace. Despite their higher rates of lung cancer screening death, vulnerable populations such as racially and ethnically diverse and medically underserved populations have not participated in lung cancer screening in large numbers, primarily due to experiencing barriers at multiple levels of access, all the way from eligibility to screening to diagnosis to treatment. Disparities exist not only in access to screening, but also in receipt of treatment. For example, many studies have demonstrated persistent underuse of surgical resection in black individuals with early stage non-small cell lung cancer compared with whites. Lung cancer is the leading cause of cancer death for both men and women in the United States. Here, we see racial disparities in both the prevalence, the incidence, and the mortality. We see that for black males, they have the highest lung cancer incidence and mortality. Black individuals are more likely to start smoking at later ages, smoke fewer cigarettes per day, and are less likely to quit compared with white individuals. Hispanic populations have lower smoking prevalence and lung cancer incidence and mortality than black and white populations. American Indian and Alaska Native populations have high prevalence of smoking, yet lower lung cancer incidence and mortality rates than black and white populations. Asian and Pacific Islander populations have lower smoking prevalence and lower lung cancer incidence and mortality rates than black and white populations. As you just heard, this is a summary of the United States Preventative Services Task Force guidelines comparing 2013 with 2021. The 2013 recommendations included those ages 55 to 80 years who currently or formerly smoked at least 30 pack years. In 2020, the USPSTF had drafted new recommendations to lower the age to 50 and to reduce the pack years from 30 to 20. These changes were based on a systematic review, which also assessed, as we just heard, the benefits and harms of screening across subgroups, specifically across race and ethnic subgroups as well as by sex. In 2021, the guidelines were officially changed with the updated criteria, which resulted in an additional 6.4 million individuals eligible for screening. It's important to note that the US Preventative Services Task Force statement recommends using age in smoking and not risk prediction models because there's insufficient evidence on whether or not the risk models actually impact outcomes. Some have raised concerns that the 2013 USPSTF lung cancer screening eligibility criteria created racial and ethnic disparities. So the 2013 guidelines didn't consider racial, ethnic, socioeconomic, or sex-based differences in smoking behaviors or lung cancer risk. From 2019 to 2020, as part of an ATS statement, Rivera and colleagues assessed the extent to which high-risk populations who didn't meet the USPSTF 2013 eligibility criteria were disproportionately comprised of minority and low SES groups. As you can see from the table here, while the percent eligible, the percent of preventable deaths, and the percent of life years gain increased across all race and ethnic groups from the 2013 to the proposed 2020 criteria, this change in eligibility is not likely to eliminate disparities. You can see that the percentages are still highest in white populations. Data also show that African American individuals exhibit higher smoking-adjusted risk of lung cancer, despite smoking less intensely than white individuals. Aldrich and colleagues used the Southern Community Cohort Study to compare eligibility for lung cancer screening among those diagnosed with lung cancer by race. The Southern Community Cohort Study includes data from 12 southern states and about 1,200 among 48,000 ever-smokers aged 40 to 79 years with 12 years of follow-up. These figures show the distribution of smoking pack years among lung cancer cases by race. African Americans who smoked and were diagnosed with lung cancer had significantly lower median pack years compared with whites who smoked, 25.8 pack years for African Americans compared with 48 pack years for whites. In the same study, Aldrich and colleagues found African Americans who smoked tended to be diagnosed with lung cancer at an earlier age compared to whites. Among the lung cancer cases, significantly fewer African-Americans were eligible for screening based on the 2013 criteria compared with whites. Thirty-two percent of African-Americans were eligible for screening compared to about 56 percent of white patients. Landy and colleagues also used data from the 2015 National Health Interview Survey to compare the risk of lung cancer death and life years gained under three different scenarios. In these figures, the blue bar represents the USPSTF 2013, the green bar the 2020 criteria, and then in the red bar the 2020 criteria plus a risk model. In terms of life years gained with lung cancer screening, they found among eligible white versus African-Americans a disparity of 15 percent for the 2013 criteria. They also found in green in that first upper left figure that there was a 16 percent disparity with the 2020 criteria, and then when they augmented the 2020 criteria with high-risk individuals that were selected using the risk model that this nearly eliminated disparities for African-Americans. What I think is striking about these figures is that we do not see the same pattern for Hispanic-Americans or Asian-Americans. Even with the 2020 criteria plus the risk models, we still see that there are evident disparities. The bottom panel shows that there are similar patterns seen for preventable deaths. Several studies have also assessed the impact of changes in eligibility criteria. So there's several in the literature, and I'm just going to review some of these here to give you a sampling of what has been published so far. Using the 2017 to 2018 behavioral risk factor surveillance system data, which actually includes 19 states and about 40,000 respondents who have a history of smoking and are aged 50 to 80 years, we see here a comparison from the 2013 criteria in red with the 2020 criteria in blue. And you can see here differences, for example, in men. Under the 2013 criteria, about 29.4 percent were eligible for screening, and then with the expanded 2020 criteria, this increased to 38.3 percent. So you see consistently an increase in those eligible over from the 2013 to the 2020 criteria. In the Reasons for Geographic and Racial Differences in Stroke study of about 14,000 participants, 19 percent of black and 27 percent of white individuals were eligible under the 2013 USPSTF criteria. With the updated criteria in 2021, about 29 percent of black individuals and 35 percent of white individuals were eligible for screening. So this is a difference of about eight percentage points in the 2013 criteria in contrast with a smaller difference of about five or six percentage points in the 2021 criteria. What is interesting about this study is that they made some adjustments to these models, and in particular, you can see in the table that they included adjustments for things like educational level, household income, social network size. And when they made these adjustments, the differences between eligibility between black and white participants actually increased. Another study evaluated the 2021 USPSTF criteria compared with several other criteria, including the 2013 criteria, the NCCN Group 2 criteria, which includes patients greater than 50 years with a smoking history of at least 20 PAC years who also have one additional lung cancer risk factor. They also compared with the PLCO modified 2012 risk prediction model. This study used the INHALE cohort, which includes individuals with and without lung cancer in the Detroit metropolitan area from 2012 to 2018. If the patients with lung cancer were used to evaluate sensitivity, then about 65 percent would have been eligible with the 2021 criteria. The PLCO model increased this to 68 percent, whereas the 2013 USPS criteria was about 49 percent, and the NCCN Group 2 had about 62 percent eligible. When racial subgroups were compared, the 2013 criteria selected 52 percent of white and 42 percent of African American individuals. And this disparity is still present with the NCCN Group 2 criteria, but in contrast, both the USPSTF 2021 and the PLCO model criteria mitigated the gap between African Americans and white patients. And then in the second table on this slide, which looks at the specificity, so if the control group is used to evaluate specificity, then the 2013 criteria excluded the most individuals, about 65 percent. Racial disparities were still present with the 2013 criteria, the NCCN Group 2 criteria, with fewer whites excluded compared with African Americans. So it's important to note that broader inclusion criteria may increase the sensitivity, but it does come at the cost of reduced specificity. It's also possible that lung cancer screening eligibility criteria may contribute to disparities not just by race, but also by sex. This study used the Chicago race eligibility — sorry, I clicked something by accident. This study used the CREST study, which is 883 lung cancer cases, and they sought to compare the sensitivity of the USPSTF criteria versus the risk prediction model, the PLCO. From the figure, you can see that the sensitivity is consistently higher for males versus females, and in addition, the sensitivities using the model are higher than using the USPSTF, excuse me, criteria. Women were more likely to be ineligible according to the USPSTF criteria because their smoking exposure was less than 20 pack years. So women just tend to smoke less than men. So the studies that I've reviewed so far have estimated the population-level outcomes associated with lung cancer screening eligibility criteria changes, or they've applied the criteria to existing cohorts of lung cancer cases and controls. This study is slightly different. In this study, they actually looked at differences in who was screened in practice. So they used data from one urban academic medical center from March of 2021 to December of 2021. They categorized patients into a 2013 cohort, and then they also categorized patients into a 2021 cohort. So the 2021 cohort just includes those patients who are in the newly eligible criteria, ages 50 to 54, or they had 20 to 29 pack years, or they also did have to quit less than 15 years ago, which is consistent. And so in this study, the numbers are fairly small because they don't have a lot of follow-up since the criteria changed recently, but you can see here, highlighted in red, the factors that seem to be different. They seem to have screened a higher proportion of African American patients with the 2021 criteria. There's a higher proportion who are currently smoking. There's a lower proportion with COPD and a higher proportion who were enrolled in Medicaid. So data from the Behavioral Risk Factor Surveillance System indicate that lung cancer screening uptake is quite low, less than 20%. We know that implementation of lung cancer screening has faced substantial barriers. Identification of the screening eligible population is challenging and requires knowledge of smoking history, which is inconsistently documented in the EHR. There's also substantial stigma associated with smoking that often makes it challenging for these patients to get care. In terms of insurance, Medicaid enrollees are at increased risk of lung cancer due to their higher cigarette smoking rates. About 25% of Medicaid patients smoke, compared with about 10% of Medicare or private insured patients. Yet we know that Medicaid programs across the country do not all cover lung cancer screening. Those living in these states could face more financial barriers to accessing lung cancer screening. There are also barriers to lung cancer screening in terms of geographic access, with those who live in rural areas may being less likely to access screening, but also less likely to access comprehensive lung cancer screening programs. So while expanded eligibility criteria are an important step to equity, barriers to lung cancer screening highlight the complex implementation challenges that we face. In summary, I think we've seen that there are changes in lung cancer screening eligibility criteria that have the potential to reduce lung cancer morbidity and decrease disparities, but this is not necessarily something that is going to happen. I think there are many aspects of this. Beyond age and PAC years and time since quit, there is the possibility of risk prediction models, but as I mentioned earlier, the USPSTF does not endorse this because there is currently insufficient evidence. Expanding the eligibility criteria alone is not enough to reduce the disparities. We must address disparities and implementation challenges to get lung cancer screening to areas in need. Okay, so I'm going to close it out and talk a little bit about 10 years of implementation of lung cancer screening and describe what I think are wins and losses. Many of these topics have been covered, so I'll, in the interest of time, try to go through quickly. My disclosures have to do with research funding, and these are, this is my brief outline. We'll talk about wins, losses, and challenges, including geographic location. I'll spend some time talking about adherence to lung screening, and Louise just did a great job highlighting the disparities. So I think the wins are that we definitely have evidence. We have definitive evidence. I don't think there's an argument to that at this point, especially as seen in these two trials and others. And this is the timeline that we've just been taken through as to when the national lung screening trial was first published in 2011, when the first task force recommendations came out in 2013, when CMS approved it as a benefit for its beneficiaries in 2015, and now we have the task force recommendations that have been updated to increase eligibility. I think the most important thing to highlight here as you talk about lung cancer screening is that bottom block, is that when you're screening patients, you want people that are asymptomatic and well and don't have competing comorbidities, which is a very difficult thing, I think, for us to conceptualize as physicians and also difficult to relate to patients. Exactly how do you tell someone, you know, you're not sick, you're pretty sick, and screening might not benefit you. But we do know that lung cancer, this is also another win, that lung cancer screening results in a stage shift. If you look here at this graph, in blue are the stage one patients versus the stage three and stage four patients. When you look at the screening trials across the bottom there, you can see that there was a higher proportion of patients with early disease, and this is what we want from lung cancer screening versus the general population prior to the advent of lung screening. So the wins are, we have recommendations in favor for screening, and for the most part, there's coverage for lung cancer screening, and with time, we have seen a stage shift. I think guidance for implementation is another big win. This was a great undertaking and has involved a lot of work and effort, and so one of the guidelines that have come out that were just highlighted were the screening for lung cancer guidelines that are a living guideline that change with time. The components necessary for high quality lung screening is often quoted and looked at as folks try to implement screening across the country. There are ATS and ACCP statements looking at the implementation about how to plan for implementation, how to implement and then maintain a program that are available, but there have been some losses, and you know, I think we can look at losses as uptake, but we can also, I like to look at it on the other side and say, well, this is a challenge and what can be done. So some of the challenge might be considered losses, but let's try to take a positive spin. So we know that lung cancer screening is efficient. You see the graph here from the National Lung Screening Trial that shows the mortality reduction, and when you compare it to the number needed to screen to prevent one lung cancer death, and you compare it to other commonly screened for cancers, you can see that lung cancer screening is highly efficient. The number needed to screen is quoted as 320, compared to breast cancer screening where the number needed to screen to prevent one death is over 1,900. But based on the guidelines in 2013, which have been discussed, with these types of recommendations also discussed, lung cancer screening was not initially well accepted, and why is that? Well, originally it was not endorsed by the American Academy of Family Physicians, and this is one of the largest primary care providers nationwide. There was some concerns about the generalizability of National Lung Screening Trial results, difficulty in identifying eligible patients through the electronic medical record, the false positive rate for some was considered to be too high, especially in areas of endemic mycoses with lots of pulmonary nodules. The shared decision-making was considered to be way too onerous, and the harm-to-benefit ratio was unclear in a real-world population. There's also concerns for low screening efficiency, and why recommend after only one randomized control trial? So these are the things that kind of plagued us at the beginning, and this shows you how uptake has been quoted as low. And so this is the, if you look at the slope of the lines for the other lung, I'm sorry, screening tests that are out there for cervical cancer and breast cancer, colorectal cancer, you can see that inflection point right around the time recommendations were made, and it went up pretty steeply. But then you look here at lung cancer screening, and the uptake was quite low. Looking at a variety of studies using the National Health Interview Survey, they estimated that uptake was around 4%, but then there was a blip looking at the behavioral risk factor surveillance system in 2019, which included 20 states, and it was estimated that about 20% of those eligible were screened in 2019. Obviously, there are some limitations to that number, but we sure hope that we're starting to see an uptick. These guidelines were already discussed in great detail, and I think the important thing to highlight here is that it will increase the number of eligible patients for screening in the United States by 81%. And so I think challenge is really access, and so when you talk about access to healthcare, you think about availability, accessibility, acceptability, and affordability, just to name a few. And this was already alluded to, however, there are some issues with rurality. There are certainly disparities in lung cancer incidents between black and whites that have already been highlighted, but this is even worse in rural versus urban areas. The prevalence of cigarette smoking, which we could argue is the highest risk factor for developing lung cancer, is much higher in rural communities where smokers consume greater than a pack a day versus metropolitan areas. So too, adolescents in rural counties start smoking earlier. My state is South Carolina, and it's not very uncommon for me to hear that an individual started smoking at age 10. Lung cancer incidence is 20% higher in rural versus urban areas, and rural areas with poverty have a higher lung cancer incidence and mortality rates. The most important thing in all of this is that rural areas are actually less likely to have accredited comprehensive lung cancer screening programs. And so I bring this slide to your attention, and this is really something that was described back in 1971, the law of inverse care, which essentially says that the availability of good medical care varies inversely with the need for it. So there's a high prevalence and a need, but the resources are not there. And this is true as it relates to lung screening. The highest distribution of people that currently smoke or would be at risk of lung cancer is down here in the southeast, and this is where the lowest number of lung cancer screening programs are. So it's that inverse ratio that really highlights the issues with access. This is another study that was published in the JNCI that looked at the variation in low-dose CT scans for lung cancer screening in the United States. In the South and the West, it was estimated that less than 4% had been screened, while in the Northeast, the rates were higher. Kentucky was the one exception, with high lung cancer rates as well as high screening rates, and I'll talk a little bit about what they did in Kentucky that is a bit exemplar. This is a color-coded graph that was published by Jan Erberth, and it's a little bit hard to decipher, so I've highlighted some things and put some happy faces. So this is the area with high access to screening. If you're in that light green, you also have a high incidence, so you want the areas with high incidence of lung cancer to have access to screening. The sad face, or the red blot there, is the state where it's the highest, I think that's Arkansas, the highest mortality from lung cancer, but also the lowest access to screening, and I think this ties into Medicaid expansion, which Dr. Henderson mentioned earlier. So access to care is covered for Medicare beneficiaries. We are age 65, CMS will cover it, but this is not for all Medicaid beneficiaries. There are 13 states that did not accept Medicaid expansion, and this is determined at the state level, and what we do know is that low-income individuals who often depend on Medicaid are also those that use the highest amount of tobacco. So racial and ethnic minority populations are also often, more often to be uninsured or underinsured, and what we have seen is that greater than 50 percent of patients meeting task force eligibility have Medicaid or are underserved. So as Dr. Henderson mentioned, and I can really just say this a bit once again because I think it's very important, dropping that age from 55 to 50 might increase the eligibility, but it's not necessarily going to increase the access because many of these individuals who are age 50 don't have insurance, they're uninsured, low socioeconomic status, don't have a primary care provider, so it's hard to know. So how do we increase lung cancer screening uptake? This is really the first time we're targeting a poor health habit for screening, right? It's not just that you're 50 and you should go get a colonoscopy, it's you smoked cigarettes, shame on you, and so I think we really need to look at destigmatizing lung cancer as it relates to smokers. We can look to other screening tests and look what works, and so in the breast cancer community, screening via outreach in community settings where you establish a local champion and outreach at churches and meet the people where they are has been a successful technique in increasing uptake and something I think we need to take advantage of as we try to improve uptake for lung screening. Public service announcements and media campaigns, we all know in October the NFL wears pink, this year they've switched that to screening uptake in general, which is a nice change, but to have that level of advocacy and outreach and information, I think we need to do a better job of educating our providers with tools in the EMR to help them dissuade the confusion about eligibility criteria. Other thoughts about this is improved access to screening, so I just went through the morality and how it's difficult, and so could we bring screening to the people? There's a paradigm for breast cancer screening with mobile mammography buses that I'm sure many of you have seen, and in the UK at the Yorkshire Lung Screening Trial, they developed what was called a lung health check where they had a mobile unit that worked in the mall parking lot with some great advertising and there was some really good uptake, and so these are some examples of the mobile screening units. The one on the right there is at the Levine Cancer Foundation in Charlotte, North Carolina. There are various ways to do this. There's the Winnebago model and then the self-contained chassis. For those of you that you don't know, in Chattanooga, Tennessee, there's a really robust mobile lung screening program that's been in place where they've been checking and developing the protocols for what a unit should look like, so I think there will be more to come on this, and it's really important as we try to reach those that may not have access, and I did mention a little bit about the Kentucky Leeds Collaborative. This was a really multi-pronged approach to educating providers where they were and accounts for that 15 percent uptake in lung screening in that state, and so I think it's something that we should all look to as we try to improve uptake in our areas. So what about adherence? I would say this is the most important challenge and a drum that I've been beating for some time now, and this is what Peter referred to as compliance. So in the National Lung Screening Trial over three years, there was 95 percent adherence. We know that was a randomized trial with clinical coordinators and money for patients to come back, not the real world, and so what is adherence? This is just the definition. Basically, it means coming back within the specified time in the context of an established screening program, and it's really needed to ensure that the potential benefits of a screening test are recognized. So the National Lung Screening Trial defined adherence as 15 months of baseline, and as I mentioned, it was 95 percent across all three rounds. CISNET is the Cancer Intervention and Surveillance Modeling Network, responsible for many of the modeling exercises that led us to determine age and smoking history, and when they did their calculations for the task force, they assumed 100 adherence, which would result in 497 fewer lung cancer deaths, but we know we're not going to achieve 100 adherence. So one CISNET group, and there are many, went back and modeled it differently, and what they found is if the adherence was only 46 percent, the mortality benefit of lung screening, which we all know is 20 percent, is cut in half, and the reason for this is that most cancers during lung screening, especially in the National Lung Screening Trial, were not found on the baseline scan. They were found on scan two and scan three, and so this really highlights the importance of coming back. When we look at adherence in other screen tests, you can see it's all over the board, and so I don't know what makes us think that we're going to achieve that 95 percent in lung screening. This also impacts the cost effectiveness. We talked about the quality adjusted life year for lung screening as being $81,000, which is acceptable in the United States, but again, that's based on a 95 percent adherence. If you start to model that adherence at a lower rate, what you find is the cost for screening goes up exponentially. So what about types of programs? There are centralized lung cancer screening programs, which involve a dedicated navigator with a software platform that manage the findings, confirm eligibility, conduct shared decision making, and there are decentralized lung screening programs, which is what we think of with any type of screening test. You see your primary care provider. They say you're eligible. You go for your test, and then there are hybrid programs where you might have a centralized screening program or the provider can do the screening on their own, and so what if the published adherence numbers look like based on program type, and I think Peter alluded to this earlier, but when you look at the centralized programs, and I think the VA is a really good example of this in their demonstration project, which was eight sites. It was 82 percent at year one. In Seattle, there was a multi-site study with a strict definition of adherence of 15 months that showed a 47 percent adherence. A prosper lung multi-site study showed 76 percent, so what I really want to highlight here is that a centralized program tends to be best. When we look at the national VA cohort of all patients screened adherence at one year was 63 percent, and that at our own institution at the Medical University of South Carolina, where we have a hybrid program, we looked at overall adherence, which was 56 percent, but when we stratified by who ordered the scan, whether it came from the centralized program or a primary care provider, we found that the adherence was much better in the central program, and I would point out that with the decentralized component, we are tweaking the primary care providers via the electronic medical record to remind them to bring their patients back, so even with that intervention, the adherence is still different, and then as we look at decentralized programs, you can see it's again as low as 30 percent, which argues why even bother screening if you're not going to bring folks back. When we look at predictors of non-adherence in lung screening program, I think it's factors that will come as no surprise to the group here. What we see is that female sex, age 65, a dedicated program, individuals who previously smoked, these are all patients that are likely to come back, and I would point out that age of 65, so that's the Medicare age where you have coverage for screening, so it's not a burden for your pocket to come back and have it annually. When we look at racial disparities in adherence to annual lung screening, in this study that was published by Anil Vachani and colleagues, what you see is that when the decentralized screened patients, African Americans were less likely to come back, but there was no difference in the centralized program, so the centralized types of programs not only improve adherence, but also diminish that disparity, which is a nice thing to see. So other predictors of lung cancer screening adherence that we have found by looking at our VA cohort, not surprising if you have a nodule, you're more likely to come back. Drive time distance, the further away you live from your center, the less likely you come back, so this kind of speaks to bringing screening to the people. The facility conducting screening and patient level factors, multi-substance abuse or PTSD were associated with less likely to come back. So some thoughts on improving adherence. I think this needs to be a multi-level type of intervention. I think dedicated navigators and tracking systems are a start. Actively contacting patients is a great idea, but sometimes the programs are busy and they don't have the capacity to call everyone. A way to integrate into the electronic medical record is a thought which can become very daunting as we all know tobacco pack year history is not very accurate. Provider education and patient level interventions are certainly needed. So in summary, there is definitely solid evidence that lung cancer screening reduces mortality in high-risk patients. We know that the task force recommendations are increasing the number of eligible, but it's really hard to say that that's actually going to increase uptake. Lung cancer screening is extremely complex. I wish I could just say it was one scan, but it's not. There are very many barriers that I think the three of us have highlighted this morning, but continued efforts to develop and support interventions to address these barriers are needed because we know lung cancer screening really does save lives. And with that I'll stop and thank you guys all for your kind attention and your attendance.

lung cancer screening

guidelines

evidence base

benefits and harms

reduction in mortality

potential harms

disparities in screening

increase uptake

improve adherence