Insights on Lung Cancer Diagnosis Trends and Screening Challenges in Community Oncology
In this interview, Jessica Paulus, ScD, senior director of real-world research at Ontada, discusses trends in advanced non–small cell and small cell lung cancer diagnoses, the impact of the COVID-19 pandemic on screening, and the urgent need for improved access to preventive care and public health initiatives.
Please introduce yourself by stating your name, title, organization, and relevant professional experience.
My name is Jess Paulus. I'm the senior director of real-world research at Ontada, which is a business of McKesson. Ontada is a real-world data business that analyzes data generated from the community oncology setting to help address unsolved or outstanding questions in cancer care. I am trained as an epidemiologist, specifically a cancer epidemiologist. I also have specialized training in the analysis of observational data and the associated methods for causal inference as well as kind of predictive modeling, using observational data.
Prior to joining Ontada, I was vice president of research at Om Research—another real-world data company headquartered in Boston—where I led a research team in these kinds of analyses. Before having these adventures in real-world data in the private sector, I was a faculty member at Tufts Medical School for about a dozen years, where I did independent research and mentored clinician scholars in clinical research methods.
What led you and your colleagues to conduct this study?
Lung cancer at an early stage is critical for maximizing the best outcomes for patients. That's true for virtually every cancer I can think of and it's important for lung cancer. The majority of disease for lung cancer—both small cell and non–small cell lung cancer—is diagnosed at an advanced stage (3 or 4). This is because that's typically when symptoms lead an individual to present to the health care system to be diagnosed.
There has been incredible investment in screening technologies for lung cancer over the last 20-plus years. In particular, in 2011 and 2013, the United States Preventive Services Task Force (USPSTF) and the American Cancer Society (ACS), and the National Comprehensive Cancer Network introduced guidance statements or guideline documents recommending low dose CT screening for patients or for individuals at higher risk of lung cancer, generally current and former smokers. Also, in 2020 there was the advent of the COVID-19 pandemic. This caused major health care utilization disruptions and delays in screening for many solid tumors because of decisions that patients, caregivers, and health societies and systems were making around slowing the pandemic spread.
Following these events in clinical medicine, we had curiosity about whether we would see a shift in the distribution of the stage of disease at presentation to our community oncology network. We wanted to leverage our real-world data to complement other sources of evidence in this area to provide insight into whether this diagnostic landscape is potentially changing.
We had the opportunity to do that in our data from Ontada, which is data generated from the community oncology setting within The US Oncology Network. This is a network of oncology providers that span a huge sector of the community oncology care setting in the US. We evaluated electronic health record (HER) data that was emanating from that network setting from the years 2013 to 2023. The important events in lung cancer that I noted earlier took place during this period. We have seen stage shifts or changes in the distribution of stage for many cancers around the time of the pandemic because of the delay in care. So, we were curious about all of those things, and that was the rationale for this analysis.
What were the main findings of the study?
One of the powerful things about this analysis is because we were leveraging data from this large community oncology setting. We had over a hundred 1000 patients with either non–small cell lung cancer (NSCLC) or small cell lung cancer (SCLC). We had a lot of statistical power and ability to generalize to the community oncology setting in the US.
Our main finding is that we saw a greater proportion of advanced disease at diagnosis. By that I mean stage 3 or stage 4 cancer increasing over the last decade over that 10-year interval I described as our study design. We also saw a marked uptick immediately following the COVID-19 pandemic. We saw this increase in the proportion of patients with stage 3 or stage 4 at presentation to the community oncology network.
Interestingly, we saw this both for NSCLC and SCLC. We did stratified analyses and saw similar trends for each. Not surprisingly, roughly between around two-thirds to three-quarters of the patients were stage 3 or stage 4 at diagnosis, and that corresponds to what we unfortunately know about this condition, which is that the majority of the time it is diagnosed at advanced stage. Of course, that's the problem we are trying to solve with the introduction of these screening guidelines as well as kind of promotion of screening care.
I want to mention as a caveat, is that we in The US Oncology Network only see patients who are diagnosed with lung cancer and who present to our network for care by medical oncologists. We do not have access to the general population of well individuals who may be at risk for lung cancer. So, this is a snapshot of those patients who already have cancer and who are presenting to our networks.
Did the results indicate any significant differences in advanced NSCLC/SCLC diagnoses across different demographics, such as age, gender, or geographic location? If so, what implications might this have for targeted interventions?
As I mentioned earlier, we did see similar trends in patterns in the burden of both NSCLC and SCLC over time. We saw this uptick over the last decade in increasing proportions of patients with advanced disease. Similarly, for both cancers, as part of the study, we didn't explore whether the trends in that increasing burden of disease in the community oncology setting varied by patient age or by geography or other types of demographic factors. However, geography can often be a source of variation in practice and also in patient behaviors. That would be an interesting analysis that we could look at that could shed light on possible drivers of this finding. But that was not part of the scope of this work.
The study highlights an increase in advanced-stage diagnosis of lung cancer from 2013 to 2023. What do you believe are the key factors contributing to this trend?
I'll start with the little bump that we saw around the COVID-19 pandemic. Overall, we saw a shift in more advanced stage of presentation to The US Oncology Network. As I mentioned earlier, that bump is reasonable given what we know about delays and care around repurposing of hospital resources and clinical resources to deal with the pandemic.
It’s less likely that access to oncology care just because the patient had lung cancer, in particular at advanced stages, was really a health crisis. Although we can imagine that there were some delays in provision of care or the appetite of patients to go present to one of these health systems when they might have symptoms of lung cancer that could be confused with COVID-19 infection, for example.
The changes we see around the pandemic could plausibly be attributed to these phenomenon around delays in screening, or some delays in access to care in terms of the overall trend, though over the last decade of that shifting toward more advanced disease. Candidly, that was not what we expecting to see. Clinical trials and other data show that these screenings and screening technologies (such as CT) for lung cancer does shift patients toward an earlier stage of disease at presentation and is associated with reductions in mortality.
However, one thing to keep in mind about the study design that I mentioned earlier, is that we can only make conclusions about the burden of advanced stage disease in the community oncology setting. Given the study and its design, we don't see that underlying population. We don't see patients who are being screened and presenting at earlier stages because this is was not designed to be a study of screening. In other words, we're only seeing patients once they get cancer. So, we can't speak to the role of screening and helping shift disease to earlier stages at diagnosis. Although that would have been what I would have expected given what we know about CT screening efficacy.
So, why are we seeing the numbers we are seeing? It could be that patients who are being screened and presenting at earlier stages are less likely to present to the community oncology setting versus other places to get their care. There is also an increased burden of disease in the community oncology setting. It's also important to note that because we're just seeing the patients who already have lung cancer within this one network, it could be that there have been changes around referrals that are, referral patterns, catchment populations that are funneling patients at earlier stages of disease to other health care settings outside the community oncology setting. We don't have access to the data that would allow us to really investigate these leads.
We can't rule out potential explanations that are driven by issues around real-world data. The capture of staging information and real-world data comes from both structured and unstructured real-world data. Over the past decade, we have been getting better and better at data quality and comprehensive capture and reducing missingness of stage over time. Data quality is also improving. When I say we, I mean the investments that Ontada has made in this area, and as a real- world data, oncology epidemiology community.
Based on the study's findings, what are the key recommendations for health care providers to improve early detection and reduce the incidence of advanced stage diagnoses in the future?
This study is important because investments and resources need to be allocated at the public health level. As I mentioned earlier, I work in a setting where we're taking care of patients and studying data from patients who already have lung cancer. But the single most important thing that we need to do collectively as a society is to promote access to treatment and health literacy around the criticality of screening programs for lung cancer (CT screenings) for those who are at higher risk, like individuals with a tobacco or smoking history.
According to the ACS and other data, not enough patients who are eligible for this type of screening are actually getting it right now, even though CT screening works to reduce lung cancer mortality. I saw some data recently that estimated fewer than 10% of patients who are eligible to be screened for lung cancer are getting this potentially lifesaving intervention. To put that in context, when we think about breast cancer screening or mammography, it depends a little bit on the part of the country we're talking about and the age group we're talking about. But we typically see among eligible women rates far in excess of 50% for annual mammography or mammography at the right cadence. So, sometimes 60%, sometimes 70% of women who are eligible to be screened are going through with their mammogram. They're getting access to it. They are getting that important preventive care. These kinds of victories in preventive care come with investment of time and resources from both governmental and non-governmental organizations.
Lung cancer is a disease that's unfortunately been stigmatized at a lot of levels, including disparities and funding. Lung cancer is responsible for more cancer deaths than any other disease. A quarter of cancer deaths in general are lung cancer deaths. But when we talk about federal funding for research, lung cancer gets less than 10% of the cancer research funding from the US government. As a society, we have to change this approach. For example, pink ribbon campaigns have been transformational for breast cancer research investment and moving the needle on survival. Lung cancer needs that ribbon. It needs that attention and investment.
In addition, I don't want to lose sight of even further upstream things we can do around lung cancer prevention and stage shift. At the public health level we need to continue to advocate and lobby for reducing our population's risk and exposures to known lung cancer, carcinogens. The big ones are, of course, tobacco smoking, radon exposure, and then air pollution. Tobacco smoking has the lion's share of attribution for lung cancer, but radon and air pollution are other exposures that are causally related to lung cancer and can benefit from policy work and other types of interventions to reduce our exposure to those agents.