By Chris Wood
Lung cancer remains the leading cause of cancer death in the United States, despite meaningful advances in targeted therapies, immunotherapies, and other treatments for advanced disease. That reality points to an opportunity that deserves attention as the next breakthrough therapy, which is finding lung cancer earlier and managing suspicious lung nodules more effectively.
The history of cancer care consistently demonstrates the power of early detection. When cancer is identified while it is still localized, treatment is more likely to be curative. For lung cancer, the difference is particularly significant. Patients diagnosed with localized disease have substantially better survival outcomes than those diagnosed after the cancer has spread, the prominent issue being that we are not reaching enough people early enough.
Lung cancer screening is one of the clearest tools we have for shifting diagnosis toward earlier stages, yet screening participation remains remarkably low. The CDC estimates that only about 18% of US adults who meet screening recommendations have been screened. While we work to bring compliance up, there is also increasing interest in expanding the guidelines for who should receive screening. This is because much more than half of those diagnosed with lung cancer do not meet the screening criteria. One group recently highlighted for possible inclusion are women with breast cancer (the most diagnosed cancer in women). These women are at increased risk for subsequent lung cancer, and lung cancer screening could provide protection against follow-on cancers.
In addition, patients who undergo CT scans for reasons that have nothing to do with lung cancer screening, such as those more commonly done for abdominal pain, cardiac evaluation, infections, or other conditions, tend to be at high risk for lung cancer. This is important to note, as these incidentally detected nodules can also represent an important pathway to earlier lung cancer diagnosis, if appropriately risk-stratified and followed.
To be clear, a nodule identified on a scan is not by any means a cancer diagnosis, as most nodules are benign. Yet determining which nodules warrant surveillance, additional imaging, biopsy, or referral requires clinical judgment and, often, follow-up. Unfortunately, many providers have found that following the enormous volume of lung nodules is nearly impossible in this resource-strained environment. While up to 40% of chest CT scans identify a nodule, only 22% of patients with small nodules actually follow up on time. Even if perfect follow-up adherence was obtainable, 50% of lung cancers “upstage” during the interval time before the follow-up scan.
The goal is to identify nodules of concern without creating unnecessary testing and burden for patients with low-risk findings. This is where AI can make a meaningful difference, not by replacing clinicians, but by helping health care systems manage an enormous volume of information more consistently.
Proactivity vs Reactivity
AI-based computer-aided radiomic diagnostic technologies can analyze imaging data and identify characteristics associated with malignancy risk. When used appropriately, these tools have the potential to help clinicians distinguish patients who warrant greater attention from those who may be appropriate for routine surveillance.
Today, lung nodule programs can generate substantial workloads for radiologists,
pulmonologists, and many other primary care teams and nurse navigators. Each
potentially actionable finding can create a cascade of tasks, such as imaging review, patient communication, determining appropriate follow-up intervals, ordering additional testing, and tracking and reassessing the nodule over time.
Radiomics and AI can help prioritize the patients most likely to benefit from additional evaluation while reducing unnecessary follow-up for patients whose nodules are less concerning. That can ease the burden on already stretched clinical teams and allow navigators to spend more time on the patients who need active intervention. Most importantly, it can help make lung nodule management more sustainable.
The future of lung cancer detection cannot depend solely on adding more people to an already constrained health care workforce. If every additional screening or incidental finding requires a proportional increase in manual review and navigation, programs will struggle to scale.
The opportunity is to build systems in which imaging, risk assessment, clinical guidelines, tracking, and patient navigation work together, and AI is becoming one critical component of that infrastructure, helping transform lung nodule management from a largely reactive process into a more proactive and risk-informed one. This does not mean that AI should make the final clinical decision, but rather that we should use technology where it can augment clinical expertise, reduce variability, and help health care professionals focus time where it has the greatest impact.
As lung cancer is the most dangerous disease of its kind, research will, and should, continue pursuing better drugs and more effective treatments. That said, if we want to make a meaningful dent in mortality, we need to sharpen our focus on increasing proactive measures.
Improving screening participation, capturing incidental nodules, identifying which patients are most likely to have cancer, and ensuring that appropriate follow-up actually happens represents one of the most consequential opportunities and pain points in lung cancer care today. The next and currently developing major advancement in lung cancer survival may not come from a new drug alone but from getting the right patient to the right diagnosis earlier, and building a health care system capable of doing that consistently at scale for other diagnostics.
— Chris Wood is the CEO of RevealDX.