Emerging strategies and technologies aid detection and classification of lung nodules

A vast majority of patients with pulmonary nodules do not have lung cancer; however, it is critical that these nodules be detected and classified early to improve patient outcomes.

During AABIP: Spotlight on Nodules: Unraveling the Future of Detection and Tracking, at the CHEST Annual Meeting in Honolulu, three leading experts covered a broad array of tools and technologies to accomplish this objective.

Building an incidental nodule program

Susan Garwood, MD
Susan Garwood, MD

Susan Garwood, MD, Physician Director of Pulmonary Services at HCA Enterprise and Advanced Bronchoscopist at Interventional Pulmonary of Nashville, described how she built a program across HCA’s 180 hospitals that systematized lung nodule tracking and lung cancer screening.

When Dr. Garwood started this initiative, she put together a stakeholder group across health system management and specialties to help develop an execution plan.

Her team built its own nodule navigation system based on natural language processing and purchased care coordination software. The navigation software identified nodules greater than 6 mm that had certain keyword characteristics.

Once these patients were identified, their data went to the navigator/care coordinator who classified them as “eligible” or “ineligible.” Eligible patients’ charts were referred to a physician who would determine next steps. Those undergoing low-dose CT screening were also assigned to navigation if they were considered high risk (Lung-RADS category 3 or 4).

In 2021 alone, 1.4 million patients were screened for incidental nodules, 20,000 had nodules greater than 6 mm, 63% of these patients agreed to be navigated, and 1,400 were found to have lung cancer. Cancer was found in a much smaller percentage of patients who went through traditional low-dose CT screening.

“The most important thing you can take away is to go back to your health care system and talk about not just screening but incidental [nodule detection],” she told session attendees.

Unraveling the data

Akrum Al-Zubaidi, DO, FCCP
Akrum Al-Zubaidi, DO, FCCP

Akrum Al-Zubaidi, DO, FCCP, Founder and CEO of Eon Health, discussed the important role of metadata in pulmonary nodule detection and tracking.

In health care, metadata—data about data—fall into several categories, including descriptive, administrative, and structural. New types of metadata are also emerging.

“Managing metadata is the difference between a well-oiled data strategy vs a cluttered repository of data,” he said. “The biggest question that you need to ask yourself is, ‘Is your data working for you, or are you working for your data?’”

Data that affect the management of pulmonary nodules are currently isolated. These siloes, which contain siloes of their own, include linguistics, image analysis, devices and procedures, pathology and biomarkers, and electronic health records (EHRs).

“Every single EHR probably has five separate places where smoking history could be found, with probably different variable answers,” he said. “When you try to leverage that data, it becomes extremely, painstakingly labor intensive, which costs money and doesn’t allow you to scale your programs.”

Dr. Al-Zubaidi anticipates that as the cost of each data source and workflow disruption decreases over time, fully integrated data ecosystems will be overlaid on EHR systems to help manage not just pulmonary nodules but all abnormalities that need follow-up, including chronic disease.

Next steps after detection

Avrum Spira, MD, MSc
Avrum Spira, MD, MSc

Avrum Spira, MD, MSc, Global Head of the Lung Cancer Initiative at Johnson & Johnson and Professor of Medicine, Pathology, and Bioinformatics at Boston University Chobanian and Avedisian School of Medicine, focused on the stratification of nodules once they are detected.

Dr. Spira described the ways nodules found on CT scans are risk stratified today: subjective clinical assessment combined with guidelines and clinical risk models. Based on current tools, many nodules fall under a wide-ranging “indeterminate” categorization of cancer probability—10% to 65% pretest risk of disease.

“Quite honestly, the guidelines are not that helpful,” he said. “And there’s a lot of heterogeneity in how we manage these patients today.”

There is a rapidly evolving landscape of diagnostic tests. While some are already in clinical use or in the late stages of testing, most have not yet been used clinically to stratify indeterminate nodules because they have not been clinically validated. Tests currently in use or development include radiomics (imaging matched with machine learning), plus molecular biomarkers such as blood proteomics, blood microRNA, circulating tumor cells and tumor DNA, airway transcriptomics, and volatile organic compounds (breath).

To be useful, Dr. Spira posited, biomarkers should reduce the time to lung cancer diagnosis without increasing unnecessary invasive procedures or reduce unnecessary invasive procedures without increasing time to lung cancer diagnosis. He expects future nodule stratification modalities to integrate multiple markers for improved performance, use markers to predict the risk of future lung cancer, and assess ground glass opacities for a chance to intercept the disease.

For more information on using biomarkers in lung cancer management, check out CHEST’s webinar series.