Association of Progressive PPFE and Mortality in Lung Cancer Screening Cohorts

Shahab Aslani, Mehran Azimbagirad, Daryl Cheng, Daisuke Yamada, Ryoko Egashira, Adam Szmul, Justine Chan-Fook, Robert Chapman, Alfred Chung Pui So, Shanshan Wang, John McCabe, Tianqi Yang, Jose M Brenes, Eyjólfur Gudmundsson, The Summit Consortium, Susan M. Astley, Daniel C. Alexander, Sam M. Janes, Joseph Jacob

TL;DR

This study demonstrates that the progression of pleuroparenchymal fibroelastosis (PPFE)—a specific type of upper-lung scarring—is a potent independent predictor of mortality and respiratory illness in lung cancer screening populations. By leveraging deep learning to quantify these changes on CT scans over time, researchers have identified a clinically relevant imaging biomarker that could help clinicians identify high-risk individuals long before they become symptomatic.

What Problem Does This Paper Address?

Lung cancer screening (LCS) programs generate millions of low-dose CT (LDCT) scans. While the primary goal is to find nodules, these scans often reveal "incidental" findings—abnormalities that aren't cancer but may still impact health. One such finding is Pleuroparenchymal Fibroelastosis (PPFE), a form of lung fibrosis (scarring) that primarily affects the upper lobes.

While PPFE is known to be deadly in patients with established interstitial lung diseases, its significance when found incidentally in a "healthy" screening population (typically heavy smokers or former smokers) was previously unclear. Specifically, the authors wanted to know: Does it matter if this scarring gets worse over time, and can we measure that progression accurately enough to predict a patient's risk of death or hospitalization?

Methods

The researchers employed a longitudinal approach, comparing "baseline" scans to "follow-up" scans to measure the rate of change.

  • Automated Quantification: Instead of relying on subjective visual scores by radiologists, the team used a deep learning algorithm (based on a U-Net/V-Net architecture) to segment and calculate the volume of PPFE in the upper lung zones.
  • Defining Progression: They established a "distribution-based threshold" for progression. If the PPFE volume increased by more than 0.41 cm³/year, the participant was classified as having "progressive PPFE."
  • Statistical Modeling: They used Cox proportional hazards models to link this progression to mortality, adjusting for age, sex, smoking history, and even the presence of other lung abnormalities (like Interstitial Lung Abnormalities, or ILA).

Datasets

The study validated its findings across two massive, geographically distinct cohorts:

  1. NLST (National Lung Screening Trial): A US-based cohort of 7,980 participants with scans from 2002–2007 and up to ten years of survival data.
  2. SUMMIT Study: A more recent UK-based cohort (London) of 8,561 participants. This dataset provided richer clinical data, including hospital admissions and medication use.

The AI model was trained on 100 manually annotated scans from the SUMMIT dataset before being deployed across the full cohorts.

Key Insights and Ideas

  • Progression is the Key: While having some scarring at the start (baseline) is a risk factor, the rate at which it grows is a much more consistent and powerful predictor of death.
  • Independence from Other Diseases: Progressive PPFE isn't just a side effect of general lung aging or other types of fibrosis. The association with mortality remained significant even after accounting for "standard" interstitial lung abnormalities (ILA) and lung function (FVC).
  • Upper-Lobe Significance: Most common lung fibroses (like IPF) start at the bottom of the lungs. PPFE is unique because it hits the top, and this study confirms that upper-lobe monitoring is critical in screening programs.
  • Morbidity Link: In the SUMMIT cohort, progressive PPFE wasn't just linked to death; it was a massive red flag for "respiratory morbidity"—meaning more hospital stays and a higher reliance on steroids and antibiotics.

Key Results

  • Mortality Risk: Progressive PPFE was independently associated with a higher risk of death in both groups. In the NLST, the Hazard Ratio (HR) was 1.25, while in the SUMMIT cohort, it was a staggering 3.14.
  • Hospitalizations: Participants in the SUMMIT study with progressive PPFE had nearly three times the rate of respiratory-related hospital admissions (Incidence Rate Ratio [IRR] = 2.79).
  • Medication: These individuals were 55% more likely to require increased use of antibiotics or steroids for chest symptoms.
  • Prevalence: "Progressors" made up about 5.4% of the NLST cohort and 1.5% of the SUMMIT cohort.

Limitations and Future Work

The authors acknowledge several limitations:

  • Observational Nature: The study shows a strong association but cannot definitively prove that PPFE causes the increased mortality (though the link to respiratory-specific deaths suggests a strong causal pathway).
  • Scan Resolution: LDCT scans used in screening have lower resolution than diagnostic-grade CTs, which might lead to underestimating very subtle, early-stage lesions.
  • Biological Mechanism: The underlying "trigger" that causes PPFE to progress in some smokers but not others remains unknown.

Future Work: The researchers suggest that quantitative AI tools for PPFE should be integrated into lung cancer screening workflows. This would allow for "risk stratification," where patients showing rapid progression could be referred to lung specialists (pulmonologists) earlier for intervention.

Summary updated

Abstract

Background: Pleuroparenchymal fibroelastosis (PPFE) is an upper lobe predominant fibrotic lung abnormality associated with increased mortality in established interstitial lung disease. However, the clinical significance of radiologic PPFE progression in lung cancer screening (LCS) populations remains unclear. Methods: We analysed longitudinal low-dose CT scans and clinical data from two LCS studies: National Lung Screening Trial (NLST; n=7,980); SUMMIT study (n=8,561). An automated algorithm quantified PPFE volume on baseline and follow-up scans. Annualised change in PPFE was derived and dichotomised using a distribution-based threshold to define progressive PPFE. Associations between progressive PPFE and mortality were evaluated using Cox proportional hazards models adjusted for demographic and clinical variables. In SUMMIT cohort, associations between progressive PPFE and clinical outcomes were assessed using incidence rate ratios (IRR) and odds ratios (OR). Findings: Progressive PPFE independently associated with mortality in both LCS cohorts (NLST: Hazard Ratio (HR)=1.25, 95% Confidence Interval (CI): 1.01--1.56, p=0.042; SUMMIT: HR=3.14, 95% CI: 1.66--5.97, p<0.001). Within SUMMIT, progressive PPFE was strongly associated with higher respiratory admissions (IRR=2.79, p<0.001), increased antibiotic and steroid use (IRR=1.55, p=0.010), and showed a trend towards higher modified medical research council scores (OR=1.40, p=0.055). Interpretation: Radiologic PPFE progression independently associates with mortality across two large LCS cohorts, and associates with adverse clinical outcomes. Quantitative assessment of PPFE progression may provide a clinically relevant imaging biomarker to identify individuals at increased risk of respiratory morbidity within LCS programmes.
BibTeX
@misc{aslani2026association,
  author = {Shahab Aslani and Mehran Azimbagirad and Daryl Cheng and Daisuke Yamada and Ryoko Egashira and Adam Szmul and Justine Chan-Fook and Robert Chapman and Alfred Chung Pui So and Shanshan Wang and John McCabe and Tianqi Yang and Jose M Brenes and Eyjólfur Gudmundsson and The SUMMIT Consortium and Susan M. Astley and Daniel C. Alexander and Sam M. Janes and Joseph Jacob},
  title = {{Association of Progressive PPFE and Mortality in Lung Cancer Screening Cohorts}},
  year = {2026},
  eprint = {2603.09531},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2603.09531},
}
Figure