Patient Position Alters Quantitative CT Measurements of Interstitial Lung Abnormalities

Standardizing the Quantitative Assessment of Interstitial Lung Abnormalities

The management of progressive fibrotic lung diseases increasingly relies on high-resolution computed tomography to quantify parenchymal changes and predict clinical decline [1]. While clinical practice guidelines provide a framework for managing complex pulmonary conditions, they often emphasize that technical factors can influence diagnostic accuracy [2, 3]. Advanced imaging techniques have highlighted the need for rigorous standardization in how thoracic data is captured and interpreted [4]. In the context of interstitial lung abnormalities, which may represent early stages of more severe disease, the precision of these measurements is paramount for accurate longitudinal tracking [1, 5]. To address this, a new prospective study evaluates how patient positioning during imaging affects the automated quantification of these early pulmonary findings.

Prospective Evaluation of Positional Variability

To determine whether patient orientation affects the digital measurement of pulmonary findings, researchers conducted a prospective study between February 2024 and February 2025. The investigation focused on the positional variability of quantitative computed tomography (qCT), an automated method that uses software to measure the volume and characteristics of lung tissue in patients with interstitial lung abnormalities. By comparing same-day scans, the study sought to determine if measurements of these early parenchymal changes remain consistent when a patient moves from a supine to a prone position. The protocol required participants to undergo sequential non-contrast supine and prone CT scans using identical acquisition parameters. This ensured that any observed differences were attributable to positioning rather than hardware settings or temporal disease progression. Of the 47 participants who initially provided consent, 38 individuals were included in the final analysis. This cohort had a mean age of 70.9 ± 6.4 years and included 27 men. By utilizing this standardized approach, the researchers could precisely isolate how gravity-dependent anatomical shifts influenced the software's ability to quantify both fibrotic and nonfibrotic lung components.

Automated Quantification and Statistical Rigor

The researchers utilized commercially available deep learning-based software to automate the quantification of interstitial lung abnormality components, moving beyond subjective visual assessment to a standardized digital metric. This software categorized findings into two primary groups. The first group comprised fibrotic components, which include reticulation and honeycombing. The second group comprised nonfibrotic components, specifically ground-glass opacities. These classifications were strictly defined in accordance with Fleischner Society definitions, ensuring that the digital measurements aligned with established international radiological standards for identifying early interstitial disease. To evaluate the reliability of these measurements across different patient positions, the study employed a rigorous statistical framework. The researchers used paired t tests to compare the mean differences between supine and prone measurements. They also applied Bland-Altman analysis with 95% limits of agreement (LOA), a statistical method used to determine the range within which most differences between two clinical measurements fall. Furthermore, concordance correlation coefficients (CCC) were calculated to assess the degree of agreement between the supine and prone qCT values, providing a metric for how closely the two sets of data matched along a line of perfect agreement. This multi-layered statistical approach allowed the authors to pinpoint exactly where the automated software struggled to maintain consistency when the patient changed orientation.

Discrepancies in Fibrotic and Total ILA Extent

The analysis demonstrated that patient orientation significantly impacts the automated quantification of lung disease. The mean total extent of interstitial lung abnormalities was greater on supine scans than on prone scans (1.76% versus 1.39%, p = 0.02). This discrepancy was primarily driven by the fibrotic component of the disease. The mean extent of fibrotic interstitial lung abnormalities was significantly higher on supine images compared to prone images (1.43% versus 1.12%, p = 0.007). For clinicians tracking disease progression, these findings suggest that a change in patient position alone could be misinterpreted as a true clinical change in lung volume. Statistical measures of agreement further highlight the lack of interchangeability between these two positions. The researchers found that the 95% limits of agreement between supine and prone scans for total interstitial lung abnormality extent ranged from -1.49% to 2.23%, indicating a wide margin of potential measurement error. When isolating the fibrotic component, the 95% limits of agreement ranged from -0.97% to 1.58%. While the agreement between positions was moderate for the fibrotic component (CCC = 0.770), the variability remains high enough to complicate longitudinal assessments. These data underscore the necessity of maintaining consistent positioning during follow-up imaging to ensure that changes in quantitative metrics reflect actual biological progression rather than positional artifacts.

Poor Reproducibility in Nonfibrotic and Dependent Zones

The study identified even greater instability when quantifying nonfibrotic components of interstitial lung abnormalities, such as ground-glass opacities. For these nonfibrotic elements, the 95% limits of agreement between supine and prone scans ranged from -0.89% to 1.03%. This variance is clinically significant because the agreement between positions was poor for the nonfibrotic component (CCC = 0.431). Such a low coefficient indicates that automated measurements of ground-glass opacities are highly sensitive to patient orientation, making it difficult for clinicians to distinguish between actual changes in inflammatory activity and simple positional shifts in lung density. A primary driver of this measurement instability is the effect of gravity on lung tissue. The researchers noted that the lowest reproducibility was observed in dependent lung zones, which are the anatomical regions most affected by gravity and prone to atelectasis or increased capillary blood volume when a patient is supine. Overall, the measurement variability of qCT results for interstitial lung abnormalities between supine and prone scans was approximately 1.9%. For the practicing clinician, this margin of error confirms that supine and prone scans are not interchangeable. To ensure accurate longitudinal tracking of early fibrotic lung disease, patients must be positioned consistently across all serial imaging sessions to prevent positional artifacts from masking or mimicking true disease progression.

References

1. Raghu G, Blink BVD, Hamblin M, et al. Effect of Recombinant Human Pentraxin 2 vs Placebo on Change in Forced Vital Capacity in Patients With Idiopathic Pulmonary Fibrosis: A Randomized Clinical Trial. 2018. doi:10.1001/jama.2018.6129

2. Galiè N, Humbert M, Vachiéry J, et al. 2015 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension. European Respiratory Journal. 2015. doi:10.1183/13993003.01032-2015

3. Patterson TF, Thompson GR, Denning DW, et al. Practice Guidelines for the Diagnosis and Management of Aspergillosis: 2016 Update by the Infectious Diseases Society of America. Clinical Infectious Diseases. 2016. doi:10.1093/cid/ciw326

4. Abadi E, Segars WP, Tsui B, et al. Virtual clinical trials in medical imaging: a review. Journal of Medical Imaging. 2020. doi:10.1117/1.jmi.7.4.042805

5. Bradley JS, Byington CL, Shah SS, et al. The Management of Community-Acquired Pneumonia in Infants and Children Older Than 3 Months of Age: Clinical Practice Guidelines by the Pediatric Infectious Diseases Society and the Infectious Diseases Society of America. Clinical Infectious Diseases. 2011. doi:10.1093/cid/cir531