Journal of Cardio-Thoracic Medicine

Journal of Cardio-Thoracic Medicine

The PLEURA I Study: Development of a new Multisite Ultrasound-Based Model for Estimation of Pleural Fluid Volume and Paul’s formula

Document Type : Original Article

Authors
1 MOSC MEDICAL MISSION HOSPITAL, THRISSUR, KERALA, INDIA- 680503
2 MOSC MEDICA MISSION HOSPITAL, KUNNAMKULAM, THRISSUR, KERALA, INDIA- 680503
10.22038/jctm.2026.98242.1538
Abstract
Introduction

From 2010 onwards, guidelines including that of British Thoracic Society (BTS) recommends the use of ultrasound in all the pleural fluid evaluation procedures. There are various established formulas to estimate the pleural fluid volume using the ultrasonography. The two formulas by Goecke and Schwerk, Einberger’s formula, Balik’s formula and Usta’s formula are some of the most commonly used ones. But considering only a single or couple of measurements to estimate the total volume of pleural effusion, especially moderate to large effusions will definitely have its own disadvantages. Objective of PLEURA 1 study was to develop a new multi- site ultrasound based model for a more accurate estimation of the pleural fluid volume and to compare it with the existing formulae.

Material and Methods

PLEURA 1, a prospective observational study was conducted in a tertiary care centre in South India. A novel multisite parameter (Mean9) was defined using nine different ultrasonographic dimensions. Pearson correlation coefficients (r) were calculated to assess the strength of association between predicted and actual pleural fluid volume. Model performance was assessed using the coefficient of determination (R²), mean absolute error (MAE), and root mean square error (RMSE) and compared with that of the existing erect position formulas.

Results

Study cohort consisted of twenty-four patients with pleural effusion. Mean9 demonstrated a strong linear relationship with actual pleural fluid volume (Pearson r = 0.853, 95% CI 0.68–0.94; p < 0.0001). ‘Paul’s formula’, ‘V = 41.4 + 270.8 × Mean9’ was derived using linear regression. Paul’s formula maintained the highest predictive performance among all evaluated models (LOOCV r = 0.820, LOOCV R² = 0.672, MAE = 422 mL, RMSE = 504 mL), and demonstrated better correlation to actual pleural fluid volume than the established erect-position formulas.

Conclusion

PLEURA 1 study and Paul’s formula suggest that incorporating multiple dimensions of the pleural effusion into a predictive model may improve the estimation of pleural fluid volume. PLEURA 2 validation study will be the second phase of the study were Paul’s formula will be applied along with other established formulae in an entirely new and larger cohort.
Keywords


Articles in Press, Accepted Manuscript
Available Online from 22 September 2026