Comparison of Race-Specific versus Race-Neutral Pulmonary Function Predicted Values in Patients Undergoing Lung Cancer Resection

R. Rajaram¹, A. Sheshadri², A. Baugh³, L. Bonnell⁴, L. Li², A. Vaporciyan⁵, M. Block⁶, E. A. David⁷, R. Habib⁴, D. Ost^2
1University of Texas-MD Anderson Cancer Center, Houston, Texas ²University of Texas MD Anderson Cancer Center, Houston, Texas ³University of California San Francisco, San Francisco, California ⁴The Society of Thoracic Surgeons, Chicago, Illinois ⁵The University of Texas MD Anderson Cancer Center, Houston, Texas ⁶Memorial Healthcare System, Hollywood, Florida ⁷University of Colorado, Aurora, Colorado

Purpose: Percent-predicted forced expiratory volume in 1 second (FEV1pp) is used for risk assessment in patients considered for lung resection. FEV1pp is adjusted for race despite concerns this may lead to inaccurate estimations of lung health. We compared prediction of race-specific versus race-neutral FEV1pp for pulmonary complications following lung cancer resection.

Methods: We identified patients who underwent lung cancer resection in the Society of Thoracic Surgeons General Thoracic Surgery Database (STS-GTSD) between 2002 and 2008. Both actual FEV1 and FEV1pp were used to generate race-specific and race-neutral FEV1pp for each patient using Global Lung Initiative reference equations, with other/miscellaneous used for race-neutral calculations. Patients with missing or implausible FEV1 and those missing race/ethnicity were excluded. We compared race-specific and -neutral model performance to predict a composite measure of pulmonary complications, defined as air leak >5 days, pneumonia, ARDS, bronchopleural fistula, pulmonary embolism, initial ventilator support>48 hours, tracheostomy, reintubation, or other pulmonary complications. Estimations of predicted postoperative FEV1 (ppoFEV1) were generated in patients who underwent left-sided lobectomies and pneumonectomies. Based on their ppoFEV1, patients were categorized into low (>60%), intermediate (30-60%), and high ( < 30%) risk groups per previously published guidelines. Pulmonary complications and operative mortality were compared between patients in discordant risk categories.

Results: There were 24,276 patients identified from 125 STS-GTSD participating centers. Most patients were non-Hispanic White (NHW) (n=21,130; 87.0%) or Black (n=1,912; 7.9%). Lobectomy was the most common operation performed (n=12,695; 52%), followed by sublobar resection (n=10,117; 41.7%), and bilobectomy/pneumonectomy (n=1,464; 6.0%). The mean FEV1pp for the entire cohort was higher using race-neutral (76.9%) compared to race-specific (72.7%) equations. Using race-specific equations resulted in a mean FEV1pp reduction of 5.3% for NHW patients and a mean FEV1pp increase of 6.2% for Black patients (Figure). Prediction of pulmonary complications using race-specific and race-neutral equations was nearly identical (C-statistic 0.70 and Brier score 0.12 for both equations).

Of the 5,422 patients with calculable ppoFEV1, 617 (11.4%) were recategorized into either a higher (n=65) or lower (n=552) surgical risk group when using race-neutral equations. Of those moving into a lower risk group, 98.0% were NHW. All patients reclassified into a higher risk group were Black (14.1% of all Black patients) with a pulmonary complication and mortality rate of 23.1% and 5.4%, respectively, higher than in those Black patients in whom race-specific and race-neutral risk categories were concordant (15.6%, 3.2%) though not statistically significant (p=0.135, 0.414).

Conclusion: Race-neutral FEV1pp performed equally well as race-specific equations in predicting pulmonary complications following lung cancer resection. Use of race-specific pulmonary equations should be reconsidered when evaluating surgical risk within this population.

Identify the source of the funding for this research project: The data for this research were provided by The Society of Thoracic Surgeons’ National Database Access and Publications Research Program.