Tuberculosis and Respiratory Diseases

  • 제86권3호
  • /
  • Pages.203-215
  • /
  • 2023
  • /
  • 1738-3536(pISSN)
  • /
  • 2005-6184(eISSN)
대한결핵및호흡기학회 (The Korean Academy of Tuberculosis and Respiratory Diseases)
권호 표지
DOI QR코드

DOI QR Code

Prediction of Postoperative Lung Function in Lung Cancer Patients Using Machine Learning Models

  • Oh Beom Kwon (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Solji Han (Department of Applied Statistics, Yonsei University) ;
  • Hwa Young Lee (Division of Allergy, Department of Internal Medicine, The Catholic University of Korea) ;
  • Hye Seon Kang (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Sung Kyoung Kim (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Ju Sang Kim (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Chan Kwon Park (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Sang Haak Lee (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Seung Joon Kim (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Jin Woo Kim (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea) ;
  • Chang Dong Yeo (Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, College of Medicine, The Catholic University of Korea)
  • 투고 : 2022.04.08
  • 심사 : 2023.04.10
  • 발행 : 2023.07.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

Background: Surgical resection is the standard treatment for early-stage lung cancer. Since postoperative lung function is related to mortality, predicted postoperative lung function is used to determine the treatment modality. The aim of this study was to evaluate the predictive performance of linear regression and machine learning models. Methods: We extracted data from the Clinical Data Warehouse and developed three sets: set I, the linear regression model; set II, machine learning models omitting the missing data: and set III, machine learning models imputing the missing data. Six machine learning models, the least absolute shrinkage and selection operator (LASSO), Ridge regression, ElasticNet, Random Forest, eXtreme gradient boosting (XGBoost), and the light gradient boosting machine (LightGBM) were implemented. The forced expiratory volume in 1 second measured 6 months after surgery was defined as the outcome. Five-fold cross-validation was performed for hyperparameter tuning of the machine learning models. The dataset was split into training and test datasets at a 70:30 ratio. Implementation was done after dataset splitting in set III. Predictive performance was evaluated by R2 and mean squared error (MSE) in the three sets. Results: A total of 1,487 patients were included in sets I and III and 896 patients were included in set II. In set I, the R2 value was 0.27 and in set II, LightGBM was the best model with the highest R2 value of 0.5 and the lowest MSE of 154.95. In set III, LightGBM was the best model with the highest R2 value of 0.56 and the lowest MSE of 174.07. Conclusion: The LightGBM model showed the best performance in predicting postoperative lung function.

키워드

과제정보

This study was supported by a 2020 grant from The Korean Academy of Tuberculosis and Respiratory Diseases.

참고문헌

  1. Global Burden of Disease Cancer Collaboration; Fitzmaurice C, Dicker D, Pain A, Hamavid H, Moradi-Lakeh M, et al. The global burden of cancer 2013. JAMA Oncol 2015;1:505-27.
  2. Young RP, Hopkins RJ, Christmas T, Black PN, Metcalf P, Gamble GD. COPD prevalence is increased in lung cancer, independent of age, sex and smoking history. Eur Respir J 2009;34:380-6. https://doi.org/10.1183/09031936.00144208
  3. Barta JA, Powell CA, Wisnivesky JP. Global epidemiology of lung cancer. Ann Glob Health 2019;85:8.
  4. Venuta F, Diso D, Onorati I, Anile M, Mantovani S, Rendina EA. Lung cancer in elderly patients. J Thorac Dis 2016;8(Suppl 11):S908-14. https://doi.org/10.21037/jtd.2016.05.20
  5. Donington J, Ferguson M, Mazzone P, Handy J Jr, Schuchert M, Fernando H, et al. American College of Chest Physicians and Society of Thoracic Surgeons consensus statement for evaluation and management for high-risk patients with stage I non-small cell lung cancer. Chest 2012;142:1620-35. https://doi.org/10.1378/chest.12-0790
  6. Oswald NK, Halle-Smith J, Mehdi R, Nightingale P, Naidu B, Turner AM. Predicting postoperative lung function following lung cancer resection: a systematic review and meta-analysis. EClinicalMedicine 2019;15:7-13. https://doi.org/10.1016/j.eclinm.2019.08.015
  7. Lim E, Baldwin D, Beckles M, Duffy J, Entwisle J, FaivreFinn C, et al. Guidelines on the radical management of patients with lung cancer. Thorax 2010;65 Suppl 3:iii1-27. https://doi.org/10.1136/thx.2010.145938
  8. Wyser C, Stulz P, Soler M, Tamm M, Muller-Brand J, Habicht J, et al. Prospective evaluation of an algorithm for the functional assessment of lung resection candidates. Am J Respir Crit Care Med 1999;159(5 Pt 1):1450-6. https://doi.org/10.1164/ajrccm.159.5.9809107
  9. Wu MT, Pan HB, Chiang AA, Hsu HK, Chang HC, Peng NJ, et al. Prediction of postoperative lung function in patients with lung cancer: comparison of quantitative CT with perfusion scintigraphy. AJR Am J Roentgenol 2002;178:667-72. https://doi.org/10.2214/ajr.178.3.1780667
  10. Cukic V. Preoperative prediction of lung function in pneumonectomy by spirometry and lung perfusion scintigraphy. Acta Inform Med 2012;20:221-5. https://doi.org/10.5455/aim.2012.20.221-225
  11. Batihan G, Ceylan KC, Usluer O, Kaya SO. Video-assisted thoracoscopic surgery vs thoracotomy for non-small cell lung cancer greater than 5 cm: is VATS a feasible approach for large tumors? J Cardiothorac Surg 2020;15:261.
  12. Landreneau RJ, Hazelrigg SR, Ferson PF, Johnson JA, Nawarawong W, Boley TM, et al. Thoracoscopic resection of 85 pulmonary lesions. Ann Thorac Surg 1992;54:415-20. https://doi.org/10.1016/0003-4975(92)90430-C
  13. Zeiher BG, Gross TJ, Kern JA, Lanza LA, Peterson MW. Predicting postoperative pulmonary function in patients undergoing lung resection. Chest 1995;108:68-72. https://doi.org/10.1378/chest.108.1.68
  14. Kushibe K, Kawaguchi T, Kimura M, Takahama M, Tojo T, Taniguchi S. Exercise capacity after lobectomy in patients with chronic obstructive pulmonary disease. Interact Cardiovasc Thorac Surg 2008;7:398-401. https://doi.org/10.1510/icvts.2007.165696
  15. Chen JH, Asch SM. Machine learning and prediction in medicine: beyond the peak of inflated expectations. N Engl J Med 2017;376:2507-9. https://doi.org/10.1056/NEJMp1702071
  16. Harford S, Darabi H, Del Rios M, Majumdar S, Karim F, Vanden Hoek T, et al. A machine learning based model for Out of Hospital cardiac arrest outcome classification and sensitivity analysis. Resuscitation 2019;138:134-40. https://doi.org/10.1016/j.resuscitation.2019.03.012
  17. Hirano Y, Kondo Y, Sueyoshi K, Okamoto K, Tanaka H. Early outcome prediction for out-of-hospital cardiac arrest with initial shockable rhythm using machine learning models. Resuscitation 2021;158:49-56. https://doi.org/10.1016/j.resuscitation.2020.11.020
  18. Toste PA, Lee JM. Limited resection versus lobectomy in early-stage non-small cell lung cancer. J Thorac Dis 2016;8:E1511-3. https://doi.org/10.21037/jtd.2016.11.71
  19. Kodama K, Doi O, Higashiyama M, Yokouchi H. Intentional limited resection for selected patients with T1 N0 M0 non-small-cell lung cancer: a single-institution study. J Thorac Cardiovasc Surg 1997;114:347-53. https://doi.org/10.1016/S0022-5223(97)70179-X
  20. Durham AL, Adcock IM. The relationship between COPD and lung cancer. Lung Cancer 2015;90:121-7. https://doi.org/10.1016/j.lungcan.2015.08.017
  21. Shin TR, Oh YM, Park JH, Lee KS, Oh S, Kang DR, et al. The prognostic value of residual volume/total lung capacity in patients with chronic obstructive pulmonary disease. J Korean Med Sci 2015;30:1459-65. https://doi.org/10.3346/jkms.2015.30.10.1459
  22. Matsumoto R, Takamori S, Yokoyama S, Hashiguchi T, Murakami D, Yoshiyama K, et al. Lung function in the late postoperative phase and influencing factors in patients undergoing pulmonary lobectomy. J Thorac Dis 2018;10:2916-23. https://doi.org/10.21037/jtd.2018.05.27
  23. Kwon OB, Yeo CD, Lee HY, Kang HS, Kim SK, Kim JS, et al. The value of residual volume/total lung capacity as an indicator for predicting postoperative lung function in non-small lung cancer. J Clin Med 2021;10:4159.
  24. Kwak SK, Kim JH. Statistical data preparation: management of missing values and outliers. Korean J Anesthesiol 2017;70:407-11. https://doi.org/10.4097/kjae.2017.70.4.407
  25. Khan SI, Hoque AS. SICE: an improved missing data imputation technique. J Big Data 2020;7:37.
  26. Zhang Z. Missing values in big data research: some basic skills. Ann Transl Med 2015;3:323.