Model for Predicting the Risk of Bronchopleural Fistula After Pneumonectomy for Destructive Pulmonary Tuberculosis

Introduction: Predicting various events based on influencing factors is important for statistical analysis in medical research. Unfortunately, mathematical models are rarely built on the identified factors.

Objective: To develop a model to predict the risk of bronchopleural fistula after pneumonectomy for destructive pulmonary tuberculosis.

Materials and methods: We analyzed medical records of 198 patients who underwent pneumonectomy. Of them 6 patients (3%) developed a bronchopleural fistula. We used machine learning algorithms such as ridge regression, support vector machine, random forest, and CatBoost, the Jupyter open­source development environment, and Python 3.6 to build prediction models. ROC analysis was used to evaluate the quality of the binary classification.

Results: We built 4 models to predict the risk of bronchopleural fistula. Their ROC AUC were as follows: ridge regression – 0.88, support vector machine – 0.87, CatBoost – 0.75, and random forest – 0.74. The model based on the ridge regression showed the best ROC AUC. Based on the coordinates of the ROC curve, the threshold value of 1.9% provides the maximum total sensitivity and specificity (100% and 68.8%, respectively).

Conclusions: The developed model has a high predictive ability, which allows focusing on the patient group with an increased risk of bronchopleural fistula and justifying the need for preventive measures.

1. Somocurcio JG, Sotomayor A, Shin S, et al. Surgery for patients with drug-resistant tuberculosis: report of 121 cases receiving community-based treatment in Lima, Peru. Thorax. 2007; 62(5):416–421. PMID: 16928717. PMCID: PMC2117182. https://doi.org/10.1136/thx.2005.051961

2. Wang H, Lin H, Jiang G. Pulmonary resection in the treatment of multidrug-resistant tuberculosis: a retrospective study of 56 cases. Ann Thorac Surg. 2008;86(5):1640–1645. PMID: 19049764. https://doi.org/10.1016/j.athoracsur.2008.07.056

3. Bai L, Hong Z, Gong C, Yan D, Liang Z. Surgical treatment efficacy in 172 cases of tuberculosis-destroyed lungs. Eur J Cardiothorac Surg. 2012;41(2):335–340. PMID: 21684172. https://doi.org/10.1016/j.ejcts.2011.05.028

4. ShiraishiY, NakajimaY, KoyamaA, Takasuna K, Katsuragi N, Yoshida S. Morbidity and mortality after 94 extrapleural pneumonectomies for empyema. Ann Thorac Surg. 2000;70(4):1202–1207. PMID: 11081871. https://doi.org/10.1016/s0003-4975(00)01612-x

5. Kim YT, Kim HK, Sung SW, Kim JH. Long-term outcomes and risk factor analysis after pneumonectomy for active and sequela forms of pulmonary tuberculosis. Eur J Cardiothorac Surg. 2003;23(5):833–839. PMID: 12754042. https://doi.org/10.1016/s1010-7940(03)00031-9

6. Byun CS, Chung KY, Narm KS, Lee JG, Hong D, Lee CY. Early and long-term outcomes of pneumonectomy for treating sequelae of pulmonary tuberculosis. Korean J Thorac Cardiovasc Surg. 2012;45(2):110–115. PMID: 22500281. PMCID: PMC3322180. https://doi.org/10.5090/kjtcs.2012.45.2.110

7. Marfina GY, Vladimirov KB, Avetisian AO, Starshinova AA, Kudriashov GG, Sokolovich EG, Yablonskii PK. Bilateral cavitary multidrug- or extensively drug-resistant tuberculosis: role of surgery. Eur J Cardiothorac Surg. 2018;53(3):618–624. PMID: 29040413. https://doi.org/10.1093/ejcts/ezx350

8. Yablonskiy P, Vasilev I, Kirjuhina L, et al. Immediate results of pneumonectomies in patients with unilateral localization of destructive pulmonary tuberculosis. Results of the prospective, nonrandomized study. Medical Alliance. 2017;(4):103–111 (In Russ.).

9. Serezvin I, Avetisyan A, Kudriashov G, Yablonskiy P. Efficacy and safety of pneumonectomy in the comprehensive treatment of patients with destructive pulmonary tuberculosis. Medical Alliance. 2022;10(1):47–57 (In Russ). https://doi.org/10.36422/230763482022-10-1-47-57

10. Seely AJ, Ivanovic J, Threader J, et al. Systematic classification of morbidity and mortality after thoracic surgery. Ann Thorac Surg. 2010;90(3):936–942. PMID: 20732521. https://doi.org/10.1016/j.athoracsur.2010.05.014

11. Aarts EHL, Korst JHM, Van Laarhoven PJM. Simulated annealing. In: Aarts E, Lenstra K, eds. Local Search in Combinatorial Optimization. Princeton University Press; 2018:91–120.

12. Browne MW. Cross-validation methods. J Math Psychol. 2000;44(1):108–132. PMID: 10733860. https://doi.org/10.1006/jmps.1999.1279

13. Kohavi R. A study of cross-validation and bootstrap for accuracy estimation and model selection. In: IJCAI-95: Proceedings of the Fourteenth International Joint Conference on Artificial Intelligence. Vol 2. IJCAII; 1995:1137–1143.

14. Hastie T, Tibshirani R, Friedman J. The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Second Edition. Springer; 2009.

15. Grigoryev SG, Lobzin YuV, Skripchenko NV. The role and place of logistic regression and ROC analysis in solving medical diagnostic task. Journal Infectology. 2016;8(4):36–45. (In Russ.). https://doi.org/10.22625/2072-6732-2016-8-4-36-45

16. Mudrov VA. ROC curve analysis algorithm in biomedical research using SPSS software package. Transbaikalian Medical Bulletin. 2021;(1):148–153. (In Russ.). https://doi.org/10.52485/19986173_2021_1_148

17. Yablonsky PK, Atiukov MA, Pischik VG, Bulisnitsa AL. Choice of treatment for the first episode of primary spontaneous pneumothorax. Vestnik of Saint Petersburg University. Medicine. 2010;(1):118–129. (In Russ.).

18. Berkasova IV, Vereshchagin EI, Mitrofanov IM. Risk on forecasting of perioperative complications in reconstructive surgery of gullet. Medicine and Education in Siberia. 2013;(2):39. (In Russ.).

19. Boyakova NV, Zukov RA, Slepov EV, Petrova EO, Vinnik YuS. Mathematical forecasting models of postoperative infectious and inflammatory complications it patients with stomach cancer. Surgical Practice (Russia). 2016:(1):31–35. (In Russ.).

20. Sato T, Kondo H, Watanabe A, et al. A simple risk scoring system for predicting acute exacerbation of interstitial pneumonia after pulmonary resection in lung cancer patients. Gen Thorac Cardiovasc Surg. 2015;63(3):164–172. PMID: 25352311. https://doi.org/10.1007/s11748-014-0487-6

21. Meyer A, Zverinski D, Pfahringer B, et al. Machine learning for real-time prediction of complications in critical care: a retrospective study. Lancet Respir Med. 2018;6(12):905–914. PMID: 30274956. https://doi.org/10.1016/S2213-2600(18)30300-X