Investigation of the relationship between microbiological features and mortality in patients with hematological malignancies who developed febrile neutropenia using machine learning models

Year 2025, Volume: 5 Issue: 3, 80 - 88, 29.07.2025

Şebnem Çalık , Oktay Bilgir , Deniz İlhan Topcu , Selma Tosun , İsmail Demir

https://doi.org/10.56016/dahudermj.1644331

https://izlik.org/JA43AN98XS

Abstract

Background: This study aimed to examine the relationship between microbiological features and mortality in hematological malignancy patients who develop febrile neutropenia using machine learning algorithms.
Methods: Patients with hematological malignancies who developed febrile neutropenia between 2011 and 2015 in a training and research hospital were included. The PyCaret low-code Python library was used to streamline the machine-learning workflow. Two separate models were developed to predict early and late mortality. The following machine learning algorithms were evaluated during the modeling process: Ridge Classifier, Random Forest Classifier, Linear Discriminant Analysis, Light Gradient Boosting Machine, Logistic Regression, Gradient Boosting Classifier, and Extra Trees Classifier. Accuracy and area under the receiver operating characteristic curve (AUC-ROC) metrics were calculated to evaluate the models’ predictive capability for both early and late mortality predictions. All analyses were performed using Python 3.12 and the PyCaret 3.0 library.
Results: The dataset used in this study consisted of 159 patients. For early mortality prediction, the Ridge Classifier demonstrated the best performance with a test set accuracy of 0.92 and an AUC of 0.94. For late mortality prediction, the Random Forest Classifier achieved the highest accuracy of 0.94 and an AUC of 0.98. For both models, ICU admission was identified as the most important feature, with a relative importance of 23.6% for early mortality prediction and 25.3% for late mortality prediction. Other key variables included pneumonia, renal function, and the duration of neutropenia.
Conclusion: Machine learning models can be applied and improved on more patient data, helping traditional statistical methods in medical research.

Keywords

Febrile neutropenia , machine learning , mortality , risk factors

Ethical Statement

The study protocol was approved by the Institutional Ethics Committee of the Bozyaka Training and Research Hospital Ethics Committee (03.09.2024/ 137) and has been performed in accordance with the ethical standards of the 1964 Helsinki Declaration and its later amendments.

Supporting Institution

-

Thanks

-

Hematolojik maligniteli hastalarda ateşli nötropeni gelişmesi üzerine makine öğrenimi modelleri kullanılarak mikrobiyolojik özellikler ile mortalite arasındaki ilişkinin araştırılması

https://izlik.org/JA43AN98XS

Abstract

Amaç: Bu çalışmanın amacı, makine öğrenimi algoritmaları kullanılarak ateşli nötropeni geliştiren hematolojik malignite hastalarında mikrobiyolojik özellikler ile mortalite arasındaki ilişkiyi incelemektir.
Yöntemler: 2011-2015 yılları arasında bir eğitim ve araştırma hastanesinde ateşli nötropeni geliştiren hematolojik maligniteli hastalar çalışmaya dahil edildi. Makine öğrenimi iş akışını kolaylaştırmak için PyCaret düşük kodlu Python kütüphanesi kullanıldı. Erken ve geç mortaliteyi tahmin etmek için iki ayrı model geliştirildi. Modelleme süreci sırasında şu makine öğrenimi algoritmaları değerlendirildi: Ridge Sınıflandırıcı, Rastgele Orman Sınıflandırıcı, Doğrusal Ayırım Analizi, Hafif Gradyan Artırma Makinesi, Lojistik Regresyon, Gradyan Artırma Sınıflandırıcı ve Ekstra Ağaçlar Sınıflandırıcı. Doğruluk ve alıcı işletim karakteristik eğrisi altındaki alan (AUC-ROC) ölçümleri, modellerin hem erken hem de geç mortalite tahminleri için tahmin yeteneğini değerlendirmek için hesaplandı. Tüm analizler Python 3.12 ve PyCaret 3.0 kütüphanesi kullanılarak gerçekleştirildi.
Sonuçlar: Bu çalışmada kullanılan veri seti 159 hastadan oluşuyordu. Erken ölüm tahmini için Ridge Sınıflandırıcısı 0,92'lik bir test seti doğruluğu, 0,94'lük AUC ile en iyi performansı gösterdi. Geç ölüm tahmini için Random Forest Sınıflandırıcısı 0,94'lük en yüksek doğruluğa, 0,98'lik AUC'ye ulaştı. Her iki model için de yoğun bakım ünitesine yatış en önemli özellik olarak belirlendi ve erken ölüm tahmini için %23,6 ve geç ölüm tahmini için %25,3'lük bir göreceli öneme sahipti. Diğer önemli değişkenler arasında zatürre, böbrek fonksiyonu ve nötropeni süresi yer aldı.
Sonuç: Makine öğrenimi modelleri daha fazla hasta verisine uygulanabilir ve iyileştirilebilir ve tıbbi araştırmalarda geleneksel istatistiksel yöntemlere yardımcı olabilir.

Keywords

Ateşli nötropeni , makine öğrenimi , mortalite , risk faktörleri

Ethical Statement

The study protocol was approved by the Institutional Ethics Committee of the Bozyaka Training and Research Hospital Ethics Committee (03.09.2024/ 137) and has been performed in accordance with the ethical standards of the 1964 Helsinki Declaration and its later amendments.

Supporting Institution

-

Thanks

-

References

• Stohs EJ, Abbas A, Freifeld A. Approach to febrile neutropenia in patients undergoing treatments for hematologic malignancies. Transpl Infect Dis. 2024 Apr; 26(2): e14236. doi: 10.1111/tid.14236.
• Contejean A, Abbara S, Chentouh R, Alviset S, Grignano E, Gastli N, Casetta A, Willems L, Canouï E, Charlier C, Pène F, Charpentier J, Reboul-Marty J, Batista R, Bouscary D, Kernéis S. Antimicrobial stewardship in high-risk febrile neutropenia patients. Antimicrob Resist Infect Control. 2022 Mar; 11(1) :52. doi: 10.1186/s13756-022-01084-0.
• Contejean A, Maillard A, Canouï E, Kernéis S, Fantin B, Bouscary D, Parize P, Garcia-Vidal C, Charlier C. Advances in antibacterial treatment of adults with high-risk febrile neutropenia. J Antimicrob Chemother. 2023 Sep;78(9):2109-2120. doi: 10.1093/jac/dkad166.4.
• Jiang Y, Luo J, Huang D, Liu Y, Li DD. Machine learning advances in microbiology: a review of methods and applications. Front Microbiol 2022 May; 13: 925454. https://doi. org/10.3389/fmicb.2022.925454
• Sidey-Gibbons JAM, Sidey-Gibbons CJ. Machine learning in medicine: a practical introduction. BMC Med Res Method. 2019 Mar; 19(1): 64. https://doi.org/10.1186/s12874- 019-0681-4.
• Calik S, Ari A, Bilgir O, Cetintepe T, Yis R, Sonmez U, Tosun S. The relationship between mortality and microbiological parameters in febrile neutropenic patients with hematological malignancies. Saudi Med J 2018 Sep; 39(9): 878-885. doi: 10.15537/smj.2018.9.22824.
• Ali M. PyCaret: An open source, low-code machine learning library in Python. Available at: https://www.pycaret.org. Accessed: February 18, 2024.
• Bachlitzanaki M, Aletras G, Bachlitzanaki E, Messaritakis I, Koukias S, Koulouridi A, Bachlitzanakis E, Kaloeidi E, Vakonaki E, Kontopodis E, Androulakis N, Chamilos G, Mavroudis D, Ioannou P, Kofteridis D. Evalution of febrile neutropenia in hospitalized patients with neoplasia undergoing chemotherapy. Microorganisms. 2023 Oct; 11(10): 2547. doi: 10.3390/microorganisms11102547.
• Gedik H, Simşek F, Kantürk A, Yildirmak T, Arica D, Aydin D, Demirel N, Yokuş O. Bloodstream infections in patients with hematological malignancies: which is more fatal - cancer or resistant pathogens? Ther Clin Risk Manag. 2014 Sep; 10: 743–752. doi: 10.2147/TCRM.S68450.
• Kang CI, Chung DR, Ko KS, Peck KR, Song JH, Korean Network for Study of Infectious Diseases. Risk factors for infection and treatment outcome of extended-spectrum β-lactamase-producing Escherichia coli and Klebsiella pneumoniae bacteremia in patients with hematologic malignancy. Ann Hematol. 2012 Jan; 91: 115–121. doi: 10.1007/s00277-011-1247-7.
• Cornejo-Juárez P, Vilar-Compte D, Pérez-Jiménez C, Ñamendys-Silva SA, Sandoval-Hernández S, Volkow-Fernández P. The impact of hospital-acquired infections with multidrug-resistant bacteria in an oncology intensive care unit. Int J Infect Dis. 2015 Feb; 31: 31–34. doi: 10.1016/j.ijid.2014.12.022.
• Kim SH, Kwon JC, Choi SM, Lee DG, Park SH, Choi JH, Yoo JH, Cho BS, Eom KS, Kim YJ, Kim HJ, Lee S, Min CK, Cho SG, KimDW, Lee JW, Min WS. Escherichia coli and Klebsiella pneumoniae bacteremia in patients with neutropenic fever: factors associated with extended-spectrum β-lactamase production and its impact on outcome. Ann Hematol. 2013 Apr; 92: 533–541. doi: 10.1007/s00277-012-1631-y.
• Padmanabhan R, Elomri A, Taha RY, Omri HE, Elsabah H, El Omri A. Prediction of Multiple Clinical Complications in Cancer Patients to Ensure Hospital Preparedness and Improved Cancer Care. Int J Environ Res Public Health 2022 Dec; 20(1): 526. doi: 10.3390/ijerph20010526.
• Venäläinen MS, Heervä E, Hirvonen O, Saraei S, Suomi T, Mikkola T, Bärlund M, Jyrkkiö S, Laitinen T, Elo LL. Improved risk prediction of chemotherpy induced neutropenia-model development and validation with real world data. Cancer Med. 2022 Feb; 11(3): 654-663. doi: 10.1002/cam4.4465.
• Du X, Min J, Shah CP, Bishnoi R, Hogan WR, Lemas DJ. Predicting in hospital mortality of patients with febrile neutropenia using machine learning models. Int J Med Inform. 2020 Jul; 139:104140. doi: 10.1016/j.ijmedinf.2020.104140.
• Cho BJ, Kim KM, Bilegsaikhan SE, Suh YJ. Machine learning improves the prediction of febrile neutropenia in Korean inpatients undergoing chemotherapy for breast cancer. Sci Rep. 2020 Sep; 10(1): 14803.
• Gallardo-Pizarro A, Peyrony O, Chumbita M, Monzo-Gallo P, Aiello TF, Teijon-Lumbreras C, Gras E, Mensa J, Soriano A, Garcia-Vidal C. Improving management of febrile neutropenia in oncology patiemts: the role of artificial intelligence and machine learning. Expert Rev Anti Infect Ther. 2024 Apr; 22(4): 179-187.
• Artificial intelligence, machine learning and deep learning: Potential resources for the infection clinician. J Infect. 2023; 87(4): 287-294. doi: 10.1016/j.jinf.2023.07.006.