Pediatrik apandisit tanı ve tedavisinde gelişmiş sonuçlar için makine öğreniminden yararlanma

Öz

Pediatrik apandisit, kritik bir durum olarak, sunumundaki değişkenlik ve hem tanı hem de sonuç tahmini için spesifik bir biyobelirtecin olmaması nedeniyle hem tanı hem de tedavi yönetiminde klinik zorluklar sunar. Makine öğrenimi (ML) algoritmalarından yararlanan bu çalışma, Almanya, Regensburg'daki Çocuk Hastanesi St. Hedwig'den kapsamlı klinik veriler ve geniş bir hasta demografisi yelpazesi içeren sağlam bir veri setini kullanarak tanı doğruluğunu ve tedavi stratejilerini iyileştirmeyi amaçlamaktadır; pediatrik apandisitin tanısını, yönetimini ve ciddiyetini değerlendirmek için 10 katlı çapraz doğrulama kullanarak Çok Katmanlı Sinir Ağları (MLNN), Destek Vektör Makineleri (SVM) ve Doğrusal Ayırıcı Analiz (LDA) dahil olmak üzere üç ML tekniğinin verimliliğini değerlendirdik. Bulgular, SVM'nin mükemmel sınıflandırma puanları elde etme ölçütlerindeki üstünlüğünü, ardından MLNN'nin güçlü performansını ortaya koymaktadır. Tersine, doğrusal yapısı nedeniyle LDA, karmaşık veri setinde bulunan karmaşık ve doğrusal olmayan ilişkileri ele almak için yetersiz olduğu kanıtlanmıştır. Çalışma, pediatrik apandisit tedavisinin yönetimine bütünsel bir yaklaşım sağlayan ML destekli klinik karar destek sistemlerinin kullanılma potansiyelini vurgulamaktadır

Anahtar Kelimeler

• Pediatrik Apandisit
• Klinik Karar Desteği
• Makine Öğrenimi

Leveraging machine learning for improved outcomes in pediatric appendicitis diagnosis and management

Öz

Pediatric appendicitis, as a critical condition, represents clinical challenges in both diagnostic and treatment management due to the variability in its presentation and the absence of a specific biomarker for both diagnosis and outcome prediction. Leveraging Machine Learning (ML) algorithms, this study aims to improve diagnostic accuracy and treatment strategies utilizing a robust dataset from the Children’s Hospital St. Hedwig in Regensburg, Germany, containing extensive clinical data and a broad spectrum of patient demographics. We evaluated the efficiency of three ML techniques, including Multilayer Neural Networks (MLNN), Support Vector Machines (SVM), and Linear Discriminant Analysis (LDA), using 10-fold cross-validation to assess the diagnosis, management, and severity of pediatric appendicitis. The findings reveal SVM’s consistently strong performance across all metrics, achieving highly accurate classification results, followed by the competitive performance of MLNN. Conversely, LDA demonstrated limitations due to its linear nature, proving insufficient for handling the intricate and nonlinear relationships present in the complex dataset. The study highlights the potential of using ML-powered clinical decision support systems, providing a holistic approach to the treatment management of pediatric appendicitis.

Anahtar Kelimeler

• Pediatric Appendicitis
• Clinical Decision Support
• Machine Learning

Kaynakça

1. Andersson RE (2007) The Natural History and Traditional Management of Appendicitis Revisited: Spontaneous Resolution and Predominance of Prehospital Perforations Imply That a Correct Diagnosis is More Important Than an Early Diagnosis. World Journal of Surgery 31(1):86–92. https://doi.org/10.1007/s00268-006-0056-y
2. Marcinkevics R, Reis Wolfertstetter P, Wellmann S et al (2021) Using Machine Learning to Predict the Diagnosis, Management and Severity of Pediatric Appendicitis. Frontiers in Pediatrics 9:1–12. https://doi.org/10.3389/fped.2021.662183
3. Acharya A, Markar SR, Ni M, Hanna GB (2017) Biomarkers of acute appendicitis: systematic review and cost–benefit trade-off analysis. Surgical Endoscopy 31(3):1022–1031. https://doi.org/10.1007/s00464-016-5109-1
4. Shommu NS, Jenne CN, Blackwood J et al (2018) The Use of Metabolomics and Inflammatory Mediator Profiling Provides a Novel Approach to Identifying Pediatric Appendicitis in the Emergency Department. Scientific Reports 8(1):4083. https://doi.org/10.1038/s41598-018-22338-1
5. Svensson J, Hall N, Eaton S et al (2012) A Review of Conservative Treatment of Acute Appendicitis. European Journal of Pediatric Surgery 22(03):185–194. https://doi.org/10.1055/s-0032-1320014
6. Svensson JF, Patkova B, Almström M et al (2015) Nonoperative Treatment With Antibiotics Versus Surgery for Acute Nonperforated Appendicitis in Children. Annals of Surgery 261(1):67–71. https://doi.org/10.1097/SLA.0000000000000835
7. Samuel M (2002) Pediatric appendicitis score. Journal of Pediatric Surgery 37(6):877–881. https://doi.org/10.1053/jpsu.2002.32893
8. Alvarado A (1986) A practical score for the early diagnosis of acute appendicitis. Annals of Emergency Medicine 15(5):557–564. https://doi.org/10.1016/S0196-0644(86)80993-3

9. Andersson M, Andersson RE (2008) The Appendicitis Inflammatory Response Score: A Tool for the Diagnosis of Acute Appendicitis that Outperforms the Alvarado Score. World Journal of Surgery 32(8):1843–1849. https://doi.org/10.1007/s00268-008-9649-y
10. Lam A, Squires E, Tan S et al (2023) Artificial intelligence for predicting acute appendicitis: a systematic review. ANZ Journal of Surgery 93(9):2070–2078. https://doi.org/10.1111/ans.18610
11. Aydin E, Türkmen İU, Namli G et al (2020) A novel and simple machine learning algorithm for preoperative diagnosis of acute appendicitis in children. Pediatric Surgery International 36(6):735–742. https://doi.org/10.1007/s00383-020-04655-7
12. Alpaydın E (2010) Introduction to Machine Learning. MIT Press, London
13. Mani S, Ozdas A, Aliferis C et al (2014) Medical decision support using machine learning for early detection of late-onset neonatal sepsis. Journal of the American Medical Informatics Association 21(2):326–336. https://doi.org/10.1136/amiajnl-2013-001854
14. Pediatric Appendicitis Dataset (2024) Children’s Hospital St. Hedwig in Regensburg, Germany
15. Muller K-R, Mika S, Ratsch G et al (2001) An introduction to kernel-based learning algorithms. IEEE Transactions on Neural Networks 12(2):181–201. https://doi.org/10.1109/72.914517
16. Cetin O (2023) Accent Recognition Using a Spectrogram Image Feature-Based Convolutional Neural Network. Arabian Journal for Science and Engineering 48 (2):1973–1990. https://doi.org/10.1007/s13369-022-07086-9
17. Gorur K, Bozkurt MR, Bascil MS, Temurtas F (2020) Comparative Evaluation for PCA and ICA on Tongue-Machine Interface Using Glossokinetic Potential Responses. Celal Bayar University Journal of Science 16(1):35–46. https://doi.org/10.18466/cbayarfbe.571994
18. Temurtas F, Gorur K, Cetin O, Ozer I (2023) Machine learning for thyroid cancer diagnosis. In: Comput. Intell. Cancer Diagnosis. Elsevier. pp 117–145.
19. Cortes C, Vapnik V (1995) Support-vector networks. Machine Learning 20(3):273–297. https://doi.org/10.1007/BF00994018
20. Liu J, Song S, Sun G, Fu Y (2019) Classification of ECG Arrhythmia Using CNN, SVM and LDA. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 11633 LNCS (2016):191–201. https://doi.org/10.1007/978-3-030-24265-7_17
21. Ozer I, Cetin O, Gorur K, Temurtas F (2021) Improved machine learning performances with transfer learning to predicting need for hospitalization in arboviral infections against the small dataset. Neural Computing and Applications 33(21):14975–14989. https://doi.org/10.1007/S00521-021-06133-0/TABLES/7
22. Cetin O, Temurtas F (2021) A comparative study on classification of magnetoencephalography signals using probabilistic neural network and multilayer neural network. Soft Computing 25(3):2267–2275. https://doi.org/10.1007/S00500-020-05296-7/FIGURES/4
23. Gorur K, Bozkurt MR, Bascil MS, Temurtas F (2018) Glossokinetic potential based tongue–machine interface for 1-D extraction using neural networks. Biocybernetics and Biomedical Engineering 38(3):745–759. https://doi.org/10.1016/j.bbe.2018.06.004
24. Shafi I, Ahmad J, Shah SI, Kashif FM (2006) Impact of Varying Neurons and Hidden Layers in Neural Network Architecture for a Time Frequency Application. In: 2006 IEEE Int. Multitopic Conf. IEEE. pp 188–193.
25. Temurtas H, Yumusak N, Temurtas F (2009) A comparative study on diabetes disease diagnosis using neural networks. Expert Systems with Applications 36(4):8610–8615. https://doi.org/10.1016/j.eswa.2008.10.032
26. Kiliçarslan S, Közkurt C, Baş S, Elen A (2023) Detection and classification of pneumonia using novel Superior Exponential (SupEx) activation function in convolutional neural networks. Expert Systems with Applications 217, 119503. https://doi.org/10.1016/j.eswa.2023.119503
27. Közkurt C, Kiliçarslan S, Baş S, Elen A (2023) α­SechSig and α­TanhSig: two novel non-monotonic activation functions. Soft Computing 27(24):18451–18467. https://doi.org/10.1007/s00500-023-09279-2
28. Zhu D, Lu S, Wang M et al. (2020) Efficient Precision-Adjustable Architecture for Softmax Function in Deep Learning. IEEE Transactions on Circuits and Systems II: Express Briefs 67(12):3382–3386. https://doi.org/10.1109/TCSII.2020.3002564