Slice-Level Classification of Kidney Organ Presence Using CNN–ViT Features: Toward Clinical Pre-Screening

Öz

Kidney stone disease is a major global health concern due to its rising prevalence and related complications. Although computed tomography (CT) is highly sensitive for diagnosis, its volumetric nature increases radiologists’ workload and review time. This study proposes a lightweight and low-cost deep learning-based pre-classification model to distinguish between CT slices containing the kidney and those that do not, as a preliminary step for kidney stone detection systems. The model aims to eliminate irrelevant slices and direct only meaningful images to both the second-stage AI model and the radiologist. Vision Transformer (ViT) was used to capture the global spatial context of the kidney, while ResNet18 extracted local features. These features were fused and classified using a shallow neural network. The model was tested within an interactive interface built using real patient data and integrated into a pilot application. Results showed that the proposed system achieved an average of 64.1% time saving per patient (~24 minutes) by filtering out non-relevant slices, with 89.4% accuracy, 89.4% recall, and 89.5% specificity. These findings suggest that the model is a practical and efficient pre-screening component for clinical workflows and highlights the importance of pilot testing and expert feedback before real-world deployment of AI systems.

Anahtar Kelimeler

• Kidney detection
• Abdominal CT
• Slice-level classification
• Deep learning
• Clinical decision support.

CNN–ViT Özniteliklerine Dayalı Kesit Bazlı Böbrek Organı Varlığı Sınıflandırması: Klinik Ön-Eleme Süreçlerine Yönelik Bir Yaklaşım

Öz

Böbrek taşı, dünya genelinde artan prevalansı ve yol açtığı komplikasyonlar nedeniyle dikkatle izlenmesi gereken önemli bir sağlık problemidir. Tanıda yaygın olarak kullanılan bilgisayarlı tomografi (BT), yüksek duyarlılık sağlasa da hacimsel veri yapısı nedeniyle radyologların inceleme süresini uzatmakta ve iş yükünü artırmaktadır. Bu çalışmada, böbrek taşı tespitine yönelik geliştirilecek karar destek sistemlerine öncülük edecek, böbrek içeren ve içermeyen BT kesitlerini ayırt edebilen, sade ve düşük maliyetli bir derin öğrenme tabanlı ön sınıflandırma modeli önerilmiştir. Modelin amacı, böbrek içermeyen kesitleri dışlayarak yalnızca ilgili görüntülerin ikinci kademe yapay zekâ modeline ve radyoloğa yönlendirilmesini sağlamaktır. Bu kapsamda, böbreğin global konumsal bağlamı için Vision Transformer (ViT), yerel öznitelikleri için ResNet18 mimarisi kullanılmış; öznitelikler kaynaştırılarak optimize edilmiş sığ bir yapay sinir ağıyla sınıflandırılmıştır. Model, gerçek hasta verileriyle geliştirilen etkileşimli bir arayüzde test edilerek pilot uygulama için entegre edilmiş ve kullanıcı senaryolarına göre değerlendirilmiştir. Önerilen model, doğru negatif sınıflamalarıyla hasta başına ortalama %64.1 oranında (~24 dakika) etiketleme süresinden tasarruf sağlamış; %89.4 doğruluk, duyarlılık ve %89.5 özgüllük ile yüksek sınıflandırma başarımı sunmuştur. Bulgular, modelin klinik entegrasyona uygun, zamandan ve işlem gücünden tasarruf sağlayan etkili bir ön sınıflandırma aracı olduğunu göstermektedir. Çalışma ayrıca, yapay zekâ sistemlerinin klinik kullanımı öncesinde pilot testlerle ve uzman katkısıyla değerlendirilmesinin önemini vurgulamaktadır.

Anahtar Kelimeler

• Böbrek tespiti
• Abdominal BT
• Dilim düzeyinde sınıflandırma
• Derin öğrenme
• Klinik karar desteği.

Kaynakça

1. [1] S. R. Khan et al., “Kidney stones”, Nat. Rev. Dis. Primer, 2(1), 1-23, (2016).
2. [2] K. Stamatelou and D. S. Goldfarb, “Epidemiology of Kidney Stones”, Healthcare, 11(3): 424, (2023).
3. [3] T. Ates, I. H. Sukur, F. Ok, M. G. Arikan, and N. Akdogan, “Global research trends in minimally invasive treatments for kidney stones: A bibliometric analysis (2015–2024)”, Urolithiasis, 53(1): 116, (2025).
4. [4] A. Y. Muslumanoglu et al., “Updated epidemiologic study of urolithiasis in Turkey. I: Changing characteristics of urolithiasis”, Urol. Res., 39(4), 309–314, (2011).
5. [5] H. Bostan et al., “The prevalence and associated risk factors of detectable renal morphological abnormalities in acromegaly”, Pituitary, 27(1), 44–51, (2024).
6. [6] N. Aiumtrakul et al., “Global Trends in Kidney Stone Awareness: A Time Series Analysis from 2004–2023”, Clin. Pract., 14(3): 3, (2024).
7. [7] A. Pietropaolo et al., “Economic Burden of Imaging and Interventions in Endourology: A Worldwide Cost Analysis from European Association of Urology Young Academic Urology Endourology and Urolithiasis Working Party”, J. Endourol., 39(4), 389–398, (2025).
8. [8] R. M. Geraghty, P. Cook, V. Walker, and B. K. Somani, “Evaluation of the economic burden of kidney stone disease in the UK: a retrospective cohort study with a mean follow-up of 19 years”, BJU Int., 125(4), 586–594, (2020).

9. [9] E. M. Worcester and F. L. Coe, “Clinical practice. Calcium kidney stones”, N. Engl. J. Med., 363(10), 954–963, (2010).
10. [10] W. Brisbane, M. R. Bailey, and M. D. Sorensen, “An overview of kidney stone imaging techniques”, Nat. Rev. Urol., 13(11), 654–662, (2016).
11. [11] Z. Sözen and N. Barışçı, “Derin Öğrenme ile Hücre Görüntülerinin Tespiti ve Sayımı,” Journal of Polytechnic, 28(3), 909–921, (2025).
12. [12] Ö. Dündar and S. Koçer, “Pneumonia Detection from Pediatric Lung X-Ray Images Using Artificial Neural Networks,” Journal of Polytechnic, 27(5), 1843–1852, (2024).
13. [13] E. Ekinci, Z. Garip, and K. Serbest, “Electromyography based hand movement classification and feature extraction using machine learning algorithms,” Journal of Polytechnic, 26(4), 1621–1633, (2023).
14. [14] O. Pauly et al., “Fast Multiple Organ Detection and Localization in Whole-Body MR Dixon Sequences,” in Medical Image Computing and Computer-Assisted Intervention – MICCAI 2011, G. Fichtinger, A. Martel, and T. Peters, Eds., Berlin, Heidelberg: Springer, 239–247, (2011).
15. [15] M. Hammami, D. Friboulet, and R. Kechichian, “Cycle GAN-Based Data Augmentation For Multi-Organ Detection In CT Images Via Yolo”, 2020 IEEE International Conference on Image Processing (ICIP), Abu Dhabi, United Arab Emirates, 390-393, (2020).
16. [16] H.-C. Shin, M. R. Orton, D. J. Collins, S. J. Doran, and M. O. Leach, “Stacked Autoencoders for Unsupervised Feature Learning and Multiple Organ Detection in a Pilot Study Using 4D Patient Data”, IEEE Trans. Pattern Anal. Mach. Intell., 35(8), 1930–1943, (2013).
17. [17] W. Touhami, D. Boukerroui, and J.-P. Cocquerez, “Fully automatic kidneys detection in 2D CT images: a statistical approach”, Med. Image Comput. Comput.-Assist. Interv. MICCAI Int. Conf. Med. Image Comput. Comput.-Assist. Interv., 262–269, (2005).
18. [18] C. Raynaud et al., “Multi-organ Detection in 3D Fetal Ultrasound with Machine Learning,” in Fetal, Infant and Ophthalmic Medical Image Analysis, M. J. Cardoso, T. Arbel, A. Melbourne, H. Bogunovic, P. Moeskops, X. Chen, E. Schwartz, M. Garvin, E. Robinson, E. Trucco, M. Ebner, Y. Xu, A. Makropoulos, A. Desjardin, and T. Vercauteren, Eds., Cham: Springer International Publishing, pp. 62–72, (2017).
19. [19] A. Mansoor, A. R. Porras, and M. G. Linguraru, “Region Proposal Networks with Contextual Selective Attention for Real-Time Organ Detection”, in 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019), Venice, Italy, 1193-1196, (2019).
20. [20] X. Xu, F. Zhou, B. Liu, D. Fu, and X. Bai, “Efficient Multiple Organ Localization in CT Image Using 3D Region Proposal Network”, IEEE Trans. Med. Imaging, 38(8), 1885–1898, (2019).
21. [21] T. Les, T. Markiewicz, M. Dziekiewicz, J. Gallego, Z. Swiderska-Chadaj, and M. Lorent, “Localization of spleen and kidney organs from CT scans based on classification of slices in rotational views”, Sci. Rep., 13(1): 5709, (2023).
22. [22] P. Y. Anari et al., “Using YOLO v7 to Detect Kidney in Magnetic Resonance Imaging”, arXiv: arXiv:2402.05817., Feb. 12, (2024).
23. [23] M. Ghahremani, B. R. Ernhofer, J. Wang, M. Makowski, and C. Wachinger, “Organ-DETR: Organ Detection via Transformers”, IEEE Trans. Med. Imaging, 44(6), 2657–2671, (2025).
24. [24] K. He, X. Zhang, S. Ren and J. Sun, "Deep Residual Learning for Image Recognition", 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 770-778, (2016).
25. [25] A. Dosovitskiy et al., “An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale,” arXiv: arXiv:2010.11929. June 03, (2021).
26. [26] “Streamlit • A faster way to build and share data apps.” Accessed: July 22, 2025. [Online]. Available: https://streamlit.io/