A Pilot Study on YOLO26-Based Detection and Segmentation of Nasal Septum Deviation and Related Findings on CBCT

Abstract

Objectives: The aim of this study was to develop and preliminarily evaluate a deep learning–based model for the automatic detection and segmentation of selected nasal structures and maxillary sinus findings on cone-beam computed tomography (CBCT) images. Material and Methods: A total of 110 expert-selected coronal CBCT slices with a wide field of view were retrospectively obtained from a larger dataset of 1000 scans. Images were preprocessed and randomly divided into training (70%), validation (15%), and test (15%) sets. A YOLOv12n-seg architecture was implemented for detection and segmentation tasks. Model performance was assessed using accuracy and area under the receiver operating characteristic curve (AUC). Results: The model achieved an accuracy of 87.5% for nasal septum deviation (NSD) and inferior concha hypertrophy (ICH). For maxillary sinus mucosal thickening (MSMT), the accuracy was 81.3%. The AUC values were 0.84 for NSD, 0.85 for ICH, and 0.68 for MSMT, indicating relatively lower performance for soft-tissue-related findings. Conclusion: The proposed deep learning model demonstrated promising performance for the detection of anatomically distinct nasal structures on selected CBCT slices. However, the limited sample size, slice-based design, and lack of external validation restrict the generalizability of the findings. This study should be considered a preliminary pilot investigation, and further research using larger, patient-based datasets and clinical comparisons is required before routine clinical application.

Keywords

• Deep Learning
• Inferior concha hypertrophy
• Nasal septum
• Maxillary sinüs
• Cone-beam computed tomography

Supporting Institution

No funding resource.

Ethical Statement

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions

Thanks

NO

KIBT Görüntülerinde Nazal Septum Deviasyonu ve İlişkili Bulguların YOLO26 Tabanlı Tespiti ve Segmentasyonu: Bir Pilot Çalışma

Abstract

Amaç: Bu çalışmanın amacı, konik ışınlı bilgisayarlı tomografi (KIBT) görüntülerinde seçilen nazal yapıları ve maksiller sinüs bulgularının otomatik tespiti ve segmentasyonu için derin öğrenme tabanlı bir model geliştirmek ve ön değerlendirmesini yapmaktır. Gereç ve Yöntemler: 1000 taramadan oluşan daha büyük bir veri setinden, uzmanlar tarafından seçilen toplam 110 geniş görüş alanlı koronal KIBT kesiti retrospektif olarak elde edildi. Görüntüler ön işleme tabi tutuldu ve rastgele olarak eğitim (%70), doğrulama (%15) ve test (%15) setlerine ayrıldı. Tespit ve segmentasyon görevleri için YOLOv12n-seg mimarisi uygulandı. Model performansı, doğruluk ve alıcı işletim karakteristik eğrisi altındaki alan (AUC) kullanılarak değerlendirildi. Bulgular: Model, nazal septum deviasyonu (NSD) ve inferior konka hipertrofisi (IKH) için %87,5 doğruluk elde etti. Maksiller sinüs mukozal kalınlaşması (MSMK) için doğruluk %81,3 oldu. AUC değerleri NSD için 0,84, IKH için 0,85 ve MSMK için 0,68 olup, yumuşak dokuyla ilgili bulgular için nispeten daha düşük performans göstermektedir. Sonuç: Önerilen derin öğrenme modeli, seçilen KIBT kesitlerinde anatomik olarak farklı burun yapılarının tespiti için umut verici bir performans sergilemiştir. Bununla birlikte, sınırlı örneklem boyutu, kesit tabanlı tasarım ve dış doğrulama eksikliği, bulguların genelleştirilebilirliğini kısıtlamaktadır. Bu çalışma, ön bir pilot araştırma olarak değerlendirilmelidir ve rutin klinik uygulamadan önce daha büyük, hasta tabanlı veri kümeleri ve klinik karşılaştırmalar kullanılarak daha fazla araştırma yapılması gerekmektedir.

Keywords

• Derin Öğrenme
• İnferior konka hipertrofisi
• Nazal septum
• Maksiller sinüs
• Konik ışınlı bilgisayarlı tomografi

Supporting Institution

Hiçbir fon kaynağı yok.

Ethical Statement

Uygulanan tüm prosedürler, insan deneyleri konusunda yetkili komitenin (kurumsal ve ulusal) etik standartlarına ve 1964 Helsinki Deklarasyonu'na ve daha sonraki versiyonlarına uygun olarak gerçekleştirilmiştir.

Thanks

YOK

References

1. Shetty SR, Al Bayatti SW, Al-Rawi NH, et al. Effect of concha bullosa and nasal septal deviation on palatal dimensions: a CBCT study. BMC Oral Health. 2021; 21: 607. doi: 10.1186/s12903-021-01974-6
2. Moshfeghi M, Abedian B, Ghazizadeh Ahsaie M, et al. Prevalence of nasal septum deviation using cone-beam computed tomography: a cross-sectional study. Contemp Clin Dent. 2020; 11: 223-8. doi: 10.4103/ccd.ccd_110_19
3. Kwon O, Yong TH, Kang SR, et al. Automatic diagnosis for cysts and tumors of both jaws on panoramic radiographs using a deep convolution neural network. Dentomaxillofac Radiol. 2020; 49: 20200185. doi: 10.1259/dmfr.20200185
4. Lee HW, Yang HJ, Kim H, et al. Deep learning with chest radiographs for making prognoses in patients with COVID-19: retrospective cohort study. J Med Internet Res. 2023; 25: E42717. doi: 10.2196/42717
5. Osie G, Darbari Kaul R, Alvarado R, et al. A scoping review of artificial ıntelligence research in rhinology. Am J Rhinol Allergy. 2023; 37: 438-48. doi: 10.1177/19458924231162437
6. Shetty S, Mubarak AS, R David L, et al. The application of Mask R-CNN-based detection of nasal septal deviation using CBCT: A proof-of-concept study. JMIR Form Res. 2024; 8: E57335. doi: 10.2196/57335
7. Hung KF, Ai QYH, King AD, et al. Automatic detection and segmentation of morphological changes of the maxillary sinus mucosa on cone-beam computed tomography images using a three-dimensional convolutional neural network. Clin Oral Investig. 2022; 26: 3987-98. doi: 10.1007/s00784-021-04365-x
8. Seol YJ, Kim YJ, Kim YS, et al. A Study on 3D deep learning-based automatic diagnosis of nasal fractures. Sensors (Basel). 2022; 22: 506. doi: 10.3390/s22020506

9. Codari M, Zago M, Guidugli GA, et al. The nasal septum deviation index (NSDI) based on CBCT data. Dentomaxillofac Radiol. 2016; 45: 20150327. doi: 10.1259/dmfr.20150327
10. Alsowey AM, Abdulmonaem G, Elsammak A, et al. Diagnostic performance of multidetector computed tomography (MDCT) in diagnosis of sinus variations. Pol J Radiol. 2017; 82: 713-25. doi: 10.12659/PJR.903684
11. Shetty SR, Al-Bayatti SW, Al Kawas S, et al. A study on the association between the inferior nasal turbinate volume and the maxillary sinus mucosal lining using cone beam tomography. Heliyon. 2022; 8: E09053. doi: 10.1016/j.heliyon.2024.e34459
12. Shetty S, Al-Bayatti S, Alam MK, et al. Analysis of inferior nasal turbinate volume in subjects with nasal septum deviation: a retrospective cone beam tomography study. PeerJ. 2022; 10: E14032. doi: 10.7717/peerj.14032
13. Akoğlu E, Karazincir S, Balci A, et al. Evaluation of the turbinate hypertrophy by computed tomography in patients with deviated nasal septum. Otolaryngol Head Neck Surg. 2007; 136: 380-4. doi: 10.1016/j.otohns.2006.09.006
14. Kapusuz Gencer Z, Ozkırış M, Okur A, et al. The effect of nasal septal deviation on maxillary sinus volumes and development of maxillary sinusitis. Eur Arch Otorhinolaryngol. 2013; 270: 3069-73. doi: 10.1007/s00405-013-2435-y
15. Ren S, Zhao H, Liu J, et al. Significance of maxillary sinus mucosal thickening in patients with periodontal disease. Int Dent J. 2015; 65: 303-10. doi: 10.1111/idj.12186
16. Cui Z, Fang Y, Mei L, et al. A fully automatic AI system for tooth and alveolar bone segmentation from cone-beam CT images. Nat Commun. 2022; 13: 2096. doi: 10.1038/s41467-022-29637-2
17. Shujaat S, Jazil O, Willems H, et al. Automatic segmentation of the pharyngeal airway space with convolutional neural network. J Dent. 2021; 111: 103705. doi: 10.1016/j.jdent.2021.103705
18. He J, Sun M, Huo Y, et al. A platform combining automatic segmentation and automatic measurement of the maxillary sinus and adjacent structures. Clin Oral Investig. 2025; 29: 88. doi: 10.1007/s00784-025-06191-x
19. Altun O, Özen DÇ, Duman ŞB, et al. Automatic maxillary sinus segmentation and pathology classification on cone-beam computed tomographic images using deep learning. BMC Oral Health. 2024; 24: 1208. doi: 10.1186/s12903-024-04924-0
20. Cai Z, Zhou K, Liao Z. A systematic review of YOLO-based object detection in medical imaging: advances, challenges, and future directions. Comput Mater Contin. 2025; 85: 2255-303. doi: 10.32604/cmc.2025.067994
21. Mori M, Ariji Y, Katsumata A, et al. A deep transfer learning approach for the detection and diagnosis of maxillary sinusitis on panoramic radiographs. Odontol. 2021; 109: 941-8. doi: 10.1007/s10266-021-00615-2
22. Kim Y, Lee KJ, Sunwoo L, et al. Deep Learning in Diagnosis of Maxillary Sinusitis Using Conventional Radiography. Invest Radiol. 2019; 54: 7-15. doi: 10.1097/RLI.0000000000000503
23. Murata M, Ariji Y, Ohashi Y, et al. Deep-learning classification using convolutional neural network for evaluation of maxillary sinusitis on panoramic radiography. Oral Radiol. 2019; 35: 301-7. doi: 10.1007/s11282-018-0363-7
24. Kuwana R, Ariji Y, Fukuda M, et al. Performance of deep learning object detection technology in the detection and diagnosis of maxillary sinus lesions on panoramic radiographs. Dentomaxillofac Radiol. 2021; 50: 20200171. doi: 10.1259/dmfr.20200171
25. Serindere G, Bilgili E, Yesil C, et al. Evaluation of maxillary sinusitis from panoramic radiographs and cone-beam computed tomographic images using a convolutional neural network. Imaging Sci Dent. 2022; 52: 187-95. doi: 10.5624/isd.20210263
26. Kim D-E, Yang S, Kang S, et al. SinusNet: Label-free segmentation of maxillary sinus lesions. 5th Medical Imaging with Deep Learning. July 2022, Zurich, Switzerland.