跳到正文
原文
Nature 健康护理主题·· 16 天前AI 评分40

深度学习在齿状突骨折检测中的应用:患者级分割与Grad-CAM评估的重要性

Deep learning for odontoid fractures: the critical role of patient-level partitioning and Grad-CAM assessment in model generalization

AI 导读

一项研究评估了六种卷积神经网络架构在严格零泄漏协议下的齿状突骨折检测性能,使用患者级分割和Grad-CAM可解释性AI验证诊断可靠性。DenseNet121、GoogLeNet和ResNet18在未接触测试集中表现出相似的高AUC值(0.830、0.829、0.829),而SqueezeNet 1.1和VGG19则完全失效。研究强调了严格患者级隔离和可解释性AI评估对临床转化的重要性。

正文

Scientific Reports (2026) Cite this article

We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.

Abstract

Missed diagnoses of odontoid fractures on Computed Tomography (CT) scans can lead to severe morbidity. While Deep Learning (DL) shows promise, many studies report inflated performance metrics due to hidden data leakage (image-level splitting). To evaluate six Convolutional Neural Network (CNN) architectures under a strict zero-leakage protocol for odontoid fracture detection, validating diagnostic reliability and anatomical focus using Explainable AI (XAI). A retrospective cohort of 540 cervical spine CT images from 96 patients was evaluated under a strict zero-leakage protocol. Data was partitioned at the patient level into three disjoint subsets: Training (60 patients, 326 images), Validation (16 patients, 92 images for early stopping), and an independent held-out Test set (20 patients, 122 images touched only once for final evaluation). Six CNN architectures were benchmarked. Statistical robustness was validated via patient-level bootstrapping (N = 1000 iterations) with Holm-Bonferroni family-wise error rate correction, paired with Grad-CAM visual interpretability. On the untouched test cohort, DenseNet121, GoogLeNet, and ResNet18 achieved similar, leading discriminative performance with AUCs of 0.830 (95% CI: 0.718–0.933), 0.829 (95% CI: 0.714–0.929), and 0.829 (95% CI: 0.697–0.934), respectively, with DenseNet121 demonstrating a balanced diagnostic profile (Accuracy: 73.8%, Sensitivity: 73.0%, Specificity: 75.8%). In stark contrast, SqueezeNet 1.1 and VGG19 suffered complete collapse to majority-class prediction (Specificity: 0.0%) under the evaluated training configuration. Pairwise bootstrap hypothesis testing confirmed that DenseNet121 showed a borderline statistically significant difference over SqueezeNet 1.1 after multiple testing correction (\(\:{p}_{\text{raw}}=0.0090\), Holm-adjusted \(\:p=0.0450\) on \(\:n=20\) patients), while differences among DenseNet121, GoogLeNet, and ResNet18 remained statistically non-significant (\(\:{p}_{\text{adj}}=0.9840\)). Architectures incorporating deep feature reuse or residual connections exhibit robust generalization under strict patient-level isolation, whereas compact models and unregularized heavyweight architectures can struggle under small-sample constraints. Rigorous zero-leakage methodology and qualitative XAI assessment are essential prerequisites prior to clinical translation of spinal CAD models.

Abbreviations

AI:

Artificial intelligence

AUC:

Area under the curve

CAD:

Computer-aided diagnosis

CI:

Confidence interval

CNN:

Convolutional neural network

CT:

Computed tomography

DICOM:

Digital imaging and communications in medicine

DL:

Deep learning

FN:

False negative

FP:

False positive

Grad-CAM:

Gradient-weighted class activation mapping

HU:

Hounsfield unit

MRI:

Magnetic resonance imaging

MSFF:

Multi-scale feature fusion

NLP:

Natural language processing

ROC:

Receiver operating characteristic

ROI:

Region of interest

RSNA:

Radiological Society of North America

TL:

Transfer learning

TN:

True negative

TP:

True positive

ViT:

Vision transformer

XAI:

Explainable artificial intelligence

Funding

This research did not receive external funding.

Author information

Authors and Affiliations

  1. Department of Computer Engineering, Graduate School of Natural and Applied Sciences, Selcuk University, 42130, Konya, Türkiye

    Yunus Emre Kıymaz

  2. Department of Computer Engineering, Faculty of Technology, Selcuk University, 42130, Konya, Türkiye

    Şakir Taşdemir

  3. Division of Brain and Nerve Surgery, Department of Surgical Medicine, School of Medicine, Selcuk University, Konya, Türkiye

    Burak Gezer

  4. Department of Computer Engineering, Rafet Kayış Faculty of Engineering, Alanya Alaaddin Keykubat University, Antalya, Türkiye

    Kübra Uyar Pınar

Authors

  1. Yunus Emre Kıymaz
  2. Şakir Taşdemir
  3. Burak Gezer
  4. Kübra Uyar Pınar

Corresponding author

Correspondence to Yunus Emre Kıymaz.

Ethics declarations

Ethical approval

The study was conducted with the formal approval of the Selçuk University Faculty of Medicine Non-Interventional Clinical Research Ethics Committee (Date: 06.06.2023, Decision No: 2023/304). Due to the retrospective nature of this study and the use of fully anonymized radiological data, the requirement for obtaining informed consent from all patients, including parents and/or legal guardians for participants who are minors, was formally waived by the Selçuk University Faculty of Medicine Non-Interventional Clinical Research Ethics Committee. All procedures were performed in accordance with the Declaration of Helsinki.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kıymaz, Y.E., Taşdemir, Ş., Gezer, B. et al. Deep learning for odontoid fractures: the critical role of patient-level partitioning and Grad-CAM assessment in model generalization. Sci Rep (2026). https://doi.org/10.1038/s41598-026-77495-5

Download citation

  • Received: 14 July 2026

  • Accepted: 20 October 2026

  • Published: 04 November 2026

  • DOI: https://doi.org/10.1038/s41598-026-77495-5

Keywords

来源:Nature 健康护理主题 · nature.com