Li Yuanlin, Ma Chenglong, Li Zilong, Wang Zhen, Han Jing, Shan Hongming, Liu Jiannan
Department of Oral Maxillofacial Head and Neck Oncology, Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine, College of Stomatology, Shanghai Jiao Tong University, National Center for Stomatology, National Clinical Research Center for Oral Diseases, Shanghai Key Laboratory of Stomatology & Shanghai Research Institute of Stomatology, Shanghai 200011, China.
Institute of Science and Technology for Brain-inspired Intelligence, Fudan University, Shanghai 200433, China.
Eur J Radiol. 2025 Jun;187:112087. doi: 10.1016/j.ejrad.2025.112087. Epub 2025 Mar 31.
To develop a semi-supervised domain adaptation technique for metal artifact reduction with a spatial-frequency transformer (SFTrans) model (Semi-SFTrans), and to quantitatively compare its performance with supervised models (Sup-SFTrans and ResUNet) and traditional linear interpolation MAR method (LI) in oral and maxillofacial CT.
Supervised models, including Sup-SFTrans and a state-of-the-art model termed ResUNet, were trained with paired simulated CT images, while semi-supervised model, Semi-SFTrans, was trained with both paired simulated and unpaired clinical CT images. For evaluation on the simulated data, we calculated Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM) on the images corrected by four methods: LI, ResUNet, Sup-SFTrans, and Semi-SFTrans. For evaluation on the clinical data, we collected twenty-two clinical cases with real metal artifacts, and the corresponding intraoral scan data. Three radiologists visually assessed the severity of artifacts using Likert scales on the original, Sup-SFTrans-corrected, and Semi-SFTrans-corrected images. Quantitative MAR evaluation was conducted by measuring Mean Hounsfield Unit (HU) values, standard deviations, and Signal-to-Noise Ratios (SNRs) across Regions of Interest (ROIs) such as the tongue, bilateral buccal, lips, and bilateral masseter muscles, using paired t-tests and Wilcoxon signed-rank tests. Further, teeth integrity in the corrected images was assessed by comparing teeth segmentation results from the corrected images against the ground-truth segmentation derived from registered intraoral scan data, using Dice Score and Hausdorff Distance.
Sup-SFTrans outperformed LI, ResUNet and Semi-SFTrans on the simulated dataset. Visual assessments from the radiologists showed that average scores were (2.02 ± 0.91) for original CT, (4.46 ± 0.51) for Semi-SFTrans CT, and (3.64 ± 0.90) for Sup-SFTrans CT, with intra correlation coefficients (ICCs)＞0.8 of all groups and p < 0.001 between groups. On soft tissue, both Semi-SFTrans and Sup-SFTrans significantly reduced metal artifacts in tongue (p < 0.001), lips, bilateral buccal regions, and masseter muscle areas (p < 0.05). Semi-SFTrans achieved superior metal artifact reduction than Sup-SFTrans in all ROIs (p < 0.001). SNR results indicated significant differences between Semi-SFTrans and Sup-SFTrans in tongue (p = 0.0391), bilateral buccal (p = 0.0067), lips (p = 0.0208), and bilateral masseter muscle areas (p = 0.0031). Notably, Semi-SFTrans demonstrated better teeth integrity preservation than Sup-SFTrans (Dice Score: p < 0.001; Hausdorff Distance: p = 0.0022).
The semi-supervised MAR model, Semi-SFTrans, demonstrated superior metal artifact reduction performance over supervised counterparts in real dental CT images.
开发一种用于减少金属伪影的半监督域适应技术，该技术采用空间频率变压器（SFTrans）模型（半监督SFTrans），并在口腔颌面CT中定量比较其与监督模型（监督SFTrans和ResUNet）及传统线性插值金属伪影减少方法（LI）的性能。
监督模型，包括监督SFTrans和一种名为ResUNet的先进模型，使用配对的模拟CT图像进行训练，而半监督模型半监督SFTrans则使用配对的模拟CT图像和未配对的临床CT图像进行训练。为了对模拟数据进行评估，我们计算了通过四种方法（LI、ResUNet、监督SFTrans和半监督SFTrans）校正后的图像的峰值信噪比（PSNR）和结构相似性指数测量值（SSIM）。为了对临床数据进行评估，我们收集了22例带有真实金属伪影的临床病例以及相应的口内扫描数据。三名放射科医生使用李克特量表对原始图像、监督SFTrans校正后的图像和半监督SFTrans校正后的图像上的伪影严重程度进行视觉评估。通过使用配对t检验和威尔科克森符号秩检验，测量感兴趣区域（ROI）（如舌头、双侧颊部、嘴唇和双侧咬肌）的平均亨氏单位（HU）值、标准差和信噪比（SNR），进行定量金属伪影减少评估。此外，通过将校正后图像的牙齿分割结果与从配准的口内扫描数据得出的真实分割结果进行比较，使用骰子分数和豪斯多夫距离来评估校正后图像中的牙齿完整性。
在模拟数据集上，监督SFTrans的表现优于LI、ResUNet和半监督SFTrans。放射科医生的视觉评估显示，原始CT的平均评分为（2.02±0.91），半监督SFTrans CT的平均评分为（4.46±0.51），监督SFTrans CT的平均评分为（3.64±0.90），所有组的组内相关系数（ICC）＞0.8，组间p＜0.001。在软组织上，半监督SFTrans和监督SFTrans均显著减少了舌头（p＜0.001）、嘴唇、双侧颊部区域和咬肌区域的金属伪影（p＜0.05）。在所有ROI中，半监督SFTrans实现的金属伪影减少效果优于监督SFTrans（p＜0.00 1）。SNR结果表明，在舌头（p = 0.0391）、双侧颊部（p = 0.0067）、嘴唇（p = 0.0208）和双侧咬肌区域（p = 0.0031），半监督SFTrans和监督SFTrans之间存在显著差异。值得注意的是，半监督SFTrans在保持牙齿完整性方面表现优于监督SFTrans（骰子分数：p＜0.001；豪斯多夫距离：p = 0.0022）。
在真实牙科CT图像中，半监督金属伪影减少模型半监督SFTrans在减少金属伪影方面表现优于监督模型。
