Department of Radiology, The First Affiliated Hospital of Anhui Medical University, Hefei, China.
College of Medical Information Engineering, Anhui University of Chinese Medicine, Hefei, China.
Med Phys. 2024 Oct;51(10):7464-7478. doi: 10.1002/mp.17316. Epub 2024 Jul 17.
X-ray radiography is a widely used imaging technique worldwide, and its image quality directly affects diagnostic accuracy. Therefore, X-ray image quality control (QC) is essential. However, subjectively assessing image quality is inefficient and inconsistent, especially when large amounts of image data are being evaluated. Thus, subjective assessment cannot meet current QC needs.
To meet current QC needs and improve the efficiency of image quality assessment, a complete set of quality assessment criteria must be established and implemented using artificial intelligence (AI) technology. Therefore, we proposed a multi-criteria AI system for automatically assessing the image quality of knee radiographs.
A knee radiograph QC knowledge graph containing 16 "acquisition technique" labels representing 16 image quality defects and five "clarity" labels representing five grades of clarity were developed. Ten radiographic technologists conducted three rounds of QC based on this graph. The single-person QC results were denoted as QC1 and QC2, and the multi-person QC results were denoted as QC3. Each technologist labeled each image only once. The ResNet model structure was then used to simultaneously perform classification (detection of image quality defects) and regression (output of a clarity score) tasks to construct an image QC system. The QC3 results, comprising 4324 anteroposterior and lateral knee radiographs, were used for model training (70% of the images), validation (10%), and testing (20%). The 865 test set data were used to evaluate the effectiveness of the AI model, and an AI QC result, QC4, was automatically generated by the model after training. Finally, using a double-blind method, the senior QC expert reviewed the final QC results of the test set with reference to the results QC3 and QC4 and used them as a reference standard to evaluate the performance of the model. The precision and mean absolute error (MAE) were used to evaluate the quality of all the labels in relation to the reference standard.
For the 16 "acquisition technique" features, QC4 exhibited the highest weighted average precision (98.42% ± 0.81%), followed by QC3 (91.39% ± 1.35%), QC2 (87.84% ± 1.68%), and QC1 (87.35% ± 1.71%). For the image clarity features, the MAEs between QC1, QC2, QC3, and QC4 and the reference standard were 0.508 ± 0.021, 0.475 ± 0.019, 0.237 ± 0.016, and 0.303 ± 0.018, respectively.
The experimental results show that our automated quality assessment system performed well in classifying the acquisition technique used for knee radiographs. The image clarity quality evaluation accuracy of the model must be further improved but is generally close to that of radiographic technologists. Intelligent QC methods using knowledge graphs and convolutional neural networks have the potential for clinical applications.
X 射线摄影是一种在全球范围内广泛使用的成像技术，其图像质量直接影响诊断准确性。因此，X 射线图像质量控制（QC）至关重要。然而，主观评估图像质量效率低下且不一致，尤其是在评估大量图像数据时。因此，主观评估无法满足当前的 QC 需求。
为了满足当前的 QC 需求并提高图像质量评估的效率，必须使用人工智能（AI）技术建立和实施一套完整的质量评估标准。因此，我们提出了一种多标准 AI 系统，用于自动评估膝关节 X 光片的图像质量。
开发了一个包含 16 个“采集技术”标签的膝关节 X 光片 QC 知识图谱，这些标签代表 16 种图像质量缺陷，以及五个“清晰度”标签，代表五个清晰度等级。十位放射技师根据该图谱进行了三轮 QC。单人 QC 结果表示为 QC1 和 QC2，多人 QC 结果表示为 QC3。每位技师仅对每张图像进行一次标注。然后使用 ResNet 模型结构同时执行分类（检测图像质量缺陷）和回归（输出清晰度评分）任务，以构建图像 QC 系统。使用 4324 张前后位和侧位膝关节 X 光片的 QC3 结果进行模型训练（占图像的 70%）、验证（10%）和测试（20%）。使用 865 个测试集数据评估 AI 模型的有效性，并在训练后由模型自动生成 AI QC 结果 QC4。最后，高级 QC 专家使用双盲法，参考 QC3 和 QC4 的结果对测试集的最终 QC 结果进行复查，并将其作为参考标准，以评估模型的性能。使用精度和平均绝对误差（MAE）来评估所有标签相对于参考标准的质量。
对于 16 个“采集技术”特征，QC4 的加权平均精度最高（98.42%±0.81%），其次是 QC3（91.39%±1.35%）、QC2（87.84%±1.68%）和 QC1（87.35%±1.71%）。对于图像清晰度特征，QC1、QC2、QC3 和 QC4 与参考标准之间的 MAE 分别为 0.508±0.021、0.475±0.019、0.237±0.016 和 0.303±0.018。
实验结果表明，我们的自动质量评估系统在对膝关节 X 光片使用的采集技术进行分类方面表现良好。模型的图像清晰度质量评估准确性还需要进一步提高，但总体上接近放射技师的水平。使用知识图谱和卷积神经网络的智能 QC 方法具有临床应用的潜力。
