Diagnosis of Acute Aortic Syndromes on Non-Contrast CT Images with Radiomics-Based Machine Learning

Ma, Zhuangxuan; Jin, Liang; Zhang, Lukai; Yang, Yuling; Tang, Yilin; Gao, Pan; Sun, Yingli; Li, Ming

doi:10.3390/biology12030337

Cited by 5 publications

(5 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“… 57 Ma et al evaluated the potential value of radiomics features obtained from non-contrast computed tomography (NCCT) to detect acute aortic syndrome (AAS) images; they found nine radiomic features, which also included wavelet LLH mean feature, and SVM algorithm showed best performance 0.997 (95% CI, 0.992–1). 58 Furthermore, our findings were consistent with the study conducted by Wang et al, where they also identified several important features, including small dependence low gray-level emphasis, that can be utilized to predict the pretreatment CD8+ T-cell infiltration status in primary head and neck squamous cell carcinoma. 59 Ren et al conducted a study which showed that multimodal radiomics can distinguish between true tumor recurrence (TuR) and treatment-related effects (TrE) in glioma patients with high accuracy (AUC: 0.965 ± 0.069).…”

Section: Discussionsupporting

confidence: 91%

The Potential of a CT-Based Machine Learning Radiomics Analysis to Differentiate Brucella and Pyogenic Spondylitis

Yasin,

Mardan,

Abliz

et al. 2023

JIR

View full text Add to dashboard Cite

Background Pyogenic spondylitis (PS) and Brucella spondylitis (BS) are common spinal infections with similar manifestations, making their differentiation challenging. This study aimed to explore the potential of CT-based radiomics features combined with machine learning algorithms to differentiate PS from BS. Methods This retrospective study involved the collection of clinical and radiological information from 138 patients diagnosed with either PS or BS in our hospital between January 2017 and December 2022, based on histopathology examination and/or germ isolations. The region of interest (ROI) was defined by two radiologists using a 3D Slicer open-source platform, utilizing blind analysis of sagittal CT images against histopathological examination results. PyRadiomics, a Python package, was utilized to extract ROI features. Several methods were performed to reduce the dimensionality of the extracted features. Machine learning algorithms were trained and evaluated using techniques like the area under the receiver operating characteristic curve (AUC; confusion matrix-related metrics, calibration plot, and decision curve analysis to assess their ability to differentiate PS from BS. Additionally, permutation feature importance (PFI; local interpretable model-agnostic explanations (LIME; and Shapley additive explanation (SHAP) techniques were utilized to gain insights into the interpretabilities of the models that are otherwise considered opaque black-boxes. Results A total of 15 radiomics features were screened during the analysis. The AUC value and Brier score of best the model were 0.88 and 0.13, respectively. The calibration plot and decision curve analysis displayed higher clinical efficiency in the differential diagnosis. According to the interpretation results, the most impactful features on the model output were wavelet LHL small dependence low gray-level emphasis (GLDN). Conclusion The CT-based radiomics models that we developed have proven to be useful in reliably differentiating between PS and BS at an early stage and can provide a reliable explanation for the classification results.

show abstract

Section: Discussionsupporting

confidence: 91%

The Potential of a CT-Based Machine Learning Radiomics Analysis to Differentiate Brucella and Pyogenic Spondylitis

Yasin,

Mardan,

Abliz

et al. 2023

JIR

View full text Add to dashboard Cite

show abstract

“…Therefore, in future research, more advanced automatic annotation methods can be considered to reduce the time and cost of data annotation and ensure the quality of data. Secondly, unlike deep learning, due to the limited amount of data, the classification model used in this paper does not require GPU support [15], so the model can be trained and applied in a central processing unit (CPU) environment. This advantage makes the algorithm easier to implement and promote, and reduces the implementation cost, so it is more suitable for practical applications in the medical field.…”

Section: Discussionmentioning

confidence: 99%

Vascular segment bifurcation detection and key point calculation based on CT radiomics

Chang,

Diao,

Wang

et al. 2024

Preprint

View full text Add to dashboard Cite

Hepatic vascular neoplasms often occur at the bifurcation of hepatic blood vessels, making them vulnerable to ruptures and bleeding, which can be life-threatening. Therefore, accurately detecting and localizing hepatic vascular bifurcation is crucial for diagnosing hepatic vascular tumors and planning surgeries. However, existing research methods have limited sensitivity to noise and vascular deformations. To address this issue, we propose a classification algorithm based on radiomic feature extraction and machine learning (ML) techniques. We begin by segmenting and processing the original hepatic vascular CT images using Gaussian filtering to remove small connected regions. This process generates vessel mask images through alignment. Next, these images are annotated, and 1042 features are extracted through radiomic feature extraction. To simplify the model, we employ Lasso regression for feature selection, resulting in the identification of 82 optimal features. We then establish and experimentally validate a machine learning classification model. Our experimental results demonstrate that the SVM classification model performs the best, achieving a validation accuracy of 0.932, precision of 0.900, recall of 0.936, F1-score of 0.915, and AUC value of 0.987. Additionally, we compute the key points of vessel segments using both the projection method and the direct thinning method. Experimental comparisons reveal that the direct thinning method yields more accurate results.In conclusion, our proposed classification algorithm, utilizing radiomic feature extraction and machine learning techniques, is crucial for rapidly and accurately detecting hepatic vascular bifurcation. These findings have significant implications for improving the diagnosis and treatment of hepatic vascular tumors.

show abstract

“…Radiomic features exhibit high correlation with each other or irrelevant to stenosis detection. Both can lead to over-fit of the model [12]. Therefore, the necessity for feature selection and reduction becomes apparent.…”

Section: Radiomics Feature Reductionmentioning

confidence: 99%

The diagnostic value of a coronary computed tomography angiography scan-based radiomics model for coronary stenosis

Niu,

Chen,

Fan

et al. 2024

Third International Conference on Biomedical and Intelligent Systems (IC-BIS 2024)

View full text Add to dashboard Cite

Coronary artery disease (CAD) is a cardiovascular disease characterized by coronary stenosis or occlusion due to atherosclerosis, which may result in a number of symptoms, including myocardial ischemia, angina and heart failure. Coronary computed tomography angiography (CCTA) is a diagnostic assessment for CAD. Radiology encompasses a vast amount of quantitative, high-dimensional features and transform medical images into a rich dataset that can be explored for insights. This study introduces an approach leveraging radiology features for the automated detection of coronary artery stenosis. We extract curved planar reconstruction (CPR) images along with the segmentation of the coronary arteries from three-dimensional CCTA images and extract radiomic features from the segmented regions of interest. Considering the high-dimensional nature of radiology features, we utilize techniques like LASSO regression to reduce the dimensionality of these features. We construct a graph convolutional network (GCN) block to fuse radiomic features and deep features embed this block within an encoder-decoder network. In the visualization analysis of coronary radiology features, there is a qualitative distinction between lipid and calcification regions, demonstrating the diagnostic value of radiology in coronary stenosis detection.

show abstract

Diagnosis of Acute Aortic Syndromes on Non-Contrast CT Images with Radiomics-Based Machine Learning

Cited by 5 publications

References 24 publications

The Potential of a CT-Based Machine Learning Radiomics Analysis to Differentiate Brucella and Pyogenic Spondylitis

The Potential of a CT-Based Machine Learning Radiomics Analysis to Differentiate Brucella and Pyogenic Spondylitis

Vascular segment bifurcation detection and key point calculation based on CT radiomics

The diagnostic value of a coronary computed tomography angiography scan-based radiomics model for coronary stenosis

Contact Info

Product

Resources

About