| 摘要: |
| 摘 要
目的 应用机器学习方法评价M型超声在超声引导下甲状腺结节细针穿刺细胞学检查(US-G FNAC)过程中结节平均位移距离与其良、恶性之间关系，评价其鉴别结节病理性质的价值。方法：收集整理病理明确诊断的甲状腺结节145例，其中恶性结节84例，良性结节61例。上述患者均行超声引导下结节细针穿刺，操作过程中进行动态存图。所有图像均使用Free-Xros M后处理，测量穿刺频率及其位移大小，用支持向量机（SVM）来分析平均结节位移距离与其良、恶性的关系。结果：甲状腺FNA结节大小、形态、位置、血流以及FNA穿刺频率等在良恶性结节中差异均无明显统计学意义（P均＞0.05）。而FNA穿刺过程中结节平均位移在诊断结节良恶性有很好的价值，其敏感性为85.96%，特异性为85.57%，阳性预测值为79.03%，阴性预测值为90.24%，准确性为85.51%， ROC曲线下面积0.855；结论：M超在超声引导下甲状腺结节FNA的平均位移是一个辅助诊断结节良恶性指标，有望作为病理学不明确结节的有效定性指标。 |
| 关键词: 甲状腺结节，超声检查，机器学习，细针穿刺活检 |
| DOI： |
| 投稿时间：2021-06-19修订日期：2022-03-21 |
| 基金项目: |
|
| Diagnostic value of mean moving distance of thyroid nodules measured by? M-mode ultrasound in FNAC in benign and malignant thyroid nodules using machine learning algorithm |
| yangjinfeng,zhang chao |
| () |
| Abstract: |
| Abstract
Objective To evaluate the relationship between the mean moving distance (MMD) of thyroid nodule in FNA measured by M-mode ultrasound and its pathological result. Besides, evaluating the differentiating value of MMD in different pathological types. Methods: 145 cases of thyroid nodules diagnosed by pathology were collected, including 84 cases of malignant nodules and 61 cases of benign nodules. All the above patients underwent fine needle puncture of nodules under the guidance of ultrasound, and dynamic mapping was performed during the operation. All images were post-processed by Free-Xros M, and the puncture frequency and its displacement were measured. Support vector machine (SVM) was used to analyze the relationship between the average nodule displacement distance and its benign and malignant.Results: No statistically significant difference in size, shape, location, blood flow and FNA puncture frequency between benign and malignant thyroid nodules were found (all P >0.05). The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of FNA were 85.96%, 85.57%, 79.03%, 90.24%, 85.51% and 0.855 respectively. Conclusion: The average displacement of FNA of thyroid nodules under ultrasound guidance is an auxiliary index for diagnosing benign and malignant nodules, and is expected to be an effective qualitative index for pathologically ambiguous nodules. |
| Key words: Thyroid nodules, ultrasound, machine-learning, fine needle biopsy |
