AI辅助诊断系统联合超声造影对甲状腺TI-RADS 4类结节的诊断价值
本研究纳入140例患者共148个常规超声评估为ACR TI-RADS 4类的甲状腺结节,均有病理或FNA结果,比较AI、CEUS及两者联合诊断的效能。联合诊断敏感度显著高于单独AI或CEUS(P<.05),联合诊断AUC为0.859,高于单独AI和CEUS。AI联合CEUS可显著提高诊断敏感度和阴性预测值,有助于恶性结节的早期检出。
要点
- 联合诊断敏感度显著优于单独AI或CEUS,有助减少TI-RADS 4类结节漏诊。
- 联合诊断AUC为0.859,高于单独AI和CEUS,提示综合判断价值更高。
- AI与CEUS单独诊断效能无显著差异,但两者NPV差异有统计学意义。
核心结论
- AI联合CEUS可显著提高ACR TI-RADS 4类甲状腺结节诊断的敏感度和阴性预测值,联合诊断AUC为0.859,高于单独AI和CEUS。
- AI具有较高诊断效能,有助于放射科医师快速评估;联合CEUS有利于恶性结节的早期检出。
研究设计与样本
- 研究纳入常规超声评估为ACR TI-RADS 4类且有病理或FNA结果的甲状腺结节,共140例患者148个结节(恶性58个、良性89个)。
- 所有结节均行CEUS检查和AI分析,比较AI、CEUS及联合诊断的敏感度、特异度、准确率、PPV和NPV。
- 参考标准为病理或细针穿刺(FNA)结果。
诊断标准与关键数值
- 联合诊断敏感度显著高于单独AI或CEUS(P<.05)。
- AI、CEUS和联合诊断的NPV差异有统计学意义(P<.05);AI与联合诊断之间NPV差异有统计学意义(P<.05)。
- AI与CEUS之间诊断效能差异无统计学意义(P>.05)。
- 联合诊断AUC为0.859,高于单独AI和CEUS。
临床应用与阅片要点
- 对常规超声评估为ACR TI-RADS 4类的甲状腺结节,可考虑AI联合CEUS进行综合评估,以提高恶性检出敏感度。
- 联合诊断阴性预测值较高,有助于减少不必要的穿刺或随访,但需结合临床。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVES: This study evaluated the diagnostic value of artificial intelligence-assistant diagnostic system combined with contrast-enhanced ultrasound in The American College of Radiology Thyroid Imaging, Reporting and Data System (ACR TI-RADS) 4 category thyroid nodules.
METHODS: Thyroid nodules that were evaluated as ACR TI-RADS 4 by conventional ultrasound were selected, all of which had pathological or fine needle aspiration (FNA) results. All nodules were examined by contrast-enhanced ultrasound (CEUS) and artificial intelligence (AI) analysis. The sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) of AI, CEUS and their combined diagnosis were compared; Analyzed and compared the diagnostic efficiency of AI, CEUS and their combined diagnosis.
RESULTS: A total of 148 thyroid nodules were included in 140 patients, including 58 malignant nodules and 89 benign nodules. The sensitivity of combined diagnosis was significantly higher than that of AI or CEUS alone (P < .05). The NPV of AI, CEUS and combined diagnosis were statistically significant (P < .05). There was no significant difference in the diagnostic efficacy between AI and CEUS (P > .05), but there was a significant difference in NPV between AI and combined diagnosis (P < .05). The AUC of the combined diagnosis was 0.859, which was higher than that of AI, CEUS alone.
CONCLUSIONS: AI has a high diagnostic efficiency, which was helpful for radiologists to make rapid assessment. AI combined CEUS can significantly improve the diagnostic sensitivity and NPV, which was beneficial for the early detection of malignant nodules.
引用
- Gong, Xin, Yin, et al. Diagnostic Value of Artificial Intelligence-Assistant Diagnostic System Combined With Contrast-Enhanced Ultrasound in Thyroid TI-RADS 4 Nodules. J Ultrasound Med. 2023. PMID: 36723397. doi:10.1002/jum.16170
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。