基于计算机辅助诊断技术的超声图像处理软件对甲状腺结节诊断效能的初步研究

目的通过比较基于计算机辅助诊断技术（CAD）的甲状腺超声图像处理软件（安克侦）与超声医师对甲状腺结节进行甲状腺影像报告和数据系统（TI-RADS）评分的差别，初步探讨安克侦CAD软件在甲状腺结节良恶性鉴别诊断中的价值。 方法选取2015年5月至2016年10月就诊于中山大学附属肿瘤医院的194例甲状腺结节患者的病例资料。每个病例选取一个结节进行分析，所有结节均经手术或超声引导下细针穿刺活检（US-FNA）病理证实。由两位年资不同的超声医师评估结节图像，分别对结节的5个特征（成分、回声、形态、边缘、强回声）进行评分，得出总的TI-RADS评分并分类。安克侦CAD软件则通过自动分析结节图像，得出基于美国放射学会（ACR）标准的TI-RADS评分。 结果安克侦CAD软件的TI-RADS评分与高年资医师比较，差异无统计学意义（Z=0.964，P=0.335），与低年资医师比较，差异具有统计学意义（Z=5.593，P<0.001）。以TR5为恶性诊断界值，安克侦CAD软件的诊断敏感度与高年资医师比较，差异无统计学意义（84.62% vs 86.54%，P=0.815），且其略高于低年资医师，但差异无统计学意义（84.62% vs 73.08%，P=0.052）。安克侦CAD软件、低年资医师、高年资医师评分的诊断特异度分别为65.56%、87.78%、82.22%，安克侦CAD软件明显低于低年资和高年资医师（65.56% vs 87.78%，χ²=12.893，P<0.01；65.56% vs 82.22%，P=0.004）。CAD软件的ROC曲线下面积均低于高年资及低年资医师（0.735 vs 0.921，Z=4.537，P<0.0001；0.735 vs 0.898，Z=4.033，P=0.0001）。 结论安克侦CAD软件对甲状腺结节的诊断敏感度与高年资及低年资医师相比，并无显著差异，但诊断特异度及诊断准确性均低于高年资及低年资医师。其综合诊断效能仍需进一步深入研究。

Computer-aided diagnostic system for assessment of thyroid nodules on ultrasound: diagnostic performance compared with radiologist-based clinical assessments

ObjectiveTo evaluate the difference in assessment of the American College Radiology thyroid imaging-reporting and data system score between a computer-aided diagnostic system (Am CAD-UT Dection) and radiologists and to discuss its value in the differential diagnosis of benign and malignant thyroid nodules on ultrasound. MethodsThis retrospective study analyzed 194 patients with thyroid nodules at Sun Yat-sen University Cancer Center between May 2015 and October 2016. Only one nodule was chosen in a patient and was proven by fine needle aspiration (FNA) cytology or thyroidectomy. A senior radiologist with 16 years of working experience and a junior radiologist with 4 years of working experience assessed the five major ultrasound characteristics of thyroid nodules including composition, echogenicity, shape, margin and echogenic foci separately, while the CAD system analyzed the same thyroid nodules automatically based on the ACR standard. ResultsThe CAD system showed a similar performance to the senior radiologist (P=0.335) but performed better than the junior radiologist (P<0.001). Using TR5 as the diagnostic standard for malignancy, the CAD system was significantly consistent with the senior doctor in sensitivity (84.62% vs 86.54%, P=0.815) but was better than the junior doctor without a significant difference (84.62% vs 73.08%, P=0.052). The specificity of the CAD system was lower than those of the above two radiologists (65.56% vs 87.78%, 82.22%, P<0.01). ConclusionThe CAD system exhibits a higher sensitivity but lower specificity and diagnostic performance than radiologist-based clinical assessments in the differential diagnosis of benign and malignant thyroid nodules.