深度学习在结直肠癌患者分期CT肝脏病灶分割与分类中的多中心技术验证研究
本研究在272例公开肝脏肿瘤CT上训练UNet深度学习模型,并在单中心2014—2019年220例结直肠癌分期CT上测试。模型检出884个病灶中的743个(84%),<10 mm病灶检出率75%;病灶级分类敏感度99.5%、特异度65.7%;患者级分类敏感度100%、特异度27.1%、NPV 100%。作者认为模型检出与分割表现良好,尤其对亚厘米病灶,100% NPV提示其作为结直肠癌分期筛查工具的潜力。
要点
- 病灶检出率随径线增大而升高:<10 mm为75%,10–20 mm为91.3%,>20 mm为96%。
- 分割DSC中位数<10 mm为0.76、10–20 mm为0.83、>20 mm为0.85。
- 病灶级分类敏感度99.5%、特异度65.7%;患者级敏感度100%、特异度27.1%。
核心结论
- 模型检出与分割表现良好,尤其对亚厘米病灶;分类准确度中等,但100% NPV提示其作为结直肠癌分期筛查工具的潜力。
研究设计与样本
- 基于UNet的深度学习模型,在272例公开肝脏肿瘤CT上训练,在单中心2014—2019年220例结直肠癌分期CT上测试。
诊断标准与关键数值
- 病灶检出:884个病灶中检出743个(84%),<10 mm检出率75%、10–20 mm为91.3%、>20 mm为96%。
- 分割DSC中位数:<10 mm为0.76(95% CI 0.72–0.80),10–20 mm为0.83(95% CI 0.79–0.86),>20 mm为0.85(95% CI 0.82–0.88)。
- 体积测量:Bland-Altman平均体积偏倚-0.12 cm3(LOAs -1.68至+1.43 cm3),ICC 0.81。
- 病灶级分类(仅检出病灶):敏感度99.5%、特异度65.7%、PPV 53.8%、NPV 99.7%、准确率75.4%。
- 患者级分类:敏感度100%、特异度27.1%、PPV 59.2%、NPV 100%、准确率64.5%。
临床应用与阅片要点
- 作者认为100% NPV提示该模型作为结直肠癌分期筛查工具的潜力。
- 作者指出未来研究将评估其对放射科医师表现与效率的影响。
局限性
- 原文未报告局限性。
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
AIM: To validate a liver lesion detection and classification model using staging computed tomography (CT) scans of colorectal cancer (CRC) patients.
MATERIALS AND METHODS: A UNet-based deep learning model was trained on 272 public liver tumour CT scans and tested on 220 CRC staging CTs acquired from a single institution (2014-2019). Performance metrics included lesion detection rates by size (<10 mm, 10-20 mm, >20 mm), segmentation accuracy (dice similarity coefficient, DSC), volume measurement agreement (Bland-Altman limits of agreement, LOAs; intraclass correlation coefficient, ICC), and classification accuracy (malignant vs benign) at patient and lesion levels (detected lesions only).
RESULTS: The model detected 743 out of 884 lesions (84%), with detection rates of 75%, 91.3%, and 96% for lesions <10 mm, 10-20 mm, and >20 mm, respectively. The median DSC was 0.76 (95% CI: 0.72-0.80) for lesions <10 mm, 0.83 (95% CI: 0.79-0.86) for 10-20 mm, and 0.85 (95% CI: 0.82-0.88) for >20 mm. Bland-Altman analysis showed a mean volume bias of -0.12 cm3 (LOAs: -1.68 to +1.43 cm3), and ICC was 0.81. Lesion-level classification showed 99.5% sensitivity, 65.7% specificity, 53.8% positive predictive value (PPV), 99.7% negative predictive value (NPV), and 75.4% accuracy. Patient-level classification had 100% sensitivity, 27.1% specificity, 59.2% PPV, 100% NPV, and 64.5% accuracy.
CONCLUSION: The model demonstrates strong lesion detection and segmentation performance, particularly for sub-centimetre lesions. Although classification accuracy was moderate, the 100% NPV suggests strong potential as a CRC staging screening tool. Future studies will assess its impact on radiologist performance and efficiency.
引用
- Bashir U, Wang C, Smillie R, et al. Deep learning for liver lesion segmentation and classification on staging CT scans of colorectal cancer patients: a multi-site technical validation study. Clin Radiol. 2025. PMID: 40327945. doi:10.1016/j.crad.2025.106914
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。