多层螺旋CT特征列线图模型在预测胰腺实性假乳头状肿瘤侵袭性中的应用
回顾性分析2016年1月至2021年1月114例胰腺实性假乳头状肿瘤(pSPN)患者的临床、病理及影像资料,侵袭组42例、非侵袭组72例,比较两组MSCT特征并构建列线图模型。模型预测侵袭的AUC为0.888,校准曲线与标准曲线吻合良好,基于MSCT特征的列线图模型对术前预测pSPN侵袭性有较高应用价值。
要点
- 肿瘤形态、生长方式、边缘、环形强化、包膜及动脉期和静脉期CT值是侵袭组与非侵袭组间有显著差异的MSCT特征。
- 生长方式、环形强化、包膜及动脉期病灶CT值是pSPN侵袭的独立预测因素。
- 基于上述MSCT特征构建的列线图模型预测侵袭的AUC为0.888,校准预测曲线与标准曲线吻合良好。
核心结论
- 基于MSCT特征的列线图模型对术前预测胰腺实性假乳头状肿瘤的侵袭性有较高应用价值。
研究设计与样本
- 回顾性分析2016年1月至2021年1月114例pSPN患者的临床特征、病理及影像数据,其中侵袭组42例、非侵袭组72例,比较两组MSCT特征并构建列线图模型。
关键影像表现/征象
- 两组在肿瘤形态、生长方式、边缘、环形强化、包膜、动脉期CT值及静脉期CT值方面差异有统计学意义(P < 0.05)。
- 两组在肿瘤位置、生长角度、钙化、浮云征、胰胆管扩张、远端胰腺萎缩、囊变、囊实性比例及平扫CT值方面差异无统计学意义(P > 0.05)。
诊断标准与关键数值
- pSPN侵袭的独立预测因素包括生长方式、环形强化、包膜及动脉期病灶CT值。
- 列线图模型预测pSPN侵袭的受试者工作特征曲线下面积为0.888,校准预测曲线与标准曲线吻合良好。
临床应用与阅片要点
- 阅片时可重点关注肿瘤生长方式、环形强化、包膜及动脉期病灶CT值,这些是pSPN侵袭的独立预测因素。
- 可参考基于MSCT特征构建的列线图模型对pSPN侵袭性进行术前预测,模型AUC为0.888。
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
OBJECTIVE: To evaluate the value of multi-slice computed tomography (MSCT) characteristic nomogram model in predicting invasion of pancreatic solid pseudopapillary neoplasms (pSPN).
METHODS: From January 2016 to January 2021, the clinical characteristics, pathology and imaging data of 114 patients with pSPN were retrospectively analyzed. 42 cases were classified in invasive group while 72 cases in non-invasive group. The tumor location, shape, growth pattern, growth angle, margin calcification, floating cloud sign, annular enhancement, pancreatic and bile duct dilation, distal pancreatic atrophy, capsule, cystic degeneration and cystic to solid ratio between the two groups were analyzed. The maximum diameter and the CT value of the solid parts in the lesion were measured.
RESULTS: There were significant differences in tumor shape, growth pattern, margin, annular enhancement, capsule, CT value in arterial phase and venous phase (P < 0.05). There was no significant difference in tumor location, growth angle, calcification, floating cloud sign, pancreatic and bile duct dilation, distal pancreatic atrophy, cystic degeneration, cystic-solid ratio and CT value of lesion in plain scan (P > 0.05). The independent predictors of pSPN invasion included growth pattern, annular enhancement, capsule, and CT value of lesion in arterial phase. A nomogram model was successfully established to predict the invasion of pSPN. The area under the receiver operating characteristic curve was 0.888, and the calibration prediction curve was in good agreement with the standard curve.
CONCLUSION: The nomogram model based on MSCT features has high application value in preoperative prediction of tumor invasion of pancreatic solid pseudopapillary neoplasms.
引用
- Li, Pan, Wang, et al. Application of MSCT characteristic nomogram model in predicting invasion of pancreatic solid pseudopapillary neoplasms. Eur J Radiol. 2022. PMID: 35180581. doi:10.1016/j.ejrad.2022.110201
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。