多参数光谱CT对胰腺导管腺癌病理分级的术前预测增量价值
本研究回顾性分析125例病理确诊的胰腺导管腺癌患者,评估多参数光谱CT定量参数对肿瘤分化程度的术前预测价值。低分化肿瘤在40keV VMI、碘浓度、有效原子序数和光谱曲线斜率上显著低于非低分化肿瘤,三相联合模型AUC为0.75,简化静脉期列线图AUC为0.68。光谱CT多参数方法具有中等鉴别能力,优于单参数分析,但不足以独立决策,可作为风险分层的辅助工具。
要点
- 低分化PDAC的40keV VMI、碘浓度、Zeff和λHU值显著低于非低分化肿瘤。
- 三相联合多参数模型预测分化程度的AUC为0.75,优于单参数。
- 简化静脉期列线图AUC为0.68,校准良好,可辅助风险分层。
关键影像表现/征象
- 低分化肿瘤的40keV VMI、碘浓度、Zeff和λHU值显著低于非低分化肿瘤(P < .05)。
- 单光谱参数显示中等诊断能力,三相联合模型预测性能最高(AUC = 0.75)。
诊断标准与关键数值
- 三相联合多参数模型预测PDAC分化程度的AUC为0.75。
- 基于四个独立预测因子的简化静脉期列线图AUC为0.68,校准良好(Hosmer-Lemeshow P = .61)。
临床应用与阅片要点
- 光谱CT多参数定量分析可用于术前辅助评估PDAC分化程度,帮助风险分层。
- 结果需结合临床和病理评估谨慎解读,不能单独用于治疗决策。
局限性
- 回顾性研究设计。
- 单中心研究。
- 多参数模型AUC仅为中等(0.75),简化模型AUC为0.68,鉴别能力有限。
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: Preoperative assessment of the pathological grade of pancreatic ductal adenocarcinoma (PDAC) is critical for determining therapeutic strategies, yet conventional imaging often lacks sensitivity. This study investigated the incremental value of multiparametric spectral computed tomography (CT) imaging in the noninvasive prediction of PDAC differentiation.
MATERIALS AND METHODS: We retrospectively analysed 125 patients with pathologically confirmed PDAC (77 poorly differentiated, 48 nonpoorly differentiated). All patients underwent triple-phase contrast-enhanced spectral CT. Quantitative parameters, including virtual monoenergetic images (VMI) at 40 to 140 keV, iodine concentration (IC), effective atomic number (Zeff), and the slope of the spectral curve (λHU), were measured. A multiparametric logistic regression model was developed to integrate these biomarkers, and a nomogram was constructed to visualise the predictive performance.
RESULTS: Poorly differentiated tumours exhibited significantly lower values in 40 keV VMI, IC, Zeff, and λHU compared to nonpoorly differentiated tumours (P < .05). While individual spectral parameters showed moderate diagnostic ability, the exploratory three-phase combined model achieved the highest predictive performance (area under the curve [AUC] = 0.75). For clinical application, a simplified venous-phase nomogram based on four independent predictors was constructed, yielding an AUC of 0.68 with good calibration (Hosmer-Lemeshow P = .61).
CONCLUSION: Spectral CT-derived quantitative parameters differ significantly between PDAC differentiation grades. A multiparametric approach integrating these parameters yields moderate discriminative performance (AUC 0.75) for preoperative grading-superior to single-parameter analysis but insufficient for stand-alone decision-making. Thus, it may serve as a complementary tool to assist risk stratification, requiring cautious interpretation alongside established clinical and pathological assessments.
引用
- Wang B, Wei W. Preoperative prediction of pathological grading in pancreatic ductal adenocarcinoma: the incremental value of multiparametric spectral computed tomography. Clin Radiol. 2026. PMID: 42950464. doi:10.1016/j.crad.2026.107492
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。