The Newcastle Upon Tyne Hospitals NHS Foundation Trust, UK.
The Newcastle Upon Tyne Hospitals NHS Foundation Trust, UK.
Clin Radiol. 2022 Aug;77(8):e568-e575. doi: 10.1016/j.crad.2022.04.011. Epub 2022 May 28.
To assess the local diagnostic accuracy and interobserver agreement of dynamic susceptibility contrast-enhanced magnetic resonance perfusion (DSC MRP) reporting in differentiating between disease progression and pseudoprogression (PP) at a tertiary UK centre.
This retrospective study included adults with histology-proven glioblastoma who underwent an index DSC MRP examination following treatment. Each index examination was evaluated by three reporters independently, including qualitative assessment and measurement of mean regional cerebral blood volume (rCBV) ratios. Consensus opinion was used as the reference standard and considered clinical, radiological and histological follow-up information. Examination reports were compared to each other and to the consensus opinion.
Thirty-two cases were included (19 progression, 13 pseudoprogression). Interobserver agreement was fair for qualitative opinion (κ=0.58, 95% confidence interval [CI] 0.40-0.76) and good for rCBV ratio measurement (intraclass correlation coefficient [ICC, two-way random effects model] 0.63, 95% CI=0.43-0.78). Qualitative opinion showed diagnostic accuracies of 77.1% (95% CI=67.4-85.1) for progression and 75% (95% CI=65.1-83.3) for pseudoprogression. rCBV ratios were higher for progression (6.85 ± 3.98) than pseudoprogression (3.71 ± 3.40); a 3.0 threshold value maximised the sum of sensitivity (91.1%) and specificity (69.7%) on receiver operating characteristic analysis.
DSC MRP and rCBV ratio measurement aid differentiation between progression and pseudoprogression following treatment for glioblastoma. Measurement of the rCBV ratio shows good interobserver agreement and can change opinion and improve confidence in DSC MRP reporting. Individual centres should validate their own threshold rCBV ratio values to optimise diagnostic accuracy.
评估在英国一家三级中心，动态磁敏感对比增强磁共振灌注（DSC-MRP）报告在区分疾病进展和假性进展（PP）中的局部诊断准确性和观察者间一致性。
这项回顾性研究纳入了经组织学证实的胶质母细胞瘤患者，这些患者在治疗后进行了 DSC-MRP 指数检查。每个指数检查均由三位报告员独立进行评估，包括定性评估和测量平均区域性脑血容量（rCBV）比值。共识意见被用作参考标准，并考虑了临床、影像学和组织学随访信息。将检查报告彼此进行比较，并与共识意见进行比较。
共纳入 32 例病例（19 例进展，13 例假性进展）。定性意见的观察者间一致性为中等（κ=0.58，95%置信区间[CI]0.40-0.76），rCBV 比值测量的一致性较好（双向随机效应模型的组内相关系数[ICC]0.63，95%CI=0.43-0.78）。定性意见对进展的诊断准确性为 77.1%（95%CI=67.4-85.1），对假性进展的诊断准确性为 75%（95%CI=65.1-83.3）。进展的 rCBV 比值（6.85±3.98）高于假性进展（3.71±3.40）；受试者工作特征分析显示，3.0 的阈值可最大限度地提高敏感性（91.1%）和特异性（69.7%）的总和。
DSC-MRP 和 rCBV 比值测量有助于在胶质母细胞瘤治疗后区分进展和假性进展。rCBV 比值的测量具有良好的观察者间一致性，可以改变意见并提高 DSC-MRP 报告的信心。各个中心应验证自己的 rCBV 比值阈值，以优化诊断准确性。
