双参数与多参数前列腺MRI的诊断准确性:临床实践中对比剂增益评估
研究纳入264例接受前列腺MRI的患者,先行T2和DWI判读,再行完整多参数MRI判读(Likert 1-5分)。双参数与多参数MRI在86%病例评分一致,阳性阈值Likert≥3时一致率92.8%;多参数MRI假阳性更少(11.4% vs 18.9%)、Likert 3分率更低(8.3% vs 17%),特异度更高(74% vs 67%),但敏感度(95% vs 94%)与ROC-AUC(90% vs 89%)相似。
要点
- 双参数MRI与多参数MRI在86%病例评分一致,阳性阈值Likert≥3时一致率92.8%。
- Likert≥4为阳性阈值时,多参数MRI敏感度高于双参数(89% vs 80%),检出更多临床显著前列腺癌(89.2% vs 79.6%)。
- 读片者回顾性认为DCE在28.4%病例中有帮助,其中30.6%仅在阅片后才意识到其价值。
核心结论
- 双参数MRI与多参数MRI的癌症检出率相当;但多参数MRI的Likert 3分召回率更低、特异度更高,且读片者主观认为28.4%的病例中DCE有益。
研究设计与样本
- 研究纳入264例接受前列腺MRI的患者,先行T2和DWI序列判读,再行完整多参数MRI判读,采用Likert 1-5分评分。
- 在复习DCE序列前前瞻性给出对比剂可能获益的意见,并在完整多参数MRI回顾后回顾性评估。
- 以前列腺靶向和/或系统穿刺后的最终组织学结果为结局参考。
关键影像表现/征象
- 双参数与多参数MRI在86%的病例中评分一致;二分为阴性或阳性MRI(Likert评分≥3)时,一致率增至92.8%。
- Likert评分≥3时,双参数MRI检出所有癌症的89.9%、临床显著前列腺癌的93.5%;多参数MRI分别为90.7%和94.6%。
- 多参数MRI假阳性少于双参数MRI(11.4% vs 18.9%),Likert 3分率更低(8.3% vs 17%)。
- Likert≥4为阳性阈值时,多参数MRI敏感度高于双参数MRI(89% vs 80%),检出更多临床显著前列腺癌(89.2% vs 79.6%)。
诊断标准与关键数值
- 阳性MRI阈值定义为Likert评分≥3;双参数MRI敏感度94%、特异度67%、ROC-AUC 89%;多参数MRI敏感度95%、特异度74%、ROC-AUC 90%。
- Likert评分≥3时,双参数MRI检出全部癌症的89.9%、临床显著前列腺癌的93.5%;多参数MRI分别为90.7%、94.6%。
- Likert≥4为阳性阈值时,多参数MRI敏感度89%、双参数MRI敏感度80%;多参数MRI检出临床显著前列腺癌89.2%、双参数MRI 79.6%。
临床应用与阅片要点
- DCE前瞻性被认为可能获益的病例占27.3%,但读片者仅会召回11%的患者加做DCE序列,主要用于评估Likert 3分的外周带病灶。
- 完整多参数MRI回顾后,DCE被认为在28.4%的病例中有帮助;其中23/75(30.6%)的病例仅在复习序列后才意识到其价值,原因包括增加信心、存在「安全网」病灶或炎性病灶。
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
PURPOSE: To assess the added value of dynamic contrast-enhanced (DCE) in prostate MR in clinical practice.
METHODS: Two hundred sixty-four patients underwent prostate MRI, with T2 and DWI sequences initially interpreted, prior to full multiparametric magnetic resonance imaging (mpMRI) interpretation using a Likert 1-5 scale. A prospective opinion was given on likely benefit of contrast prior to review of the DCE sequence, and retrospectively following full mpMRI review. The final histology result following targeted and/or systematic biopsy of the prostate was used for outcome purposes.
RESULTS: Biparametric magnetic resonance imaging (bpMRI) and mpMRI were assigned the same score in 86% of cases; when dichotomising to a negative or positive MRI (Likert score ≥ 3), concordance increased to 92.8%. At Likert score ≥ 3 bpMRI detected 89.9% of all cancers and 93.5% clinically significant prostate cancers (csPCa) and mpMRI 90.7% and 94.6%, respectively. mpMRI had fewer false positives than bpMRI (11.4% vs 18.9%) and a lower Likert 3 rate (8.3% vs 17%), conferring higher specificity (74% vs 67%), but similar sensitivity (95% versus 94%) and ROC-AUC (90% vs 89%). At a positive MRI threshold of Likert ≥ 4, mpMRI had a higher sensitivity than bpMRI (89% versus 80%) and detected more csPCa (89.2% versus 79.6%). DCE was prospectively considered of potential benefit in 27.3%, but readers would only recall 11% of patients for DCE sequences, mainly to assess score 3 peripheral zone lesions. Following full mpMRI review, DCE was considered helpful in 28.4% of cases; in 23/75 (30.6%) of these cases this only became apparent after reviewing the sequence, reasons included increased confidence, presence of "safety-net" lesions or inflammatory lesions.
CONCLUSION: BpMRI has equivalent cancer detection rates to mpMRI; however, mpMRI had fewer Likert 3 call rates and increased specificity and was subjectively considered of benefit by readers in 28.4% of cases.
KEY POINTS: • bpMRI has similar cancer detection rates to the full mpMRI protocol at a positive MRI threshold of Likert 3. • mpMRI had fewer intermediate category 3 calls (8.3%) than bpMRI (17%) and fewer false positives than bpMRI (11.4% vs 18.9%), conferring higher specificity (74% vs 67%). • Readers considered DCE beneficial in 28.4% of cases, but in a relatively high number (30.6%) this only became apparent after reviewing the sequence.
引用
- Zawaideh JP, Sala E, Shaida N, et al. Diagnostic accuracy of biparametric versus multiparametric prostate MRI: assessment of contrast benefit in clinical practice. Eur Radiol. 2020. PMID: 32166495. doi:10.1007/s00330-020-06782-0
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。