T2加权MRI诊断移行区前列腺癌:主观特征与定量形态分析的比较
回顾性比较22例移行区前列腺癌与30例良性前列腺增生结节,两位盲法放射科医师主观评估形态、边缘及T2W低信号环,并提取定量形态特征。主观分析形态、边缘及无T2W环在两组间差异有统计学意义(p<0.0001),诊断移行区前列腺癌的AUC:形态0.88±0.05、边缘0.89±0.04、T2W环0.91±0.04;圆形度AUC 0.98±0.01,敏感度90%、特异度96%。
要点
- 完整T2W低信号清晰环准确诊断良性前列腺增生。
- 圆形形态准确诊断良性前列腺增生,可用圆形度定量评估。
- 透镜形准确诊断移行区前列腺癌,可用凸度定量评估。
核心结论
- 主观分析T2W-MRI可准确诊断移行区前列腺癌,定量形态分析也显示高准确度,可能有助于未来移行区肿瘤的影像基因组学分析。
研究设计与样本
- 经IRB批准,回顾性比较22例连续移行区前列腺癌与30例连续良性前列腺增生结节(15例间质型、15例腺体型),以根治性前列腺切除术MRI定位图诊断;两位盲法放射科医师主观评估形态、边缘及T2W低信号环,并分割病灶提取定量形态特征;统计采用卡方检验、Mann-Whitney U检验、Cohen kappa/Bland-Altman及ROC分析。
关键影像表现/征象
- 主观分析形态、边缘及无T2W环在移行区前列腺癌与良性前列腺增生间差异有统计学意义(p<0.0001),观察者间一致性中等至几乎完全一致(kappa:形态0.56、边缘0.72、T2W环0.97)。
- 诊断移行区前列腺癌的AUC±标准误:形态0.88±0.05、边缘0.89±0.04、T2W环0.91±0.04。
- 主观判断形态特异度高(R1 100%、R2 94%),敏感度低至中等(R1 55%、R2 88%)。
- 圆形度与凸度在两组间差异有统计学意义(p<0.001),拓扑/骨架分支差异无统计学意义(p=0.31)。
诊断标准与关键数值
- 诊断移行区前列腺癌的AUC±标准误、敏感度/特异度:圆形度0.98±0.01、90%/96%;凸度0.85±0.06、68%/97%。
- 圆形度AUC高于主观分析(p=0.01和0.26)。
- 完整T2W低信号清晰环准确诊断良性前列腺增生;圆形形态准确诊断良性前列腺增生;透镜形准确诊断移行区前列腺癌。
临床应用与阅片要点
- 阅片时可关注病灶形态(圆形/卵圆形 vs 透镜形)、边缘(清晰 vs 模糊/不清)及有无完整T2W低信号环,以鉴别移行区前列腺癌与良性前列腺增生。
- 定量圆形度与凸度可用于辅助评估形态特征,圆形度诊断移行区前列腺癌的AUC达0.98±0.01。
- 主观判断形态特异度高但敏感度低至中等,需结合其他特征综合判断。
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 T2-weighted (T2W) MRI to differentiate transition zone (TZ) prostate cancer (PCa) from benign prostatic hyperplasia (BPH).
MATERIALS AND METHODS: With IRB approval, 22 consecutive TZ PCa were retrospectively compared with 30 consecutive BPH (15 stromal, 15 glandular) nodules diagnosed using radical prostatectomy MRI maps. Two blinded radiologists (R1/R2) subjectively assessed the shape (round/oval vs. lenticular) and margin (circumscribed vs. blurred/indistinct) and for a T2W hypointense rim. Both radiologists segmented lesions extracting quantitative shape features (circularity, convexity and topology/skeletal branching). Statistical tests were performed using chi-square (subjective features), Mann-Whitney U (quantitative features), Cohen's kappa/Bland-Altman and receiver-operator characteristic analysis.
RESULTS: There were differences in the subjective analysis of the shape, margin and absence of a T2W-rim comparing TZ PCa with BPH (p < 0.0001) with moderate to almost perfect agreement [kappa = 0.56 (shape), 0.72 (margin), 0.97 (T2W-rim)]. Area under the curve (AUC ± standard error) for diagnosis of TZ PCas was shape = 0.88 ± 0.05, margin = 0.89 ± 0.04, and T2W-rim = 0.91 ± 0.04. Shape, judged subjectively, was specific (100%/94% R1/R2) with low-to-moderate sensitivity (55%/88% R1/R2). Circularity and convexity differed between groups (p < 0.001) with no difference in topology/skeletal branches (p = 0.31). Agreement in measurements was substantial for significant quantitative variables and AUC ± SE, sensitivity and specificity for diagnosis of TZ PCa were: circularity = 0.98 ± 0.01, 90%/96%; convexity = 0.85 ± 0.06, 68%/97%. AUCs for circularity were higher than for subjective analysis (p = 0.01 and 0.26).
CONCLUSION: Subjective analysis of T2W-MRI accurately diagnoses TZ PCa with high accuracy also demonstrated for quantitative shape analysis, which may be useful for future radiogenomic analysis of transition zone tumors.
KEY POINTS: • Presence of a complete T2-weighted hypointense circumscribed rim accurately diagnoses BPH. • Round shape accurately diagnoses BPH and can be assessed quantitatively using circularity. • Lenticular shape accurately diagnoses TZ PCa and can be assessed quantitatively using convexity.
引用
- Krishna S, Schieda N, McInnes MD, et al. Diagnosis of transition zone prostate cancer using T2-weighted (T2W) MRI: comparison of subjective features and quantitative shape analysis. Eur Radiol. 2018. PMID: 30105411. doi:10.1007/s00330-018-5664-z
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。