文献速递泌尿生殖腹部MR

基于双参数MRI的前列腺癌深度学习检测系统与放射科阅片对比评估

Evaluation of a deep learning prostate cancer detection system on biparametric MRI against radiological reading.
PubMed 文献 · Eur Radiol · 2024-12-19

本研究在4381例双参数MRI训练集上训练3D nnU-Net检测临床显著前列腺癌(GGG≥2),在PROSTATEx数据集328例测试集上评估,模型AUC为0.83(95%CI:0.80,0.87),在等效假阳性率下真阳性率高于非专家放射科医生,但小病灶检测仍具挑战。

要点

  • 模型在独立测试集检测GGG≥2前列腺癌的AUC为0.83(95%CI:0.80,0.87)。
  • 病灶水平敏感度在0.5 FP/volume时为0.85(95%CI:0.82,0.94),在1 FP/volume时为0.88(95%CI:0.79,0.92)。
  • 等效假阳性率下模型真阳性率高于放射科医生,但小病灶检测仍困难。

核心结论

  • 深度学习模型在独立测试队列检测临床显著前列腺癌表现良好,超越放射科解读,显示AI可能提高非专家放射科医生的诊断准确性,但小病灶检测仍具挑战。

研究设计与样本

  • 训练集含三大洲4381例双参数MRI病例(3800例阳性、581例阴性),80%以PI-RADS标注、20%以Gleason评分标注;测试集为PROSTATEx数据集328例(34%阳性GGG≥2、66%阴性);训练3D nnU-Net用于病灶检测,以病理为基础标注评估,并与使用多参数MRI的非专家放射科医生比较。

诊断标准与关键数值

  • 临床显著前列腺癌定义为Gleason Grade Group(GGG)≥2;模型AUC为0.83(95%CI:0.80,0.87);病灶水平敏感度在0.5 FP/volume时为0.85(95%CI:0.82,0.94),在1 FP/volume时为0.88(95%CI:0.79,0.92);平均精度为0.55(95%CI:0.46,0.64);对大于650 mm³病灶敏感度超过0.90,各GGG敏感度均超过0.85。

临床应用与阅片要点

  • 在等效假阳性率下,模型对PI-RADS≥3病灶的真阳性率为0.93、对PI-RADS≥4病灶为0.79,放射科医生分别为0.87和0.68(p≤0.05);作者认为AI可能提高非专家放射科医生的诊断准确性。

局限性

  • 检测小病灶仍具挑战。

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

OBJECTIVES: This study aims to evaluate a deep learning pipeline for detecting clinically significant prostate cancer (csPCa), defined as Gleason Grade Group (GGG) ≥ 2, using biparametric MRI (bpMRI) and compare its performance with radiological reading.

MATERIALS AND METHODS: The training dataset included 4381 bpMRI cases (3800 positive and 581 negative) across three continents, with 80% annotated using PI-RADS and 20% with Gleason Scores. The testing set comprised 328 cases from the PROSTATEx dataset, including 34% positive (GGG ≥ 2) and 66% negative cases. A 3D nnU-Net was trained on bpMRI for lesion detection, evaluated using histopathology-based annotations, and assessed with patient- and lesion-level metrics, along with lesion volume, and GGG. The algorithm was compared to non-expert radiologists using multi-parametric MRI (mpMRI).

RESULTS: The model achieved an AUC of 0.83 (95% CI: 0.80, 0.87). Lesion-level sensitivity was 0.85 (95% CI: 0.82, 0.94) at 0.5 False Positives per volume (FP/volume) and 0.88 (95% CI: 0.79, 0.92) at 1 FP/volume. Average Precision was 0.55 (95% CI: 0.46, 0.64). The model showed over 0.90 sensitivity for lesions larger than 650 mm³ and exceeded 0.85 across GGGs. It had higher true positive rates (TPRs) than radiologists equivalent FP rates, achieving TPRs of 0.93 and 0.79 compared to radiologists' 0.87 and 0.68 for PI-RADS ≥ 3 and PI-RADS ≥ 4 lesions (p ≤ 0.05).

CONCLUSION: The DL model showed strong performance in detecting csPCa on an independent test cohort, surpassing radiological interpretation and demonstrating AI's potential to improve diagnostic accuracy for non-expert radiologists. However, detecting small lesions remains challenging.

KEY POINTS: Question Current prostate cancer detection methods often do not involve non-expert radiologists, highlighting the need for more accurate deep learning approaches using biparametric MRI. Findings Our model outperforms radiologists significantly, showing consistent performance across Gleason Grade Groups and for medium to large lesions. Clinical relevance This AI model improves prostate detection accuracy in prostate imaging, serves as a benchmark with reference performance on a public dataset, and offers public PI-RADS annotations, enhancing transparency and facilitating further research and development.

来源:Eur Radiol,查看原文。版权归原作者所有。

引用

  1. Debs N, Routier A, Bône A, et al. Evaluation of a deep learning prostate cancer detection system on biparametric MRI against radiological reading. Eur Radiol. 2024. PMID: 39699671. doi:10.1007/s00330-024-11287-1

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。