最大强度投影与深度学习对PET/CT中[18F]FDG和[68Ga]Ga-PSMA高摄取病灶全自动分割的增益价值
本研究评估最大强度投影(MIP)图像用于深度学习(DL)全自动分割[18F]FDG和[68Ga]Ga-PSMA PET/CT高摄取病灶的增益价值。结果显示,[18F]FDG PET/CT中联合MIP-DL可显著提升分割性能、降低假阳性病灶并改善肿瘤负荷定量,而[68Ga]Ga-PSMA PET/CT中未见额外获益。
要点
- [18F]FDG PET/CT中联合MIP-DL方法在内部和外部测试集均显著优于标准DL。
- [18F]FDG PET/CT内部测试集标准DL病灶假发现率中位12%,联合方法为0%。
- [68Ga]Ga-PSMA PET/CT中MIP-DL及联合方法较标准DL无显著性能提升。
核心结论
- 全身或全身[18F]FDG PET/CT病灶全自动分割可能受益于在标准DL基础上加入MIP-DL分割,可减少假阳性病灶并改善患者肿瘤负荷定量。
- [68Ga]Ga-PSMA PET/CT中,与标准DL分割相比未观察到获益。
研究设计与样本
- 研究使用489例[18F]FDG PET/CT分期扫描(黑色素瘤、淋巴瘤或肺癌患者;391例用于训练、98例用于内部测试),另用117例淋巴瘤患者[18F]FDG PET/CT分期扫描作为外部测试集(另一中心、2台扫描仪)。
- [68Ga]Ga-PSMA方面使用355例前列腺癌患者全身[68Ga]Ga-PSMA PET/CT扫描(285例用于训练、70例用于测试)。
- 所有扫描均有对应的专家分割作为参考标准;每种放射性药物采用三种全自动分割方法:直接应用于PET图像的三维U-Net(标准DL)、应用于多角度MIP图像的三维U-Net(MIP-DL)以及联合方法(标准DL+MIP-DL)。
- 性能通过与参考标准分割比较进行评估,指标包括病灶检测评分、体素级分割重叠指标和临床相关影像特征定量。
主要发现
- [18F]FDG PET/CT中,MIP-DL方法病灶假发现率低于标准DL方法,但在内部和外部测试集中均未达统计学显著。
- [18F]FDG PET/CT中病灶检测灵敏度无显著差异,体素级指标观察到下降(内部测试集中位Dice系数0.65 vs 0.80)。
- [18F]FDG PET/CT中联合方法在两个测试集均获得显著提升的性能;内部测试集病灶假发现率中位0%(标准DL为12%),病灶特征一致性明显提高(组内相关系数范围:标准DL 0.42-0.94,联合方法0.80-0.94),外部集观察到相似结果。
- [68Ga]Ga-PSMA PET/CT中,MIP-DL和联合方法较标准DL无显著性能提升,标准DL在病灶可检测性方面已表现突出。
关键数值与推荐
- [18F]FDG PET/CT内部测试集:标准DL病灶假发现率中位12%,联合方法中位0%。
- [18F]FDG PET/CT内部测试集:标准DL中位Dice系数0.80,MIP-DL中位Dice系数0.65。
- [18F]FDG PET/CT内部测试集病灶特征组内相关系数范围:标准DL 0.42-0.94,联合方法0.80-0.94。
临床怎么用
- 原文未据此给出推荐。
- 作者认为,全身或全身[18F]FDG PET/CT病灶全自动分割可能受益于在标准DL基础上加入MIP-DL分割。
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)。
原文
This study investigated the added value of using maximum-intensity projection (MIP) images for fully automatic segmentation of lesions using deep learning (DL) in [18F]FDG and [68Ga]Ga-prostate-specific membrane antigen (PSMA) PET/CT scans. Methods: We used 489 staging [18F]FDG PET/CT scans from patients diagnosed with melanoma, lymphoma, or lung cancer (391 scans for training and 98 for internal testing). As an external test set, 117 staging [18F]FDG PET/CT scans from lymphoma patients (another center, 2 scanners) were used. For [68Ga]Ga-PSMA, 355 whole-body [68Ga]Ga-PSMA PET/CT scans from patients with prostate cancer were used (285 scans for training and 70 scans for testing). All scans had corresponding expert-based segmentation (ground truth). Three approaches per radiopharmaceutical were used for fully automatic segmentation: 3-dimensional U-Net applied directly on PET images (standard-DL-based), 3-dimensional U-Net applied on multiangle MIP images (MIP-DL-based), and a combined approach (standard-DL+MIP-DL-based). The performance was evaluated in comparison with ground truth segmentation through lesion detection scores, voxelwise segmentation overlap metrics, and quantification of clinically relevant imaging features. Results: For [18F]FDG PET scans, the MIP-DL-based method showed a lower lesion false-discovery rate than did the standard-DL-based approach, although not significant in internal and external test sets. Sensitivity in lesion detection did not vary significantly, and a reduction in voxelwise metrics was observed (median Dice coefficient of 0.65 vs. 0.80 in the internal test set). Significantly increased performance was obtained with the combined approach in both test sets. In the internal test set, the median false-discovery rate was 0% (12% using the standard-DL), and a considerable increase in the agreement of lesion features was observed (intraclass correlation coefficient range, 0.42-0.94 for standard-DL-based and 0.80-0.94 for the combined approach). Similar results were observed in the external set. Regarding [68Ga]Ga-PSMA scans, there was no significant increase in the performance of MIP-DL-based and combined approaches compared with standard-DL, which was already outstanding in lesion detectability. Conclusion: Fully automatic segmentation of lesions in whole-body or total-body [18F]FDG PET/CT scans may benefit from the addition of the MIP-DL-based segmentation compared with the standard-DL-based method. It reduces the number of false-positive lesions and improves the patients' tumor burden quantification. In [68Ga]Ga-PSMA PET/CT scans, no benefits were observed compared with standard-DL-based segmentation.
引用
- Constantino CS, Oliveira FPM, Machado M, et al. The Use of Maximum-Intensity Projections and Deep Learning Adds Value to the Fully Automatic Segmentation of Lesions Avid for [18F]FDG and [68Ga]Ga-PSMA in PET/CT. J Nucl Med. 2025. PMID: 40081959. doi:10.2967/jnumed.124.269067
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。