个性化自动分割用于大体积脑转移瘤磁共振引导自适应放疗
本研究基于177个脑转移瘤的多序列图像训练基础DLAS模型,并利用前瞻性试验中20个大体积脑转移瘤的741幅每日在线MR图像开发个性化模型。个性化模型在MRgART末次分次的平均DSC为0.90,优于基础模型的0.86(p=0.01)和刚性配准的0.88(p=0.05)。DLAS辅助勾画较手动勾画平均DSC为0.89 vs 0.85(p=0.001),勾画时间平均减少53.5%(193秒 vs 424秒,p<0.001)。
要点
- 个性化DLAS模型在MRgART末次分次平均DSC达0.90,优于基础模型和刚性配准。
- DLAS辅助勾画较手动勾画平均DSC提高至0.89(p=0.001)。
- DLAS辅助勾画平均节省53.5%时间(193秒 vs 424秒,p<0.001)。
核心结论
- 个性化DLAS模型可能提高手动勾画的准确性和效率,以优化大体积脑转移瘤MRgART的工作流程。
研究设计与样本
- 研究基于177个脑转移瘤的多序列图像训练基础DLAS模型。
- 从前瞻性试验中收集20个大体积脑转移瘤的741幅每日在线MR图像,用于开发个性化模型。
- 在8例于MRgART期间接受对比增强T1重扫描的患者中比较手动勾画与DLAS辅助勾画的准确性和效率。
主要发现
- 在MRgART末次分次(肿瘤体积显著缩小时),个性化DLAS模型表现显著优于基础模型和刚性配准。
- 基础模型、刚性配准和个性化模型的平均DSC分别为0.86(p=0.01)、0.88(p=0.05)和0.90。
- DLAS辅助勾画较手动勾画显著提高勾画准确性,平均DSC为0.89 vs 0.85(p=0.001)。
- DLAS辅助勾画平均减少勾画时间53.5%(193秒 vs 424秒,p<0.001)。
临床怎么用
- 原文未据此给出推荐。
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
BACKGROUND AND PURPOSE: Magnetic resonance-guided adaptive radiotherapy (MRgART) may improve the efficacy of large brain metastases (BMs)(≥2 cm), whereas the workflow requires optimized. This study develops a two-stage, personalized deep learning auto-segmentation (DLAS) model to assist online delineation of large BMs. MATERIALS AND METHODS: Multi-sequences images from 177 BMs were trained to develop the basic DLAS model. Then, 741 daily online MR images of 20 large BMs from a prospective trial were collected for developing a personalized model. The dice similarity coefficient (DSC) was evaluated across three methods: basic model, rigid registration and personalized model, at intervals of every five fractions. The accuracy and efficiency were compared between manual delineation (MD) and DLAS assistant delineation (DLAS-AD) in 8 patients who underwent contrast T1 re-scan during MRgART. RESULTS: The personalized DLAS model demonstrated significantly better performance compared to the basic model and rigid registration during the last fraction of MRgART (when the tumor volume achieved a significant reduction). The mean DSC for basic model vs. rigid registration vs. personalized model were 0.86 (p = 0.01) vs. 0.88 (p = 0.05) vs. 0.90, respectively. The DLAS-AD significantly improved contouring accuracy compared to MD, with a mean DSC of 0.89 vs. 0.85 (p = 0.001), and reduced contouring time by an average of 53.5 % (193 s vs. 424 s, p < 0.001). CONCLUSION: Personalized DLAS model may increase the accuracy and efficiency of MD to optimize the workflow of MRgART for large BMs.
引用
- Ma Y, Ma X, Li C, et al. Personalized auto-segmentation for magnetic resonance imaging-guided adaptive radiotherapy of large brain metastases. Radiother Oncol. 2025. PMID: 39914742. doi:10.1016/j.radonc.2025.110773
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。