肿瘤负荷评估对于放射治疗（RT）疗效评估和临床决策至关重要，但手动肿瘤描绘因放射学复杂性而仍然费力和具有挑战性。本研究的目的是探讨基于nnUNet算法的预训练深度学习模型集合 HD-GLIO 工具在肿瘤分割和反应预测方面的可行性，以及其在临床部署方面的潜力。我们分析了来自23名4级胶质瘤患者的49个多参数磁共振检查中 HD-GLIO 输出预测的强化对比度（CE）和非强化（NE）情况。在前瞻性测试 HD-GLIO 输出在 RT 环境下的可行性之前，将其体积与2名独立操作员的手动描绘进行了回顾性比较。对于 CE，操作员1 和操作员2 的中位 Dice 分数分别为 0.81（95％ CI 0.71-0.83）和 0.82（95％ CI 0.74-0.84）。对于 NE，中位 Dice 分数分别为 0.65（95％ CI 0.56-0.69）和 0.63（95％ CI 0.57-0.67）。比较体积大小，对于 CE，我们发现了优秀的组内相关系数，分别为 0.90（P <.001）和 0.95（P <.001），对于 NE，分别为 0.97（P <.001）和 0.90（P <.001）。此外，神经肿瘤学体积和 HD-GLIO 体积的反应评估之间存在强相关性（P <.001，Spearman R2 = 0.83）。CE 和 NE 体积的纵向生长关系能够区分临床反应：对于反应者，CE 和 NE 体积的 Pearson 相关系数为 0.55（P = .04），对于非反应者为 0.91（P> .01），对于中间/混合型反应者为 0.80（P = .05）。 HD-GLIO 对于 RT 靶区描绘和 MRI 肿瘤体积评估是可行的。CE/NE 肿瘤区域生长的相关性显示了预测治疗临床反应的潜力。©2023年作者及牛津大学出版社、神经肿瘤协会和欧洲神经肿瘤协会共同发表。

Tumor burden assessment is essential for radiation therapy (RT), treatment response evaluation, and clinical decision-making. However, manual tumor delineation remains laborious and challenging due to radiological complexity. The objective of this study was to investigate the feasibility of the HD-GLIO tool, an ensemble of pre-trained deep learning models based on the nnUNet-algorithm, for tumor segmentation, response prediction, and its potential for clinical deployment.We analyzed the predicted contrast-enhanced (CE) and non-enhancing (NE) HD-GLIO output in 49 multi-parametric MRI examinations from 23 grade-4 glioma patients. The volumes were retrospectively compared to corresponding manual delineations by 2 independent operators, before prospectively testing the feasibility of clinical deployment of HD-GLIO-output to a RT setting.For CE, median Dice scores were 0.81 (95% CI 0.71-0.83) and 0.82 (95% CI 0.74-0.84) for operator-1 and operator-2, respectively. For NE, median Dice scores were 0.65 (95% CI 0.56-0,69) and 0.63 (95% CI 0.57-0.67), respectively. Comparing volume sizes, we found excellent intra-class correlation coefficients of 0.90 (P < .001) and 0.95 (P < .001), for CE, respectively, and 0.97 (P < .001) and 0.90 (P < .001), for NE, respectively. Moreover, there was a strong correlation between response assessment in Neuro-Oncology volumes and HD-GLIO-volumes (P < .001, Spearman's R2 = 0.83). Longitudinal growth relations between CE- and NE-volumes distinguished patients by clinical response: Pearson correlations of CE- and NE-volumes were 0.55 (P = .04) for responders, 0.91 (P > .01) for non-responders, and 0.80 (P = .05) for intermediate/mixed responders.HD-GLIO was feasible for RT target delineation and MRI tumor volume assessment. CE/NE tumor-compartment growth correlation showed potential to predict clinical response to treatment.© The Author(s) 2023. Published by Oxford University Press, the Society for Neuro-Oncology and the European Association of Neuro-Oncology.