Hannisdal Marianne H, Goplen Dorota, Alam Saruar, Haasz Judit, Oltedal Leif, Rahman Mohummad A, Rygh Cecilie Brekke, Lie Stein Atle, Lundervold Arvid, Chekenya Martha
Department of Oncology, Haukeland University Hospital, BergenNorway.
University of Bergen, Bergen, Norway.
Neurooncol Adv. 2023 Apr 13;5(1):vdad037. doi: 10.1093/noajnl/vdad037. eCollection 2023 Jan-Dec.
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 ( < .001) and 0.95 ( < .001), for CE, respectively, and 0.97 ( < .001) and 0.90 ( < .001), for NE, respectively. Moreover, there was a strong correlation between response assessment in volumes and HD-GLIO-volumes ( < .001, Spearman's R = 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 ( = .04) for responders, 0.91 ( > .01) for non-responders, and 0.80 ( = .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.
肿瘤负荷评估对于放射治疗（RT）、治疗反应评估及临床决策至关重要。然而，由于放射学的复杂性，手动肿瘤勾画仍然费力且具有挑战性。本研究的目的是探讨HD-GLIO工具（一种基于nnUNet算法的预训练深度学习模型集合）在肿瘤分割、反应预测及其临床应用潜力方面的可行性。
我们分析了23例4级胶质瘤患者49次多参数MRI检查中HD-GLIO预测的对比增强（CE）和非增强（NE）输出。在对HD-GLIO输出应用于RT设置的临床应用可行性进行前瞻性测试之前，将这些体积与两名独立操作人员相应的手动勾画进行回顾性比较。
对于CE，操作人员1和操作人员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，我们发现CE的类内相关系数分别为0.90（<.001）和0.95（<.001），NE的类内相关系数分别为0.97（<.001）和0.90（<.001）。此外，体积反应评估与HD-GLIO体积之间存在强相关性（<.001，Spearman相关系数R=0.83）。CE体积与NE体积之间的纵向生长关系按临床反应区分患者：反应者CE体积与NE体积的Pearson相关系数为0.55（P=.04），无反应者为0.91（P>.01），中度/混合反应者为0.80（P=.05）。
HD-GLIO在RT靶区勾画和MRI肿瘤体积评估方面是可行的。CE/NE肿瘤区生长相关性显示出预测治疗临床反应的潜力。
