基于体素剂量学预测树脂基钇90放射栓塞治疗肝细胞癌患者的治疗反应及相关毒性:一项前瞻性单臂研究
本研究基于前瞻性单臂临床试验NCT04172714,评估基于体素剂量学预测肝细胞癌患者接受树脂基钇90放射栓塞后治疗反应与不良事件的效果。结果显示,平均肿瘤剂量253 Gy预测客观缓解(灵敏度92%、特异度83%,AUC=0.929,p<0.001),平均肿瘤剂量337 Gy预测完全缓解(灵敏度83%、特异度89%,AUC=0.845,p<0.001);非段切除队列中平均非肿瘤肝剂量81 Gy和87 Gy分别在3个月和6个月预测3级不良事件(灵敏度100%、特异度100%)。结论认为树脂基钇90治疗肝细胞癌存在直接影响结局的剂量反应和剂量毒性阈值。
要点
- 平均肿瘤剂量253 Gy预测客观缓解,灵敏度92%、特异度83%(AUC=0.929,p<0.001)。
- 平均肿瘤剂量337 Gy预测完全缓解,灵敏度83%、特异度89%(AUC=0.845,p<0.001)。
- 非段切除队列中平均非肿瘤肝剂量81 Gy和87 Gy分别在3个月和6个月预测3级不良事件(灵敏度100%、特异度100%)。
核心结论
- 在接受树脂基钇90治疗的肝细胞癌患者中,存在直接影响结局的剂量反应和剂量毒性阈值。
研究设计与样本
- 本相关性研究基于前瞻性单臂临床试验NCT04172714,该试验评估低剂量/侦察活性(555 MBq)树脂基钇90用于治疗计划的有效性。
- 非段切除治疗采用分区模型,目标为肿瘤剂量(TD)>200 Gy且非肿瘤肝剂量(NTLD)<70 Gy;段切除采用单 compartment 剂量200 Gy。
- 给予处方钇90活性减去侦察活性作为治疗性钇90,随后行钇90-PET/CT,使用Sureplan®(MIM Software,克利夫兰,俄亥俄州)进行剂量学分析。
- 共治疗30例患者、33个肿瘤(19个段切除和14个非段切除);1例患者在首次成像前死亡,其临床随访被排除在本分析之外。
- 治疗反应在3个月和6个月评估,采用受试者工作特征曲线确定预测客观缓解(OR)和完全缓解(CR)的TD反应阈值以及预测不良事件(AEs)的NTLD阈值。
主要发现
- 总体而言,26个(81%)肿瘤达到客观缓解,23个(72%)达到完全缓解。
- 平均TD 253 Gy预测客观缓解,灵敏度92%、特异度83%(AUC=0.929,p<0.001)。
- 平均TD 337 Gy预测完全缓解,灵敏度83%、特异度89%(AUC=0.845,p<0.001)。
- 在非段切除队列中,平均NTLD 81 Gy和87 Gy分别在钇90治疗后3个月和6个月预测3级不良事件,灵敏度100%、特异度100%。
关键数值与推荐
- 非段切除治疗采用分区模型,目标为TD>200 Gy且NTLD<70 Gy;段切除采用单 compartment 剂量200 Gy。
- 平均TD 253 Gy预测客观缓解的灵敏度92%、特异度83%(AUC=0.929,p<0.001)。
- 平均TD 337 Gy预测完全缓解的灵敏度83%、特异度89%(AUC=0.845,p<0.001)。
- 非段切除队列中平均NTLD 81 Gy和87 Gy分别在3个月和6个月预测3级不良事件,灵敏度100%、特异度100%。
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: There is an increasing body of evidence indicating Y90 dose thresholds for tumor response and treatment-related toxicity. These thresholds are poorly studied in resin Y90, particularly in hepatocellular carcinoma (HCC). PURPOSE: To evaluate the efficacy of prospective voxel-based dosimetry for predicting treatment response and adverse events (AEs) in patients with HCC undergoing resin-based Y90 radioembolization. MATERIALS AND METHODS: This correlative study was based on a prospective single-arm clinical trial (NCT04172714), which evaluated the efficacy of low/scout (555 MBq) activity of resin-based Y90 for treatment planning. Partition model was used with goal of tumor dose (TD) > 200 Gy and non-tumoral liver dose (NTLD) < 70 Gy for non-segmental therapies. Single compartment dose of 200 Gy was used for segmentectomies. Prescribed Y90 activity minus scout activity was administered for therapeutic Y90 followed by Y90-PET/CT. Sureplan® (MIM Software, Cleveland, OH) was used for dosimetry analysis. Treatment response was evaluated at 3 and 6 months. Receiver operating characteristic curve determined TD response threshold for objective response (OR) and complete response (CR) as well as non-tumor liver dose (NTLD) threshold that predicted AEs. RESULTS: N = 30 patients were treated with 33 tumors (19 segmental and 14 non-segmental). One patient died before the first imaging, and clinical follow-up was excluded from this analysis. Overall, 26 (81%) of the tumors had an OR and 23 (72%) had a CR. A mean TD of 253 Gy predicted an OR with 92% sensitivity and 83% specificity (area under the curve (AUC = 0.929, p < 0.001). A mean TD of 337 Gy predicted a CR with 83% sensitivity and 89% specificity (AUC = 0.845, p < 0.001). A mean NTLD of 81 and 87 Gy predicted grade 3 AEs with 100% sensitivity and 100% specificity in the non-segmental cohort at 3- and 6-month post Y90, respectively. CONCLUSION: In patients with HCC undergoing resin-based Y90, there are dose response and dose toxicity thresholds directly affecting outcomes. CLINICAL TRIAL NUMBER: NCT04172714.
引用
- Kokabi N, Arndt-Webster L, Chen B, et al. Voxel-based dosimetry predicting treatment response and related toxicity in HCC patients treated with resin-based Y90 radioembolization: a prospective, single-arm study. Eur J Nucl Med Mol Imaging. 2023. PMID: 36650357. doi:10.1007/s00259-023-06111-9
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。