摘要目的:探讨人工智能(AI)辅助联合低剂量扇形束计算机断层扫描(FBCT)引导的在线适应性放疗(OART)治疗宫颈癌的可行性及安全性。方法:在联影uCT-ART平台予11名宫颈癌(10名术后辅助，1名根治性放疗)患者行高年资放疗医师主导触发的OART。分析AI辅助低剂量FBCT引导的OART治疗宫颈癌的可行性，包括评估自动分割轮廓质量、自动放射治疗计划、OART在线剂量学分析、OART流程时长；以及11名宫颈癌患者的放疗相关不良反应分析。结果:在297个分次治疗中，经高年资放疗医师判断，共启动81次OART，人均启动OART 7.4次。OART流程平均总时长为18.97 min OART调整平均时长为15.87 min。11例患者在定位CT上经AI辅助勾画工具自动分割感兴趣区域(ROI)，得到ROI auto经高年资放疗医师修改及审核后得到ROI edit，其中临床靶区(CTV)的Dice相似系数为0.85 ± 0.04，不劣于前期模型0.89 ± 0.02( P>0.05)，95%豪斯多夫距离为(5.64 ± 1.60)mm，优于前期模型构建的(6.28 ± 2.31)mm( t＝－2.34， P<0.05)。OART启动后轮廓勾画策略为优先采用CTV刚性拷贝/OAR自动分割＋高年资放疗医师修改。OART计划的计划靶区(PTV)剂量分布更为紧凑，且剂量整体更接近处方剂量；OART计划中OAR剂量控制更贴近临床要求，OAR受照剂量显著低于影像引导放疗(IGRT)计划；适形指数CI与均匀性指数HI优于手动计划。在OART过程中，分次间靶区体积变化范围主要集中在±5%的范围内，而OAR体积变化范围波动较大，且无明显规律。消化系统、泌尿系统、造血系统发生急性不良反应均为PRO-CTCAE 1～2级，未见3级及以上的不良反应发生。 结论:本研究描述了基于uCT-ART的OART系统在宫颈癌放疗的成功实施，证实了OART临床应用的可行性与安全性。

abstractsObjective:To explore the feasibility and safety of artificial intelligence (AI)-assisted, low-dose fan beam CT (FBCT)-guided online adaptive radiotherapy (OART) for cervical cancer.Methods:A total of 11 cervical cancer patients (10 treated with postoperative adjuvant therapy, and 1 with radical radiotherapy) underwent OART led by senior radiation oncologists (RO) on a uCT-ART platform. The feasibility of AI-assisted, low-dose FBCT-guided OART in the treatment of cervical cancer was analyzed, including the assessment of automatic contouring quality, automatic radiotherapy planning, OART online dosimetry analysis, and OART durations, as well as the analysis of radiotherapy-associated adverse reactions in the 11 cervical cancer patients.Results:According to the senior ROs, OART was initiated 81 times in 297 fractions, with an average of 7.4. The average total OART duration was 18.97 min, and the average OART adjustment duration was 15.87 min. The regions of interest (ROIs) of 11 patients were automatically segmented based on CT images for positioning using the AI-assisted contouring tool. Then, the obtained ROI autowere modified and audited by ROs, yielding ROI edit. The dice similarity coefficient of the clinical target volumes (CTVs) was 0.85 ± 0.04, which was not inferior to that derived from previous models (0.89 ± 0.02; P > 0.05). The 95% Hausdorff distance was (5.64 ± 1.60) mm, which was better than that constructed using previous models [(6.28 ± 2.31) mm; t = -2.34, P < 0.05]. After the initiation of OART, a contouring strategy involving CTV rigid copy/automatic segmentation of organs at risk (OARs) + RO modification was preferred. In the OART plans, doses to the planning target volumes (PTVs) were more compact and closer to the prescribed doses overall. The dose control of OARs in the OART plans was more consistent with the clinical requirements, with the radiation doses to OARs significantly lower than those of image-guided radiation therapy plans. The conformity index and homogeneity index of the OART plans exhibited superiority over those of manual plans. In the course of OART, the target volumes between fractions fluctuated within ±5%, while the OAR volumes changed greatly without following evident laws. Regarding adverse reactions, the acute adverse reactions in the digestive, urinary, and hematopoietic systems were all of grades 1 to 2 according to the Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events, with no adverse reactions of grades 3 or above occurring. Conclusions:This study presents the successful implementation of the uCT-ART-based OART system in radiotherapy for cervical cancer, confirming the feasibility and safety of the clinical applications of OART.