多模态图像自动配准融合技术联合手术机器人在面侧深区肿瘤诊疗中的应用

Application of Multi-Mode Image Automatic Registration and Fusion Technology Combined withSurgical Robot in the Diagnosis and Treatment of Face-Side Deep Tumors

目的探讨CT及MRI等多模态图像自动配准融合技术联合手术机器人在面侧深区肿瘤诊疗中的意义。资料与方法回顾性分析2019年8月—2021年8月解放军总医院第一医学中心收治的面侧深区肿瘤患者45例,术前行机器人辅助CT、MRI图像自动配准融合并计算配准误差,术中行机器人导航下穿刺活检术及肿瘤切除手术,术后行病理检查验证穿刺准确度。分析机器人自动配准融合技术引导下的穿刺准确度、穿刺并发症、配准融合误差等。结果45例患者CT、MRI图像全部实现自动配准融合,基于互信息类方法评价融合误差绝对值为0.88±0.06。良、恶性肿瘤的不同部位肿瘤例数比较,差异无统计学意义(χ^(2)=6.515,P>0.05)。机器人导航穿刺活检术取材45例,恶性肿瘤与良性肿瘤/病变分别为29例和16例;肿瘤切除手术取材37例,恶性肿瘤和良性肿瘤/病变分别为23例和14例;根据病理诊断,穿刺准确率为97.3%(36/37)。结论基于多模态CT、MRI图像自动配准融合技术联合机器人导航系统可提高面侧深区肿瘤诊断的准确性,有助于向临床提供更合理的治疗选择。

Purpose To explore the significance of multimodal image automatic registration and fusion technology via CT and MRI in the diagnosis and treatment of deep tumor in facial and lateral areas.Materials and Methods A retrospective analysis was performed on 45 patients with deep facial tumors who admitted from August 2019 to August 2021 in the First Medical Center,Chinese PLA General Hospital.Robot-assisted CT and MRI images were automatically registered and fused before surgery and registration errors were calculated.Robotguided puncture biopsy and tumor resection were performed during surgery,and postoperative pathological examination was performed to verify puncture accuracy.The puncture accuracy,puncture complications and registration and fusion errors under the guidance of robot automatic registration and fusion technology were analyzed.Results CT and MRI images of 45 patients were all automatically registered and fused,and the absolute value of fusion error was 0.88±0.06 based on mutual information class method.There was no significant difference in the number of different sites between the benign and malignant tumors(χ^(2)=6.515,P>0.05).There were 45 cases with robot guided puncture biopsy,29 cases with malignant tumor and 16 cases with benign tumor/lesion.There were 37 cases with tumor resection,including 23 cases with malignant tumor and 14 cases with benign tumor/lesion.According to pathological diagnosis,puncture accuracy was 97.3%(36/37).Conclusion Automatic registration and fusion technology based on multi-mode CT and MRI combined with robot navigation system can improve the accuracy of the diagnosis of face-side deep tumors,and contribute to providing more reasonable treatment options for the clinic.