Objective: To investigate the role of minimally invasive crenel lateral lumbar interbody fusion (CLIF) in the decision of fusion level in posterior correction for severe adult degenerative scoliosis. Methods: This is prospective study.Patients with level Ⅴ and Ⅵ of Lenke-Silva classification who were treated in department of spine surgery,orthopedics center,the Second Affiliated Hospital, School of Medicine, Zhejiang University from June 2016 to March 2019 were included.First,the enrolled patients completed the preoperative clinical and imaging examination,the Lenke-Silva classification was evaluated,the surgical segments in first-stage CLIF was determined and the fusion segments required for single-stage posterior correction was predicted.After the first-stage CLIF,patients received reassessment of Lenke-Silva classification and global coronal and sagittal balance.Patients were divided into two groups:the effective group (level of Lenke-Silva classification decreased) and the ineffective group (level of Lenke-Silva classification unchanged).Second-stage posterior surgery was performed based on the results of reassessments.The fusion segment,Cobb angle,parameters of global coronal and sagittal plane,visual analogue pain score (VAS) and Oswestry disability index (ODI) were compared between the two groups preoperatively,after first-stage CLIF,second-stage posterior fixation and at the final follow-up.The potential factors associated with the decrease of the level of Lenke-Silva classification were recorded and compared between the two groups.Independent sample t test,repeated measure analysis of variance,rank sum test,χ² test or Fisher exact method were used to compare the difference among groups. Results: Fifty-four patients were enrolled,including 8 males and 46 females,aged (68.8±5.8) years (range:56 to 77 years).Preoperatively,26 patients were classified as level Ⅴ by Lenke-Silva classification,28 cases were grade Ⅵ.CLIF was performed in 194 segments,with 114 (58.8%) segments receiving anterior column realignment (ACR) and 15 (7.7%) segments using hyperlordotic cages.After first-stage CLIF,22 patients with level Ⅴ and 10 patients with Ⅵ of Lenke-Silva classification decreased and were classified into effective group.The level of the remaining 4 patients with level Ⅴ and 18 patients with grade Ⅵ unchanged and were classified into ineffective group.Preoperatively,the apical vertebrae was below L₁ in all 32 patients of effective group and 18 (81.8%,18/22) patients of ineffective group.The difference was statistically significant (P=0.023).There were 7 (31.8%,7/22) patients had continuous osteophyte in front of the intervertebral space in ineffective group,while none patient had continuous osteophyte in front of the intervertebral space in effective group,and the difference was statistically significant (P=0.001).In first-stage CLIF,more intraoperative ACR(71.2% vs.39.5%,χ²=20.66,P<0.01)and hyperlordotic cage (12.7$ vs.0,P=0.001)were used in the effective group,while there was less severe cage subsidence after the operation (5.9% vs.15.8%,χ²=4.793,P=0.029) in effective group.After first-stage CLIF,there was no difference in the Cobb angle between the two groups.While,lumbar lordosis (LL) in effective group (34.0±8.3)° was greater than that of the ineffective group (25.5±9.7)° (t=3.478,P=0.001),and the difference between the pelvic incidence (PI) and LL in effective group (15.7±4.6)°was significantly smaller than ineffective group(20.0±10.8)° (t=-2.129,P=0.038).The posterior fusion levels was less,the rate of fusion to thoracic spine region and the actual fusion segment was less than that of single-stage posterior correction in effective group (all P<0.01).All patients were follow-up for 24 to 45 months.There was no significant difference in radiological and clinical results between the two groups after first-,second-stage surgery and at the final follow-up (all P>0.05). Conclusions: First-stage CLIF decreased the Lenke-Silva classification of some patients with severe degenerative scoliosis.Combined with the reassessment of Lenke-Silva classification and global coronal and sagittal plane,it helps to accurately determine the fusion segment.Decrease of Lenke-Silva classification was associated with the preoperative level of apical vertebrae,continuous osteophytes in front of the intervertebral space,intraoperative use of ACR and hyperlordotic cage and the degree of cage subsidence postoperatively.

目的： 探讨微创猫眼侧方腰椎融合术（CLIF）在重度成人退变性脊柱侧凸后路矫正融合节段选择中的作用。 方法： 本研究为前瞻性研究。纳入2016年6月至2019年3月浙江大学医学院附属第二医院骨科中心脊柱外科收治的Lenke-Silva分级为Ⅴ级和Ⅵ级的重度成人退变性脊柱侧凸患者。入组患者首先完成术前临床和影像学检查，评估Lenke-Silva分级，确定Ⅰ期CLIF手术节段，同时预判如行单纯后路矫形需融合节段数；Ⅰ期术后再次评估Lenke-Silva分级，将患者分为有效组（Lenke-Silva分级降低）和无效组（Lenke-Silva分级不变），根据二次评估的结果制定Ⅱ期后路手术方式。比较两组患者实际融合节段与预计行单纯后路矫形需融合节段的差值、分期手术前后及末次随访时侧凸Cobb角、冠状面与矢状面影像学参数、疼痛视觉模拟评分（VAS）和Oswestry功能障碍指数（ODI），分析Lenke-Silva分级降低的影响因素。组间比较采用独立样本t检验、重复测量方差分析、秩和检验、χ²检验或Fisher确切概率法。 结果： 共54例患者纳入本研究。男性8例，女性46例，年龄（68.8±5.8）岁（范围：56~77岁），术前Lenke-Silva分级Ⅴ级26例，Ⅵ级28例。行Ⅰ期CLIF 194个节段，其中114个节段58.8%行前柱松解重建术（ACR），15个节段7.7%使用大角度融合器。Ⅰ期术后，32例患者（22例Ⅴ级、10例Ⅵ级）Lenke-Silva分级降低，纳入有效组；22例患者（4例Ⅴ级、18例Ⅵ级）Lenke-Silva分级无变化，纳入无效组。有效组中32例患者术前侧凸顶椎位置均在L₁以下，无效组中18例（81.8%，18/22）患者术前侧凸顶椎位置在L₁以下，差异有统计学意义（P=0.023）；有效组患者椎间隙前方均无连续性骨赘，无效组中7例椎间隙前方有连续性骨赘（31.8%，7/22），差异有统计学意义（P=0.001）。有效组患者Ⅰ期术中行ACR（71.2%比39.5%，χ²=20.66，P<0.01）和使用大角度融合器的节段更多（12.7%比0，P=0.001），而术后严重融合器下沉节段较少（5.9%比15.8%，χ²=4.793，P=0.029）。Ⅰ期术后，两组患者腰椎Cobb角无差异，有效组腰椎前凸角为（34.0±8.3）°大于无效组的（25.5±9.7）°（t=3.478，P=0.001），而骨盆投射角与腰椎前凸角之差为（15.7±4.6）°小于无效组的（20.0±10.8）°（t=-2.129，P=0.038）。有效组与无效组患者相比，Ⅱ期手术融合节段更少、上端固定椎终止于胸椎例数更少，实际融合节段较预判行单纯后路矫形所需融合节段减少更多（P值均<0.01）。所有患者均获得随访，随访时间24~45个月，两组患者Ⅰ期、Ⅱ期术后和末次随访时的影像学结果和临床效果的差异无统计学意义（P值均>0.05）。 结论： Ⅰ期CLIF能降低部分严重退变性脊柱侧凸患者的Lenke-Silva分级，Ⅰ期CLIF术后Lenke-Silva分级降低可能与术前顶椎位置、椎间隙前方连续性骨赘，术中使用ACR和大角度融合器，以及术后融合器下沉程度等因素相关。Ⅱ期术前进行脊柱力线再评估，有助于融合节段的准确选择，减少手术创伤。.