全髋关节置换人工智能三维术前规划的临床应用

Application of artificial intelligence hip system for preoperative planning total hip arthroplasty

摘要:

[目的] 探讨人工智能THA 术前规划系统（artificial intelligence hip, AIHIP） 在全髋置换术（total hip arthroplasty,THA） 应用的临床价值。[方法] 选择2020 年3 月—2022 年1 月拟行单侧初次THA 的112 例患者，依据术前医患沟通结果，74 例使用AIHIP 系统（智能组），38 例使用二维模板测量（常规组） 进行术前规划。比较两组围手术期、随访及影像结果。[结果] 智能组在手术时间[(78.4±29.6) min vs (89.1±34.3) min, P＜0.05]、髋臼假体符合率（98.6% vs 78.9%, P＜0.05）、股骨假体符合率（100.0% vs 84.2%, P＜0.05） 均显著优于常规组。两组切口长度、术中失血量、切口愈合等级、下地时间、住院时间的差异均无统计学意义（P＞0.05）。所有患者均获随访12 个月以上，智能组恢复完全负重活动时间显著早于常规组[(21.8±4.8) d vs (23.0±5.2)d, P＜0.05]。随时间推移，两组患者髋关节Harris 评分、伸屈ROM、VAS 评分均显著改善（P＜0.05）；术后1、6 个月，智能组Harris 评分和伸屈ROM、VAS 评分均显著优于常规组（P＜0.05），但术后12 个月时，两组上述指标的差异已无统计学意义（P＞0.05）。影像方面，术后两组髋臼假体外翻角、前倾角、LLD、股骨柄大小合适程度及放置位置的差异均无统计学意义（P＞0.05）。但是，智能组术后纵向髋旋转中心偏移[(1.0±0.0) mm vs (2.0±0.0) mm, P＜0.05] 和横向髋旋转中心偏移[(2.0±0.0) mm vs (3.0±0.0)mm, P＜0.05] 均显著小于常规组。[结论] AIHIP 系统辅助下THA 术前规划准确率高，能有效帮助重建髋关节旋转中心，缩短手术时间，早期恢复髋关节功能。

Abstract:

[Objective] To evaluate the clinical value of artificial intelligence hip (AIHIP) system used in preoperative planning for totalhip arthroplasty (THA). [Methods] A total of 112 patients who were undergoing unilateral primary THA from March 2020 to January 2022were selected. According to preoperative doctor-patient communication, AIHIP system was used in 74 cases (intelligent group) and 2D tem-plate measurement was used in 38 cases (conventional group) for preoperative planning. The documents of perioperative period, follow-upand imaging were compared between the two groups. [Results] The intelligent group proved significantly superior to the conventional groupin terms of operative time [(78.4±29.6) min vs (89.1±34.3) min, P＜0.05], accordance rate of acetabular component (98.6% vs 78.9%, P＜0.05)and accordance rate of femoral component (100.0% vs 84.2%, P＜0.05), despite of insignificant differences in incision length, intraoperativeblood loss, incision healing grade, ambulation time, and hospital stay between the two groups (P＞0.05). All patients were followed up formore than 12 months, and the intelligent group resumed full weight-bearing activities significantly earlier than the conventional group[(21.8±4.8) days vs (23.0±5.2) days, P＜0.05]. The Harris hip score, flexion-extension ROM and VAS scores significantly improved in bothgroups over time (P＜0.05). The intelligent group was significantly better than the conventional group in terms of Harris scores, extensionROM and VAS scores 1 and 6 months after operation (P＜0.05), but there was no statistically significant difference in abovementioned itemsbetween the two groups 12 months after surgery (P＞0.05). Radiographically, there were no statistically significant differences between thetwo groups in terms of acetabular abduction and anteversion, as well as leg length discrepancy (LLD), femoral stem size and its placement (P＞0.05). However, the intelligent group was significantly superior to the conventional group in deviation values of longitudinal hip center of rota-tion [(1.0±0.0) mm vs (2.0±0.0) mm, P＜0.05] and transverse hip center of rotation [(2.0±0.0) mm vs (3.0±0.0) mm, P＜0.05] postoperatively.[Conclusion] The AIHIP system has high accuracy in preoperative planning for THA, which can effectively help to reconstruct hip rotation center, shorten operation time and restore hip function in early stage.