基于三维重建软件探究人工智能术前规划系统辅助全髋关节置换的准确性及相关影响因素

Exploration of the accuracy and related influencing factors of artificial intelligence-assisted preoperative planning for total hip arthroplasty based on three-dimensional reconstruction software

摘要

目的 基于三维重建软件分析人工智能术前规划（AIHIP）系统辅助全髋关节置换术（THA）的假体预测准确性，并探究影响其规划精度的相关因素。方法 回顾性分析2022年4月至2024年5月在解放军陆军第七十二集团军医院骨科使用生物型假体行初次THA的75例患者的临床资料。术前所有患者均行常规二维模板术前规划和基于AIHIP系统的人工智能术前规划，术后比较不同术前规划方法预测假体型号准确性的差异。同时利用Mimics三维重建软件对髋关节解剖参数进行测量，探究解剖差异与AIHIP系统规划准确性间的相关性。结果 AIHIP系统预测髋臼侧、股骨侧假体完全匹配的概率显著高于二维模板组的（76.0%比34.7%，86.7%比32.0%），差异有统计学意义（χ²_髋臼侧=25.92，χ²_股骨侧=46.45，均P<0.001）。比较不同疾病患者髋臼侧术前规划的准确性，对于诊断结果为股骨头坏死（ONFH）及髋关节骨性关节炎的患者，AIHIP系统规划结果准确性显著高于二维模板测量组（90.9%比52.7%，83.3%比16.7%；均P<0.05）；对于诊断结果为发育性髋关节发育不良（DDH）的患者，AIHIP组与二维模板组髋臼侧规划结果的准确性差异无统计学意义（均P>0.05）。比较不同股骨髓腔形态患者股骨侧假体规划的准确性，对于髓腔形态为Dorr A型的患者，两种术前规划方法的准确性差异无统计学意义（P=0.273）。对于髓腔形态为Dorr B型及Dorr C型的患者，AIHIP系统术前规划结果的准确率高于二维模板测量组（89.1%比58.7%，78.3%比52.2%；均P<0.05）。分析影响AIHIP系统假体预测准确性的因素，在髋臼侧，双侧股骨头中心至泪滴线的距离差值是导致AIHIP系统规划准确性降低的因素（OR=4.332，P=0.026）。在股骨侧，双下肢长度差（OR=1.620，P=0.029）和术前诊断为DDH（OR=2.433，P<0.001）为导致AIHIP系统术前规划准确性降低的因素；而诊断为ONFH是提高AIHIP系统股骨侧假体规划准确性的因素（OR=0.410，P<0.001）。结论 AIHIP与传统二维模板规划相比具有更高的准确性与可靠性。对利用AIHIP系统进行术前规划的患者，术前应注意是否存在影响其准确性的危险因素，包括：术前诊断为DDH、双侧股骨头中心至泪滴线的距离差值大、双下肢长度差异大。手术时术者应根据术前规划结果配合术中实际情况进行适度调整，以提高手术的安全性，减少术后并发症的发生。

Abstract

Objective To investigate the accuracy of prothesis prediction of the artificial intelligence preoperative planning (AIHIP) system for total hip arthroplasty and identify factors associated with its planning precision using three-dimensional reconstruction software. Methods Clinical data of 75 patients who underwent primary artificial hip arthroplasty using cementless prosthesis from April 2022 to May 2025 in the Department of Orthopedics, the 72nd Group Army Hospital of Army Ground Force were retrospectively analyzed. Conventional 2D template measurements and artificial intelligence preoperative planning based on the AIHIP system were performed in all patients preoperatively, and the differences in the accuracy of predicting the prosthesis type between different preoperative planning methods were compared postoperatively. The anatomical parameters of the hip joint were also measured using Mimics3D reconstruction software to explore the correlation between anatomical differences and the planning accuracy of the AIHIP system. Results The AIHIP system predicted a significantly higher probability of complete matching for both the acetabular and femoral prostheses compared with the two-dimensional templating group (76.0% vs. 34.7%, 86.7% vs. 32.0%), and the differences were statistically significant (χ²_acetabular=25.92, χ²_femoral=46.45, both P<0.001). When comparing the accuracy of preoperative planning for patients with different diseases, the accuracy of planning results of the AIHIP system was significantly higher than that of the two-dimensional template measurement group for patients diagnosed with osteonecrosis of the femoral head (ONFH) and osteoarthritis of the hip (90.9%vs. 52.7%, 83.3% vs. 16.7%, all P<0.05). For patients diagnosed with developmental dyspia of the hip (DDH), there was no statistically significant difference in the accuracy of acetabular side planning results between the AIHIP system and the 2D template group (P>0.05). When comparing the accuracy of femoral prosthesis planning in patients with different femoral canal morphology, there was no statistically significant difference in the accuracy of the two preoperative planning methods in patients with Dorr A medullary pattern (P=0.273). For patients with medullary canal morphology of Dorr type B and Dorr type C, the accuracy of preoperative planning results of the AIHIP system was higher than that of the two-dimensional template measurement group (89.1% vs. 58.7%, 78.3% vs. 52.2%; all P<0.05). Analyzing the factors affecting the accuracy of prosthesis prediction by the AIHIP system, on the acetabular side, the difference in the distance from the center of the femoral head to the teardrop line on both sides was a factor leading to decreased planning accuracy of the AIHIP system (OR=4.332, P=0.026). On the femoral side, bilateral lower limb length difference (OR=1.620, P=0.029) and preoperative diagnosis of DDH (OR=2.433, P<0.001) were factors leading to decreased preoperative planning accuracy of the AIHIP system; whereas a diagnosis of ONFH was a factor for improved planning accuracy of the AIHIP system for femoral side prostheses (OR=0.410, P<0.001). Conclusions AIHIP system has higher accuracy and reliability compared to traditional two-dimensional template planning. For patients using the AIHIP system for preoperative planning, preoperative attention should be paid to the presence of risk factors affecting its accuracy, including: preoperative diagnosis of DDH, large difference in the distance from the center of the femoral head to the teardrop line on both sides, and large bilateral lower limb length difference. Surgeons should make appropriate adjustments during surgery based on the preoperative planning results combined with the actual intraoperative situation to improve surgical safety and reduce the occurrence of postoperative complications.

关键词

人工智能 / 关节成形术，置换，髋 / 术前规划 / 三维重建

Key words

Artificial intelligence / Arthroplasty, replacement, hip / Preoperative planning / Three-dimensional reconstruction

参考文献

[1] Saiz A M, Lum Z C, Pereira G C, et al. Etiology, evaluation, and management of dislocation after primary total hip arthroplasty [J]. JBJS Rev, 2019, 7(7): e7. DOI: 10.2106/JBJS.RVW.18.00165.

[2] 李允龙. 人工髋关节置换术的生物力学研究进展[J]. 医用生物力学, 2023, 38: 636.

[3] Brenneis M, Braun S, van Drongelen S, et al. Accuracy of preoperative templating in total hip arthroplasty with special focus on stem morphology: a randomized comparison between common digital and three-dimensional planning using biplanar radiographs [J]. J Arthroplasty, 2021, 36(3): 1149-1155. DOI: 10.1016/j.arth.2020.10.016.

[4] 席立成, 方怀玺. 全髋关节置换人工智能三维术前规划的临床应用[J]. 中国矫形外科杂志, 2024, 32(1): 18-23. DOI: 10.3977/j.issn.1005-8478.2024.01.03.

[5] 徐征宇, 杜俊炜, 姜瑶, 等. 三维数字规划和二维胶片模板测量辅助全髋关节置换的效果比较[J]. 生物骨科材料与临床研究, 2021, 18(2): 17-22. DOI: 10.3969/j.issn.1672-5972.2021.02.004.

[6] 吴东, 刘星宇, 张逸凌, 等. 人工智能辅助全髋关节置换术三维规划系统的研发及临床应用研究[J]. 中国修复重建外科杂志, 2020, 34(9): 1077-1084. DOI: 10.7507/1002-1892.202005007.

[7] 丁冉, 王淇, 刘烨, 等. 人工智能三维术前规划在全髋关节置换术中的应用和准确性分析[J]. 生物骨科材料与临床研究, 2022, 19(2): 33-38. DOI: 10.3969/j.issn.1672-5972.2022.02.007.

[8] Huo Jiabang, Huang Guangxin, Han Dong, et al. Value of 3D preoperative planning for primary total hip arthroplasty based on artificial intelligence technology [J]. J Orthop Surg Res, 2021, 16(1): 156. DOI: 10.1186/s13018-021-02294-9.

[9] 郭峰, 张志峰, 孙智文, 等. 基于术前规划导航系统辅助全髋关节置换：一项随机对照临床研究[J]. 中国组织工程研究, 2023, 27(27): 4324-4331.

[10] Charnley J. Total hip replacement by low-friction arthroplasty [J]. Clin Orthop Relat Res, 1970, 72: 7-21. PMID: 5459811.

[11] Müller M E. Lessons of 30 years of total hip arthroplasty [J]. Clin Orthop Relat Res, 1992, (274): 12-21. PMID: 1728996.

[12] Sinclair V F, Wilson J, Jain N P, et al. Assessment of accuracy of marker ball placement in pre-operative templating for total hip arthroplasty [J]. J Arthroplasty, 2014, 29(8): 1658-1660.

[13] Eggli S, Pisan M, Müller M E. The value of preoperative planning for total hip arthroplasty [J]. J Bone Joint Surg Br, 1998, 80(3): 382-390. DOI: 10.1302/0301-620x.80b3.7764.

[14] 吴云起, 林万忠, 王东, 等. 人工髋关节置换术下肢长度平衡方法的研究现状及进展[J]. 中外医疗, 2018, 37(31): 196-198. DOI: 10.16662/j.cnki.1674-0742.2018.31.196.

[15] 夏天卫, 刘星宇, 刘金柱, 等. 人工智能术前规划系统辅助人工全髋关节置换术治疗成人Crowe Ⅳ型先天性髋关节发育不良的疗效研究[J]. 中国修复重建外科杂志, 2021, 35(10): 1265-1272. DOI: 10.7507/1002-1892.202106103.

[16] Ding Xuzhuang, Zhang Bingshi, Li Wenao, et al. Value of preoperative three-dimensional planning software (AI-HIP) in primary total hip arthroplasty: a retrospective study [J]. J Int Med Res, 2021, 49(11): 3000605211058874. DOI: 10.1177/03000605211058874.

[17] Robinson M, Bornstein L, Mennear B, et al. Effect of restoration of combined offset on stability of large head THA [J]. Hip Int, 2012, 22(3): 248-253. DOI: 10.5301/HIP.2012.9283.

[18] Mahmood S S, Mukka S S, Crnalic S, et al. Association between changes in global femoral offset after total hip arthroplasty and function, quality of life, and abductor muscle strength. A prospective cohort study of 222 patients [J]. Acta Orthop, 2016, 87(1): 36-41. DOI: 10.3109/17453674.2015.1091955.

[19] 杨龙, 王建吉, 刘国勇, 等. 3D打印技术在髋臼发育不良髋关节置换中的初步应用[J]. 中国矫形外科杂志, 2016, 24(17): 1550-1553. DOI: 10.3977/j.issn.1005-8478.2016.17.03.

[20] 任鹏, 柴伟, 周勇刚, 等. 人工全髋关节置换治疗Crowe Ⅱ、Ⅲ型发育性髋关节发育不良——非骨水泥臼杯高旋转中心安放的短期随访结果[J]. 中华骨与关节外科杂志, 2015, 8(2): 145-150. DOI: 10.3969/j.issn.2095-9985.2015.02.012.

[21] 蔡宏, 赵然, 刘延青, 等. 行初次全髋关节置换国人正常股骨髓腔近端解剖参数研究及其影响因素分析[J]. 中国矫形外科杂志, 2017, 25(5): 466-470. DOI: 10.3977/j.issn.1005-8478.2017.05.16.

[22] 甄平, 李旭升, 周胜虎, 等. 生物型短柄假体在年轻患者Dorr C型股骨髓腔全髋关节置换中的应用[J]. 中国矫形外科杂志, 2018, 26(3): 193-197. DOI: 10.3977/j.issn.1005-8478.2018.03.01.

[23] 张凯, 郭卓涛, 马桥桥, 等. 人工智能三维规划系统在全髋关节置换中的准确性及其影响因素[J]. 中国组织工程研究, 2024, 28(12): 1863-1868. DOI: 10.12307/2024.026.

[24] 吴迪, 贾其余, 胡联英, 等. 人工智能三维术前规划在全髋关节置换术中的临床应用[J]. 实用骨科杂志, 2023, 29(3): 202-207. DOI: 10.13795/j.cnki.sgkz.2023.03.007.

[25] 王文哲, 李涛, 张子安, 等. 国产全流程辅助髋关节置换机器人假体定位精准性研究[J]. 中华骨与关节外科杂志, 2023, 16(4): 303-309. DOI: 10.3969/j.issn.2095-9958.2023.04.03.

[26] 王坤正. 国产人工智能与骨科机器人在关节置换手术中的应用现状与前景[J]. 中华骨科杂志, 2023, 43(1): 1-4. DOI: 10.3760/cma.j.cn121113-20221209-00697.

[27] 王一寒, 李杨, 张玲, 等. 数字骨科三维可视化技术在股骨转子间骨折复位内固定中的应用[J]. 中国组织工程研究, 2021, 25(24): 3816-3820.

[28] 杨明辉, 彭卫东. 努力拓展智能骨科在老年髋部骨折治疗中的应用[J]. 骨科临床与研究杂志, 2024, 9(4): 193-196. DOI: 10.19548/j.2096-269x.2024.04.001.

[29] 刘鐘阳, 孔祥朋, 柴伟, 等. 国产关节置换手术机器人临床应用现状及展望[J]. 骨科临床与研究杂志, 2025, 10(1): 52-56. DOI: 10.19548/j.2096-269x.2025.01.007.