Simulated operation combined with patient-specific instrumentation technology is superior to conventional technology for supramalleolar osteotomy: A retrospective comparative study CURRENT STATUS: UNDER REVIEW

Over the past seven years, our team has designed a simulated operation combined with patient-specific instrumentation (SO-PSI) assisted supramalleolar osteotomy (SMOT) method and applied it in the clinic. This study aimed to evaluate the differences between SO-PSI technology and conventional operation (CO) technology for SMOT in preoperative planning, intraoperative application, and postoperative curative effect. We retrospectively analyzed SMOT data collected from our hospital between October 2014 and December 2018. Patients (n = 28) were enrolled and divided into CO (n = 17) and SO-PSI (n = 11) groups; mean follow-up time was 33.4 (range, 13 to 59) months. We statistically analyzed and compared perioperative data, accuracy of preoperative planning, and intraoperative application; difference between pre- and post-operative radiologic ankle angles; and changes in American Orthopaedic Foot & Ankle Society (AOFAS) score, visual analogue scale (VAS) score, range of ankle motion, and Takakura stage after surgery. recovered Chi-squared tests were used to analyze differences in sex, etiology, status, stage, number of plates, specific operation steps, and complication rate. The accuracy of pre-operative plans in both groups was analyzed by linear correlation analysis. Significance was defined as p < 0.05. Statistical analyses were performed using 20.0 software (SPSS Inc., Chicago,

compared with conventional technology, SO-PSI has advantages for preoperative planning, intraoperative application, and postoperative curative effect. Background Supramalleolar osteotomy (SMOT) is an effective procedure for treatment of midstage asymmetric ankle osteoarthritis [1][2][3][4]. Studies have demonstrated that the key factors affecting the postoperative curative effect of SMOT are postoperative ankle Takakura stage [5], lower limb power line [6], and ankle angles [7]; however, SMOT using conventional operation (CO) methods often does not perfectly accomplish these surgical aims, leading unsatisfactory outcomes for some patients [8,9]. Reasons for these issues include the dependence on freehand drawing for preoperative planning, and surgeons determining the operating procedure, which depend on the surgeon's experience and fluoroscopy imaging [10].
Some studies have demonstrated that simulated operation combined with patient-specific instrumentation (SO-PSI) can be used to optimize preoperative planning and the operative process, allowing personalized and precision surgery [11][12][13]. Over the past seven years, our team has designed a SO-PSI assisted SMOT method and applied it in the clinic. Surgeons can complete preoperative evaluation and surgery simulation using a three-dimensional (3D) digital model, to optimize planning by repeated confirmation, then accurately execute the preoperative planning with the help of PSI during surgery.
Here we report collected data on preoperative planning, intraoperative application, and postoperative follow-up of SMOT performed by CO and SO-PSI technology, and a retrospective comparative study to determine whether CO or SO-PSI was more effective. (1) refused to participate in the study; (2) Charcot arthropathy, rheumatoid arthritis, or ankle infection; (3) physical activity disorders, caused by other disease, such as neurologic disorders; (4) mental illness.

Study design and patients
In total, 28 patients were included and divided into CO (n = 17) and SO-PSI (n = 11) groups. The basic characteristics of the patients are presented in Table 1. There were no significant differences in demographic factors or preoperative characteristics between the two groups. Follow-up was conducted at 3, 6, and 12 months postoperatively, with at least one follow-up after 12 months postoperative. Mean follow-up time was 33.4 (range, 13 to 59) months.

Preoperative evaluation and planning
In the CO group, preoperative planning was based on radiography images and the surgeon's experience. First, the anatomical axis of the tibia and angles of the ankle were drawn (Figs. 1a and 1b), including the tibial anterior surface angle (TAS), tibiotalar tilt angle (TTA), malleolar angle (MA), and tibial lateral surface angle (TLS). Then an osteotomy plan was drawn on the X-ray. The aim of osteotomy was to generate ideal ankle angles, while retaining the correct force line [7,[14][15][16][17] ( Fig. 1c); however, only target values of TAS and TLS could be accurately designed in the preoperative plan.
In the SO-PSI group, initially, to achieve a quick understanding of the disease condition, the anatomical axis of the tibia and ankle angles were drawn, as for the CO group (Figs. 1a and 1b). Next, computed tomography (CT) data were entered into E-3D V17.08 software (Huiqing, LTD., Nanjing, China) and incorporated into a 3D digital model. Then simulated surgery was conducted on the digital model (Fig. 1d). The simulated operation was usually divided into five steps: (1) The axis and angles were drawn on the digital model to further understand the disease condition. (2) Osteotomy, reduction, and other specific operative steps were conducted on the digital model. At this point, the osteotomy, reduction, and specific operative steps could be repeatedly adjusted to optimize the parameters (Fig. 1c). (3) One to four suitable plates and appropriate screws were chosen from our virtual database of internally fixed models, then the virtual internal fixation was installed on the digital model. (4) PSI was designed to assist with the operation. Usually, an osteotomy guide plate was constructed to assist in osteotomy, as well as a temporary fixation guide plate to maintain the normal position of bone blocks. (5) Repeated checks were conducted to determine whether the simulated operation was satisfactory, and active modifications conducted until it was optimal. Finally, after a satisfactory simulated operation had been conducted, all procedures, tools and internal fixations were recorded and prepared accordingly. Nylon material was used to construct the PSI and life-sized 3D model, using selective laser sintering (SLS) (Fig. 1e).

Operation and rehabilitation
In the CO group, as for conventional SMOT [1,2,7,[15][16][17], the medial point of the osteotomy line was approximately 3-5 cm above the joint surface. According to the preoperative plan, osteotomy and other steps were conducted to optimize the ankle alignment and position, assisted by multiple intraoperative fluoroscopies.
In the SO-PSI group, most steps were as same as those for the CO group; however, there were differences, as follows: (1) The position of the osteotomy line was determined according to the osteotomy guide plate, allowing osteotomy to be performed accurately, without repeated confirmations and attempts (Fig. 1f); (2) Optimal axes and angles were obtained with the assistance of a temporary fixation guide plate; when the plate was installed on bone blocks and then used to reciprocally form a rigid whole, the ankle alignment and position were consistent with the preoperative plan (Figs. 1d and 1 g); and (3) Suitable plates and appropriate screws were installed in the skeleton according to the preoperative plan (Fig. 1d). Ankle alignment and position in the mortise were verified fluoroscopically before and after final fixation.
For both groups, the patient wore a plaster slab until the stitches were removed from the wound, then a below-knee protective plaster cast was worn. Patients were encouraged to mobilize, with partial weight-bearing for the first 3-6 weeks; subsequently, full weight-bearing, without a plaster cast, began after the osteotomy site achieved bony union, according to radiography.

Measurement and follow-up
To compare parameters between the two groups during the perioperative period, we carefully searched and collected the following data: pre-operative planning time, number of plates, operationspecific steps, operation-time, number of fluoroscopic examinations, operative blood loss, reduction of albumin (ALB, the difference between the albumin value two days pre-op and two days post-op), total hospital stay, hospitalization expenses, complication rate, and union time.
To evaluate the accuracy of preoperative planning and intraoperative application, actual imaging data for preoperative planning and at three to six months postoperatively were compared. In the CO group, TAS and TLS values were compared, while in SO-PSI group, the values compared were TAS, TTA, MA, TLS, and tibial ankle center discrepancy (TACD; the distance between tibial ankle center in the preoperative SO plan and the actual postoperative position on three to six month CT) (Figs. 1h to 1j).
To compare postoperative follow-up between the two groups, we carefully searched and collected imaging data to evaluate changes in the ankle angle and Takakura stage. Moreover, the pre-and post-operation American Orthopaedic Foot & Ankle Society (AOFAS) ankle-hindfoot score [18], ankle range of motion (ROM) [19], and visual analogue scale (VAS) [20] were used to assess functional outcomes.

Statistical analysis
An independent Samples t-test was used to evaluate differences in age, body mass index, AOFAS score, VAS, ankle ROM, follow-up time, pre-operative planning time, operating time, number of fluoroscopies, operative-blood loss, reduction of ALB, hospital stay, pre-operative hospital stay, post-operative hospital stay, hospitalization expenses, union time, and changes in angles before and after surgery. Chi-squared tests were used to analyze differences in sex, etiology, diabetes, smoker status, Takakura stage, number of plates, specific operation steps, and complication rate. The accuracy of pre-operative plans in both groups was analyzed by linear correlation analysis. Significance was defined as p < 0.05. Statistical analyses were performed using SPSS 20.0 software (SPSS Inc., Chicago, IL, USA).

Perioperative outcomes
Mean pre-operative planning time for the SO-PSI group (289.1 ± 44.4 min) was significantly greater than that for the CO group (31.5 ± 9.1 min) (P < 0.0001). The mean operating time for the SO-PSI group (101.8 ± 16.0 min) was significantly less than that for the CO group (134.1 ± 23.2 min) (P < 0.0001). Patients in the SO-PSI group underwent significantly fewer fluoroscopy examinations (mean 7.4 ± 2.1) than those in the CO group (10.3 ± 3.3) (P = 0.008). Mean ALB reduction in the SO-PSI group (1.3 ± 0.3 g/L) was significantly less than that in the CO group (2.7 ± 1.1 g/L) (P < 0.0001). The mean pre-operative hospital stay of patients in the SO-PSI group (5.5 ± 1.4 days) was significantly longer than those in the CO group (2.0 ± 0.8 days) (P < 0.0001); however, the post-operative hospital stay of patients in the SO-PSI group (6.6 ± 1.8 days) was significantly shorter than that of the CO group (9.2 ± 1.7 days) (P = 0.001). The mean hospitalization expenses for the SO-PSI group (11369.1 ± 1700.3 dollars) was significantly greater than for the CO group (9888.2 ± 1472.6 dollars) (P = 0.028). No significant differences were detected in other parameters between the two groups ( Table 2).  to 3f). Moreover, the mean TACD in the SO-PSI group was 1.86 ± 1.06 mm (range, 0.51 to 4.2 mm) ( Table 3).

Follow-up outcomes
All patients were followed up for radiologic evaluation at one year post-operative and all included parameters improved significantly in both groups; however, the improvement in TAS and TTS were more obvious in the SO-PSI than the CO group (TAS, P = 0.006; TTS, P = 0.005) ( Table 3). The AOFAS scores for SO-PSI and CO groups increased from 51.7 ± 10.9 and 52.5 ± 10.6 before surgery to 88.2 ± 5.7 and 83.1 ± 4.5, respectively, at final follow-up. Further, the AOFAS scores for the SO-PSI group were better than those for the CO group at six months (P = 0.014) and one year (P = 0.048) postoperative, as well as at final follow-up (P = 0.024) (Fig. 4a). The VAS scores for the SO-PSI and CO groups decreased from 7.3 ± 1.6 and 7.7 ± 1.4 before surgery to 3.7 ± 0.7 and 3.1 ± 0.9, respectively, at final follow-up; the VAS score for the SO-PSI was only significantly better than that for the CO group at six months post-operatively (P = 0.006) (Fig. 4b). The mean ankle ROM values in the SO-PSI and CO groups increased from 31.8 ± 8.4 and 31.2 ± 6.8 before surgery to 37.8 ± 4.4 and 35.9 ± 5.5, respectively, at final follow-up; ankle ROM did not differ significantly between the two groups at any time point (Fig. 4c). Takakura stage improved after surgery in both groups; however, the improvement in the SO-PSI group was more marked than that in CO group. In the CO group, four patients improved their Takakura stage by two levels (23.5%), nine patients by one level (52.9%), three patients had similar stage (17.6%), and one patient became worse (5.9%) at final follow-up, relative to preoperative (Fig. 5a). In the SO-PSI group six patients improved their Takakura stage by two levels (54.5%), four patients by one level (36.4%), and one patient was at a similar stage (9.1%) at final follow-up, relative to preoperative (Fig. 5b).

Discussion
This study aimed to determine whether there is a difference between CO and SO-PSI technology in terms of preoperative planning, intraoperative application, and postoperative follow-up for SMOT.
Reference to relevant studies combined with our own experience, strongly suggests that accurate preoperative analysis of the condition and optimal postoperative ankle angles, line of force, and arthritis stage are vital factors influencing SMOT, and have become the focus of current research [1, 2, 6-10, 15-17, 21].
The results of our study demonstrate that SO-PSI technology has advantages over CO technology in preoperative planning, intraoperative application, and postoperative follow-up. There are four reasons for these findings, as follows: (1) Preoperative 3D computer analysis can assist in accurate understanding of the original state of the disease, and is considerably more reliable than conventional, hand-drawn, two-dimensional analysis, and simple observation of three-dimensional reconstruction of CT images. Numerous studies have reached similar conclusions, and it is generally considered that preoperative 3D digital analysis will substantially improve the surgeons' grasp of the original disease state [22,23]. (2) Also consistent with similar studies, the surgeons performed repeated simulated operations on a digital model prior surgery; hence, they were required to confront difficulties and prepare details in advance; hence, the feasibility and optimization of preoperative planning were superior because they were well thought through [24]. (3) Compared with conventional technology, the use of PSI during surgery can speed up the process of traditional osteotomy and alignment and make them more accurate. PSI can reduce the number of surgical procedures and fluoroscopic examinations conducted during surgery, and many procedures do not require repeated attempts when PSI is used [11]. (4) Another possible reason is that the rich digital orthopedic surgery experience of our team provided good personnel, technical reserves, and the ability to execute this new technology [12].
Nevertheless, SO-PSI technology also has some shortcomings. First, the preoperative planning time was much longer for the SO-PSI group than that for the CO group, due to the addition of numerous preoperative steps, including copying the original CT data, 3D digital model reconstruction, completion of simulated surgery, and PSI 3D printing, which are very time-consuming [25]. Second, the average hospitalization expenses of the SO-PSI group were 1,400 dollars higher than those of the CO group, because the fees for design labor and 3D printing production were high, and are not included in medical insurance in our country; however, with the improvement in health care policies, software upgrades, and the rapid development of 3D printing technology [26,27], completion of these steps is gradually becoming easier and faster for doctors, and the corresponding costs will also gradually reduce.
This study had some limitations. First, designing the PSI is a complex procedure, and surgeons usually need a technologist who is skilled at using the software to assist with this part of the process. Second, it will be necessary to research the survival rate of the ankle joints in each group after surgery, based on long-term observation in the future. Third, this study was retrospective; a prospective study should be conducted, or potentially a randomized controlled trial.

Conclusions
Our results indicate that SO-PSI technology can be used to accurately and rapidly conduct preoperative planning for SMOT. In general, compared with conventional technology, SO-PSI has advantages for preoperative planning, intraoperative application, and postoperative curative effect.     Bar charts showing that, compared with the CO group, AOFAS score, VAS, and ankle ROM of the SO-PSI group gradually improved following surgery. Specifically: (A) the AOFAS score for the SO-PSI group was better than that for the CO group at each time point; (B) the VAS for the SO-PSI group was better than that for the CO group at six months postoperative; and (C) there was no significant difference in ankle ROM between the two groups at each time point.

Figure 5
Graph showing individual changes in Takakura stage for (A) the CO group and (B) the SO-PSI group with time. Each circle represents one ankle (patient). Takakura stage improved in both groups following surgery; however, the improvement in the SO-PSI group was more marked than that in the CO group.