Virtual and Traditional Surgical Planning in Orthgnathic Surgery– Systematic Review and Meta-analysis

Traditional surgical planning (TSP) and virtual surgical planning (VSP) have been used in bimaxillary osteotomy planning. The time is taken in the planning time and operating stages, and the working/doctor/total time of either approach are useful determinants of the efficiency of the operating method and quality of care. This systematic review and meta-analysis examined if VSP has a comparative advantage over TSP in the bimaxillary osteotomy. Cochrane Library, PubMed, EMBASE, and Google Scholar were used as databases to collect studies that met the outlined inclusion criteria based on PRISMA. Eight of 759 studies were considered to meet the eligibility criteria, and six fit for meta-analysis. The findings demonstrated significant VSP advantage over TSP in planning time (Z = 3.97 (P < 0.00001), WMD = -5.29 (CI -7.90 to -2.68)). While more time-efficient than TSP, the difference with VSP was not significant during surgery (Z = 0.44 (P = 0.66), WMD = -0.10 (CI -0.51 to 0.34)). The study used random effects due to the high I 2 of the planning mean differences. The continued evolution of VSP and improved application knowledge will be important in reducing the time of planning and surgery, thus improving the outcomes of the complex bimaxillary osteotomy. The current evidence shows that VSP significantly performs better than TSP in reducing the bimaxillary osteotomy planning time, but the timing difference is not significant during surgery. Future analysis will benefit from using studies with standard research and reporting metrics and procedures, thus improving evidence-based clinical practice.


Introduction
Orthognathic surgery is a delicate process that relies on rigorous preoperative planning, accurate implementation of the selected operative plan, and postoperative care to be successful 1 . Over six decades, orthognathic surgery has evolved to become safer, faster, less costly, and more successful 2,3 . Traditional surgical planning (TSP) such as manual model surgery, the use of photographs, and two-dimensional radiographs have been conventionally used at the preoperative stage of orthognathic surgery 4 . Virtual surgical planning (VSP) approaches have increasingly been used as an alternative to TSP. The computer-aided surgical simulations used in VSP have offered surgeons a three-dimensional (3D) facial skeleton, soft tissue, and dentition representation to facilitate virtual diagnosis and surgery 2,5 .
The comparison between VSP and TSP is not straightforward across different outcomes. While VSP has been presented as a superior alternative to TSP, researchers have questioned whether the VSP technique's accuracy is higher than TSP 3 . TSP has, however, been identified to be superior from a cost perspective. HOWEVER, the TSP cost advantage was present after the initial fixed cost investment in VSP 1 . In their systematic review and metaanalysis, Chen et al. 1 found that VSP and TSP are comparable in accuracy prediction in the sagittal plane. The findings emphasize the need to evaluate specific outcomes of VSP and TSP when used in a specific osteotomy.
Bimaxillary osteotomy is a common yet complex surgical procedure that requires rigorous planning and accurate operation. Accuracy, time, and cost are key considerations when evaluating the treatment procedure in bimaxillary osteotomy 6,7 . The time that has to be considered in comparing the effectiveness of VSP and TSP includes the planning time, operating time, and the working/doctor/total time 2,7,8 . Ideally, shorter operating times are preferred as they ensure lower risk and less anaesthesia and blood loss, while shorter planning time ensures that necessary or urgent care is not delayed 1 . The total time helps determine the efficiency and cost of the approach used in bimaxillary osteotomy 2,9,10,11 .
The purpose of this study is to systematically review and perform meta-analysis on orthognathic surgery literature to determine if VSP is superior to TSP in terms of timeconsumption in bimaxillary orthognathic surgery.

Protocol
The Preferred Reporting Items for Systemic Reviews and Meta-Analyses (PRISMA) was the protocol of this systematic review and meta-analysis. The guidelines provided on PRISMA were followed in defining the inclusion criteria for this systematic review and metaanalysis. Two independent reviewers searched, reviewed, and extracted eligible articles based on the inclusion criteria. The author, however, had the final decision-making responsibility in the areas that had contention or disagreement.

Search Strategy
The articles considered and used in the systematic review and meta-analysis were extracted from Cochrane Library, PubMed, EMBASE, and Google Scholar. The bias standards outlined in PRISMA were adhered to in the search to ensure the credibility of the sources. The medical subject heading (MesSH) terms that were used as search keywords include "virtual surgery planning," "VSP," "traditional surgical planning," "conventional surgical planning," "computer-, "three-dimensional surgical planning," "3D surgical planning", "two-dimensional surgical planning," "2D surgical planning", "bimaxillary osteotomy," "bimaxillary orthognathic surgery," "double jaw surgery," "surgical time," "operating time," "total time," and "doctor time." These keywords were used in different combinations across all databases.
The search was limited to studies published between 2014 and 2021. Non-English, editorial perspective, editorial letter, and literature reviews were excluded from the study. The studies that focused solely on single jaw osteotomy were also excluded. A summary of the selected studies was provided with the author's name, year of publication, topic, location, sample size, nature and focus of study, primary outcomes, and secondary outcomes is presented as descriptive data.

Eligibility Criteria
The inclusion criteria focused on the studies that compared VSP and TSP the planning, operating, and total time in the bimaxillary osteotomy. The studies that explicitly mentioned bimaxillary osteotomy or double-jaw surgery were included. The studies that focused on single jaw osteotomy were excluded. The studies solely focused on the ischemia time were also excluded due to scarcity and inability to meet the earlier set bimaxillary osteotomy criteria. The studies with history of previous osteotomy were also excluded.
While the study prioritized random control trials (RTCs), retrospective cohort studies and other cohort studies were included in the search. The full-text analysis was used to determine consistency in meta-analysis's measured outcomes and metrics.

Data Extraction
The selected studies were assessed for risk bias based on the Cochrane Handbook of Systematic Reviews of Interventions 10 . The studies were assigned equal weight and assessed based on the randomization process, missing outcome data, deviations from intended intervention, selection of the reported result, and the measurement of the outcome. The extraction identified the primary and secondary outcomes of each included study. The study's sample sizes, mean, median, standard deviations (SD), and interquartile ranges were extracted for analysis. This systematic review and meta-analysis use Hozo et al. 11 to convert interquartile ranges (IQR) to standard deviation with the median held as the mean. The other extracted information includes demographics, interventions, and the study groups.

Data Analysis
The collected data were analysed to compare the planning, operation, and total time for bimaxillary orthognathic surgery in VSP and TSP. Additionally, statistical analysis was performed by pooling the studies and examining the two groups (VSP and TSP) for the surgical and planning time outcomes. Only one paper studied the difference in total time 2 , thus making the statistical analysis through forest plots ineffective. Review Manager Software version 5.4 (RevMan 5.4) was used in the statistical analysis. The I 2 statistic in the generated forest plots was used to examine the heterogeneity among the selected studies, and accordingly either fixed or random-effects approach was applied. The forest plot was used to determine the distribution of the studies.

Study Selection
A total of 759 studies were screened and considered for inclusion in the systematic review and meta-analysis. Of these, 74 were excluded for duplication. An additional 299 were excluded for being case reports (51), literature analysis (31), reviews (44), literature reviews (17), systematic reviews and meta-analysis (89), and commentaries (67). Of the remaining studies, 207 were excluded for using the intervention that was not targeted, measuring other outcomes, or including patients in only the single jaw osteotomy. The full-text screening was performed on the remaining sources leading to the exclusion of another 103 studies for the lack of full-text publications. Additional, 67 studies were excluded for lacking control groups and incomplete reporting of critical measures. A total of 6 studies were left after the screening to be used in the statistical meta-analysis section, and another 2 used in the narrative analysis section of this study ( Table 1) (Table 1). (Table 2) provides the statistical measures of the targeted outcomes.

Quality of Studies
The included studies were evaluated for their usefulness in this systematic analysis and meta-analysis. Only one was reviewed by an independent data monitoring committee of the included studies. None of the studies included an enrolment flow diagram, though they all described their eligibility criteria and included and excluded patients. The Cochrane Library Risk of Bias (RoB) tool evaluated the studies. The risk of bias was classified as either high, low, or uncertain. Overall, most studies were of low risk, with data and methodological concerns used only in the narrative analysis. The RoB tool used the measures of the randomization process, missing outcome data, deviations from intended intervention, selection of the reported result, and the measurement of the outcome to evaluate the risks (Figure 2).

Statistical Results for the Planning Time
Six studies discussed the planning time, however only four of them have sufficient data for meta-analysis 6,7,9,12 . One of the four studies have measured the planning time in three different groups and were included as three different measures. The diamond shape on the forest plot in ( Figure 3) does not touch the line of no effect and is on the left side. This implies that VSP is significantly favoured for its time-saving ability in the planning phase compared to TSP. Individually, all statistically analysed studies 6, 7, 9, 12 reported a significant advantage of VSP compared to TSP in saving planning time.  Figure   3). Hence, the random effects rather than fixed effects treatment are applied to the analysis.
The meta-analysis showed that the planning time in VSP is significantly shorter than TSP (P < 0.00001). The weighted mean difference (WMD) was -5.29 (CI -7.90 to -2.68).

Statistical Results for the Surgical Time
The sample results of the included two studies with regards to surgical time were used in the meta-analysis since they were highly homogenous given their low level of heterogeneity. The I 2 of 0% is evidence of the high homogeneity. In the case of the comparison for the surgical time, the I 2 = 0% Heterogeneity: Tau² = 0.00; Chi² = 0.58, df = 1 (P = 0.44). At 95% confidence level, the overall effect of Z = 0.44 (P = 0.66) with a WMD of -0.10 (CI -0.51 to 0.34) for surgical timing. The diamond shape on the forest plot in (Figure 4) touches the line of no effect.
This implies that the difference between the groups that went through surgery after VSP and those in the TSP condition was insignificant (P = 0.66). These statistics demonstrate the forest plots' fit in comparing the surgical times for VSP and TSP (Figure 4).

Discussion
In this systematic review, eight studies were reviewed to compare VSP and TSP as methods of planning bimaxillary orthognathic surgery.
Despite articles covering other areas such as soft tissue or hard tissue accuracy and precision, this analysis narrowed to three aspects: planning time, surgery time, and total time.
CT scan data in the reconstruction of 3D maxillofacial images has been in place since the 1980s 1 . The improved quality of VSP has followed over the last two decades as technology has advanced. The slow practical application of these technological advancements has been identified to contribute to the slight differences in the time advantage of VSP over TSP in some studies. Wrzosek et al. (2016) 12 demonstrated that other variables might determine the planning time. With the study using post-graduate trainees, the time was significantly prolonged given limited experience using these technologies. The time in both VSP and TSP conditions was long in the study when compared to other studies. This could be explained by the fact that most of the planning steps in the TSP were performed by residents under supervision of their surgeon, and in VSP the planning was performed through an online discussion between the clinician and a software engineer.
VSP is not an entirely new technology, the learning curve can be instrumental in reducing the time in using it in the planning 1,15 . This systemic review and meta-analysis sought to determine the significance of the theorized advantage of VSP and evaluate if the learning curve reduces the ability to realize these gains.
The studies used in the statistical analysis of the comparative surgery timing advantage of VSP over TSP were both RCTs 2,8 . The statistical analysis showed that TSP was a significant time-consumer relative to VSP in planning. VSP outperformed TSP during surgery, the difference was not significant. The insights gained from the individual studies and pooled effects demonstrate a strong case for the preference towards VSP over TSP.

Planning Time
The studies defined the TSP is the process required to complete the preoperative planning (including radiographs, 2d analysing, model surgery and surgical splints) while the VSP was defined as the tool for predicting complex surgical movement in three dimensions (including radiographs, clinical photographs, production and scanning the plaster models, virtual analysis and 3D printing of the surgical splints) 6, 7, 12 .
Our meta-analysis showed that VSP is significantly favoured for its time-saving ability in the planning phase compared to TSP. All of the included studies individually showed that less time was required when VSP method was adopted compared to TSP 6,7,9,12 . Wrzosek et al. 12 found that the advantage persisted irrespective of the complexity of the case with savings of 2.19 ± 0.93 h (30.1%), 2.22 ± 1.74 h (29.0%), and 1.98 ± 0.80 h (26.3%) in simple cases, complex cases, and multisegmented cases respectively. The findings were also consistent in the studies by Resnick 7 who also showed that the advantage persisted in symmetric, nonsegmental, asymmetric, and segmental groups. Hanafy et al. 14 found that the time between the virtual plan to STL export was 113 while TSP took 192 minutes from maxillary incision to fixation.
The planning timing studies agreed on the significant advantage of VSP. It is important to understand the area that accounts for the planning advantage in VSP compared to TSP. Van et al. 13 found statistically significant differences in soft tissue planning, but the difference was not significant in complex tissue planning.

Surgery Time
The surgery time should ideally be short to limit the loss of blood, use of anaesthesia, and risk of errors. In (Figure 4) the forest plot showed that the difference between the groups that went through surgery after VSP and those in the TSP condition was insignificant. Schwartz 2 found that the surgical times were the same at 250 minutes for both VSP and TSP. The researcher, however, noticed that residents were responsible for most surgeries. This finding, along with Wrzosek et al. (2016) 12 , provides an important caveat when evaluating the difference between VSP and TSP time. The potential impact of inexperience can account for the insignificant differences in VSP and TSP time. The findings are different from the recent study by Chen et al that reported a significant difference in surgery time with TSP taking longer than VSP 1 . The difference in the included studies contributes to this difference is significance. Including studies that looked at both maxillary and bimaxillary surgeries could explain the difference.
Hanafy et al. 14 demonstrated that VSP was particularly advantageous over TSP especially for trainees and junior surgeons. This is because it can precisely detect bone/root morphology compared to TSP. The finding illustrates that both TSP and VSP are beneficial in offering the guidance necessary to complete the surgery. The finding also emphasizes the need for examining VSP as an approach that might save time by reducing the errors during surgery 16,17 . Such advantage is likely to reduce for experienced practitioners who are accustomed to TSP and less prone to making errors irrespective of the method used.

Total Time
Two studies 2,9 looked at the total time from planning through the surgery to the postoperative stages. While the studies could not be statistically analysed due to missing values, narrative analysis can be performed on the researchers' findings. In their study, Steinhuber 9 found that the total working time reduction was significant when moving from TSP to VSP in double-jaw surgeries. The reduced time difference in the double-jaw surgeries compared to the single jaw surgery is evidence of the complexity of bimaxillary surgeries, thus the need for efficient processes. Significant time savings in VSP emerged from the surgeon and technician laboratory work in the total time.

Limitations
The limited number of RCTs is a legitimate concern when examining the comparative advantage of VSP over TSP. With most studies being retrospective studies, there is a risk that researcher selection bias might affect the quality of the results. This study prioritized RCTs but confronted the reality of limited RCT design in the current literature focused on time. The studies used in this study had small sample sizes. Larger samples in future studies will also be necessary for performing a comparative analysis of VSP and TSP.

Conclusion
VSP significantly performs better than TSP in reducing the bimaxillary osteotomy planning time, but the timing difference is not significant during surgery. While the statistical analysis shows that bimaxillary osteotomy planning time can be improved through improved knowledge among healthcare practitioners on how to use the relatively new technologies.