After sagittal split osteotomy, the mandibular distal and proximal fragments do not always align themselves passively to one another, resulting in bony interferences and subsequent anomalous settlement of the condyles. Predicting these interferences could be an important ancillary procedure for avoiding intra- and postoperative surgical complications, rendering orthognathic surgery more effective and safer. This study evaluated the relevance of virtual surgical planning in assessing the displacement of the proximal segments after virtual distal segment repositioning, for predicting bony interferences between the segments and thus avoiding related intra- and postoperative surgical complications. The presence of interferences between the distal and proximal segments was compared between virtually predicted (computer-assisted simulation surgery, Dolphin software) and real cases in 100 consecutive patients diagnosed with dentofacial deformities who underwent orthognathic surgery with mandibular repositioning (using a short lingual osteotomy (SLO)). The results indicated that clockwise rotation of the mandible was the mandibular movement most prone to segment interference. Furthermore, virtual planning was sensitive (100%) but had low specificity (51.6%) in predicting proximal and distal segment interferences. This low specificity was due to the software-based automated design of the mandibular osteotomy, where the length of the distal segment was longer than the real SLO, and the mandibular ramus sagittal split was located just behind Spix’s spine. Thus, more precise simulated osteotomies are needed to further validate the accuracy of virtual planning for this purpose.
The bilateral sagittal split ramus osteotomy (BSSRO) is a technique used widely in orthognathic surgery for the correction of mandibular deformities. It has advantages compared to other mandibular osteotomies due to the wide overlapping cutting surface, which enables a variable range of three-dimensional (3D) movements of the distal segment (DS), ensuring broad enough surface contact for osteosynthesis and bone healing .
Unfortunately, the distal and proximal fragments do not always align themselves passively to one another, resulting in interferences between them: (1) forward or backward movement of the DS of the mandible may lead to external rotation of the proximal segments (PS) in the pitch and roll axes; (2) lateral shifting of the midline causes one DS to rotate laterally while the other is rotated medially in the roll and yaw axes; (3) occlusal cant modifications can produce a gap at the lower margin on one side and at the upper margin on the other, in the vertical and transverse axes; and (4) occlusal plane changes may require pitch adaptation of the PS .
Thus, inadequate accommodation of the PS may result in anomalous positioning of the condyles. If left uncorrected, this could cause temporomandibular joint (TMJ) disorders, impaired bone healing, inferior alveolar nerve (IAN) overextension, and a tendency to relapse, as well as unaesthetic outcomes .
Nowadays, 3D virtual surgical planning (VSP) has become an increasingly used tool, allowing more precise orthognathic surgery and final outcomes that come as close as possible to the intended outcomes. An improved virtual anatomical study is possible using this technique, with better symmetry axes and the anticipation of surgical complications such as the osteotomized bone segment interferences mentioned above . Predicting these interferences could be an important ancillary procedure for avoiding intra- and postoperative surgical complications, rendering orthognathic surgery more effective and safer . Few studies have addressed this issue in the literature to date. The aim of the present study was therefore to evaluate the relevance of 3D VSP in assessing displacement of the PS after virtual BSSRO, predicting bony interferences between the segments and thus avoiding related intraoperative and postoperative surgical complications.
Materials and methods
This prospective study included 100 consecutive patients diagnosed with a dentofacial deformity and subjected to orthognathic surgery with BSSRO at Teknon Medical Centre Barcelona between February 2017 and January 2018. All surgeries were virtually planned and performed by the same surgeon (FHA).
The patients were selected on the basis of the following inclusion criteria: age >18 years, dentofacial deformity in need of mandibular correction, and signed informed consent. Patients who underwent an isolated maxillary Le Fort I osteotomy were excluded, as were those presenting any craniofacial syndrome or pathological background that could compromise bone healing, and patients failing to sign the informed consent.
The study was approved by the Ethics Committee of Teknon Medical Centre (Barcelona, Spain; Ref. LO-OS) and was conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
All patients followed the standard workflow for orthognathic surgery planning and surgical splint fabrication of the department, as described elsewhere . The protocol is based on a single cone beam computed tomography (CBCT) scan (iCAT; Imaging Sciences International, Hatfield, PA, USA) of the head of the patient, with surface intraoral scanning of the dental arches using the Lava Scan ST scanner (3 M ESPE, Ann Arbor, MI, USA) for subsequent fusion of the two datasets. In addition, facial photographic records were obtained to complete the preoperative study protocol.
Virtual planning work-up
Computer-assisted simulation surgery was conducted using specific software (Dolphin 3D Orthognathic Surgery Planning Software version 11.8; Dolphin Imaging & Management Solutions, Chatsworth, U.S.A.) . The BSSRO design was generated according to the standardized Dolphin protocol, where the clinician only needs to mark the following landmarks ( Fig. 1 ): (1) in the lingual view, two landmarks are placed parallel to the occlusal plane, located slightly above Spix’s spine, on the front and back; (2) in the top view, four landmarks are used to trace the osteotomy line between the most medial landmark in the lingual view and a landmark between the first and the second molars just below the molar gingival line; (3) in the buccal view, five landmarks are placed slightly above the caudad edge of the body of the mandible: the most medial point is placed following a perpendicular line across the occlusal plane between the first and second molars, and the most distal point is located following an imaginary line parallel to the mandibular ramus through the posterior-most landmark of the lingual view.
Once the mandibular and maxillary osteotomies had been designed ( Fig. 2 ), surgical repositioning of the maxillomandibular complex was virtually simulated following the upper incisor to soft tissue plane (UI-STP) protocol ( Fig. 3 ), validated previously and described in detail elsewhere . The new mandibular DS position in turn determined 3D settlement of the mandibular PS, in order to avoid mismatches between them ( Fig. 4 ).
Finally, any observed interferences between the distal and proximal segments were noted in the surgical plan for further consideration as possible intraoperative interferences requiring an additional surgical approach, as described later in this article ( Fig. 2 ).
The patients were operated upon under general anaesthesia. In all cases, the mandible was operated on first, and the sagittal split was performed using the Hunsuck–Dal Pont–Obwegeser technique or so-called short lingual osteotomy (SLO) . Next, interferences between the distal and proximal bony segments precluding gentle settlement of the PS were checked subjectively by the main surgeon ( Fig. 5 ). In the event of any such interference, a greenstick osteotomy of the lingual cortical layer of the DS was performed, without stripping the soft tissues on the lingual surface in order to avoid IAN damage, and the osteotomized bone fragment was left in place ( Figs 2 , 6 , and 7 ). This technique, referred to as a lingual osteotomy (LO), was first described by Ellis in 2007 – the only difference being that we performed the osteotomy with a piezoelectric saw device (Implant Center 2; Satelec-Acteon Group, Tuttlingen, Germany).
This technique should enable smooth transition between the segments and proper 3D repositioning of the DS with passive accommodation of the condyle at the glenoid fossa, while increasing the contact surface between the two fragments. As a routine measure, proper seating of the condyles into the uppermost-anterior part of the fossa was ensured with a bidirectional manoeuvre. Then, rigid internal fixation with a hybrid technique (a miniplate fixed with four monocortical screws and a retromolar bicortical screw) was performed , followed by removal of the intermaxillary fixation. Before removing the intermediate splint, proper condylar positioning and intermediate occlusion were checked again. Lastly, if necessary, the upper maxilla was repositioned according to the final splint.
All patients wore a closed-circuit cold mask (17 °C) during hospital admission and were discharged 24 hours after surgery. Standard antibiotic and anti-inflammatory medications for orthognathic surgery were prescribed. Functional training with light guiding elastics was followed for 1 month, with the observation of a soft diet for the same period of time.
In order to evaluate the relevance of 3D VSP in assessing 3D displacements of the PS after mandibular distal fragment repositioning in orthognathic surgery, the following 3D surgical movements were registered in each virtually planned case: (1) B-point movements (which are to be DS movements) in all three axes: sagittal, vertical, and transverse. (2) Mandibular occlusal plane changes (which are to be DS movements) on the right and left sides. (3) PS angular movements while maintaining the condyles in place (right and left sides) for adaptation to the DS movement in all three axes: pitch, roll, and yaw ( Fig. 4 ).
Then, interferences between the distal and proximal segments were compared between the virtually predicted cases and the real cases (those that required a LO) in order to examine the true capacity of 3D VSP to predict bony interferences.
Moreover, the LO technique described above was subjectively tested by the main surgeon (FHA) in terms of fragment interference and subsequent 3D gentle settlement of the PS after performing the osteotomy.
Finally, in assessing the safety of surgery, the following conditions were considered potential complications of the procedure: IAN or lingual nerve damage, bone sequestration, intra- or postoperative malocclusion secondary to condylar sag , and TMJ symptoms at 1 year of follow-up.
A descriptive analysis was made of the study variables, with calculation of the mean, standard deviation, minimum and maximum values, and median for continuous variables. Absolute and relative frequencies (percentages) were used for qualitative variables.
A one-sample t -test was used to determine whether the change in a certain cephalometric parameter was relevant, and the kappa concordance index was used to assess agreement between planning and execution of the osteotomy procedure. In addition, simple binary logistic regression models were used to evaluate the impact of cephalometric changes upon the probability of performing an osteotomy. The level of significance was set at 5% (α = 0.05).
One hundred and twenty-seven patients were scheduled for orthognathic surgery during the study period; a total 100 were enrolled based on the inclusion and exclusion criteria. Five patients were excluded because of insufficient data, 15 because they had undergone isolated Le Fort I maxillary surgery, six because they were under-aged, and one patient because surgery was in the context of a craniofacial syndrome.
The study sample comprised 61 women (61%) and 39 men (39%), with a mean age of 27.6 years (range 18–56 years).
Virtual surgical planning
The magnitudes of the planned surgical mandibular movements are reported in Table 1 . Although movements of the DS were performed in all planes, only occlusal plane in pitch changes on both sides and sagittal B-point movements proved statistically significant ( P < 0.001) when comparing the pre- and postoperative VSP ( Figs 8 and 9 ).