Alloplastic reconstruction of the mandible after subtotal mandibulectomy for medication‐related osteonecrosis of the jaw: An update of the method

Reconstruction of continuity defects following osteonecrosis in multimorbid patients is challenging. In all cases of the predescribed palliative treatment method for alloplastic mandible reconstruction, plate fractures were detected in follow‐up. We hypothesized that a modification could avoid these fractures, leading to stable long‐term results.


| INTRODUCTION
Medication-related osteonecrosis of the jaw (MRONJ) is an infrequent but significant complication of antiresorptive medication. 1,2These drugs are most commonly used in the management of osteoporosis and cancers with bony involvement (myeloma or bone metastasis) and are therefore most commonly used in an elderly population, with multiple co-morbidities, with an event rate of up to 15%. 3 The development of MRONJ can lead to a significant reduction in quality of life due to pain, restrictions in food intake, oro-cutaneous fistula, and facial deformity. 4Stage 3 MRONJ, as classified by the American Association of Oral & Maxillofacial Surgeons review 2 is characterized by intraoral exposed bone with evidence of infection which extends beyond the alveolus, or involves a pathological fracture or an extra-oral fistula.
Such patients may require large segmental resections 5 which would normally be considered for reconstruction with a bone or composite microvascular free flap.Such extensive surgery in these often frail patients with multiple co-morbidities, and variable prognosis, may not be indicated or appropriate. 6Alloplastic reconstruction with reconstruction plate alone often leads to plate exposure or fracture, and does not permit dental rehabilitation. 7,8Because of this, we introduced in 2018 a technique for alloplastic reconstruction of the mandible in patients not suitable or willing to undergo osseous-free flap reconstruction, which aimed to preserve the form and function of the mandible. 9Although satisfactory results could be achieved in most patients during the initial follow-up period, plate fractures at the interface of implant to bone were observed over the longer term (Figure 1).
Despite this, most of the implants could remain in the situs but in some cases, revision surgery had to be performed due to mobility and resulting implant exposure.
To improve the long-term stability of the method and reduce the rate of secondary operations we modified the technique used for fixation of the Palacos mandible implant to use a custom-made 3.0 mm laser sintered CAD-CAM plate (Gebrüder Martin GmbH & Co. KG, KLS Martin Group ® , Tuttlingen).We present here the modified surgical technique and compare the outcomes to our earlier design.

| Study design and data collection
A retrospective comparative cohort observational study was performed in the Clinic of Oral and Maxillofacial Surgery of the University Clinic "Carl Gustav Carus" Dresden.
The study examined two cohorts of patients who suffered from Stage 3 MRONJ, and required a segmental resection of the mandible due to pain, infection, and/or pathological fracture of the mandible.
The first cohort was treated according to the method described by Bräuer et al 9 The second cohort was treated with an update of the method.Both methods are described below.All patients with a minimum follow-up of 6 months were included.
Demographic data including underlying pathology, ASA-score, classification of the resulting defect, 10 length of defect, and types of wound closure were recorded.
Parameters observed post-operatively at 2 weeks, 1 month, and subsequently semi-annually were failure of osteosyntheses, including plate fractures, exposure of the implant extra-or intraorally, oral food intake, and the possibility for dental rehabilitation.Postoperative radiographs served as a control of osteosynthesis 10 days after surgery and semi-annually.

| Statistical analysis
Data processing and statistical analysis were carried out using IBM ® SPSS ® Statistics, version 27.Data were expressed as mean ± standard deviation.For statistical investigations, descriptive statistics and independent Student t-tests were performed.For correlation analysis, the rank correlation coefficient according to Spearman was calculated.For survival time analyses, the Kaplan-Meier estimator with log-rank test was used.A p-value < 0.05 was considered statistically significant.

| Operative technique-First cohort
An individual 3D model of the mandible (Phacon GmbH, Leipzig) based on the preoperative CT scan dataset of the patient was used.The resection borders were defined and marked on the 3D model.In dental laboratory, a mold was created from sterilizable, light-curing modeling resin (FREEFORM ® , DETAX GmbH & Co.KG, Ettlingen) that encases the defect in the original shape of the mandible.Two 2.0 mm titanium mini plates (Gebrüder Martin GmbH & Co. KG, KLS Martin Group ® , Tuttlingen) were precontoured on the model and the mold was adopted.Via submandibular approach, subtotal mandibulectomy was performed under general anesthesia.The inner part of the mold served as a retainer to preserve the dimension of the mandible.After surgical resection, the two precontoured 2.0 mm Mini Plates were fixed with miniscrews to the mandibular stumps.The outer part of the mold was temporarily attached, and the defect was filled in with PALACOS ® LV + G Bone Cement.Thus, the mini plates in the defect area were completely covered by bone cement and the mandible was reconstructed in the former dimensions and contour. 9

| Operative technique-Second cohort
The preparation of the mold for the cement and the surgical preparation and resection were undertaken as described above for the first cohort.Utilizing the KLS-Martin IPS-Gate ® protocol the surgical resection margins were defined and a custom-made 3.0 mm laser sintered CAD-CAM plate (Gebrüder Martin GmbH & Co. KG, KLS Martin Group ® , Tuttlingen) was ordered.
After subtotal mandibulectomy via a submandibular approach, the CAD-CAM plate is fixed in its pre-planned position using 2.7 mm bicortical locking screws (Figure 2).The defect is filled in with polymethyl methacrylate (Palacos ® LV + G bone cement, Heraeus Medical GmbH, Wertheim, Germany; Figure 3), the mold is detached and the implant surface is smoothed.It is imperative to pay special attention to accurate, tensionfree re-approximation of the floor of the mouth to the implant to avoid dead space, thus preventing intraoral wound dehiscence and enabling adequate function, that is, swallowing and speaking.Postoperative radiographs are taken within 10 days (Figure 4).In 2016 and 2017, seven patients (6 female, 1 male) were included in the first cohort.The demographic details, are presented in Table 1.The mean age of the patients at surgery was 76 ± 14 years (median 80 years).The incisionsuture time was 296 ± 42 min (median 286 min).The inpatient stay was 24 ± 12 days.On average, clinical follow-up was 24 ± 7 months.All patients showed regular oral food intake in the postoperative course.There were no signs of material hypersensitivity.In two cases we observed exposure of the bone cement implant within the immediate postoperative course.This was treated with local measures and did not lead to implant infection or removal of the implant.In two other cases, new bone necrosis with superinfection and subsequent exposed implant material happened more than 2 years postoperatively, which led to removal of the implant in one case.The principal problem noted, however, was plate fracture which occurred on average 17 ± 8 months postoperatively in all patients of this cohort.This led to removal of the implant with subsequent re-osteosynthesis in two cases.One implant was repaired after fracture due to blunt force trauma after a fall.All fracture lines were exclusively located at the transition from implant to bone.

| Second cohort (reconstruction using CAD-CAM Plate and Palacos ® )
The second cohort was recruited during the period from 2017 to 2020 and comprised a total of 10 patients (8 female, 2 male) with Stage 3 MRONJ of the mandible with multiple co-morbidities and not suitable or willing to undergo microvascular bony reconstruction.Three patients were excluded due to lack of follow-up.The demographic details are presented in Table 2.In one case, a microvascular fasciocutaneous flap graft had to be used to cover the large mucosal defect.At the time of surgery, the patients were on average 77 ± 7 years old (median 78 years).The mean incision-suture time was 257 ± 108 min (median 233 min).The average inpatient stay was 12 ± 2 days.Clinical follow-up was on average 32 ± 10 months (median 36 months).All patients were able to eat normally.No signs of material hypersensitivity were detected.In six cases (86%), we observed uneventful wound healing.Three patients developed a new osteonecrosis of the mandible with superinfection and subsequent intra-or extra-oral exposed bone cement implant over the time.In one case, this happened 1 year after initial surgery, and in the two other cases, more than 2 years later.No implant or plate fractures occurred in the follow-up.No implant had to be removed.

| Comparison of first and second cohort
There were significantly less plate fractures in the second cohort (p < 0.001).There was no significant difference in wound dehiscence and implant exposure due to new bone necrosis.The time to plate fracture is demonstrated in Figure 5.The inpatient stay in the second cohort was significantly shorter (p = 0.032).

| Analysis of all alloplastic reconstructions
In the entire recruitment period from 2016 to 2020, a total of 17 alloplastic mandibular reconstructions were performed.The mean patient age was 77 ± 10 years (median 80 years).The duration of surgery was mean of 286 ± 112 min, median of 251 min.Looking at the respective incision-suture times of all alloplastic reconstructions (n = 17) during the course of the study, a strong negative correlation can be seen (ρ = À0.713,p = 0.01).

| Prosthetic rehabilitation
In one patient in the second cohort, prosthetic rehabilitation was achieved with mucosa-supported complete dentures.In this particular case of medication-related osteonecrosis with a pathological fracture of the mandible, alloplastic mandible reconstruction of an L-defect 10 was performed solely via an extraoral approach.There was no intraoral mucosal lesion.Radiological evidence of significant peri-implant bone formation was detected after 12 months (Figure 6).
In cooperation with the Department of Prosthodontics of the University Hospital Dresden, the patient received new complete upper and lower dentures 22 months postoperatively (Figures 7 and 8).After the insertion of the prostheses, there was an additional follow-up period of 13 months.There were no complications, in particular, no mucosal lesions were detected.Regular follow-ups ended 3 years after initial surgery due to a severe Covid-19 infection of the patient.

| DISCUSSION
MRONJ can lead to significant patient morbidity, with defects of the oral cavity allowing repeated ingress of organisms and infection. 1,2,11ome patients can be managed with local measures, which may include local flaps, to debride and cover intraoral defect. 12In exceptional cases with large intraoral defects where the patient has few comorbidities, we also consider radial forearm flaps suitable for safe defect closure.
In advanced disease with pathological fractures and/or extra-oral fistulae aggressive surgical resection and reconstruction can be necessary.Reconstruction of mandibular continuity defects is a complex and challenging field of oral and maxillofacial surgery.Regarding the reconstruction of mandibular continuity defects in the treatment of Stage 3 MRONJ, no clear recommendations exist.Both alloplastic and microvascular bony reconstruction are mentioned as palliative long-term therapy. 2 A crucial aspect in therapy planning is the patient's quality of life, which can be considerably reduced in occurrence of MRONJ, comparable to impairments in the context of tumor therapy. 4icrosurgical bony reconstruction is considered the gold standard, most commonly with a microvascular fibula, iliac crest, or scapular graft, 13 which provides the best functional and cosmetic results.
Age, comorbidities including peripheral vascular compromise, functional status, and the presence of osseous pathology at donor sites, means that in many cases of MRONJ, free flap reconstruction cannot be considered. 14][16] In a review, the average age of 59 patients receiving microvascular reconstruction in the therapy of MRONJ was 64 years. 17The patients with multiple co-morbidities that underwent alloplastic reconstruction in this study had a significantly higher age (mean 77 ± 10 years, median 80 years).
Another problem in using microvascular bony reconstruction in MRONJ patients is that the bone at the contact between native and flap bone may have problems healing due to effects of the anti-resorptive medication on residual bone.The relatively high recurrence rate of bone necrosis with subsequent implant exposure in this study (43%, mean follow-up 32 months) is a known problem in treatment of MRONJ.Klingelhöffer et al. 18 had a recurrence rate of 72.3% 6 months after surgery.Reich et al. 19 observed an absence of MRONJ lesions in only 43% of their cases after 23.5 months of mean follow-up.
One alternative is to reconstruct the defect with only a plate.Historically plate fractures were common in the long term, and cosmetic and functional outcomes are poor. 20CAD-CAM plates, which have been increasingly used in recent years, provide higher fatigue strength, but the problem of screw loosening and breakage remains. 21lacement of an antibiotic-infused bone cement spacer to reconstruct the mandibular defect provides the best available compromise in the form and function of the mandible in this cohort of patients, in our opinion.It is noteworthy that the spacer does not become infected in the case of minor intraoral dehiscence and does not necessarily have to be replaced.Superinfected surrounding soft tissue, a known problem in MRONJ, heals properly under systemic perioperative antibiotics and the antibioticinfused bone cement.The idea of an alloplastic spacer for reconstruction of the mandible is not a new one.3][24] A disadvantage of the methods discussed in the literature appears to be the poor fit of the spacers, leading to inadequately reproduced original dimensions of the jaw.It is common when resecting MRONJ-affected bone to have to adapt from the presurgical plan, and therefore with preoperatively fabricated spacers, inaccurate fit of the spacer is a significant problem. 25With our technique for intra-operative creation of the bone spacer, 9 any change in the resection margins from the preoperative planning can be adequately addressed by oversizing the temporary mold and filling the resection defect with lowviscosity bone cement.Using a custom-shaped low viscosity bone cement, the original, three-dimensional shape of the jaw is preserved which counteracts tissue shrinkage and this reduction in soft tissue contraction has a positive effect on facial aesthetics. 22As a result, the Palacos ® implant is sized as large as the original resected bone.However, intraoperative smoothing of the implant made the alveolar process rounder and slightly smaller to avoid sharp edges and thus intraoral implant exposure.We also postulate that extraoral or intraoral exposure of the implant material, normally a common problem, 7,26 can be minimized due to the reduced adaptation required of the local soft tissues.The prevalence of rare allergic reactions to bone cement is reported in the literature to vary from 0.6% to 1.6%, 27 there was none in this study.
Stability of the implant is however required to minimize local tissue trauma.We believe that with an accurately fitting spacer, the force transmission is not exclusively through the plate and the screws, but that the bone cement-native bone interface can contribute to the force transmission and stability of the reconstruction and prevent screw loosening and fractures.
Stabilization of the implant with 2.0 mm Mini Plates has proved ineffective in the long term due to plate fractures.Despite this complication occurring in all patients eventually, 3 of the 7 patients required no intervention and a further two only required local measures.Alternatively, instead of the two Mini Plates described, a conventional reconstruction plate could be used and pre-bent on the patient-specific 3D-model.Compared to patientspecific CAD-CAM plates, the use of a conventional reconstruction plate is much more cost-effective and faster to implement.In addition, it is possible to bend the plate intraoperatively and thus react to possible intraoperative changes in the resection length.In combination with the bone cement, sufficient stability should be achieved.To the author's knowledge, there have been no studies on the fracture rates of Palacos ® encased conventional reconstruction plates.However, in the long term, plate fractures are a known problem with conventional hand-bent reconstruction plates. 28Almansoori et al. 29 found a fracture rate of 14% after 20 months of follow-up.We presented here a modified technique with a custommade 3.0 mm laser sintered CAD-CAM plate, which is not exposed to any bending stress, does not suffer from cold deformation, and is therefore significantly less prone to fracture than conventional reconstruction plates. 30sing such plates, good stability has been achieved and no plate fractures, screw fractures, or screw loosening occurred.Follow-up in this second cohort was longer than in the first cohort using 2.0 mm Mini Plates.However, a negative aspect of using these patient-specific reconstruction plates is the high price and the relatively long planning time depending on a medical device company.In addition, intraoperative bending and adjustment of the plate to a modified resection is not possible.
With this new technique, there was a low complication rate with one wound healing disorder with exposure of the implant (14%).In three cases (43%) recurrence of MRONJ led to implant exposure after a mean of 23 months, but no implants required removal.This is lower than described complications and loss rates in the literature for plate-only reconstructions.A meta-analysis found average complication rates of 40.1% and loss rates of 30.8%. 31 Ultimately, the aim of the palliative situations addressed must be to maintain or improve the patient's quality of life.The extent of the surgical intervention must be chosen in such a way that the quality of life gained through the resection of the jaw necrosis is not reduced again by the consequences of the reconstruction.Another parameter in this context is the duration of surgery.The total operative time for the alloplastic reconstructions presented was significantly lower compared to microvascular osseous reconstructions described in the literature.Ren et al. 32 found total operative times of 6-7 h for reconstructions with fibula-free flaps, Jacek et al. 33 found average times of 6.5-8.5 h for microvascular bony mandible reconstruction.On average, the operations described in this study were 2-4 h shorter, which represents a significantly lower burden for multimorbid patients.However, in situ fabrication of the implant, smoothing of the implant surface, and subtle readaptation of the surrounding tissue require a certain amount of time, so that on average a surgical time of four and a half hours was required.The alloplastic reconstructions performed showed a decrease in incision-suture time during the course of the study.This correlation can be interpreted as a learning effect, with shortened operative time coming with experience of the technique.
The prosthetic rehabilitation with mucosa-supported dentures is generally not recommended in patients with conventional alloplastic reconstruction of the mandible.The reason for this is lack of denture bearing area and vestibulum in the resected part of the mandible and the absence of bony tissue that could be used for implantsupported dentures. 34Nevertheless, in the presented study, prosthetic rehabilitation was successfully achieved in one patient.The described surgical method allows partial remodeling of the alveolar process in the resected area.As a result, the vestibulum is preserved, thus a sufficient denture bearing area can be achieved.There is however a high risk that prosthesis pressure points will lead to exposure of the implant, which can lead to a total loss of the implant.
The presented method might be considered as an option for mandible reconstruction after segmental mandibulectomy due to MRONJ in elderly and multi-morbid patients in a palliative situation.The modification of the original method resulted in a significantly improved long-term stability with a low complication rate.Since the original shape of the mandible can be reconstructed, soft tissue shrinkage is minimized and extraoral or intraoral exposure of the implant can be avoided.There is no donor site morbidity using this method of palliative surgery.Operation time is significantly shorter than in microvascular bony reconstructions.The prosthetic rehabilitation with mucosa-supported dentures may be possible in selected individual cases.

ACKNOWLEDGMENT
Open Access funding enabled and organized by Projekt DEAL.

F I G U R E 2
Defect following left segmental mandibulectomy.CAD-CAM plate is fixed and the inner part of the mold is in place.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 3 The defect is filled in with polymethyl methacrylate (Palacos ® LV + G bone cement).Both parts of the mold are removed.Implant surface is shown prior to smoothing.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 1 Plate fractures in panoramic radiograph, 13 months after right segmental mandibulectomy and alloplastic reconstruction with two 2.0 mm Mini Plates and Palacos ® .

F
I G U R E 4 Postoperative panoramic radiograph.T A B L E 1 Summary of cases using 2.0 mm Mini Plates and Palacos ® LV + G bone cement.Patient (age/sex) according to Jewer et al. 10 : C = central segment, L = lateral segment, combinations possible.T A B L E 2 Summary of cases using CAD-CAM Plate and Palacos ® LV + G bone cement.according to Jewer et al. 10 : C = central segment, L = lateral segment, combinations possible.

F
I G U R E 5 Kaplan-Meier analysis of plate fractures in both cohorts (log-rank χ 2 (1) = 13.146,p < 0.001).[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 6 Panoramic radiograph, 12 months postoperatively with periimplant bone formation (arrow).[Color figure can be viewed at wileyonlinelibrary.com]

F I G U R E 7
View enface (A) and laterally (B) before prosthetic restoration.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 8 View enface (A) and laterally (B) after insertion of the dentures.[Color figure can be viewed at wileyonlinelibrary.com]