A small number of residual teeth after the mandibular resection of oral cancer is associated with titanium reconstruction plate exposure

Abstract Objective Reconstruction plates are used to treat patients with a segmental mandibular defect after oral cancer surgery. Reconstruction plate failure analysis has rarely focused on occlusion, which conducts a mechanical force to the mandible and the plate. To determine the prognostic factors, we retrospectively evaluated patients who underwent reconstruction of a mandibular segmental defect with a reconstruction plate and assessed the number of residual paired teeth. Material and Methods From among 390 patients with oral cancer who visited University of Tsukuba Hospital (Tsukuba, Japan) between 2007 and 2017, we selected and analyzed the data of 37 patients who underwent segmental resection of the mandible and reconstruction with reconstruction plates. Prognostic factors evaluated were patient age, sex, TNM classification, plate manufacturer, treatment with radiotherapy or chemotherapy, whether the patient had diabetes or smoked, and whether the patient had a small number of residual paired teeth, plate length, and use of a fibular‐free flap. Among these 37 patients, eight reconstruction plates had intraoral or extraoral exposure and were removed in 5 years. Results Kaplan–Meier and log‐rank analyses revealed that the prognosis for the 5‐year plate exposure‐free rate was significantly poorer for patients with a small number of residual teeth than for patients with no teeth or those with a large number of residual teeth (.01). Univariate Cox regression analysis revealed that a small number of residual teeth was a significant prognostic factor in the loss of a reconstruction plate (hazard ratio: 5.63; 95% confidence interval [1.10, 25.85]; .04). Conclusions A small number of residual teeth after the segmental resection of oral cancer is significantly involved in reconstruction plate survival and may be important in predicting reconstruction plate prognosis.

KEYWORDS mandibular reconstruction, neoplasm, reconstruction plate 1 | BACKGROUND Treating advanced oromandibular cancer often requires resecting mandibular and alveolar bones and soft tissues such as the oral mucosa, muscles, or external skin of the mandible. Mandibular reconstruction with a plate, with or without a vascularized free flap, is a common treatment to deliver acceptable function and cosmesis to maintain the postoperative quality of life. The role of the reconstruction plate is very important because an unreconstructed mandibular defect may cause deformation of the remaining mandible in association with several problems for the patients such as insufficient sealing of the lips and feeding (van der Rijt, Noorlag, Koole, Abbink, & Rosenberg, 2015). However, many postsurgery complications occur with reconstruction plates, and the rates of reported complications with reconstruction plates range from 28% to 39% and include loosening of the osteosynthesis screws, plate fracture, and intraoral or extraoral exposure or infection (Kammerer, Klein, Moergel, Gemmel, & Draenert, 2014;Maurer, Eckert, Kriwalsky, & Schubert, 2010). A recent systematic review and meta-analysis (Sadr-Eshkevari et al., 2013) demonstrated that the failure rate was 30.8% at the 32-month follow-up and discussed the risk factors of reconstruction plate failure from various viewpoints. Many reports discuss the influencing factors for reconstruction plate survival and list many factors such as physical factors such as whether an individual has diabetes mellitus (van der Rijt et al., 2015), smoking habit (Maurer et al., 2010;van der Rijt et al., 2015), whether an individual has had a blood transfusion (Fanzio et al., 2015), and surgical infection site (Wood, Shinn, Amin, Rohde, & Sinard, 2018), hardware factors such as plate length and location of a mandibular defect (Arden, Rachel, Marks, & Dang, 1999;Ettl et al., 2010;Mariani, Kowalski, & Magrin, 2006;Okura, Isomura, Iida, & Kogo, 2005;Poli, Ferrari, Bianchi, & Sesenna, 2003;Prasad et al., 2018;Shibahara, Noma, Furuya, & Takaki, 2002), and treatment factors such as radiation therapy and chemotherapy (Okura et al., 2005;Ryu et al., 1995;Shibahara et al., 2002;Wang, Zhang, & Mendenhall, 2005). Most reports have focused on the shape of the mandibular bone defect after resection and hardware-related complications; however, few reports have analyzed risk factors for plate failure from the viewpoint of dental occlusion (i.e., the mechanical force exerted on the reconstructed mandible and plate).
The maxillomandibular occlusal state differs depending on the relationship between the upper and lower teeth. Disharmony of chewing motion after surgery is caused by the resection of the mandibular bone and the masticatory and facial muscles. The load of the stress then alters depending on whether a stable stop position of occlusion does or does not exist. In this study, we investigated the predictive factors of reconstruction plate exposure by using the number of paired teeth as a variable to define prognostic factors for reconstruction plate loss and to improve future treatment planning. Among them, we selected 37 patients who underwent segmental resection of mandibular and reconstruction with reconstruction plates and analyzed the data. The patients included 21 men and 16 women with a mean age (standard deviation) of 65.8 (10.4) years. Each case of oral cancer was staged, using the International Union against Cancer system (Sobin, Wittekind, & Gospodorowicz, 2009 smoked; and whether the patient had received chemotherapy or radiotherapy. Residual paired teeth meant teeth that can stably maintain a vertical dimension of occlusion. Therefore, if the mobility of the opposite tooth was substantial, which included mobility occurring with severe periodontitis, we did not include it as a residual paired tooth. By using Eichner's index, which is a system for classification of partial edentulous arches based on occlusal contact between in the premolar and molar regions, ll dental arches that had residual paired teeth were classified as Eichner's index B3 (Eichner, 1995), and dental arches that had no residual paired teeth were classified as Eichner's index C1-C3 (Eichner, 1995). The clinical features of the patients are presented in Table 1. The patients were treated by surgical excision alone (23 patients) or by surgical excision plus chemoradiotherapy (14 patients).
The 14 patients were treated by linear accelerator radiosurgery (mean dose, 52.4 ± 9.4 Gy; range, 37.3-60 Gy). The study protocol was reviewed and approved by the Research Ethics Committee of the University of Tsukuba (Tsukuba, Japan; approval no., H29-258).

| STATISTICAL ANALYSIS
For the univariate analysis, we used Kaplan-Meier analysis, evaluated by the log-rank test, and the Cox proportional hazard regression model. Statistical analyses were conducted using the software package JMP 12.0.1 for Mac (SAS Institute Inc., Cary, NC, USA).

| Univariate analysis of the 5-year exposure-free rate based on Kaplan-Meier analysis and the log-rank test
Kaplan-Meier analysis was used to estimate each factor's 5-year exposure-free rate (i.e., 5-year survival rate of the reconstruction plate). During this term, reconstruction plate exposure occurred in eight (21.6%) of 37 patients. The overall 5-year exposure-free rate was 0.71, as estimated using Kaplan-Meier analysis. Table 2 presents the 5-year exposure-free rate and log-rank test results based on Kaplan-Meier analysis. Figures 1 and S1-S13 present the survival curves for each factor. The 5-year exposure-free rate was significantly different between the null tooth group (i.e., zero teeth), small number of teeth group (i.e., 1-5 residual paired teeth), and large number of teeth group (i.e., six residual paired teeth or more); the rate was 0.81, 0.00, and 0.79 for the null group, small number of teeth group, and large number of teeth group, respectively (p = .04; Figure S5).
We then divided the tooth groups into two groups: (Arden et al., 1999) small number of teeth group and (Coletti, Ord, & Liu, 2009) large number of teeth group and null teeth group combined (i.e., other group). We divided the type of flap into two groups: (Arden et al., 1999) the fibular-free flap group and (Coletti et al., 2009)   These findings indicated that a small number of teeth may be a poor predictive factor for reconstruction plate survival.
Markwardt did not conduct a long-term observation; therefore, they may not have been able to determine a significant difference.
However, some reports (Hoefert & Taier, 2018;Park, Lee, & Noh, 2018;Yi et al., 1999) present analysis results from a biomechanical viewpoint and demonstrate that reconstruction plates and screws are often subjected to excessive stress produced by functional loading, moment, and shear forces. From these reports, the reconstruction plate bridging mandibular bone can be estimated to burden excessive load by the shape of bone defect and the loading points (Hoefert & Taier, 2018;Park et al., 2018;Yi et al., 1999). In particular, Park et al., 2018 divided patients into three groups-unilateral molar clenching, group function clenching, and incisal clenching-and measured the von Mises stress, a value used to determine if a given material will yield or fracture, occurring on the reconstruction plate. The maximum von Mises stress on the reconstruction plate of patients with unilateral molar clenching was larger than in that of group function clenching (Park et al., 2018).
Unilateral molar clenching group in their analysis resembles the small number of teeth group in our study, and the group function clenching in their study resembles large number of teeth group in our study. Our results may theoretically be supported from the viewpoint of biomechanical analysis. However, there is no point in loading an occlusal force in an edentulous jaw. The mechanical stress caused by occlusal movement such as twisting torsions may be reduced.
Our findings showed a close relationship between a small number of residual teeth and reconstruction plate prognosis. To clarify this causal relationship, a biomechanical analysis of the stress load to the mandible and reconstruction plate is needed. However, if this problem can be clarified, the prognosis of the reconstruction plate can be determined, and countermeasures to reduce the stress can be established. The stress generated by occlusion should be analyzed from multiple perspectives in future studies.

| CONCLUSIONS
The 5-year exposure-free survival of a mandibular plate was poorer for patients with a small number of residual paired teeth than for patients with no teeth or those with a large number of residual paired teeth. Thus, stress burden on the reconstruction plate appears to differ based on the number of residual paired teeth. To accurately specify the cause of this stress burden and utilize it for treatment, future in vivo studies on the stress load on the reconstruction plate are needed.

ACKNOWLEDGEMENT
This work was supported by JSPS KAKENHI (Tokyo, Japan; grant number, 19H03846).