Osseocutaneous Radial Forearm Free Tissue Transfer for Repair of Complex Midfacial Defects

Objective  To evaluate the resulting aesthetics, function, and donor site morbidity of the osseocutaneous radial forearm free flap (OCRFFF) used for midface reconstruction.

Design  Prospective case series and a retrospective review of results.

Patients  Ten patients from an academic practice who underwent reconstruction at the University of Michigan Hospitals between 1995 and 2001.

Interventions  All patients had maxillectomy defects in which the entire infraorbital rim was reconstructed with an OCRFFF. Of the 10 patients included in the study, 3 underwent a total maxillectomy with orbital exenteration, 4 had a total maxillectomy without orbital exenteration, and 3 had a limited maxillectomy that did not involve the palate. Patients with palatal defects underwent reconstruction with a prosthetic palatal obturator.

Main Outcome Measures  Facial contour and aesthetic results, speech understandability, ability to eat solid foods, oronasal separation, socializing outside the home, and return-to-work status. Flap success, donor site morbidity, and orbital complications were also studied.

Results  Mean ± SEM follow-up was 23.2 ± 5.0 months. A modified Funk facial deformity scale was used, and 7 of the 10 patients had either no deformity or minimal deformity. The mean aesthetic score for these reconstructions was 2.1 ± 0.3 on a scale of 1 to 4, with 1 representing no deformity and 4 representing a severe deformity. All patients returned to a solid diet and had understandable speech, although patients who had an orbital exenteration trended to poorer scores. All patients socialized either frequently or occasionally outside the home, and all patients not retired or disabled prior to surgery returned to work.

Conclusion  The OCRFFF reconstruction of the infraorbital rim in patients with total maxillectomy defects and obturator of the palatal defect controls orbital complications and optimizes aesthetic outcome while achieving nearly normal palatal function.

Reconstructive efforts of the midface must support the orbital contents by appropriate positioning of the bony orbital rim and floor, thus preventing changes in globe position and orbital volume, which can lead to diplopia, enophthalmos, ectropion, and dystopia. In addition, proper reconstruction of the bony palate is also necessary to preserve a platform for mastication and maintain oronasal separation, which is important for speech quality and understandability. Proper reconstructive techniques must also maintain contour symmetry with the other side of the face.

Several strategies for the reconstruction of midfacial defects with free tissue transfer have emerged in the reconstructive literature. These approaches are typically based on the extent of the maxillectomy defect and attempt to simultaneously address the aesthetic and functional requirements of both the orbital and palatal defects. For total maxillectomy defects with the loss of the infraorbital rim, some authors advocate the use of a rectus abdominis myocutaneous free flap for coverage and obliteration of the maxillary cavity and reconstruction of the infraorbital rim with a nonvascularized bone graft.^1,2 However, resorption of nonvascularized bone along the infraorbital rim can be problematic with this approach. Alternatively, the scapula osseocutaneous free flap, with appropriate osteotomies, has been used for reconstruction of the infraorbital rim and palate in extensive midfacial defects.^3,4 The short vascular pedicle and difficulty with orienting the scapular flap to address both the palate and orbit defects make this flap less than ideal in some surgeons’ hands. In addition, osseointegrated implants for palate rehabilitation can be difficult to retain in the scapular bone. Other authors have described the use of the iliac crest myo-osseous flap for the repair of total maxillectomy defects, but this approach can be limited by a shorter pedicle and difficulty with reconstructing the exenterated orbit.^5,6 Several publications have also described the use of the fibula osseocutaneous free flap to repair a variety of maxillary defects.^7,8 The fibula is excellent for the reconstruction of infrastructure maxillectomies but cannot address orbital rim or orbit defects.

Our strategy for the reconstruction and rehabilitation of total maxillectomy defects involving the orbital rim at the University of Michigan has focused on maintaining support for the globe and restoring the bony orbit. We believe that preventing periocular complications such as diplopia, enophthalmos, ectropion, and dystopia is an important first step for a successful functional and aesthetic outcome. For total maxillectomy defects that involve greater than 40% of the orbital rim and/or involve the malar eminence, we reconstruct the orbital defect with a scapular osseocutaneous free flap. For total maxillectomy defects (with or without orbital exenteration) that involve the infraorbital rim and orbital floor, we use an osseocutaneous radial forearm free flap (OCRFFF) to support the globe and reconstruct the infraorbital rim. We obturate the palate and restore the masticatory surface with a palatal prosthesis.

In the present study, we present a prospective case series and a retrospective review of results of our experience with limited and total maxillectomy defects involving the infraorbital rim that were reconstructed with an OCRFFF and obturated, when necessary, with a palatal prosthesis.

This study is a case series of 10 patients treated at the University of Michigan Health System between 1995 and 2001 who underwent OCRFFF reconstruction of the inferior orbital rim after limited or total maxillectomy and rehabilitation of the palate (when necessary) with a palatal obturator. Patients were divided into 3 groups (Table 1): total maxillectomy including the hard palate and infraorbital rim with orbital exenteration (group 1; n = 3); total maxillectomy including the hard palate and infraorbital rim without orbital exenteration (group 2; n = 4); and suprastructure maxillectomy including the infraorbital rim with preservation of the hard palate (group 3; n = 3). Medical records as well as intraoperative and postoperative photographs of these cohorts were retrospectively reviewed.

In this series, there were 5 men and 5 women. The mean ± SEM age was 46.4 ± 6.6 years (age range, 19-76 years). Tumors were of diverse types, the most common being squamous cell carcinoma and sarcoma (3 of each), followed by spindle cell carcinoma, adenoid cystic carcinoma, hemangiopericytoma, and ameloblastoma (1 of each). Tumors were either advanced stage (stage 3 or 4, 6/10) or recurrent (4/10) in this population. Three patients (30%) required orbital exenteration secondary to tumor extension. All patients underwent postoperative radiation therapy, and 2 patients underwent perioperative chemotherapy. Follow-up varied from 3 to 53 months (mean ± SEM, 23.2 ± 5.0 months). Overall survival was only 50% during the 6-year study.

We approached reconstruction of the maxillectomy defect with 6 goals: (1) to maintain or restore orbital position and function; (2) to maintain or restore the infraorbital rim and malar contour; (3) to provide separation of the oronasal cavities; (4) to restore the occlusal surface for mastication; (5) to maintain the position of the nasal ala; and (6) to maintain lateral nasal length. To achieve these goals, we were guided by the following principles: (1) If the orbit was partially or completely resected, the reconstruction was designed to be functional and aesthetic. The lids, canthi, and muscle aponeurotic system were carefully repositioned to maintain facial symmetry. (2) If the orbit was intact, an anatomic reconstruction of the infraorbital rim and orbital floor was performed. (3) If the patient was to receive radiation, vascularized bone was used for the infraorbital rim reconstruction in an effort to reduce bone resorption, decrease associated soft tissue atrophy, and reduce the likelihood of plate exposure. (4) If the hard palate was resected, a stable platform for mastication was provided, which also achieved oronasal separation.

The method of harvest has been widely described, but certain technical aspects are helpful in optimizing orbital reconstruction. Because there is limited flexibility of the bone relative to the skin paddle, the flap inset rather than the patient’s nondominant hand dictates the side of the donor site. Rather than cutting a tangent to the cross section of the radius, a wedge of bone is removed and centered on the attachment of the flexor retinaculum and the vascular supply from the radial artery. This approach allows for better control of the amount of bone being harvested. The skin paddle tends to be a trapezoid. The longer side of the trapezoid runs anteriorly from just below the malar eminence to the postsuperomedial aspect of the resection that will cover the radial bone and the exposed native bone of the ascending process of the maxilla. The shorter side of the trapezoid is posterior at or near the orbital apex and extends from the lateral orbit to the most posteromedial aspect of the reconstruction. Full arm casting for 4 weeks is followed by 2 additional weeks of forearm casting to allow gradual loading and reorientation of the stress lines to minimize the risk of a pathologic radius fracture.

A 1.7-mm titanium plate is precontoured to span the infraorbital rim resection. The orbital floor is supported with a Medpor implant (Porex Surgical Products Group, Newnan, Ga) and is sutured anteriorly to the radial bone once the bone is fixed in position. The Medpor implant cannot protrude past the radial bone in the coronal plane. The soft tissue portion of the flap must be sutured high (superiorly) in the maxillary cavity so as to not interfere with the palate obturator. The nasal septal mucosa is folded down to meet with the radial skin paddle to ensure coverage of the remaining native bone and the radial bone, thus restoring the ascending process of the maxilla. The flap must also be sutured anteriorly to cover the radial bone to decrease the risk of bone resorption. Crawford tubes are brought into the maxillary cavity.

The obturator can be in contact with the vascular pedicle as it courses posterolaterally through the resected cavity into the ipsilateral neck without deleterious consequences. In general, the pedicle is exposed between the inferior aspect of the soft tissue flap inset and the superior aspect of the buccal resection margin.

Outcome variables measured included facial contour and aesthetic results, speech understandability, ability to eat solid foods, oronasal separation, socializing outside the home, and return-to-work status. Flap success, donor site morbidity, and orbital complications were also studied. Facial contour and aesthetic results were assessed by 3 of the authors (D.B.C., J.S.M., and T.N.T.) through postoperative photographs using a modified scale originally described by Funk et al.⁹ Patients were assigned a numerical score (1, no deformity; 2, minimal deformity; 3, moderate deformity; or 4, severe deformity). Patients with no deformity had an operative side that resembled the appearance of the nonoperative side in contour and symmetry with no ectropion or enophthalmos. Minimal defects included only minor soft tissue and skeletal asymmetry with minor ectropion or enophthalmos. Moderate deformity involved a closed orbit without a nasocutaneous fistula, no exposed plating, moderate ectropion or enophthalmos, and moderate soft tissue asymmetry or skeletal deformity as compared with the nonoperative side. Severe deformity consisted of gross soft tissue asymmetry, gross skeletal deformation, nasocutaneous fistula, exposed plating, or severe ectropion or enophthalmos.

The speech understandability scale was modified from List et al¹⁰: always understandable, 5; understandable most of the time with occasional repetition necessary, 4; usually understandable but face-to-face contact necessary, 3; difficult to understand, 2; and never understandable with written communication necessary, 1.

Ability to eat solid foods was scored using the following criteria: full range of solids with no restrictions, 6; minimally restricted solids with few specific exclusions (eg, bread crumbs), 5; variety of solids taken but facilitated by increased moisture or liquid chasers, 4; minced, moist, or soft diet, 3; pureed solids, 2; and no solids, 1.

Oronasal separation was scored as follows: no evidence of velopharyngeal incompetence or nasopharyngeal reflux or nasal regurgitation of liquids, 5; mild inconsistent nasal emission, nares constriction, hypernasality, or nasal regurgitation or reflux, 4; moderate and consistent nasal emission, nares constriction, inappropriate nasality, or reflux or regurgitation, 3; severe or frequent nasal emission, nares constriction, and inappropriate nasality, 2; and constant and continuous nasal emission, nares constriction, hypernasal resonance, or nasal reflux or regurgitation, 1.

We determined whether patients had returned to work, were permanently disabled, or were retired prior to surgery. We also noted if they socialized outside the home frequently, occasionally, rarely, or never.

Data on sex, age, hospital stay, contour deformity, follow-up, survival, speech, diet, and oronasal separation were tabulated. Frequencies and basic descriptive statistics were calculated, including means, standard error of the means, and ranges. Univariate analysis was conducted using SPSS for Windows 2000, version 10.0 (SPSS Inc, Chicago, Ill). Owing to the small sample size, no multivariate analyses were attempted. Analysis of variance with post hoc analysis was used to determine statistical significance among the 3 groups listed in Table 1. To determine differences between group 1 and group 2 with regard to speech, diet, and oronasal separation, we used χ² analysis. Statistical significance was set at α<.05.

There were no flap failures in our series of 10 patients. Donor site morbidity was negligible with no pathologic fractures after removal of the OCRFFF. One (10%) of 10 patients developed osteoradionecrosis of the implanted radial bone after postoperative radiation therapy resulting in soft tissue breakdown and plate extrusion along the infraorbital rim. A surgical procedure was required to remove necrotic bone. Two (29%) of 7 patients developed postoperative diplopia requiring nonsurgical correction, while 1 (14%) of 7 developed ectropion that required surgical correction.

Assessments of facial contour and aesthetic results were performed using postoperative photographs and medical records (Table 1, Figure 1, and Figure 2). Seven (70%) of 10 patients had either no deformity or minimal deformity, and only 1 patient had a severe deformity. The mean ± SEM aesthetic contour score of this patient population was 2.1 ± 0.3. Subgroup analysis demonstrated that the presence of orbital exenteration did not significantly impact aesthetic contour scores (P = .13), but there was a trend toward a poorer score in this group (2.9 ± 0.6) than in reconstructive cases without orbital exenteration (1.5 ± 0.3).

Level of socializing and return-to-work status were also studied in this population. All patients socialized outside the home either frequently (60%, 6/10) or occasionally (40%, 4/10). Of note, 2 (66%) of 3 patients with orbital exenterations socialized only occasionally. Two (20%) of the patients in this study were retired and 1 (10%) was disabled at the time of reconstruction. All 7 patients who were not disabled or retired at the time of reconstruction (100%) returned to work.

Dental rehabilitation for all 7 patients with total maxillectomy defects was achieved with obturation using a palatal obturator and denture. Table 2 summarizes the overall favorable mean ± SEM scores resulting from this form of palate and dental rehabilitation: speech (4.8 ± 0.1) and oronasal separation (4.8 ± 0.1). The speech understandability, solid diet, and oronasal separation scores of the patients in group 3 who had an intact oral palate were compared with those of the patients in groups 1 and 2 who had a palate prosthesis. Importantly, patients in group 2 (total maxillectomy without orbital exenteration) had speech scores similar to those in group 3, who had an intact native palate. However, patients in group 1 (total maxillectomy with orbital exenteration) tended to have worse speech scores (4.3 ± 0.3) than those without exenteration (5.0 ± 0.0). This difference in speech score approached statistical significance (P = .06).

The mean ± SEM diet score for all patients with total maxillectomy defects was 5.8 ± 0.1. Subgroup analysis in patients with orbital exenteration (group 1, 5.3 ± 0.3) demonstrated a trend toward significantly worse diet scores (P = .06) than those of patients with a retained globe (group 2, 6.0 ± 0.0). Again, group 2 patients had diet scores similar to group 3 patients with intact palates. There were no significant differences in oronasal separation among the 3 groups, although the mean score in group 1 was lower.

Mean ± SEM hospital stay following reconstructive surgery was 6.7 ± 0.7 days for all patients. Patients with orbital exenteration had significantly longer hospital stays (9.3 ± 1.2 days) than those patients without orbital exenteration (5.3 ± 0.3 days) (P = .04).

While overall survival in this patient series was only 50% during the 6-year study period, survival of patients without orbital exenteration (5/7) was better than for those patients requiring orbital exenteration (0/3).

In this series of 10 patients, we investigated the aesthetic and functional outcomes of patients with limited and total maxillectomy defects whose infraorbital rims were reconstructed with an OCRFFF. In addition, we studied the impact of a palatal obturator on speech, diet, and oronasal separation for those patients who underwent total maxillectomy.

In our study, 7 of 10 patients had an intact globe. Two (29%) of 7 patients developed postoperative diplopia requiring nonsurgical correction that has since resolved with this intervention. One of 7 patients (1 of the patients with diplopia) developed ectropion that required surgical correction. There were no instances of enophthalmos. In contrast, an earlier study found that approximately 75% of patients with similar defects who underwent reconstruction with nonvascularized bone and a rectus myocutaneous free flap developed ectropion, and 10% developed dystopia.¹ Perhaps one of the reasons that our ectropion rates were lower was our use of vascularized bone grafts and the resultant lack of bone resorption in the irradiated field.

Importantly, the total maxillectomy group without orbital exenteration tended to compare favorably with those patients described in the literature who had their palates reconstructed with free tissue transfer. A study by Cordeiro and Santamaria¹ of patients with similar total maxillectomy defects reconstructed with a rectus myocutaneous free flap showed that 75% of the patients had an unrestricted diet and 25% were on a soft diet. In a study by Futran et al,⁷ who used an osteocutaneous fibula free flap to reconstruct the palate, approximately 50% of patients ate a regular diet and 50% a soft diet. While one could argue that the type of defect in these studies may have not been suitable for a maxillary prosthesis, the data would suggest when obturation is feasible (ie, when suitable supporting structures are present to maintain an obturator), it offers a comparable result to a free tissue–reconstructed palate. In addition, a palatal obturator has the added benefit of immediate dental restoration, which is particularly important in populations where survival is variable and insurance coverage of dental implants may be an issue.

The aesthetic and functional outcomes of a reconstructive process should be directly correlated with the ability of patients to return to work and resume normal social interactions. This is especially true for defects in the midface region, an area so important for sight, speech, and deglutition. In our study, all patients socialized outside the home either frequently or occasionally, and 100% of eligible patients returned to work after their reconstructive procedure. Two of 3 patients with orbital exenterations socialized only occasionally, and this may be a result of worse aesthetic and functional outcomes in this group.

The midface is a functioning anatomic unit important for orbital support, nasal support, lip position, mastication, speech, and aesthetics. Our strategy for reconstruction of the orbital component of maxillectomy defects is to use vascularized bone for the infraorbital rim and meticulous repositioning of native orbital and periorbital structures. Our strategy for repairing the secondary palate defects includes the use of a palatal obturator. Seven (70%) of our 10 patients had minimal to no facial deformity, 2 (29%) of 7 had diplopia that could be corrected with glasses. Patients without an orbital exenteration had normal speech, swallowing, and oronasal separation scores, comparable to patients with an intact palate. Patients with an orbital exenteration trended toward poorer scores with respect to palate rehabilitation but were still able to eat a solid diet, and their speech was still understandable most of the time. Importantly, 6 (60%) of 10 patients socialized frequently, and 7 (100%) of our 7 patients who were not retired prior to surgery returned to work. These data suggest that with this approach, eye rehabilitation compares favorably with other approaches, that the outcome of our palate rehabilitation with an obturator is comparable with an intact palate, and that our patients return to their pretreatment employment and socialization roles. The survival in this population is poor, and consideration of concurrent chemotherapy and radiation therapy is gaining favor. Any salvage procedures as a result of treatment failure in these patients will place even greater demands on rehabilitation of the eye.

Correspondence: Douglas B. Chepeha, MD, MSPH, Department of Otolaryngology–Head and Neck Surgery, University of Michigan Health System, 1904 Taubman Center, 1500 E Medical Center Dr, Ann Arbor, MI 48109-0312 (dchepeha@umich.edu).

Submitted for Publication: February 25, 2004; final revision received December 22, 2004; accepted February 12, 2005.