Imaging evidence of osseous fusion of patients undergoing segmental mandibular fibular free flap reconstruction – a retrospective analysis

INTRODUCTION

Segmental mandibular reconstruction is performed for defects from malignant and benign tumors, osteoradionecrosis (ORN), osteomyelitis, trauma, and congenital deformities.^[1-7] The fibular free flap (FFF) is commonly used for mandibular reconstruction due to its long supply of bone, muscle, skin, and rich arterial blood supply.^[8] Reconstruction allows for restoration of facial symmetry and mandibular function including swallowing, speech, and dental rehabilitation with low morbidity.^[9-11] While successful reconstruction with the FFF has been well established,^[12] imaging findings regarding osseous union of the flap and native mandible have not yet been described. In addition, virtually no data exist demonstrating the effect of the different indications for surgery (IFS), patient factors such as smoking, steroid use, diabetes, history of chemotherapy or radiotherapy (RT), steroids, diabetes, adjuvant chemotherapy and RT, and other factors on osseous fusion. This retrospective study evaluates the effect of these factors on imaging evidence of osseous fusion (IEOF) and the timing of fusion in patients undergoing FFF reconstruction.

MATERIAL AND METHODS

A waiver of informed consent was granted by the Institutional Review Board to evaluate patients who underwent microvascular free flap reconstruction of the mandible between 2013 and 2023. Patients were identified using oral and maxillofacial surgery (OMS) logbooks. Fixation in all cases was achieved through use of a reconstruction plate, bicortical screws in the native mandible, and monocortical screws in the FFF. Patients were NPO for 1 week following surgery utilizing a nasogastric tube. They were then advanced to a soft diet for approximately 2 months. If patients could not maintain adequate nutrition through oral intake alone, then a nasogastric tube was left in place to provide supplemental nutrition. No dental rehabilitation utilizing dental implants occurred in subjects with malignant disease. A small number of patients with benign disease received dental implants at the time of reconstructive surgery, with variable times until loading (immediate vs. 6 months). All patients underwent a medical record and imaging search. Inclusion criteria were segmental mandibular defect and reconstruction with an osseous FFF, post-operative imaging including computed tomography (CT), cone beam CT (CBCT), and/or positron emission tomography (PET)/CT. Exclusion criteria included patients undergoing initial mandibular reconstruction with a flap other than a FFF, early flap failures, absent postoperative imaging, and patients under age 18.

Imaging was reviewed on a PACS system (Visage) or on a CBCT (New Tom) system. CBCT scans were acquired by the OMS service in an outpatient clinical setting. CT and PET/CT scans were acquired according to standard radiology departmental protocols. All imaging was reviewed by a board-certified neuroradiologist with a certificate of added qualification and 13 years of post-fellowship experience.

Demographic and clinical data are recorded in Table 1. Pre-operative risk factors such as history of smoking, steroid use, history of RT, and chemotherapy were recorded. Surgical factors included anastomotic artery and vein, location of mandibular defect including the central segment, mandibular condyle, ramus and combinations thereof, as well as the number of osteotomy segments and composite versus bone-only flaps. Perioperative glucose levels and osteoporosis were not documented in many patients and were excluded. Since most patients were not diabetic, this factor was not analyzed.

The use of adjuvant RT and/or chemotherapy, steroids, and non-steroidal anti-inflammatory drugs (NSAIDs) was recorded. Perioperative complications were categorized as early (within 30 days of surgery) and late (later than 30 days) including wound infection, wound breakdown, flap failure, and combinations thereof. Complications that required early or late return to the OR were recorded, including wound infection, wound breakdown, fistula, flap, and hardware failure.

Fusion was categorized as complete, partial, or no fusion and by grouping partial and fusion versus no fusion. Although no universally radiographic definition exists, complete fusion was defined as cancellous bone bridging across the flap to the native mandible across both portions of the fusion construct.^[13] Partial fusion was defined as bridging across either segment but not both. In cases with multiple osteotomies, complete fusion was designated if all segments demonstrated osseous bridging, and partial if there was at least one segment with bridging [Figure 1].

Close

Figure 1: (A) 43 year old man with a history of oral cavity cancer. Axial non-contrast computed tomography (CT) demonstrates right mandibular body reconstruction with fibular free flap. Note the lack of bone fusion across the proximal and distal osteotomy sites (thin white arrows). (B) 58 year old male who underwent mandibular reconstruction for osteoradionecrosis. Axial non-contrast CT demonstrates partial fusion of the right fibular free flap mandibular reconstruction. Note the cancellous and cortical bone bridging at the anterior right parasymphyseal segment (thick white arrow), and the lack of bridging proximally (thin blue arrow). (C) 68 year old male who underwent mandibular reconstruction for advanced oral cavity cancer. Axial non-contrast CT demonstrates complete fusion of a 3-osteotomy mandibular reconstruction involving the central segment and anterior bodies bilaterally. Note the bridging cancellous bone at each osteotomy site (thin red arrows).

Export to PPT

Analyses were conducted in aggregate and individually for different IFS: Malignant, benign, and ORN. Trauma patients were included in aggregate analyses but not analyzed individually because of the low number of patients (n = 5). Primary endpoints were IEOF and timing of IEOF. Both endpoints were analyzed with fusion categorized as non-fusion, partial, and complete, and by combining partial and complete into non-fusion and fusion. Finally, the presence of fusion at time intervals 0–6, 6–12, and 12+ months was recorded.

RESULTS

79 patients met inclusion criteria. Thirty patients underwent reconstruction for malignant lesions, all of which were squamous cell carcinoma. There were 20 (67%) patients with T4, 3 (10%) with T3, and 7 (23%) with T2 lesions. 17 (22%) patients had benign tumors: 14 had ameloblastoma, one each with a keratocystic odontogenic tumor, juvenile ossifying fibroma, and a myxoma. 27 (34%) patients underwent reconstruction for ORN. Only 5 (6%) patients underwent reconstruction for trauma.

23 patients had reconstruction of the body, central segment, and ramus: 18 had reconstruction of the body and ramus. 13 had reconstruction of the body, ramus, and condyle. 7 had reconstruction of the body and central segment. 5 had reconstruction of the body, central segment, ramus, and at least one condyle. Three had reconstruction of the ramus and one condyle only.

24 patients received adjuvant RT. 10 patients received adjuvant chemotherapy. 51 patients were treated with NSAIDs post-operatively, and 52 received steroids perioperatively.

Of the most severe complications, 4 patients developed flap failures. Two of these patients achieved complete fusion after revision, one by undergoing a scapular free flap, and the other a deep circumflex iliac artery free flap, with average imaging follow-up of 32.6 months. The two patients without IEOF developed complications due to flap failure after reconstruction for ORN.

All but one patient with late complications achieved complete fusion with an average imaging follow-up time of 30.6 months. The one patient without IEOF underwent reconstruction for ORN and had a CBCT 13 months after surgery.

62 (78%) patients had evidence of osseous fusion. 53 (67%) had complete fusion. The imaging follow-up interval for these patients ranged from 2 to 94 months with an average of 20.6 months. 9 patients had partial fusion with imaging follow-up ranging from 5 to 59 months with an average of 17.1 months. 17 patients had no evidence of osseous fusion. Imaging follow-up for these patients ranged from 2 to 13 months, with an average of 6.4 months.

DISCUSSION

There is a lack of critical evaluation in the medical literature regarding radiographic evidence of osseous fusion when reconstructing the mandible with fibula free flaps. One study reviewed osseous union, but combined maxillary and mandibular construction and the use of patient-specific implants.^[13,14] In comparison, there are countless manuscripts regarding spinal fusion constructs.^[15,16]

As opposed to grafts, free flaps fuse with the primary mandibular bone in a single stage in a process known as osteointegration due to the blood supply from the flap.^[15-17] The FFF has both a segmental and intraosseous blood supply, allowing multiple osteotomies without devascularization.^[18-20]

Both cortical and cancellous bone are important to bone strength.^[21] Cortical bone is dense with limited porosity which can maintain heavy axial loads. After reconstruction, both cortical and cancellous bone undergo a remodeling and maturation process. Full incorporation of the flap can take 1–2 years. Surgical hardware immobilizes the junction of bone segments, providing temporary stability allowing for fusion. Failure of fusion, or nonunion, leads to low-grade mobility between the segments resulting in pseudoarthrosis which is associated with pain and impaired function.^[22] Risk factors for nonunion include osteoporosis, systemic diseases such as inflammatory arthritis, renal disease, smoking, and poor nutrition.^[23-28] Medications such as NSAIDs, steroids, immunosuppressants, and RT can impede bone healing.^[28-30]

These results demonstrate that the timing of follow-up was useful in detecting the presence of osseous fusion [Table 2]. This is likely because imaging was ordered until the patient was healed clinically and radiographically. The methodology captured patients near the point of osseous fusion to a large degree. At each time interval, patients were selected who had achieved radiographic fusion which coincided with clinical examination findings of a successful reconstruction, while patients with prolonged healing required longer follow-up. Patients with malignant lesions had longer surveillance periods, and they demonstrated fusion at later time points.

In aggregate, the length of imaging follow-up was closely correlated with fusion. The longer patients were followed, the more likely they were to have IEOF. Most patients needed at least 12 months to achieve fusion. There was a subset of patients who achieved either partial or complete fusion by 4–6 months. Statistical analyses were performed on this group of “fast fusing” patients and no significant associations were identified. However, within the benign subgroup, female patients disproportionately fused by 6 months. This may account for, in part, the accelerated group.

Across all IFS, timing of follow-up was the most significant factor regarding IEOF. Patients with malignant lesions had an average imaging follow-up time of 22 months, benign lesions 10.4 months, and ORN 16.2 months, respectively. Malignant lesions were more likely to achieve partial and complete fusion as compared to the other IFS. This may be in part because of the longer follow-up times for disease surveillance for malignant lesions versus benign indications. Within the malignant subgroup, the T4 tumors were less likely to achieve fusion relative to lower stage tumors. This may be a type 2 error due to the disproportionate number of T4 lesions. However, it may suggest that the less aggressive tumors are more easily reconstructed. Patients with benign tumors and ORN had more non-fusion than expected, which was likely due to shorter imaging follow-up. For all IFS, the non-union patients had imaging follow-up that was significantly shorter than the patients who demonstrated complete fusion.

Patients with malignant lesions who received adjuvant chemotherapy were less likely to achieve fusion and more likely to achieve non- or partial fusion. There was a similar imaging follow-up interval for the adjuvant and non-adjuvant chemotherapy groups (15.6 vs. 17.2 months). This suggests that adjuvant chemotherapy may have prolonged healing time; however, the overall rate of fusion between the two groups did not change. This suggests that adjuvant chemotherapy may delay fusion, but given enough time, these patients eventually fuse. A similar association with steroids was noted, with a transient delay in healing at 6 months for patients with reconstructions for malignancy.

Mandibular defects involving the central segment pose unique challenges for reconstruction related to functional restoration. Restoration of mandibular condyle defects poses challenges related to temporomandibular joint stability and dental occlusion.^[11] The data fail to show an association between mandibular location and outcome or timing of fusion.

RT has long been debated as a source of morbidity and surgical failure due to impaired vascularity. While perioperative RT has been associated with wound complications, previous series demonstrated no negative effect on fusion.^[29-31] The data show that prior RT resulted in delayed bone fusion at 1 year, with most patients requiring more than 12 months to achieve fusion. In addition, the prior RT group lacked the “fast fusing” patients that existed in the aggregate sample. This did not hold for patients receiving adjuvant RT. This suggests that the impact of RT is greater if received before reconstruction and not during the healing period.

The effect of post-operative NSAIDs on bony fusion in the spinal fusion literature is controversial, with studies suggesting that NSAID use can delay fusion or result in non-union.^[29] Patients with benign lesions who received NSAIDs had greater overall fusion than patients who did not. However, differences in follow-up imaging time likely explain this difference, rendering it erroneous. Patients without NSAIDs had significantly shorter imaging follow-up time (7.4 vs. 12.6 months) with 4 of 7 patients with <6 months of post-operative imaging. Patients who received NSAIDs had significantly longer times with only 3 of 10 patients with <6 months of follow-up.

This study had several limitations. It is a single-institution retrospective analysis. A specific type of patient was selected: Mandibular reconstruction with FFF with sufficient imaging follow-up to ascertain the presence and timing of fusion. This resulted in low sample sizes precluding some analyses and others that were underpowered. There were survivorship and compliance biases. Lack of standardized imaging acquisition after surgery was a limitation. Data on clinical outcomes that could be correlated with radiographic results were not collected. Data distinguishing between flap-to-flap fusion and flap to native mandible fusion were not collected. Finally, the presence of fibrous union was not analyzed.

CONCLUSION

This study demonstrates that given enough time, most patients undergoing mandibular reconstruction with FFF achieve fusion. Prior RT, adjuvant chemotherapy, and perioperative steroids may transiently delay fusion, but after 12 months, most patients fuse. Early imaging after 3–6 months may be obtained to detect potential complications, but to confirm fusion, imaging is most effective 12 months after surgery. ORN patients took longer to fuse than other reconstruction indications. These results suggest that time is the most important factor in fusion of a mandibular fusion construct, and although variability in healing exists, most patients have good radiographic outcomes.