A, Intraoperative patient positioning and flap design; B, harvested flap.
Frontal (A) and lateral (B) views of a patient with gigantiform cementoma; frontal (C) and lateral (D) view of a patient with latissimus-serratus-rib free flap.
Kim PD, Blackwell KE. Latissimus-Serratus-Rib Free Flap for Oromandibular and Maxillary Reconstruction. Arch Otolaryngol Head Neck Surg. 2007;133(8):791-795. doi:10.1001/archotol.133.8.791
To review complications and outcomes associated with latissimus-serratus-rib free flap oromandibular and midface reconstruction.
Retrospective medical record review.
Two academic tertiary care medical centers.
Twenty-eight patients with segmental resection of the mandible and 1 patient with combined resection of the mandible and maxilla after excision of neoplasms of the oral cavity, who were believed to be poor candidates for fibula free flap reconstruction, were identified.
Twenty-seven latissimus-serratus-rib osteomusculocutaneous free flap reconstructions and 2 serratus-rib osteomuscular free flap reconstructions were performed.
Main Outcome Measures
The outcome of microvascular free tissue transfer as well as short- and long-term complications were recorded.
There were no perioperative free flap failures. Delayed partial rib graft resorption occurred in 1 patient 33 months after free flap transfer for maxillary reconstruction. Among 28 cases of mandibular reconstruction, 1 case of bone graft nonunion was noted after a postoperative period of 57 months. All other cases achieved successful restoration of mandibular continuity. Donor site morbidity was well-tolerated in all patients.
Latissimus-serratus-rib osteomusculocutaneous free flaps are effective for reconstruction of composite defects of the mandible in patients who are not candidates for more commonly used vascularized bone-containing free flaps.
Restoration of mandibular continuity after resection of oral cavity neoplasms is a common indication for microvascular flap reconstruction. Until the advent of bone-containing free flaps, successful immediate reconstruction of segmental oromandibular defects was limited. Wound healing complications and bone graft resorption were common in patients who underwent immediate reconstruction using nonvascularized bone grafts and pedicled bone-containing flaps. With the increased reliability of mandibular reconstruction provided by free flaps, we now strive to provide the best possible results in terms of aesthetics, mastication, deglutition, and speech.
At many medical centers, the fibula osteocutaneous free flap is the most common flap used for reconstruction of segmental mandibular defects. Advantages of the fibula free flap include low donor site morbidity, a long bone graft and vascular pedicle length, and a reliable skin paddle. However, peripheral vascular disease sometimes precludes the use of a fibula free flap owing to occlusive disease affecting the peroneal artery or to occlusive disease affecting the anterior and posterior tibial arteries, in which case collateral circulation to the foot may be provided by the peroneal artery. There is a high prevalence of smoking in patients with head and neck cancer, which increases the risk of patients developing peripheral vascular disease that may preclude use of a fibula flap. For those patients, other reconstructive options are often required.
Other commonly used vascularized bone-containing free flaps for oromandibular reconstruction include iliac crest flaps, scapula flaps, and radial forearm flaps. Difficulties with the iliac crest free flap often arise from its bulky skin paddle, which may be ill suited for reconstruction of low-volume soft tissue defects in the oral cavity and from significant potential donor site complications that include gait disturbance, incisional hernia formation, anesthesia of the nerve distribution of the lateral femoral cutaneous nerve, and wound dehiscence.1 Harvest of free flaps perfused by the circumflex scapular artery that contain the lateral border of the scapula bone require intraoperative patient repositioning into a decubitus position. This prevents simultaneous resection of the mandible and flap harvest, thereby increasing the duration of surgery. In addition, the length of scapula that is available for bone grafting is limited to 10 to 14 cm, and the scapula bone stock that is available may be ill suited for placement of dental implants, particularly in women of small stature. The osteocutaneous radial forearm flap also offers a limited length of bone graft that is ill suited for placement of dental implants, and prior authors have reported significant potential donor site morbidity in terms of pathological fracture of the radius after flap harvest.
Latissimus-serratus-rib osteocutaneous free flap are an uncommonly used alternative for oromandibular reconstruction. The potential advantages of this flap include the supine patient position used during flap harvest that allows for simultaneous 2-team surgery, a skin paddle that can be designed to reconstruct most associated soft tissue defects in the oral cavity, low donor site morbidity, and a long vascular pedicle that contains large caliber vessels. The present study was undertaken to examine the utility of latissimus-serratus-rib free flaps for oromandibular reconstruction. To our knowledge, this study is the largest series in the literature on this topic.
A retrospective review was undertaken of 29 consecutive latissimus-serratus-rib free flap reconstructions performed by 2 surgeons between 1995 and 2005. This included 27 latissimus-serratus-rib osteomusculocutaneous free flaps and 2 serratus-rib osteomuscular free flaps. Twenty-seven patients received 1 flap, while 1 patient with polyostotic tumor involvement of the mandible and maxilla received 2 flaps by undergoing staged jaw resections and flap reconstructions. There were 19 male and 9 female patients in the series. The median age was 68 years (range, 21-82 years). In 27 patients, fibula flap reconstruction was precluded by peripheral vascular disease of the legs. In 1 patient, use of a fibula flap was precluded by a history of osteogenesis imperfecta, which caused the patient to experience more than 80 fractures of the extremities in the past. Primary tumors included 23 squamous cell carcinomas, 2 osteosarcomas, 1 gigantiform cementoma, 1 undifferentiated carcinoma, and 1 mucoepidermoid carcinoma. Patient comorbidity, which we have shown to be a significant predictor of postoperative complications in patients undergoing microvascular head and neck reconstruction, included American Society of Anesthesiologists (ASA) class 1 (n = 1), ASA class 2 (n = 12), ASA class 3 (n = 12), and ASA class 4 (n = 4). History of other treatments included prior surgery (n = 6) and/or radiation therapy (preoperative [n = 13] and postoperative [n = 9]). The extent of the mandibular resections as classified by Urken et al2 is listed in the Table. The mean follow-up period was 21 months.
The main outcome measures were any complications occurring within 30 days of surgery and wound healing complications affecting recipient or donor sites that occurred at any time during the follow-up period.
Flaps were harvested from the flank that was located ipsilateral to the mandibular resection and cervical recipient blood vessels (Figure 1). The flaps were harvested by positioning the patient supine on a bean bag with slight elevation (10º-20º) of the donor sight flank. This allowed for simultaneous surgery at the head and neck and flank, and no intraoperative repositioning of the patient was required. The ipsilateral arm and flank were prepped into the field, and an incision was made in the mid-axillary line from the axilla in an inferior direction toward the iliac crest. Dissection is deepened onto the anterior border of the latissimus dorsi muscle. The loose areolar tissues between the latissimus dorsi and serratus anterior muscles were then dissected until the dominant branch of the thoracodorsal artery to the serratus anterior muscle was identified. The serratus anterior vascular branch was then dissected, and its branching pattern was determined. The most inferior slip of serratus muscle found to be perfused by the serratus anterior branch of the thoracodorsal artery was selected to provide a periosteal blood supply to one of the ribs. This was usually the sixth or seventh rib. The intercostals muscles were divided, and the parietal pleura was exposed above and below the identified serratus slip and rib. The long thoracic nerve was transected in the intercostal space immediately above the rib graft, leaving intact the motor innervation to the more cephalic slips of serratus anterior muscle. A Doyen periosteal elevator was then used to elevate the posterior rib periosteum off of the parietal pleura. A cuff of soft tissues including the intercostal blood vessels were left attached to the rib bone graft. The rib was then divided anteriorly and posteriorly to complete elevation of the serratus-rib component of the flap. The length of the rib harvested was based on the anticipated length of the jaw bone defect. The rib graft was centered at the anterior axillary line, since this is where the serratus anterior muscle has robust insertions into the rib periosteum. On occasion, rib elevation resulted in a small laceration of the parietal pleura. However, in most cases, the pleural defect was repaired with sutures after evacuating the pneumothorax using a red rubber catheter placed on suction, and chest tubes were rarely required after surgery.
After elevation of the serratus-rib component of the flap, the latissimus dorsi myocutaneous component was then designed and elevated according to the anticipated oral cavity soft tissue defect. Soft tissue defects in this series ranged from minimal resections floor of mouth mucosa to through-and-through resections of oral mucosa and external skin. The thoracodorsal pedicle was then followed superiorly up to the take-off of the subscapular vessels from the axillary vessels. Flap harvest was completed by ligating and transecting the subscapular artery and vein after the completion of the oral cavity tumor resection.
The bone and soft tissues of the flap were trimmed to an appropriate size prior to completion of the flap insetting. The curvature of the rib graft was found to be appropriate for reconstruction of hemimandibular defects without needing to make osteotomies in the bone graft. For mandibular defects that crossed the midline, a single closing osteotomy was made through the lingual cortex of the reconstructed mandibular symphysis. Microvascular anastomoses were created after insetting the flap. Cervical recipient blood vessels were used to provide flap perfusion. Postoperative anticoagulation was used consisting of daily aspirin, 81 mg, for 7 days after surgery. Postoperatively, flaps were monitored by periodic use of a Doppler stethoscope and assessment of the flap's color, turgor, and capillary refill.
There were no postoperative deaths, and there were no cases of free flap failure. There were 14 complications that occurred during a 30-day postoperative period. Eight of the complications were related to cardiopulmonary conditions including congestive heart failure, atrial fibrillation, pneumonia, and 1 airway obstruction requiring an urgent tracheostomy. Other complications not related to wound healing included 1 pneumothorax from central catheter placement, 1 case of delirium tremens, and 1 deep vein thrombosis. Three early postoperative wound healing complications included 1 orocutaneous fistula, 1 case of partial flap skin paddle necrosis, and 1 case of necrosis of the margin of a neck skin incision. There were no significant donor site complications, although prolonged high-volume drain output from the flank wound was common, requiring insertion of closed-suction drains for as long as 3 weeks after surgery. No complaints of arm or shoulder weakness or winging of the scapula were noted. No flank seromas or hematomas were noted.
Delayed wound healing complications were noted in 4 patients. Two patients experienced intraoral extrusion of mandibular reconstruction plate. These cases were managed successfully by removing the hardware at 4 months and 13 months after free flap reconstruction, and both patient did well thereafter. At the time of mandibular hardware removal, the rib bone grafts were well incorporated without evidence of nonunion. A third patient developed hardware extrusion through her upper lip skin 22 months after reconstruction of the maxilla and palate using a latissimus-serratus-rib free flap (Figure 2). At the time of hardware removal, the rib bone graft appeared to be well healed, but she experienced slowly progressive midface retrusion over the next 10 months. Reassessment by computed tomographic scan revealed partial resorption of the rib graft that was used for reconstruction of the right maxilla. The fourth patient was seen at 56 months after hemimandible reconstruction with infected hardware. Findings at the time of hardware removal revealed the osteosynthesis between the symphysis and the rib was well healed, but there was a nonunion between the rib graft and the subcondylar segment. This will be repaired in the future by replating and a nonvascularized iliac bone graft after the infection is resolved.
There were 9 patients with intraoperative pneumothoraces, of which 5 were large enough to warrant chest tube insertion. The chest tubes were generally removed within 24 to 72 hours after surgery, since there was never a visceral pleural injury or air leak. The chest tubes never contributed to the length of stay. The mean length of stay for patients with a chest tube was 11.5 days vs 11.6 days for patients without a chest tube. Length of stay was not associated with pneumothorax or chest tubes.
An emergent tracheostomy was performed on a patient with a mucoepidermoid cancer of the parotid that was enveloping the mandible. It was a pure lateral resection of the ramus and condyle with radical parotidectomy, cheek skin resection, and lateral temporal bone resection. There was a limited resection of the buccal mucosa that was repaired primarily. We have performed many cases similar to this, and this is the only one that has ever needed a tracheostomy. These cases in our opinion do not require a tracheostomy.
With the advent of free flaps, head and neck reconstruction has seen great advances. Microvascular head and neck reconstruction has been shown to be extremely reliable and to be associated with an acceptable incidence of complications.3 Free tissue transfer currently represents the most popular and successful method for immediate mandibular reconstruction. Although the fibula free flap is the most commonly selected method of reconstruction, peripheral vascular disease and the need for a larger and bulkier skin paddle may preclude the selection of this flap in certain cases.
The combined free serratus anterior latissimus dorsi flap transferred on 1 vascular pedicle has been described previously.4- 9 The serratus anterior–rib free flap has been described for use in lower extremity reconstruction with good results.10,11 Cadaveric studies have been performed to show the reliability of the blood supply to the rib provided by the thoracodorsal artery. Hui et al12 performed arteriograms through the thoracodorsal artery, and microscopic dissections were done at the rib periosteum. The sixth through the ninth ribs showed consistent filling of their respective intercostal vessels. The rib segments near the anterior axillary line had the most abundant communicating vessels between the serratus muscle and the periosteum. Additional studies have shown that serratus muscle slips 5 through 9 are consistently supplied by a single dominant branch of the thoracodorsal artery. There is a consistent vascular pattern to each muscle slip, in which the serratus artery gives rise to common slip arteries, each of which supplies adjacent muscle slips. The mean length of a muscle slip from its origin on the rib periosteum to the division of the common slip artery is 9.6 cm. These findings imply that the slips may be separated to the level of these common slip arteries, with up to 5 slips transferred on a single neurovascular pedicle and each slip oriented independently.13
The use of latissimus-serratus-rib free flaps for oromandibular reconstruction was described originally in 1985.14 Netscher et al15 discussed the advantages of the composite latissimus-serratus-rib free flap that include the following: the rib bone may be oriented independent of the soft tissue component of the flap, the cartilaginous portion of the rib is available to reconstruct the temporomandibular joint, and high volume of soft tissue is available for reconstruction of large defects.16 Further studies found the serratus-rib free flap to be very reliable in oromandibular reconstruction in patients with.17
In this series of 29 reconstructions using a latissimus-serratus-rib free flap, there were no complete flap failures, but 14 patients (48%) experienced early postoperative complications, and delayed wound healing complications were observed in 14% of cases. Most of the early postoperative complications were related to cardiopulmonary events including congestive heart failure, atrial fibrillation, and pneumonia and were likely a reflection of preexisting patient comorbidities. Successful long-term reconstruction of jaw continuity was achieved in 27 of 29 cases, although delayed partial bone graft resorption was noted in 1 case of attempted maxillary reconstruction, and a case of bone graft nonunion was noted in 1 case of mandibular reconstruction. All 4 cases of delayed wound healing complications initially presented with hardware infection or extrusion. The cases of delayed partial maxillary bone graft resorption may have been elicited by the episode of hardware extrusion causing osteitis of the bone graft. Alternatively, the delayed bone graft resorption in 1 of the patients may have been related to the patient's underlying condition of osteogenesis imperfecta, although it should be noted that this same patient has clinical and radiographic persistence of a serratus-rib graft used for mandibular reconstruction after a follow-up period of 45 months.
Additional advantages of the latissimus-serratus-rib free flap include its long vascular pedicle that contains large-caliber vessels that are rarely involved by peripheral vascular disease, the ability for primary closure of the donor defect with limited long-term donor site morbidity, the ability to use a 2-team approach with no intraoperative repositioning, and the availability of a long bone graft that can span any length of segmental mandibular resection. The size of the latissimus myocutaneous component can be tailored to match low-volume or high-volume soft tissue defects. Previous studies have demonstrated scapular winging; however, this has been minimally noticeable, with no patient attributing work disability to the donor site.18 Our series also concurs with this observation. The main disadvantage of latissimus-serratus-rib free flaps is that they provide a relatively thin bone graft that has little cortical bone, making it ill suited for placement of osteointegrated dental implants.
In conclusion, the latissimus-serratus-rib free flap for oromandibular reconstruction is an excellent choice in patients in whom other osteocutaneous free flaps are unacceptable. This series of patients demonstrates the reliability and the utility of this flap.
Correspondence: Paul D. Kim, MD, Division of Otolaryngology–Head and Neck Surgery, 11234 Anderson St, Suite 2584, Loma Linda, CA 92354 (email@example.com).
Submitted for Publication: July 6, 2006; final revision received January 21, 2007; accepted February 12, 2007.
Author Contributions: Drs Kim and Blackwell had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Kim and Blackwell. Acquisition of data: Kim and Blackwell. Analysis and interpretation of data: Kim and Blackwell. Drafting of the manuscript: Kim. Critical revision of the manuscript for important intellectual content: Kim and Blackwell. Statistical analysis: Kim. Administrative, technical, and material support: Kim. Study supervision: Blackwell.
Financial Disclosure: None reported.