The Scapular Osteofasciocutaneous Flap: A 12-Year Experience

Objective  To elucidate the factors that play a role in the decision-making process to use the scapular donor site, we have reviewed our 15-year experience with 57 clinical cases, to our knowledge the largest case series to date.

Design  Retrospective, single-surgeon medical record review.

Patients and Methods  Retrospective review of 57 consecutive cases (53 patients) involving mandibular and maxillary reconstruction using bone-containing scapular free flaps over a 15-year period. Composite flap composition as well as donor and recipient site complications were recorded.

Results  Forty-one reconstructions were performed for mandibular defects, 11 were performed for maxillary defects, and 5 for combined defects involving the mandible and maxilla. Seven flaps were composed of 2 separate bone flaps using the angular branch and the circumflex scapular artery. A total of 6 flaps were failures in 5 patients, giving an overall success rate of 89%.

Conclusions  The subscapular system of flaps is a versatile donor site that offers distinct advantages in the older patient population as well as in patients with a preexisting gait disturbance. It is particularly advantageous in patients requiring a large surface area of soft tissue to restore their defect.

THE SCAPULAR osteofasciocutaneous flap is unique for its diversity of available skin, bone, and muscle, as well as the mobility of the soft tissue components relative to the bone. The complex vascular anatomy associated with the axillary region has resulted in a variety of flaps, all of which are based on a single vascular pedicle (Table 1).¹ Baudet et al² described the first successful transfer of the lateral thoracic flap, one of the earliest flaps harvested from the axillary region. Subsequently, dos Santos³ defined the subscapular vascular anatomy through a series of cadaveric dissections and injections of dye, resulting in the clinical application of free scapular osteocutaneous flap by Teot et al.⁴ It was not until Swartz et al⁵ published their experience with 26 clinical cases that the scapular donor site became a popular source of vascularized bone for head and neck reconstruction. Swartz et al demonstrated the ability to reliably harvest vascularized bone, multiple cutaneous paddles, and vascularized muscle, as a single composite free flap. Swartz et al applied the scapular free flap to 5 maxillary reconstructions and 21 composite mandibular defects. The mobility of the soft tissue relative to the bone facilitated the complex 3-dimensional reconstruction of a bilateral maxillectomy defect, an orbital floor defect, and a hemipalate defect. For the first time, complex composite defects could be restored in a single stage using bone from the lateral scapular border. Similarly, Swartz et al demonstrated the applicability of the scapula for the reconstruction of composite defects of the mandible by using the latissimus dorsi muscle with a split-thickness skin graft for an external cutaneous defect and the skin paddle for intraoral reconstruction.

The unique ability to manipulate the vascularized bone graft relative to the skin paddle was further enhanced in 1991 by Coleman and Sultan,⁶ who described an alternative blood supply to the tip of the scapula. For the first time, Coleman and Sultan demonstrated that the angular artery, a branch of either the thoracodorsal artery or the artery to the serratus anterior muscle, enabled the harvest of 2 separate bone segments based on 2 separate branches of the subscapular vascular system. Coleman and Sultan found that the ability to harvest independently vascularized scapular tip not only enhanced mandibular reconstruction but also proved highly useful in the reconstruction of orbital and maxillary defects.

Shortly after the report of 26 cases of osseous reconstruction in 1986 by Swartz et al, there were several smaller series reported in the late 1980s (Table 2). Aside from the experience by Sullivan et al¹⁰ with 31 cases in 1990, and Coleman and Sultan's report of 22 cases 1 year later, the scapular flap waned in popularity, relegated as a secondary and even tertiary choice for osseous reconstruction of the head and neck. The introduction of alternative bone-containing free flap donor sites, such as the iliac crest and fibula, is largely responsible for this decline. Extensive experience with the iliac crest-internal oblique osteomusculocutaneous free flap by Urken and colleagues^17,18 and application of the fibular osteocutaneous free flap to head and neck reconstruction by Hidalgo and Rekow^19,20 popularized 2 new sources of vascularized bone for head and neck reconstruction. Unlike the scapular donor site, the vascular anatomy of the iliac crest and fibula are consistent so that harvesting from these sites is straightforward.^21,22 In addition, the iliac and the fibular donor sites allow for a 2-team approach, thereby decreasing operative time. As the fibula and iliac became increasingly popular in the late 1980s and early 1990s, the shortcomings associated with the scapular donor site led to a decrease in its popularity. This is reflected in the substantial decline, over the past decade, in the number of reported series using the osteofasciocutaneous scapular flap for maxillary and mandibular reconstruction (Table 2).

We have continued to use the scapular donor site as an important tool for reconstruction of the upper and lower jaws. In an effort to elucidate the factors that play a role in the decision-making process to use this donor site, we have reviewed our 15-year experience with 57 clinical cases, to our knowledge the largest case series to date. We have retrospectively examined patient-related factors and defect-related factors to help define the current indications for the use of this flap in contemporary reconstruction of the mandible and maxilla.

A retrospective review was performed of 57 consecutive cases (53 patients) of mandibular and maxillary reconstruction using bone-containing scapular free flaps over a 15-year period. All of the cases reviewed were performed at Mount Sinai Medical Center, Department of Otolaryngology–Head and Neck Surgery, New York, NY, a tertiary referral center for otolaryngology. Any patient who had undergone a reconstruction of the mandible, maxilla, or both using a bone-containing scapular free flap was included in this study. Medical records were reviewed for age, sex, pathologic diagnosis, preexisting comorbidities, donor site and recipient site complications, and factors related to donor site choice. Defect characteristics were reviewed, including the extent of the soft tissue defect, the extent of the bony defect, and the location of these defects.

A 74-year-old man had had a history of peripheral vascular disease and biopsy-proven squamous cell carcinoma of the right mandibular alveolus. The patient underwent a composite resection (Figure 1) of the right mandibular body and floor of mouth and bilateral neck dissections. Because of the patient's history of significant peripheral vascular disease, he was not a candidate for a fibula free flap. In an effort to expedite postoperative ambulation, he was also not considered as a candidate for an iliac crest free flap. Consequently, a scapular free flap was chosen for the reconstruction. Because of the patient's body habitus and excessive subcutaneous adipose tissue, a skin paddle was undesirable for the intraoral reconstruction. Instead, the lateral border of the scapula was used to reconstruct the mandibular defect and the intraoral reconstruction was achieved with the dorsal thoracic fascia, based on the circumflex scapular artery and vein (Figure 2). Five weeks postoperatively, the dorsal thoracic fascia has begun to mucosalize intraorally (Figure 3), and the patient has achieved an acceptable cosmetic result (Figure 4). This case demonstrates the option of incorporating the vascularized dorsal thoracic fascia for intraoral reconstruction when the body habitus and skin paddle thickness are undesirable for intraoral lining.

A 52-year-old man had had a history of an extensive basal cell carcinoma of the left temporal region, involving the temporal skin and the ascending ramus of the mandible (Figure 5). The patient underwent a left lateral temporal bone resection, resection of the left ascending ramus and condyle of the mandible, and extensive resection of the temporal skin and deep soft tissues of the temporal region and the facial nerve. An 8 × 15-cm parascapular skin flap was raised with a vascularized bone segment for the restoration of the mandibular defect (Figure 6). The parascapular flap was deepithelialized and used to restore form lost as a result of the infratemporal fossa resection. A 20 × 8-cm myocutaneous latissimus dorsi flap was raised and used to resurface an extensive cutaneous scalp defect. A segment of the thoracodorsal nerve was used as a facial nerve graft. Osseointegrated implants were placed in the remaining temporal bone for the placement of an auricular prosthesis.

This case demonstrates the versatility of the subscapular system of flaps. Only the scapular donor site allows for the harvest of a vascularized bone segment and 2 separate large skin flaps. An extensive latissimus skin paddle can be reliably harvested while the deepithelialized scapular flap is used to augment the underlying soft tissue defect. Postoperatively, the patient regained facial symmetry and contour as well as complete coverage of the scalp defect (Figure 7).

A 72-year-old active female tennis player had a squamous cell carcinoma of the right alveolar ridge involving the body and retromolar trigone of the mandible. Because of this patient's active lifestyle and desire to continue playing tennis, the scapular donor site was chosen. The patient is right-handed so a left-sided scapular flap was used for the reconstruction of a right-sided composite defect of the mandible and floor of mouth (Figure 8). The patient had an excellent aesthetic and functional result (Figure 9), and by 10 weeks postoperatively the patient was able to participate in tennis without a significant residual donor site deficit (Figure 10).

This case demonstrates that the scapular donor site offers several unique advantages over the iliac crest and fibular donor sites. Because bone is not harvested from the lower extremity, patients are able to ambulate earlier in the postoperative course, decreasing the chance of a deep vein thrombosis and pulmonary embolus.

A 67-year-old man who had a history significant for childhood poliomyelitis currently suffers from postpoliomyelitis syndrome. He relies on 2 canes and a body brace for ambulation. This patient was initially seen with mandibular osteoradionecrosis and an exposed reconstruction plate (Figure 11), after being previously treated for squamous cell carcinoma of the floor of mouth and mandible with a surgical resection and external beam irradiation. Because of this patient's profound gait disturbance, he was not considered a candidate for an iliac crest or fibular free flap reconstruction. A scapular free flap was raised consisting of a vascularized bone segment and a myocutaneous latissimus dorsi flap (Figure 12). The vascualrized bone segment was used to reconstruct the mandibular defect while the latissimus dorsi muscle was placed into the neck and over the mandibular reconstruction. The skin paddle was deepithelialized and used to augment the soft tissue deficit.

This patient is presently 3 years postoperative, ambulating with 2 canes and a body brace without difficulty (Figure 13). This case demonstrates that harvesting from the scapular donor site did not destabilize the shoulder nor did it impede this patient's ability to ambulate effectively. However, patients with a preexisting gait disturbance may not compensate after harvest of a composite free flap from the lower extremity or pelvic girdle.

From March 1, 1987, to April 30, 2000, 53 patients underwent 57 bone-containing scapular free flap reconstructions. Patients ranged in age from 16 to 89 years (mean age, 57 years). There were 24 female and 29 male patients with pathologic diagnoses including squamous cell carcinoma (25 patients), osteoradiorecrosis (11 patients), sarcoma (11 patients), salivary gland carcinoma (2 patients), ameloblastoma (1 patient), fibroma (1 patient), hemangioma (1 patient), and posttraumatic deformity (1 patient). Forty-one reconstructions were performed for mandibular defects, 11 were performed for maxillary defects, and 5 were performed for combined defects of the mandible and maxilla.

All 57 scapular free flaps consisted of the lateral border of the scapula, and 7 flaps were composed of 2 separate bone flaps based on the circumflex scapular arterial perforators and the angular artery. In 4 of these cases, the scapular tip was used to reconstruct the palate, maxilla, or inferior orbital floor. Latissimus dorsi muscle was harvested as part of a composite flap in 26 cases. Three of these muscle flaps were reinnervated (thoracodorsal nerve to facial nerve) as part of a facial reanimation procedure. Nine scapular free flaps were harvested with 2 separate skin paddles (scapular and parascapular), and 4 flaps were harvested as a vascularized bone graft without a skin paddle. Osseointegrated dental implants were placed primarily in 6 patients and secondarily in 4 patients.

There were 3 recipient site wound infections, which were treated successfully with intravenous antibiotic agents and local wound care. There were 2 donor site wound infections; one required intraoperative debridement and the second was treated successfully with local wound care. Two of 57 free flaps required postoperative debridement of the skin paddle; however the bone remained vascularized, and there were 6 flap failures in 5 patients (overall flap success of 89%). Three of the 5 patients who sustained flap failure had multiple medical problems including peripheral vascular disease, documented coronary artery disease, and chronic obstructive pulmonary disease.

The application of microvascular free tissue transfer to the reconstruction of oromandibular defects has revolutionized the management of oral cancer. The early phase of this era saw an explosion in the number and variety of donor sites from which to harvest vascularized bone containing free flaps. With the introduction of each new donor site, the enthusiasm curve for each of these composite flaps started off at a feverish pitch. As experience with each of these composite flaps grew, a more reasoned view of the advantages and disadvantages became apparent and allowed surgeons to place these donor sites in their appropriate position relative to available reconstructive options. Multiple factors are at play in determining the utility of a particular composite flap donor site (Table 3).

Excluded from Table 3 is the surgeon's facility and comfort with the dissection of that donor site, as that is something that can readily be learned and mastered. As experience with a particular composite flap grows, these factors become more evident and the enthusiasm for a particular donor site can usually be reliably measured by the number and frequency of articles published in the literature. Since its introduction into the head and neck surgeon's reconstructive options, the subscapular system has been recognized for the tremendous range of soft tissue flaps that can be transferred and their 3-dimensional mobility relative to the bone. It is safe to say that this donor site offers the most versatility of any donor site. However, the inconvenience of patient positioning during harvest has led most surgeons to use other donor sites far more frequently, with a growing sense of abandonment of the subscapular system. It is for this reason that we chose to review our series of oromandibular reconstructions using this flap, which, to our knowledge, represents the largest series to date published in the literature (Table 2).

It is our belief that there are certain clinical situations where this donor site emerges as the flap of choice to restore oromandibular defects. Some of these situations are patient related, while others are dictated by the particular defect. The age range for patients afflicted with oral cancer often requires that the head and neck surgeon render treatment to patients in the seventh, eighth, and occasionally in the ninth decades of life. Patients in this age group are more difficult to mobilize after surgery for all of the reasons associated with the aging process, in addition to the comorbidities, which often coexist with the disease in the oral cavity. It is our belief that one of the primary goals in patients who undergo surgery in the latter years of life should be to mobilize them quickly in the postoperative period to help minimize further complications. Harvest of a composite flap from the subscapular system permits a far more rapid overall rehabilitation of patients than the harvest of a fibular or iliac flap. The same philosophy holds true for the patient of any age who is seen with an existing gait disturbance in whom the harvest of a composite flap from either the fibula or the ilium may compromise a fragile balance that could prove extremely difficult to rehabilitate to achieve the same level of preoperative ambulation.

Severe peripheral vascular disease identified on preoperative history or on vascular imaging, or the presence of vascular anomalies of the lower extremities, serves as a contraindication to fibular flap harvest and points toward the subscapular system as a safer donor site. Although vascular anomalies of the subscapular system are found in particular on the venous side,²¹ it is rare for the vessels of the subscapular system to be adversely affected by atherosclerosis.

There are certain features of a particular oromandibular defect that we believe lend themselves to the use of a subscapular composite flap. The range of soft tissue types and the maneuverability relative to the bone make it particularly conducive to the more complex defects of the oromandibular region. The potential surface area of the soft tissue flaps, which can be transferred, is far greater than any other available donor site. In heavyset patients with more subcutaneous fat throughout their body, cutaneous flaps are often too thick for use in the oral cavity. The subscapular system provides an opportunity to harvest either the latissimus dorsi muscle or the thoracolumbar fascia, which are both well vascularized and significantly thinner soft tissue flaps in virtually all patients.^2,23 The angular branch arising from the thoracodorsal artery or the branch to the serratus anterior muscle offers a unique opportunity to harvest a second independent vascularized segment of bone from the scapular tip. First described by Deraemaecker et al²⁴ in 1988 and later applied clinically by Coleman and Sultan,⁶ the ability to harvest 2 segments of bone that are vascularized by independent branches of the parent artery and vein is a truly unique feature of the subscapular donor site. While there are rare instances that one would need 2 independent segments of vascularized bone in oromandibular reconstruction, there are distinct advantages of a second blood supply to the scapular tip. The creation of osteotomies along the lateral scapular border is often necessary to appropriately contour the scapular border to the shape of the missing segment of the mandible. The vascular supply to the distal segments of that bone is derived from the muscle and periosteal circulation, which must be left attached to the bone while creating these osteotomies. In the original article by Swartz et al⁵ concerning application of the subscapular composite flap to oromandibular reconstruction, they reported several postoperative bone scans in which the distal segments of bone, following the creation of contouring osteotomies, did not appear to be well vascularized. The incorporation of the angular branch in situations where multiple osteotomies of the lateral scapular border are required would help to ensure the vascular supply to the entire neomandible. The length and caliber of the vascular pedicle of the subscapular flap are usually suitable for most head and neck reconstructions. However, there are clinical situations in which the recipient vessels in the neck are deficient, and in these, transfer of the scapular tip, based on the angular branch, offers a unique advantage for reconstruction. The significant additional pedicle length afforded by transferring vascularized bone supplied by the angular artery helped to avoid the use of vein grafts in one such patient in this series. The transfer of an independent segment of vascularized bone based on the angular artery also provides the maximum separation and freedom of maneuverability of the scapular and parascapular soft tissue flaps relative to the bone. There are select clinical situations in which enhanced flexibility may be advantageous, in comparison with the usual more limited relationship of the scapular and parascapular flaps to the lateral scapular border supplied by the circumflex scapular artery and vein. In the more complex palatomaxillary reconstructions, we have used 2 independent segments of bone, based on the circumflex scapular and angular branches, to achieve a more accurate restoration of this 3-dimensional anatomy. In these reconstructions, we have used the scapular tip as a horizontal shelf of the palate and the lateral scapular border to restore the vertical dimension of the maxillary buttress.²⁵ In such situations, the complete freedom of maneuverability of these 2 independent bone segments is helpful to achieve a 3-dimensional anatomical restoration of the palatomaxillary defect.

In 1994, we conducted a study to evaluate the bone stock that is available for placement of dental implants in the most common donor sites used to harvest vascularized bone-containing free flaps.²⁶ In that series we evaluated 28 cadaveric specimens and determined the mean index of implantability, that is, an index derived from height, width, and cross-sectional area associated with each donor site. The results of that study revealed that the scapular donor site ranked second behind the iliac crest for reliability of the bone stock that can be harvested and its suitability for placement of dental implants. In the current series, there were few patients undergoing oromandibular reconstruction with the scapular composite flap in whom dental rehabilitation using dental implants was achieved. The decision to use osteointegrated implants is a multifactorial one that involves not only the availability of sufficient bone stock, but also the motivation of the patients and the anticipated functional benefit to the patient with respect to the ability to achieve functional mastication following what is frequently multimodality cancer therapy. In addition, financial means to afford prosthetic restoration is also a necessary factor that must be included in the decision-making process. Many of the patients in this series were older or had more advanced disease, which made them less suitable candidates for dental rehabilitation. In analyzing the current series, it is believed that these factors were more important in explaining the relative scarcity of patients with restored dental implants compared with our larger series using the iliac and fibular donor sites as well.

In this series, the free flap failure rate was 11%. Most failures occurred early in our experience with the subscapular system of free flaps. Since gaining experience with this donor site, our success rate has vastly improved.

We have reported the largest series of scapular composite free flaps used in head and neck reconstruction. With the senior author's (M.L.U.) experience in performing more than 300 composite free flaps for oromandibular reconstruction, the current perspectives on the role of the subscapular system in contemporary head and neck surgery are provided. Despite the inconvenience of patient positioning during surgery, there are very defined indications in which the subscapular flap remains the reconstructive method of choice to restore the mandibular infrastructure in combination with the oral lining and for the overlying skin.

Accepted for publication April 6, 2001.

Presented at the annual meeting of the American Head and Neck Society, Fifth International Conference on Head and Neck Cancer, San Francisco, Calif, August 1, 2000.

Corresponding author and reprints: Mark L. Urken, MD, Box 1189, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029 (e-mail: mark.urken@mssm.edu).