A, Multiply recurrent melanoma of the scalp (arrows). B, Scalp resection, complete with superficial parotidectomy and upper (levels I, II, and III) neck dissection. C, Latissimus dorsi free flap (muscle only) repair. D, Six months postoperatively.
Beasley NJP, Gilbert RW, Gullane PJ, Brown DH, Irish JC, Neligan PC. Scalp and Forehead Reconstruction Using Free Revascularized Tissue Transfer. Arch Facial Plast Surg. 2004;6(1):16-20. doi:10.1001/archfaci.6.1.16
From the Wharton Head and Neck Centre, Princess Margaret Hospital, Toronto, Ontario.
Objective To examine the indications for, and the success of, free flap reconstruction in patients with forehead and scalp defects.
Design Case series.
Setting Two tertiary referral university teaching hospitals.
Patients Twenty-six consecutive patients, aged 31 to 85 years, presenting with 26 scalp defects, 5 forehead defects, and 1 combined defect (size, 70-672 cm2). Three patients required resection and repair of the dura at surgery.
Intervention Patients were staged according to the size of the defect and the viability of surrounding tissue; free flap reconstruction was performed where indicated.
Main Outcome Measures Flap survival, complications, and disease-free and overall survival.
Results Thirty-four free flap reconstructions were performed (24 latissimus dorsi free flaps, 4 scapular free flaps, 3 rectus abdominis free flaps, and 3 radial forearm free flaps). One failed 2 weeks postoperatively, and 2 required exploration (1 for arterial ischemia and 1 for a hematoma). There were 3 cases of donor site morbidity (2 early seromas and 1 late abdominal hernia). One patient died of a pulmonary embolus 1 week postoperatively. Disease-free survival was 48% at 5 years and overall survival was 59% at 5 years, with a median follow-up of 24 months.
Conclusions Free revascularized tissue transfer is a reliable and safe way of reconstructing large scalp or forehead defects after traumatic injury or neoplastic resection. The muscle-only latissimus dorsi free flap for scalp reconstruction and the cutaneous scapular free flap for the forehead have proved successful in selected patients with a low complication rate and satisfactory cosmesis.
Reconstruction of large scalp and forehead defects presents a particularly challenging clinical problem. While the simplest and most reliable method of reconstruction should be considered in all patients, there are a number in whom the size of the defect, the presence of infection, or previous radiation and surgery may indicate the need for a more radical approach. Scalp and forehead defects may be caused by trauma, burns, benign or malignant tumor resection, osteomyelitis, osteoradionecrosis, or congenital lesions.1-4 Some of these defects will just involve skin superficial to the pericranium; others will involve the entire thickness of the soft tissue of the scalp, including calvarial bone and dura.1, 3
The approach to scalp and forehead reconstruction starts with a careful examination of the patient and the potential defect. This includes assessment of the location and size of the defect, with radiological assessment of the depth and bony or dural defects.4 Previous surgery or radiation or ongoing infection, such as osteomyelitis, will affect the viability of the skin around the defect, reducing the available tissue for local flaps. The pathological features of the presenting lesion are important in planning the reconstruction.1
This article outlines our approach to scalp reconstruction based on our experience during the past 9 years, and describes a staging system for forehead and scalp defects that aids flap selection (Table 1). This system is based on the cause of the defect, previous treatment, planned future treatment, and the size of the defect. A series of 26 patients undergoing 34 free flap reconstructions are reviewed; outcome measures are given.
The medical records of 26 consecutive patients undergoing 34 free tissue transfers for large scalp and forehead defects between July 1, 1992, and June 30, 2001, at the Toronto General Hospital and the Toronto Sunnybrook Regional Cancer Centre, Toronto, were examined. Details were collected on the patients' age, sex, pathological features, previous treatment, and comorbidities. The defect was staged according to the system described in Table 1; the type of reconstruction, vessels used, operative time, hospital stay, perioperative complications, length of follow-up, and outcome were noted. Disease-free and overall survival were calculated actuarially, and survival curves were constructed using the Kaplan-Meier method.
Of the 26 patients, 19 were men and 7 were women; their median age was 69 years (range, 31-85 years). The original presenting pathological features included malignant disease in 22 patients (8 squamous cell carcinomas, 2 basal cell carcinomas, 4 malignant fibrous histiocytomas, 2 dermatofibrosarcomas, 2 angiosarcomas, 1 Merkel cell carcinoma, and 3 melanomas). Twelve of these patients previously underwent scalp and/or forehead radiation, and 14 previously underwent surgery. There were only 4 patients with malignant disease who had not undergone previous treatment. Four patients had benign or nontumor disease. There was 1 patient with a meningioma, and 1 with an arteriovenous malformation; both previously underwent surgery. There was 1 patient with a full-thickness electrical burn to the scalp in whom the vascularity of the calvarial bone was poor. In this patient, the dead bone was not resected but covered successfully with a muscle-only latissimus dorsi free flap. The last patient had herpes zoster, and had scratched through all the layers of the scalp, exposing a sizable area of bone.5
Four patients required a second resection and free flap reconstruction (1 for a flap failure, 1 for recurrent basal cell carcinoma, 1 for recurrent malignant fibrous histiocytoma, and 1 for osteoradionecrosis following a previous resection and radiation for a malignant fibrous histiocytoma). Two of these patients required a third resection and free flap reconstruction (1 for recurrent malignant fibrous histiocytoma and 1 for recurrent angiosarcoma).
The median size of the scalp defect was 225 cm2 (range, 70-672 cm2). All patients had a significant area of bone exposure. Five patients had dural exposure, and 3 required resection and repair of the dura at surgery.
There were 34 free flap reconstructions; the stage grouping of these patients and the type of reconstruction used are summarized in Table 2. All muscle-only flaps were covered with 1.4-mm-thick split-thickness skin grafts taken from the anterolateral thigh. The superficial temporal artery was used in 22 patients, the facial artery in 6, the occipital artery in 1, the external carotid artery in 1, the superior thyroid artery in 2, a branch of the transverse cervical artery via a vein graft in 1, and a previously connected thoracodorsal artery from the first free flap reconstruction in 1. For the venous anastomosis, the following were used: 21 superficial temporal veins, 1 occipital vein, 1 external jugular vein, 4 facial veins, 5 internal jugular veins (1 via a vein loop), 1 branch of the transverse cervical vein via a vein graft, and 1 previously connected thoracodorsal vein. The median operative time, including resection, was 10 hours (range, 5-17 hours); the median hospital stay was 10 days (range, 4-54 days).
One latissimus free flap failed 2 weeks postoperatively after the patient was discharged from the hospital. The flap was grossly infected when the patient returned to the hospital. It was not possible to determine whether the infection was the primary event causing this failure or if it was secondary to necrosis of the flap. The dead flap was resected, and a second latissimus dorsi free flap reconstruction was performed successfully. One rectus abdominis free flap required operative exploration because of arterial ischemia; this was probably due to a drain compressing the arterial pedicle. A hematoma formed under a latissimus dorsi free flap in 1 patient, requiring operative drainage. Two patients developed seromas at the donor site after latissimus dorsi free flap harvest; 1 became infected. Both were treated with drainage and antibiotics. There were 2 cerebrospinal fluid leaks after scalp reconstruction involving dural resection; both were successfully repaired. One patient died 1 week postoperatively of a pulmonary embolus despite appropriate deep venous thrombosis prophylaxis. There was 1 late complication, with a patient developing a hernia at the donor site after harvest of a rectus abdominis free flap.
The 3 patients with nontumor disease remain alive and well, with no evidence of pathological features, and the patient with a meningioma is alive, with residual disease. Of the 22 patients with malignant disease, 10 are alive and disease free, 4 died of other causes, and 7 died of disease; there were no data on the follow-up of 1 patient. The median follow-up for all patients was 24 months (range, 1 week to 108 months). When the data on the patients with malignant disease were analyzed actuarially using the Kaplan-Meier method, disease-free survival was 48% at 5 years and overall survival was 59% at 5 years.
A 69-year-old man presented with multiple small recurrences of a previously resected scalp melanoma (Figure 1, A). The defect was estimated at 300 cm2 preoperatively and was a stage II lesion. At surgery, a wide area of scalp was resected, with the superficial lobe of the parotid and the upper neck dissected (levels I, II, and III) (Figure 1, B). A free latissimus dorsi muscle–only flap was used to repair the defect (Figure 1, C). A split-thickness skin graft taken from the anterolateral thigh was then used to cover the muscle. He had a satisfactory cosmetic result 6 months postoperatively (Figure 1, D), but died of distant metastases 12 months postoperatively.
The simplest possible method of reconstruction should be considered in all patients while ensuring adequate resection of the lesion and a good functional result. For defects that are less than 3 cm in diameter, primary closure is usually achieved with undermining of the remaining forehead or scalp where necessary.4, 6 For larger defects in which the pericranium is intact, a split-thickness skin graft may be adequate to cover the defect. This allows for early detection of tumor recurrence, but postoperative radiation is not tolerated well, and it is cosmetically inferior to other methods because the color match is poor and the texture and thickness do not match the rest of the forehead and scalp.4, 7 It is, however, a useful holding maneuver, and later tissue expansion will often allow coverage of the defect.4
Excellent cosmetic results can be achieved with tissue expansion when hair-bearing skin has been lost,7-8 and repeated expansion for large defects is possible.9 The main advantage is the ability to replace one tissue with exactly the same tissue. Tissue expanders have been used in patients with scalp carcinomas,10 but are usually reserved for low-grade malignancies, such as basal cell carcinoma, or for secondary replacement of tissue because the time required for the tissue expansion to occur would otherwise delay the surgery for the primary tumor.1 Tissue expansion is not generally recommended for infected cases because of the risk of the expander becoming infected and the whole expanded flap being lost, nor is it recommended in radiated skin because of decreased compliance and a higher rate of skin necrosis.11
Various types of pericranial flaps, based on the supraorbital and supratrochlear arteries, can be used to cover a wide area of exposed bone, sometime in combination with local flaps and split-thickness skin grafts.12-13 Galeal and temporoparietal fascial flaps have a wide arc of rotation, a good blood supply, and minimal donor site morbidity. They conform well to the underlying bone, and are of the correct thickness for a good cosmetic result.4 Damage to the hair follicles while raising the flap from the donor site can result in hair loss, and damage to the frontal branch of the facial nerve has been reported.4
Carefully designed local flaps can cover up to 50% of the surface area of the scalp where bone is exposed, including defects of the calvaria and dura.7 For frontal, temporal, or occipital defects, a single flap may be used based on the superficial temporal artery or the occipital artery, while central defects can be covered with multiple rotation or transposition flaps.14-16 Other local flaps have been described, such as the vertical trapezius myocutaneous flaps for large posterior defects of the scalp17 and bipedicled fronto-occipital flaps for large defects of the lateral scalp.18 Regarded as simpler and safer than free flaps, they generally need to be longer and larger than conventional flaps because the scalp is inelastic. Although galeal relaxation incisions can be used, they add only a little to the length of the flap and they risk damage to the blood supply of the flap.7
The staging system proposed in this report indicates those patients in whom free flap reconstruction should be considered. While these numbers are somewhat arbitrary, they are, nevertheless, based on our own experience and on cumulative reports in the literature.3, 16, 18-20 However, to our knowledge, there has been no attempt in the literature to stage or categorize these defects. The staging and treatment algorithm presented herein is an attempt to provide guidelines based on this knowledge. Generally, lesions exposing bone greater than 50 cm2 on the forehead and greater than 200 cm2 on the scalp are candidates for free flap reconstruction because of their size, with those lesions greater than 600 cm2 requiring 2 flaps to cover the defect. Smaller lesions may also be candidates for free flap reconstruction. Factors that impair the viability of tissue surrounding smaller defects may make the use of a free flap necessary. Problems such as heavy trauma, previous radiation, osteoradionecrosis, osteomyelitis, or previous local flap failure6, 20-22 would severely compromise the chance of success of a local flap in this area because wide resection of tissue around the defect may be required.22 Furthermore, aesthetic considerations may lead one to choose a free flap over a local flap that will also require a significant adjacent donor defect. In cases of high-grade malignancy, the extent of resection may be difficult to judge preoperatively, and is usually determined by intraoperative frozen section. The use of a free flap reconstruction allows wide margins to be taken and margins to be revised intraoperatively if positive. These patients often also require high-dose postoperative radiation, and fresh well-vascularized tissue tolerates this well.21-22 If the calvaria is infected, it should be removed and a bony reconstruction performed at a later date.20 Calvarial or dural defects are ideal for reconstruction with free revascularized tissue because of their excellent vascularity; in a recent series23 of patients with skull base defects, free flap reconstruction resulted in better wound healing and a shorter hospital stay. Noninfected bone, even if poorly vascularized, can be left in place and covered with well-vascularized tissue.20
Many flaps have been described for use in forehead and scalp reconstruction. These include the latissimus dorsi, radial forearm, rectus abdominis, scapular, omentum, and Scarpa free adipofascial flaps.1-3,24 These may or may not include bone graft and involve dural repairs. In this series, we favored the use of the cutaneous scapular free flap for forehead reconstruction because of its excellent color match on the face and its thickness and consistency, which match surrounding tissue. Earlier patients in the series using the radial forearm free flap had not had such a good cosmetic result, and donor site morbidity can be significant, particularly when large areas are needed. The muscle-only latissimus dorsi free flap has proved to be an excellent choice for large scalp defects because of its large size, predictable blood supply, ease of harvesting, and excellent vascularity for compromised beds with a low donor site morbidity.19 The long thoracodorsal artery and vein pedicle length are usually sufficient to reach the donor vessels. It is generally used without skin because this is less bulky and contours easily. It is covered with a split-thickness skin graft taken from the anterolateral thigh,25 and preferably not meshed because the meshed skin graft retains the outlines of the meshing. This gives it the appearance of crocodile skin, or chicken skin, and is unsightly (Figure 1, D). Muscle-only rectus abdominis free flaps can be used for smaller defects, and bilateral latissimus dorsi free flaps for particularly large defects. In terms of the difference between scapular skin flaps and latissimus dorsi muscle flaps, the former is a better substitute for forehead skin because of its color, texture, and thickness. While the latissimus flap contours nicely to the forehead, the appearance of the muscle with skin graft is, in our opinion, inferior to the scapular flap. Similarly, the scapular flap seems somewhat thick when applied to the scalp, whereas the latissimus flap atrophies and looks much like normal scalp once the skin graft matures. Furthermore, the donor defect on the back can be closed easily because no skin is harvested. In this way, extremely large defects of the scalp can be reconstructed with minimal donor morbidity. Even though large scapular flaps can be harvested with direct closure of the donor defect, some of these large scalp defects would require grafting of the donor site if they were reconstructed with the scapular flap.
The superficial temporal artery and vein have been reliable, and are used if available. The facial artery and other vessels in the neck were used less frequently because of the increased pedicle length required. On 2 occasions, vein grafts were required, and are occasionally needed when there is an insufficient pedicle length or if the arteries and veins of the neck have been compromised by the ablative resection.2
Major complications were infrequent in this series, with the loss of just 1 flap and the death of 1 otherwise healthy, if elderly, patient 1 week postoperatively of a pulmonary embolus. Other complications included seroma at the donor site and tumor recurrence.2 The outcome after resection of malignant disease indicates that about half the patients remain disease free and survive long-term, with 2 patients alive after recurrence in this series.
In summary, the use of free flaps is a reliable and safe way of reconstructing large or high-risk scalp or forehead defects after traumatic injury or neoplastic resection. The muscle-only latissimus dorsi free flap for scalp reconstruction and the cutaneous scapular free flap for forehead reconstruction have proved successful in selected patients. The complication rate was low, and cosmesis was satisfactory.
Corresponding author: Peter C. Neligan, MB, Wharton Head and Neck Centre, Princess Margaret Hospital, 610 University Ave, Toronto, Ontario, Canada M5G 2M9 (e-mail: firstname.lastname@example.org).
Accepted for publication May 13, 2003.
This study was presented at the meeting of the American Academy of Facial Plastic and Reconstructive Surgery; September 6, 2001; Denver, Colo.