Endonasal endoscopic approach has been increasingly used in the operation of skull base tumors during the last decades. However, in some complex situations, such as advanced intra- and extracranial communicating tumors, open approaches are preferable or even mandatory [7]. In these cases, radical resection of tumors frequently results in complex composite defects involving the dura, brain, and surrounding areas such as the orbit, ear, maxilla, hard palate, and skin. Reliable methods of reconstruction are therefore required to achieve the following goals: (1) separation of the CNS from the external environment and the upper aerodigestive tract; (2) supporting the brain and orbit; (3) obliteration of dead space; (4) replacement of lost skin coverage; (5) restoration of the three-dimensional appearance of bony and soft tissues [8, 6, 12].
The use of local flaps such as temporalis and galeal-pericranial flaps is limited by their inadequate tissue, frequent unavailability and restricted arcs of rotation[10]. Regional pedicled flaps, such as the submental and pectoralis major flaps, have the disadvantage that the distal tip of the flap, and thus the most unreliable, is usually the part of the flap that reaches the intracranial defect [6].
In patients previously operated and/or irradiated and those planned for postoperative radiotherapy, both local and regional flaps are usually unreliable or unavailable.
Radiotherapy have been reported to result in partial or total local flap loss in up to 7–20% of patients [
13]. Therefore, when large three-dimensional composite defects are encountered, microvascular free tissue transfer may be more suitable. Free flap transfer in these cases provides adequate, well-vascularized and robust tissue coverage of the defects, with consequent lower complication rates. It has been reported that the wound related complication of locoregional flaps is much higher than that of free flaps (36% vs. 10%) [
12]. Free flaps do not have the attachment of a pedicle and have flexibility in design. Besides, free flaps have good tolerance to irradiation, and can be harvested according to the types of defect tissue, such as fasciocutaneous or myocutaneous flap [
14]. Resto et al. [
15] suggested pectoralis major flap can be an important alternative to free flap reconstruction of composite lateral skull base defect in selected cases. One patient (#5) in our cohort had a pedicled pectoralis major flap as replacement of necrotic free flap caused of arterial thrombosis.
An ideal free flap for skull base reconstruction should meet the following criteria: adequate tissue volume, availability of diverse tissue types on one pedicle with consistent anatomy, versatility in design, minimal donor-site morbidity, and feasibility of a two-team approach [9]. A variety of free flaps including anterolateral thigh flap (ALT), rectus abdominis flap (RAM), radial forearm flap, latissimus dorsi flap and fibula flap for craniofacial reconstruction have been reported [16, 9, 11, 17–21]. In earlier reviews, the RAM was the workhorse for microvascular flap reconstruction of skull base due to the large volume of soft tissue that it provides, allowing for better obliteration of complex dead space [16]. The advantages of the RAM also include excellent vascularity with long vascular pedicle of up to 15 cm, versatility to be used with or without skin paddle, primary closure of the donor site and simultaneous harvesting in a two team setting [10, 16]. Disadvantages of this flap include large abdominal scar, bulkiness sometimes and susceptibility of patients to postoperative hernia formation [18]. Currently at our center, the use of the RAM has largely been replaced by the ALT as it has many of the benefits of the RAM without the donor site morbidity. In the present series, the ALT flap was used in all but 2 of the cases to reconstruct composite skull base defects. The ALT flap has been described more frequently in the last decades’ publications [3]. The ALT flap, based on the descending branch of the lateral circumflex femoral artery, has many advantages of RAM flap such as reliability, long vascular pedicles, versatility in the components and two team approach. The thigh donor site can be closed primarily and does not violate functional motor unit and thus leads to minimal morbidity [17]. The thigh scar can be more cosmetically acceptable to patients. The use of this flap also avoids a separate donor site for harvesting of fat strips and fascia grafts. The main disadvantage to this flap is the anatomic variability of the perforating vessels. This has become less problematic with more experience of the flap harvest gained.
In the last decade, the radial forearm flap, latissimus dorsi flap and fibula flap were rarely used. The radial forearm flap is harvested as a fasciocutaneous flap making it pliable and thin. This makes it criticized for inadequate tissue volume to repair a large three-dimensional defect. The donor site requires reconstruction with split thickness skin grafting, and patients may experience functional deficits in the forearm [10]. However, it has been reported that the radial forearm flap could be reliable when used in a double-layer fashion in large skull base defect [22]. Although the latissimus dorsi flap has large amount of skin and potential muscle volume for microvascular transfer, it is not practical that the patient needs to reposition twice intraoperatively [16]. The fibula flap has become the primary flap for bony reconstruction of the head and neck since 1986, especially when postoperative radiation is expected [16, 4]. However, osseous defects left by resection of skull base tumors rarely require hard support [23]. Considering that most patients in our cohort need postoperative adjuvant radiotherapy or have received radiotherapy previously, we believe that soft tissue reconstruction is far more important than osseous reconstruction. A new classification concept of skull base defects has been raised by Yano et al. [24], and they supposed that defects of class Ib, IIb or defects of the skin or orbital contents and combined defects should be reconstructed with a free flap. Rosenthal et al. [25] proposed a classification of periauricular defects that class I should be reconstructed with regional rotation flap or radial forearm flap, class II with ALT or radial forearm flap and class III with ALT or RAM flap.
Many publications have reported postoperative complications of free flap for skull base reconstruction and its risk factors. The overall complication rate of a variety of the free flaps is 10–32%, and partial or complete flap failure occurs in 2–9% [3, 25, 26, 12, 7, 27]. The overall incidence of complications in this series was 17.6% and the rate of flap loss was 5.9%, which compared well with the literature. The use of free flaps has significantly decreased postoperative nervous complications, and reduced the incidence of CSF leaks from 25 to 6.5% [7].
However, no CSF leaks occurred in our cohort, which may be due to our careful closure of the dural defects (fascia lata graft, fat strips and vascularized flap).
It has been reported that history of radiation therapy, medical comorbidities and the extent of intracranial tumor extension were independent predictors of complication rate [
26]. Thompson et al. [
28] supposed that there was a higher wound complication rate in those with a history of smoking and diabetes. Postoperative facial fistula and recipient-site infection occurred mainly in low-BMI patients, reported by Heo et al. [
29]. Flap crisis included infection, arterial and venous thrombosis, and vascular thrombosis remains the primary reason of flap failure [
30]. In the series, one patient (#5) developed both ipsilateral cerebral infarction and flap crisis with clinical manifestations in skin temperature, color and capillary refill on the third postoperative day. After surgical exploration followed by a second flap transfer and anticoagulant therapy, hemiplegia was improved and the incision healed at the time of discharge. Yang et al. [
30] noted that free flap loss was 6.2-times more likely to develop in patients undergoing surgical exploration after 72 hours. So, the most important thing is to identify signs of flap crisis as early as possible, and then explore without hesitation. Llorente et al. [
7] supposed that all patients should receive prophylaxis with anticoagulant therapy.
In the past, in cases of advanced skull base tumors, the indication for radical surgery is limited since an adequate reconstruction is usually hard to perform. At present, free flaps transfer for skull base reconstruction allow much wider and radical tumor resections, early radiotherapy and effective chemotherapy [19]. Llorente et al. [7] reported that the 5-year local disease control rate was 43%, and the 5-year disease-specific survival rate was 36%. In this series, the median PFS was 31 months and the mean OS was 66 months. As salvage surgeries for advanced skull base tumors, the outcomes are acceptable and satisfying. Besides, the removal of the tumor reduced the patient's pain (both physical and mental) and resulted in a better quality of life over the patients’ lifetime.