Facial Artery Myomucosal Flap vs. Islanded Facial Artery Myomucosal Flap Viability: A Systematic Review

The aim of this study was to estimate the overall viability of the islanded facial artery myomucosal flap (iFAMM) compared to the facial artery myomucosal flap (FAMM). A systematic review of English articles was performed on PubMed and Cochrane Library electronic databases. Search terms included every nomenclature variant for FAMM flap and iFAMM flap. A total of 373 articles were identified, and 20 articles were considered eligible for inclusion in the review. A total of 486 flaps were evaluated (350 FAMM and 136 i-FAMM flaps). In all studies, the primary outcome assessed was flap viability, collecting the rate of total and partial flap necrosis and then comparing FAMM to i-FAMM. Overall reported partial/total necrosis rate for FAMM flap was 9.7%, 1.4% as total and 8.3% as partial necrosis. Overall partial/total reported necrosis rate for iFAMM flaps was 2.2%, 1.5% as total and 0.7% as partial necrosis. FAMM flaps, both as classical or islanded variants, are an effective option for intraoral small/medium sized defect reconstruction. Outcomes from the present review show a higher partial/total survival rate when this flap is harvested as islanded flap.


Rationale
Intraoral mucosal defects can be related to several pathological states such as cancer ablation, palatal cleft and oronasal fistula following conventional closure. Several surgical techniques, involving local, loco-regional and free flap have been described in literature in order to reconstruct intraoral defects in areas such as the tongue, the soft and hard palate, the floor of the mouth and lips [1][2][3]. Microsurgical tissue transfer is often identified as the gold standard in surgical reconstruction, although it requires prolonged operating time, longer hospitalization and, especially for intraoral defects, makes "like-with-like" reconstruction difficult. In cases of small or medium sized intraoral defects, the use of cheek mucosa flap was shown to be effective. Bozola et al. was the first to describe in 1989 an axial buccal mucosal flap based on the buccal artery for the resurface of oral mucosa defects [4]. In 1992, Pribaz combined the principles of nasolabial and buccal mucosa flaps in order to describe an axial musculomucosal flap designed on the facial artery: the facial artery myomucosal flap, also known as FAMM flap [5]. Although similar in concept to Bozola's axial flap, the course of the facial artery, and therefore the orientation of the entire FAMM flap, has several advantages over buccal artery based flaps, allowing the FAMM flap to be more versatile in closing both upper (palate and lateral oropharinx wall) and lower (floor of the mouth, tongue, etc.) intraoral defects. Moreover, the direct and indirect artery flow variants are additional surgical choices that increment the flap versatility. One of the biggest disadvantages of the FAMM flap is the need for tooth extraction or bite block in order to avoid pedicle bite before the second surgical step, usually performed 3 weeks after the preceding surgery, represented by the section of the pedicle and flap remodeling.
In 1999, Zhao et al. described the "island" variant of the FAMM, the so-called i-FAMM flap [6]. This flap has the advantage to avoid the need for temporary bite block after the operation and the need for late pedicle division.

Objectives
In the present systematic literature review, we evaluated flap viability comparing FAMM to i-FAMM flap in oral reconstruction.

Protocol and Registration
The methods and the inclusion criteria of this work were specified in advance and documented in a protocol, according to quality standards described in the PRISMA 2009 checklist [7]. The review was registered in the CRD York website PROSPERO (protocol number CRD42020210349).

Eligibility Criteria
The following focus question was developed according to the population, intervention, comparison and outcome (PICO) study design: In patients undergoing oral reconstruction surgery (P), does i-FAMM flap (I) compared to FAMM flap (C) have better viability (O)?

Information Sources
The research was carried out on electronic PubMed and Cochrane Library databases identifying articles from September 1992 to September 2020. The search was conducted up to September 4th, 2020. Article language was limited to English using database-supplied filters.

Search
The keywords were used and combined with Boolean operators, adapted for every database, both as text words and Medical Search Headings (MeSH terms) as follows:

Study Selection
The full texts of all possibly relevant studies were selected considering the following inclusion criteria: techniques studies; retrospective studies; prospective studies; articles written in English. Exclusion criteria were: articles where outcomes related to flap viability were not reported; articles where numbers of patients were not described; articles based on buccal flap harvested with no pedicle identification; cadaveric or animal studies. Case reports and case series with fewer than 10 patients were excluded due to the insufficient information provided by a limited number of patients. Review articles were excluded, but their reference lists were examined to identify other potentially pertinent studies; editorials, letters and commentaries were excluded. Two reviewers (R.F., G.L.G.) performed eligibility assessment independently. Disagreements between reviewers were resolved by consensus. When consensus was not reached, a senior member mediated (R.R.).

Data Collection Process
Two reviewers (R.F., G.L.G.) performed data extraction independently. Disagreements between reviewers were resolved by consensus. When consensus was not reached, a senior member mediated (R.R.). A standard chart form of the obtained data was prepared to facilitate comparison among the articles.

Data Items
The following data from each study were extracted: author, date, study design, number of patients, number of iFAMM and FAMM flaps performed, rate of partial and total flap necrosis.

Risk of Bias in Individual Studies
Two independent reviewers (G.L.G., R.F.) performed quality assessments of the included studies; in cases of discrepancies in the results, they consulted a third senior reviewer (R.R.). ROBINS-I tool was be used to assess non-randomized studies. Five levels (low, moderate, serious, critical or no information) were used to present the risk of bias [8]. The Robvis visualization tool web app was used to create "traffic light" plots of the domainlevel judgements for each individual result and weighted bar plots of the distribution of risk-of-bias judgements within each bias [9].

Summary Measures
Partial and total flap necrosis were expressed as integer numbers and percentage.

Additional Analyses
No additional analyses were performed.

Study Selection
The PubMed and Cochrane Library database search identified 373 articles: 156 were duplicate articles, while 217 studies were screened for title, abstract and language. Seventytwo full-text articles were finally selected for further evaluation. Of the 72 papers, 49 were excluded: 4 case reports, 13 case series with fewer than 10 patients, 3 letters, 1 review article, 1 paper in Spanish language, 21 papers reporting other kinds of flap than FAMM or i-FAMM, 5 papers not reporting flap viability outcomes and 4 papers not meeting the inclusion criteria. The selection process identified 20 articles as eligible for inclusion in the review: 12 retrospective studies, 5 case series and 3 case-control studies (Figure 1).

Study Characteristics
The review process identified a total of 20 studies: Pribaz et al.

Risk of Bias within Studies
The analysis of the paper quality assessment is presented in Figure 2; Figure 3.

Synthesis of Results
Data extraction from the 20 articles evaluated allowed us to list a total of 350 FAMM and 136 i-FAMM flaps performed. Overall reported necrosis rate for FAMM flaps was 9.7%, 1.4% (5 cases) as total and 8.3% (29 cases) as partial necrosis. Overall reported necrosis rate for iFAMM flaps was 2.2%, 1.5% (2 cases) as total and 0.7% (1 case) as partial necrosis. Results are synthesized in Table 1; Table 2.

Summary of Evidence
From the myomucosal side of the cheek, several flaps are available to reconstruct different small to medium sized defects of the oral cavity such as the floor of the mouth, tongue, hard and soft palate. Bozola was the first one identifying the myomucosal cheek as a donor site for intraoral reconstruction [4]. The author described the axial buccinator musculomucosal flap, pedicled on the buccal artery, with a posterior mucosal pivot at the maxillary tuberosity and a horizontal axis on the oral commissure. In 1991, Carstens et al., in contrast with Bozola's beliefs, reported that the buccal artery was not the main blood supply of the buccinator, but the facial artery seemed to be preeminent instead [28]. On this basis, Carstens proposed the anteriorly based buccinator myomucosal island flap based on the facial artery and vein [29,30]. In 1992, Pribaz et al. proposed the facial artery musculomucosal flap, an axial myomucosal flap centered on the course of the facial artery with an orthograde flow (inferiorly based) or a reverse flow (superiorly based). In 1999, Zhao et al. described two buccinator myomucosal island flaps, with two different vascular patterns: the buccinator myomucosal neurovascular flap posteriorly based supplied by the buccal artery and the buccinator myomucosal reversed-flow arterial island flap superiorly based supplied by the lateral nasal artery, a terminal branch of the facial artery, with a reverse flow supply [6].
The main difference between FAMM and i-FAMM flaps is represented by the need, for the former, of a temporary post-operative bite block and the need for secondary pedicle section surgery, at least in edentulous patients. i-FAMM, on the contrary, has the advantage of avoiding the aforementioned secondary surgery, although the harvesting is technically challenging compared to FAMM.
The The term buccal artery myomucosal flap is precise and unequivocal. Moreover, the acronym BAMM is similar to FAMM, and it could be successfully accepted and avoid confusion resulting from the use of incorrect or confused terms. For the aforementioned reasons, in order to perform an objective evaluation, the reviewers carefully examined the description of the harvesting technique of all selected papers. The flaps where the pedicle was not identified and where the flap was almost with random vascularization rather than axial were excluded. All BAMM flaps with the identification of the buccal pedicle were considered as FAMM.
However, during the review process, we also faced some personal modifications of the FAMM flap, such as those proposed by Lee or Ibrahim; in both cases the flaps were based on the facial artery [19,23]. Lee's modification was based on creating space in the retromolar trigone and palate to accommodate an inferiorly based pedicle in order to avoid secondary division. The modification proposed by Ibrahim had the aim of avoiding secondary surgery for pedicle interruption and flap inset. The procedure forecasts the anterior incision of the flap to be extended into the alveolar crest in order to reach the defect; the base of the flap is then dissected in a subperiosteal plane over the alveolar crest all the way to the floor of mouth. Additional mucosa can be resected posterior to the defect if needed to accommodate any redundancy. The flap, now in continuity with the defect, can be rotated in place with no intervening bridge; the gingiva is incorporated to the base of the flap. Both modifications, performed by Lee and Ibrahim, in the present review were incorporated in the FAMM flap group.
The absolute number of complication cases analyzed is low; nevertheless, the low necrosis rate reported shows how both flaps have an average good survivability. The higher partial necrosis rate of FAMM flaps could be explained by the harvesting technique: the flap is sometimes grossly harvested, solely considering the facial artery axis. Therefore, marginal tissue could keep random vascularization, and arterial blood flow may not withstand the metabolic demand of the entire area. On the other hand, in cases of island FAMM flap harvesting it is mandatory to perform a meticulous dissection of the facial artery pedicle; a careful dissection of the pedicle could explain the higher survival rate of i-FAMM vs. FAMM.

Limitations
Some limitations can be identified in this review. The wide and confused nomenclature that arose from literature analysis directly influenced the viability assessment. The so-called BAMM flap is sometimes intended as an axial flap based on buccal artery performed with the pedicle identification, while some other authors harvest it without pedicle identification, using only theoretical pedicle orientation. We carefully evaluated all the flap harvesting techniques published and excluded those with no pedicle identification; therefore, in the total count of FAMM flaps there were also some flaps identified as BAMM by the authors but with the identification of the vascular pedicle. Moreover, the realistic viability rate of this flap may be highly biased due to the intrinsic nature of surgical case reports in literature, mostly affected by positive-outcome bias. The high scores encountered during bias evaluation of patient selection, outcome measurement and selection of reported results sections uphold our concerns.

Conclusions
The iFAMM flap retains analogous total necrosis rate and a lower partial necrosis rate compared to FAMM flap. Although the technique is more difficult to execute, the advantages represented by lower complication rate, higher flexibility in defect reconstruction and absence of compulsory second surgery could widen surgeons' choices in oral defect reconstructions.