Cheek Reconstruction

Skin cancer remains the most commonly diagnosed cancer in the United States, with an estimate of over 5 million cases diagnosed annually.[1] Many of these cancers occur on the face, including the cheeks, which comprise a significant portion of the overall facial surface area. While in most cases the extensive mobility of cheek skin provides for excellent and relatively simple closures following tumor extirpation, larger defects, those located near free margins, or those within areas of decreased skin laxity may require a more complicated closure.[2, 3] This article reviews the principles of cheek reconstruction, from aesthetic and anatomic considerations (see the images below) to specific reconstructive options. Lipodystrophy related to HIV therapy can be addressed with a variety of fillers and is beyond the scope of this article.[4, 5]

In this article, cheek reconstruction is indicated for the surgical excision of a variety of cutaneous tumors. These include, but are not limited to, the following entities:

• Basal cell carcinoma

• Squamous cell carcinoma

• Melanoma[6]

• Keratoacanthoma

• Bowen disease

• Microcystic adnexal carcinoma

• Merkel cell carcinoma

• Atypical fibroxanthoma

• Malignant fibrous histiocytoma

• Dermatofibrosarcoma protuberans

• Sweat gland carcinoma

• Sebaceous carcinoma

Each cheek is a large cosmetic unit demarcated by several junction lines as follows: medially at the nasofacial groove and melolabial crease, laterally at the preauricular crease, superiorly at the orbital rim and superior zygomatic arch, and inferiorly at the border of the mandible.[7] Moreover, the cheek can be further divided into 5 subunits whose skin surface attributes are consistent within.[8] Even so, the subunit principle of reconstruction is not as critical here as it is on the nose, because cheek subunits themselves are less distinct.[9, 10] Although junctional lines can offer convenient sanctuaries in which to camouflage scars, most cheek defects lie in more aesthetically exposed regions, where placement of scars along relaxed skin tension lines is a primary principle.[11]

Cheek reconstruction requires attention to the free margins of the lower eyelid, nasal rim, upper lip, and oral commissure. Special care must be exercised during surgery near the lower eyelid in particular, owing to the relatively low tension capable of producing an immediate or delayed ectropion. Repositioning the patient to an upright position after principle suture placement is often helpful to assess the effects of gravity. Subtle distortions can be found by asking the patient to gaze upward while opening the mouth widely. A Frost suspension suture is sometimes useful in this area to counteract temporary forces such postoperative edema or bleeding.[12]

The cheek is relatively devoid of zones in which neurovascular, glandular, or mucosal tissue is at great risk of tumor infiltration or iatrogenic compromise. However, some anatomic considerations merit mention. After exiting the stylomastoid foramen and splitting the posterior auricular branch, the inferior division of the facial nerve divides into 2 main trunks within the substance of the parotid gland. These further divide into the 5 main branches of the facial nerve that supply motor innervation to the superficial muscles of facial expression: temporal, zygomatic, buccal, marginal mandibular, and cervical. These are at greatest risk as they exit the anterior border of the parotid gland, on the deep fascia of the masseter muscle, where they are only covered by superficial soft tissue. Deep excisions near the parotid gland may warrant intraoperative nerve stimulation studies by a head and neck surgeon, beyond the purview of a dermatologic surgeon.

For most individuals (approximately 85%), the zygomatic, buccal, and cervical branches of the facial nerve comprise multiple rami with considerable cross-arborization such that paresis resulting from transection of one of these branches is often temporary.[13, 14] Injury to the temporal branch, however, results in an ipsilateral inability to raise the eyebrow or open the eye widely and can eventually lead to brow ptosis. During cheek surgery, this nerve is most frequently encountered over the zygoma. The marginal mandibular branch of the facial nerve is at greatest risk where it crosses the mandible anterior to the masseter muscle. Loss of this nerve may impair the ability to depress and evert the ipsilateral lower lip, resulting in a permanent grimace.

Cheek sensation is provided mainly by the trigeminal nerve, whose branches often travel above the superficial musculoaponeurotic system (SMAS), just below subcutaneous fat. Although easily exposed to iatrogenic compromise, most resulting dysfunction is neither debilitating nor permanent.[15] Similarly, the well-arborized blood supply of the cheek minimizes any vascular compromise from transecting branches of the facial artery along the cheek’s medial border or superficial temporal artery along its lateral extent.[14] Lastly, the Stenson duct can be palpated along the anterior border of the masseter muscle during teeth clenching. The main drainage channel of the parotid, injury to the Stenson duct is rare during dermatologic surgery but may result in chronic drainage through a fistula, which often requires a reparative procedure.[10, 13]

Contraindications to performing dermatologic cheek reconstruction with the patient under local anesthesia include any underlying medical illness that would preclude safe treatment in an outpatient setting. For instance, patients with advanced cardiac or pulmonary disease, severe hypertension, or a bleeding diathesis often benefit from a monitored operating room. Those patients with psychiatric disorders that would interfere with a surgical procedure, such as tumor removal and closure, may not tolerate cheek reconstruction without sedation.

Following the excision of malignancies with increased recurrence potential, it is often wise to facilitate secondary intention healing, or cover the defect with a split-thickness skin graft, rather than proceed with a complicated reconstruction. This enables close surveillance of the area during follow-up examinations. Likewise, a simpler reconstructive option is preferred for patients whose physicians require them to remain on anticoagulants during the procedure, to minimize the risk of excessive or postoperative bleeding. Furthermore, cheek reconstruction should always be managed such that functional concerns override aesthetics.

Secondary intention healing

For selected defects within the cheek, secondary intention healing may offer acceptable functional and aesthetic results. This is particularly true for small, superficial defects located in concave areas of the cheek, such as the preauricular sulcus, melolabial crease, and facial-alar groove.[16] A large lateral cheek defect that cannot be closed primarily, in a patient who cannot tolerate complicated reconstructive options, may heal by secondary intention with a reasonable result (see the image below).

Large defects on the medial cheek, particularly when deep, often distort the free margin of the eyelid or lip in an unacceptable manner. In such cases, guiding sutures can be useful to direct the natural forces of wound contraction away from margin distortion.[17] The color of the patient’s skin should also be considered, especially with larger defects, as the mature scar formulated via secondary intention healing is often hypopigmented. An additional disadvantage is the lengthy postoperative course as the wound is granulating. Proper wound management by the patient is an important aspect of secondary intention healing because it minimizes pain, prevents infection, and enhances reepithelization.[18] A first-choice option for concave defects, secondary intention healing is largely a last-choice option for most of the cheek.

Primary closure

Primary closure is often the treatment of choice for small or medium-sized defects of the cheek, providing an expedient closure, rapid healing time, and excellent aesthetic result.[10] Resultant scars should be placed within cosmetic unit junctions or along relaxed skin tension lines. Prior to the infiltration of local anesthesia, the relaxed skin tension lines and cosmetic unit junction lines may be outlined to ensure perfect alignment at the time of closure. Moving the patient to an upright position after principle suture placement is also helpful to assess the effects of gravity.

The aging face provides multiple reservoirs of tissue redundancy that can be used in cheek reconstruction, such as the preauricular area, deepened melolabial folds, and pronounced jowls.[14] Even large cheek defects can be repaired primarily with minimal scar visibility if the patient has sufficient skin laxity and photodamage (see the image below).

Many modifications of the standard elliptical excision or standing cone repair technique can be used during primary closure to camouflage scars.[19]  A crescentic or lazy-S modification is often helpful to mimic the physiologic curve over convex surfaces and reduce standing cone formation. Historically, Z-plasty has been suggested to break long, linear scars into smaller, less visible partitions. Data from 2016, however, suggest that long, linear cheek scars may be perceived by the public more favorably than a series of zig-zags.[20] Care should be taken to attend to the principles of skin biomechanics, such as creep and stretch, because excess wound tension and compromised circulation foster the development of scar widening and boost the incidence of wound necrosis, infection, and dehiscence.[11] Free margin preservation is a primary principle of facial reconstruction. If primary side-to-side closure results in an excessively long scar, increased wound tension, or distortion of a free margin, other options should be pursued.

Skin grafting

The split-thickness skin graft provides a convenient window for tumor surveillance. However, discrepancies in skin color/texture and significant wound contracture limit the effective use of this technique either to wounds in which tumor surveillance outweighs cosmetic outcome or to those that cannot be repaired by any other method.[21] The electric Zimmer dermatome is an effective tool for harvesting a split-thickness skin graft, although other methods can be used.[22] .

Full-thickness skin grafts on the cheek generally produce an inferior aesthetic result because they are poorly matched in skin color, skin texture, and hair quality.[23] Moreover, they usually cannot fill the deep subcutaneous defect created after tumor extirpation in a convex area. In fact, patient satisfaction has been shown to be significantly higher after local flap closure of cheek defects compared with full-thickness skin grafts.[24] Occasionally, repairing part of a large wound primarily and using the excised Burrow triangle as a full-thickness graft for the remaining defect is useful. A local graft provides a better color/texture match than skin from a distant donor site. (see the image below).

Even so, contour irregularities often ensue, and they can be modified with adjunctive spot dermabrasion or intralesional steroid injections. The graft should be oversized to allow for some contracture, particularly when this technique is applied near free margins. Vascular lasers may address graft erythema or reactive telangiectasias along the graft periphery.

The initial step in creating a full-thickness skin graft is harvesting tissue from the donor site. This is often accomplished by using the defect to generate a template. The donor site is then prepared and draped in a sterile fashion and the template is used to outline the graft, which should be slightly oversized (3-5%) to allow for shrinkage.[25] Undersized grafts lead to increased tension and may ultimately result in graft failure, while excessive tissue can create a pin-cushioning effect. The donor site is then injected with local anesthesia, the graft is harvested to the level of subcutaneous fat, and the donor site closed. Although exceptionally rare, the iatrogenic spread of Merkel cell carcinoma in an immunocompromised patient to a distant graft donor site has been reported. The aggressive tumor, presumably transferred via contaminated surgical excision tools used for graft harvesting, serves as a reminder of the importance of meticulous attention to detail during cutaneous surgical procedures.[26]

The graft should be defatted before it is placed into the defect, rotated, and trimmed until a precise fit is obtained. Several interrupted sutures secure the graft into its recipient bed before the entire perimeter is meticulously approximated with epidermal stitches. Basting sutures are sometimes helpful to facilitate inosculation. Bolsters provide additional support to the graft and are commonly secured with tie-over sutures or adhesive wound closure tapes. The bolster is left in place for 5-7 days. A light nonadherent bandage and pressure dressing is placed over the bolster and removed after 24-48 hours.

Skin substitutes

The development of bioengineered skin substitutes has led to a novel reconstructive option that is both safe and effective.[27, 28] A variety of manufactured dermal substitute products are available and have been used for the reconstruction of full- or partial-thickness cutaneous defects of the cheek.[29] They are particularly useful for larger defects that cannot be repaired easily using traditional closure methods in the setting of complex clinical comorbidities.[30] Wound healing time and clinical appearance of the wound using these products are both typically improved when compared with secondary intention healing.[31] A detailed review of the application of these products is beyond the scope of this article.

Local cutaneous flaps

When primary closure of a cheek defect is not possible, a local tissue flap is usually the next best reconstructive option.[10, 11] Skin flaps can be categorized in several ways, including based on blood supply (eg, axial, random), method of flap movement (eg, advancement, rotation), or shape (eg, rhombic, bilobe).[21] The cheek is sufficiently vascular such that random pattern flaps, when designed properly, provide excellent aesthetic and functional results. In fact, all of the flaps described below are be based on deep or lateral vascular pedicles and do not necessarily require larger caliber named vessels.

Flap movement is a convenient and familiar way to divide the random pattern of local cutaneous flaps used in cheek reconstruction and is thus the categorical scheme used below. Neither hinged nor interpolation flaps are used with any frequency in cheek reconstruction and have been omitted from this discussion. Complex myocutaneous free flaps, or the use of tissue expanders, may be needed to repair extensive and full-thickness wounds but are beyond the scope of this article.[32, 33, 34]

Advancement

Advancement flaps are moved into a defect by being stretched, in a linear fashion, toward it. As such, tension is only minimally diminished, and the tension vector is not altered.[35] Tissue advancement is an effective closure method for many cheek defects.[36] The primary benefit is the redistribution of tissue redundancies or dog-ears, while minimal tension release is a secondary gain.[37] An example of this is the Burrow advancement flap, which is particularly useful for preauricular defects, because it provides a vertical incision line and adequate redundant tissue from the preauricular crease (see the image below).

Similarly, infraorbital/medial cheek defects may be reconstructed with a Burrow advancement flap that has a component of rotation. The curved superior border of the flap can be positioned along the infraorbital rim (see the image below). When possible, this flap should be extended laterally above the level of the lateral canthus to minimize the risk of ectropion. Although it is easy to separate advancement motion from rotation in principle, in reality, most rotation flaps contain a component of advancement, and advancement flaps occasionally contain some rotation. The term sliding flap is often used to describe these 2 motions collectively.[38]

Advancement flap design is uncomplicated. Parallel incisions made perpendicular to the defect allow the tissue between them to slide toward the wound. Multiple forces inhibit tissue movement, some of which can be altered. Undermining of the flap pedicle releases deep attachments and should be performed below the level of the subdermal plexus to encourage flap survival. If further movement of the flap is required, undermining beneath the base of the pedicle can be undertaken cautiously and in small increments because it clearly compromises vascular input. This is a process of diminishing returns, and, when the maximum benefit has been achieved, the surgeon may begin to consider additional release by increasing the flap length. Often, the forces inhibiting flap movement are such that lengthening the flap does not significantly increase tissue advancement. A general guideline states that flap length should not exceed width beyond a ratio of 3:1.[37]

Principle suture placement should be at the point of maximum flap tension, which is usually from the leading edge of the flap to the far edge of the defect.[39] Afterwards, the remainder of the flap is sewn into place and the resulting tissue redundancies can be either sewn out by the rule of halves or can be removed by a variety of methods, such as Burrow triangles or M-plasty.[40] Flap viability should be confirmed by observing the color at rest or testing capillary refill time.

Multiple variations of this simple tissue transfer technique have been developed. Double advancement flaps, such as the A-to-T or O-to-H, can be helpful when filling larger defects, but these are not generally used in cheek reconstruction because too many straight lines are created.[10, 39]

The V-Y advancement flap, or island pedicle flap, has been used with success in this area, particularly near the melolabial fold when 1 or 2 of the lines can be integrated into a junction line. The longer sides of the island pedicle are situated along relaxed skin tension lines, while the leading edge of the flap tends to lie perpendicular to these lines. This flap is designed with a deep vascular pedicle and all lateral dermal attachments are severed. The blood supply is from the subcutaneous tissue and musculature and is so rich that this technique may be applied successfully even in devascularized tissue, such as can be seen after radiation therapy.  

Some authors have noted a puffiness from this flap that can be difficult to correct, particularly over the convex portions of the cheek. This trap-door effect can theoretically be minimized by under-sizing the flap and placing it under some tension.[40] It cannot be reliably prevented, however, particularly in younger skin.[41] Multiple modifications of the island pedicle flap have been developed, including a bilateral version, the creation of a transposition flap, or burying the entire pedicle in a subcutaneous tunnel. Placing the incisions in a curvilinear fashion, such that both advancement and slight medial rotation occur, can be useful in cheek reconstruction. This results in a horizontal displacement of tension vectors and is less likely to distort free margins (see the image below). In 2019, a flipped variation of the island pedicle flap has been suggested to reconstruct larger defects of the preauricular cheek to minimize tissue manipulation and maintain the hairless region of skin between the tragus and sideburn in male patients.[42]

Rotation

Sometimes, altering the direction of wound closure tension is necessary when repairing a surgical defect. Rotation flaps accomplish this while recruiting tissue from nearby skin laxity. A rotation flap is designed to allow adjacent skin to rotate along a pivot point into the wound (see the image below). In doing so, a secondary defect is created that is equal in area to the primary one, but it has a long, crescentoid shape, which diffuses and redirects the stress of wound closure.[43] The design implies a curved line, which makes this flap useful when placing scars along relaxed skin tension lines during cheek reconstruction. Yet, a curved line over a convex surface may foster tethering and produce a depressed scar.[2]

When a medial cheek defect is large, some authors prefer to replace the entire cosmetic unit with a laterally and inferiorly based rotation/advancement flap.[44] These large flaps can hide incision lines along cosmetic unit junction lines such as the infraorbital crease, preauricular crease, and nasofacial junction. Ectropion, immediate or delayed, is an uncommon but possible association with such flaps and can be minimized by special techniques, including periosteal anchoring sutures, to prevent lower lid retraction.[45, 46, 47]

The arc of the rotation flap is typically 2-4 times longer than the arc of the defect it needs to close.[48] Obviously, wound closure tension disperses with increasing flap length, but the disadvantage is a longer scar. Curvature of the flap increases mobility, but it also redirects wound closure stress and reduces flap blood supply. Mobility may be further augmented by wide undermining beneath the flap pedicle and pivotal area. If additional movement is required, a back-cut can be made into the flap pedicle, but this should be done conservatively because it greatly reduces vascular input. In 2017, a variation of the conventional, anterior-based cervicofacial flap was reported to conserve the sideburn area while using the large tissue reservoir available in that location.[49] The O-to-Z flap, a double-rotation flap, recruits tissue from opposite sides of the wound and can be tissue conservative, taking the Burrow triangles created by primary closure and rotating them inward rather than excising them.[50] Rotation flaps often necessitate long incision lines for adequate movement, and, therefore, other flap options may produce superior results.

Transposition

Transposition flaps or lifting flaps also redistribute wound closure tension. The unique advantage of this tissue transfer technique is the ability to recruit tissue from distant sites. This flap is designed to use tissue from a reservoir that is lifted over uninvolved skin to cover the recipient site.[51] The incision lines from these flaps are shortened and geometric in contrast to the long, straight lines produced by sliding flaps.

Many varieties of transposition flaps have been developed, the most common being rhombic, bilobe, and Z-plasty. Inferiorly based rhombic transposition flaps use preauricular, temple, lower cheek, or cervical skin to repair a defect, in which most final incisions lie along relaxed skin tension lines (see the first image below). This flap is particularly useful in the infraorbital region to minimize an ectropion (see the second image below). Banner flaps are modified rhombic transposition flaps in which the pedicle is unusually long, usually 3:1 to 5:1.[51]

Another rhombic modification, the 30° Webster flap, is useful in dividing wound closure tension between the primary and secondary defects (see the image below). Bilobed flaps have can be helpful for preauricular, infraorbital, or central cheek defects. However, the curved incisions may prove aesthetically suboptimal.[52]

The Z-plasty has traditionally been used to revise traumatic or iatrogenic scars by lengthening them and redirecting wound closure lines. Z-plasty can also be used to reconstruct defects on the cheek that are too large for primary closure.[53] The Z-plasty may be designed so that most final incisions are aligned along relaxed skin tension lines and divide a long linear scar into a series of smaller defects (see the first image below). Larger defects may require a multiple Z-plasty or concomitant parietal defect closure, where possible (see the second image below).

An understanding of flap geometry, knowledge of primary and secondary tissue movement, and much practice are needed to consistently achieve satisfactory results with transposition flaps. The interested reader is advised to consult the textbooks listed in the references section for further information.

The complications associated with cheek reconstruction consist of the standard complications associated with other cutaneous procedures. A 2018 literature review suggests that the cheek may be associated with a slightly higher rate than other regions of the face.[54] Long-term follow up has been suggested to improve the early detection rate of delayed ectropion.[55] The most common and notable complications are infection, tissue necrosis, wound dehiscence, bleeding, and hematomas.[56] Other adverse sequelae encountered by the dermatologic surgeon include suture reactions, contact dermatitis (often to topical antibiotic postoperative therapy), postinflammatory hyperpigmentation or hypopigmentation, edema, erythema, reactive telangiectasias, widened or hypertrophic scarring, and sensory or motor nerve damage.

Some complications unique to this area of the face include eyelid swelling or bruising, ectropion or eclabium, and parotid fistula formation. A detailed description of complication recognition and management is beyond the scope of this article.

When designed properly, most cheek reconstruction repairs after tumor extirpation result in an excellent functional and aesthetic prognosis. Scar revision following cutaneous surgery is not uncommon, and a variety of methods, such as injection, debulking, and nonablative or ablative resurfacing may be undertaken to improve the final outcome.[57, 58]   A combination approach, addressing scar tension, erythema, and scar texture as distinct clinical features, is often helpful.[59] As our understanding of the subtleties that comprise a visually appealing face and those components that detract from it increases, reconstructive efforts will evolve to become more effective. A mastery of technical skill and artistic talent facilitate the reconstructive surgeon’s reapproximation of native anatomy.