oromandibular reconstruction after cancer resection

C L I N I C S I NP L A S T I C

S U R G E R Y

Clin Plastic Surg 32 (2005) 361–375

361

Oromandibular ReconstructionAfter Cancer ResectionAchilleas Thoma, MD, MSc, FRCS(C), FACS*,Carolyn Levis, MD, FRCS(C), J.E.M. Young, MD, FRCS(C)

& Epidemiology& Classifications& Preoperative assessment& Principles of reconstruction& Reconstructive options& Commonly used free flaps

Department of Surgery, McMaster University, Hamilton,* Corresponding author. Surgical Outcomes Research CeSouth, Hamilton, Ontario L8P 3A9, Canada.E-mail address: [email protected] (A. Thoma).

0094-1298/05/$ – see front matter © 2005 Elsevier Inc. All rightsplasticsurgery.theclinics.com

Fibula flapRadial forearm osteocutaneous flapScapula flapIliac crest

& Future predictions and summary& References

The oral cavity includes the buccal mucosa (ie,inside lining of the lips and cheeks), the teeth, theanterior two thirds of the tongue, the floor ofthe mouth below the tongue, the hard palate, andthe retromolar trigone (area behind the wisdomteeth) [1].For reconstructive purposes and in consideration

of the functions of this anatomic entity (ie, mas-tication, deglutition, and speech), Yousif et al [2]divided the oral cavity into 10 functional units:

Oral sphincterLingual and buccal sulciAlveolar ridgesFloor of mouthMobile tongueBase of tongueTonsillar pillarsSoft palateHard palateBuccal mucosa

From a pragmatic and reconstructive point of view,the authors classify the defects as those requiring

Lining onlyLining and boneBone aloneCover and boneLining, bone, and coverage (through and

Ontarintre,

reserve

through defects)

Epidemiology

Head and neck cancers are common in severalregions of the world where rates of tobacco useand alcohol consumption are high. The age-standardized incidence rate of head and neck can-cer in males circa 1990 exceeds 30 per 100,000in regions of France, Hong Kong, the Indian sub-continent, Central and Eastern Europe, Spain, Italy,and Brazil, and in the United States among African

o, CanadaSt. Joseph's Healthcare, 101-206 James Street

d. doi:10.1016/j.cps.2005.01.006

362 Thoma et al

Americans. The variation in incidence of cancers byanatomic site within the head and neck is mostlyrelated to the relative distribution of major riskfactors, such as tobacco or betel quid chewing,cigarette or bidi smoking, and alcohol consump-tion [3].From 1985 to 1994, the largest proportion of head

and neck cancers arose in the larynx (20.9%) and theoral cavity, including the lip (17.6%) (lip = 3.5%,oral cavity = 14.1%, oropharynx = 12.3%) [4].In a recent analysis of 221 patients with squamous

cell carcinoma (SCC) of the oral cavity, 161 patientshad cancer of the tongue, 28 had cancer of the oralcavity, 12 had cancer of the hard palate, 11 hadcancer of the buccal mucosa, and nine had cancerof the gingival area [5].The median age of patients with oral cavity can-

cers was 64.0 years. Men represented 60.2% ofpatients. Pathologic diagnosis was SCC in 86.3%of cases. African Americans (independent of in-come), lower-income patients, and patients withhigher-grade disease were seen more frequentlywith advanced-stage SCC [6].An estimated 28,260 new cases of oral cavity and

pharynx carcinoma were expected in the UnitedStates in 2004. Incidence rates are more than twiceas high in men as in women and are highest inmen older than 50 years. Incidence rates for cancerof the oral cavity and pharynx continued to de-cline in the 1990s in African American and whitemen and women. An estimated 7230 deaths fromoral cavity and pharynx cancer are expected in2004. Death rates have been decreasing since thelate 1970s, with this decline more rapid in the1990s. For all stages combined, about 84% of per-sons with oral cavity and pharynx cancers survive1 year after diagnosis. The 5-year and 10-year sur-vival rates are 57% and 45%, respectively [7].Unfortunately, this trend does not apply every-

where. Male incidence rates of head and neckcancer are rising in most regions of the world.Age-adjusted incidence rates of oral and pharyngealcancer increased after 1970 by 11% per 5-yearperiod and 14% per 5-year period, respectively.The prognosis has not improved substantiallysince the 1950s [8].Despite dramatic improvements in surgical and

reconstructive techniques in the past 3 decades, theoverall mortality rates remain relatively unchanged.The overall 5-year survival rate for persons withoral cavity and pharyngeal cancer is only 52%.When we consider that the risk factors are knownand the lesions in the oral cavity and pharynx areeasily accessible for early detection, the lack ofprogress in controlling this cancer is perplexing.The lack of awareness of the disease burden andthe risk factors, the tendency for occurrence in

lower socioeconomic classes and poorly compliantpopulations, and the lack of a simple screening testhave hindered progress [9].Scant change was seen in early detection of oral

cancer or 5-year relative survival rates between theperiods 1973 to 1984 and 1985 to 1996 in nine Sur-veillance, Epidemiology, and End Results regions.This finding suggests a deficiency in professionaland public education regarding early diagnosis oforal cancer [10].The degree of mandibular invasion influences

the survival rate of patients with SCC of the oralcavity, and this difference is not due to local fail-ure. The 5-year survival is 25.4% after segmentalmandibulectomy, as compared with 40% after rimmandibulectomy. The degree of mandibular in-volvement does not influence the local failuretreated by surgery and radiotherapy [11].The overall 5-year survival rate of oral cavity

cancer in a retrospective review of 277 patientsreceiving initial treatment at Washington Univer-sity Medical Center between 1980 and 1989 was46%. Survival rates by tumor, node, and metastasis(TNM) stage were as follows: stage I, 72%, stage II,54%, stage III, 37%, and stage IV, 29%. Whenpatients were grouped according to the clinical-severity staging system, survival rates were as fol-lows: stage I, 77%, stage II, 56%, stage III, 42%, andstage IV, 14%. The current TNM staging system fororal cavity cancer is based solely on the morpho-logic description of the tumor and disregards theclinical condition of the patient. Patient factors,such as cancer symptom severity and comorbidity,have a significant impact on survival [12].These discouraging statistics indicate that oro-

mandibular cancer, with its obligatory resectionand reconstruction, will continue to consume scarcehealth care resources in the future. Third partypayers and those in decision-making positionsneed to ensure that our institutions consider thisproblem in their budgeting plans.

Classifications

Various classification schemes for segmental man-dibular defects have been described over the years[13–16], reflecting the progress that has been madein understanding the consequences of the defectsand the technical improvements that have takenplace. One such classification by Boyd et al [16] isshown in Fig. 1. This classification is based on threeupper case and three lower case characters: H, C, Land o, m, s. H defects are lateral defects on anylength, including the condyle but not significantlycrossing the midline; L defects are the same butexclude the condyle; C defects consist of the entire

Fig. 1. The HCL classification of mandibular defects. (Adapted from Boyd JB, Gullane PJ, Rotstein LE, et al.Classification of the mandibular defects. Plast Reconstr Surg 1993;92(7):1269; with permission.)

363Oromandibular Reconstruction

central segment containing the four incisors andthe two canines. These letters may be combined(eg, LCL would represent an angle-to-angle defect).The letter o represents the absence of mucosa andskin component, s represents skin deficit, andm rep-resents mucosa deficit [16].Although some microsurgeons use these classi-

fications as a strict guide for reconstruction, theauthors believe that each defect needs to be exam-ined in terms of its anatomic extent and the effectthis will have on the four functions mentionedearlier. The reconstruction needs to be tailored toproviding the functions to be mentioned in the nextsection while, if possible, minimizing the morbid-ity of the donor site. To this end, an effort shouldbe made to reconstruct as much of the normalanatomy as possible.

Preoperative assessment

State of the art management of the oromandibulardefect endeavors to restore function in terms ofspeech, deglutition, mastication, taste, oral hygiene,oral airway maintenance, and cosmesis.

Preoperative assessment and planning for pa-tients undergoing immediate reconstruction oforomandibular defects entails a multidisciplinaryteam consisting of head and neck surgeons, plasticsurgeons with microsurgical expertise, maxillofa-cial prosthodontists, radiation oncologists, medicaloncologists, pathologists, occupational therapists,speech pathologists, internists, and psychologists[17]. The preoperative assessment usually includesa full head and neck examination, chest radiograph,biopsy of the lesion, quadroscopy, liver functiontests, and computerized tomography (CT).Preda et al [18] investigated the use of helical

CT in the pre- and postoperative management oforomandibular reconstruction of patients withoropharyngeal carcinoma using microvascular com-posite free flaps. The CT permitted adequate assess-ment of the extent of mandibular infiltration anddetected early ischemic complications and distantrecurrences. Postoperative assessment of correctflap positioning was possible and helpful in sub-sequent rehabilitation with osseointegrated im-plants [18].When patients arrive in the hospital and have

surgery on the same day, as in the authors’ institu-

364 Thoma et al

tion, it is important for the department of anesthesiato assess patients a week in advance and determinethe need for any additional investigations. Particularattention should be paid to the risk of postoperativedelirium tremens, given the high prevalence of alco-hol abuse in this patient population.When a radial forearm flap is contemplated for

reconstruction, an important preoperative measureis to inform the patient which forearm the surgeonintends to use as the donor site. Patients and theirimmediate families must ensure that no blood isdrawn or intravenous lines started from this site,because this may damage the potential veins of thefuture flap.

Principles of reconstruction

Although it is important to be aware of the ana-tomic areas (functional units) described earlier dur-ing reconstruction, from the practical point of view,the authors often see defects that include variouspermutations of those functional units. Commonly,they are found to be in continuity.To date, it is evident that the single-stage micro-

surgical reconstruction of the oromandibular defectis the standard of care in North America. Komisar[19] reported that patients undergoing conven-tional mandibular reconstructions (ie, those notusing microvascular techniques) required morehospitalization and had poorer function. Urkenet al [20] compared several functional parametersin patients who had microvascular oromandibularreconstruction with those in a group of unrecon-structed controls and found significantly improvedfunction in chewing performance, dental rehabili-tation, and bite force, but in not swallowing orspeech. Superior masticatory function was observedin patients undergoing microvascular oroman-dibular reconstruction versus a control group ofunreconstructed patients [21]. The microsurgicalreconstruction option is chosen when the defectis of such magnitude that skin grafts and localflaps will not adequately address it.In the authors’ view, the most important consid-

erations in reconstructing the floor of the mouthare (1) to allow freedom of the tongue, especiallyon the anterior floor (avoid tethering of tongue),(2) to avoid redundancy of the flap, which com-promises proper hygiene as the folds trap foodparticles, and (3) to use thin flaps, which drapegently over the alveolus (native or reconstructedmandible) and floor of the mouth and permitdenture wearing [Fig. 2].In the authors’ experience and that of other head

and neck microsurgeons, the most common defectrequiring reconstruction is one of the floor of the

mouth with a segmental defect of the mandible[15,22].Soft tissue resection at the floor of the mouth,

including various amounts of tongue and mandi-ble, has a profound, cumulative effect on functionand cosmetic appearance. Despite the advances ofthe last 3 decades, the recovery of upper aerodi-gestive tract function after reconstruction of thesegmental oromandibular defect still tends to beincomplete. Wagner et al [23] performed univariateand multivariate analyses in a prospective observa-tional study of 21 patients undergoing oromandibu-lar reconstruction. Significant univariate predictorsof outcome included tongue resection, pharynxresection, and flap skin-paddle area. Increasingsize of the skin paddle in the reconstructive flapwas an indirect indicator of larger resection. Thisfactor influenced functional outcome measuressuch as speech articulation and oral pharyngealswallowing efficiency to a greater degree in univari-ate analysis. Tongue function, however, remainedthe only significant predictor of function in multi-variate analysis [23]. Hence a whole article in thisissue has been dedicated to tongue reconstruction.In the past, particular emphasis was placed on

proper dental occlusion to achieve normal align-ment of the reconstructed mandible; this occlusionwas achieved by means of arch bars, interosseouswires, Gunning splints, or wiring of dentures of theanterior arch. The missing posterior segment wasreplaced with the bone graft, supported by smallor large reconstructive plates. The authors havefound this approach inadequate, because the free-floating posterior mandibular segment can oftenbe ’’enforced’’ to meet the demands of the anteriorocclusion in many different spatial orientations,excepting the one true orientation. This ’’true andoriginal’’ spatial orientation can only be achievedwhen the mandible is preplated before the canceris resected. This measure is feasible in the vastmajority of cases. The plate is then removed andsterilized and is ready for reapplication after can-cer extirpation. In the occasional case where thetumor is expansile and preplating is not possible,the authors apply an external fixator that by-passes the tumor. The fixator maintains the truespatial position of the condyle in its fossa untilthe reconstruction. The authors’ results have im-proved considerably in the decade since their adop-tion of this approach.The achievement of the normal contour of the

mandible almost invariably requires a number ofosteotomies; this is particularly true for a centraldefect, a central–lateral defect, or a lateral ramus–ascending ramus defect [Fig. 3]. The ’’carpentry’’of the bone may be completed either while theflap is still attached to the donor site and perfused

Fig. 2. (A) A 60-year-old woman with SCC invading the alveolus and floor of mouth. (B) Defect after oroman-dibular resection. (C ) Radial forearm osteocutaneous flap used (skin paddle 6 × 6 cm; bone 7 cm). (D) Insettingof flap at floor of mouth. (E ) Postoperative appearance of same patient. (F ) Excellent contour of the forearm flapin floor of mouth, allowing free tongue movement.


or after detachment. The authors have used bothapproaches. Performing the ’’carpentry’’ on anattached bone graft makes sense insofar as it mini-mizes ischemia; however, the authors find it sim-pler from the practical point of view to perform thebone adjustments at the extirpative site. As a result,they now divide the pedicle as soon as the resectionteam removes the tumor and prepares the recipientvessels. The ’’carpentry’’ of the neomandible is donein the head and neck area. Because this ’’carpentry’’may take up to 1 hour, the bone and pedicle shouldbe irrigated every 5 minutes to keep themmoist andavoid desiccation. The bone fixation is achievedwith reconstruction plates in the larger defects,

whereas in the small defects miniplates may alsobe used.Given a choice for a particular defect, some mem-

bers of the authors’ unit prefer to use the smallerplates, because they limit elevation of the perios-teum andminimize the risk of vascular compromiseto the bone. However, the larger reconstructionplates provide superior stability and tolerate higherloads. Therefore, they are recommended in younger,more active patients in whom the teeth are pre-served. If osseointegration is planned, all plates,especially the screws that traverse the ’’neomandi-ble,’’ may interfere with the integration and needto be removed before dental restoration. If the den-

Fig. 3. (A) Defect after oromandibular resection of SCC arising from tongue and invading the mandible in a62-year-old woman. (B) Radial forearm osteocutaneous flap used (skin paddle 7 × 6.5 cm; bone 10.5 cm). (C ) Twoosteotomies were used to reconstruct the ascending ramus and provide anterior curvature; reconstruction platewas used. (D) Postoperative appearance.

366 Thoma et al

tal implants are placed immediately at the time ofreconstruction, the position of the plates and screwsshould be considered. Some microsurgeons believethat lateral mandibulectomy defects in selectedpatients with poor prognosis can be reconstructed(traversed) with a plate alone and a myocutaneousflap, such as a pectoralis major, or fasciocutaneousfree flaps. However, when a plate and fasciocuta-neous or myocutaneous flap are employed, theanterior defects need to be managed with vascu-larized bone reconstruction to avoid the commonproblem of erosion of the plate into the oral cavity[24]. In a large, recent series where plates and freeflaps were used instead of osteocutaneous flaps, theplate exposure was 46%. As a result, a secondaryprocedure was needed in 31% of the patients [25].The philosophy of the authors’ center in the pastdecade has been that, if a patient is fit to undergothe extirpative part of the procedure, he or she isfit to undergo the free flap reconstruction of themandible. The authors prefer to reconstruct the

bone and oral lining with a vascularized osteocuta-neous flap.An effort is made to save the condyle of the

mandible when possible, because of the difficultyof its reconstruction. If, however, it needs to beresected for oncologic reasons, one method ofreconstruction uses the head of the fibula. Thismethod adds complexity to the dissection, becausethe common peroneal nerve must be protected toavoid injury, lateral knee or leg pain, or even in-stability. Furthermore, if an osteocutaneous flapis required, the skin paddle is usually located verydistal to this proximal fibular head, making thelining and bone reconstruction incompatible. Thebone may not correspond to any acceptable per-forators, and yet another problem occurs whenthe fibular head is too big. A novel technique is tofashion a condyle from the fibula edge (not includ-ing the head), then wrap the end of the ’’neocon-dyle’’ with periosteum resected beyond the requiredfibula [26,27]. Delayed reconstruction of the con-


dyle is an extremely difficult problem, because thefacial nerve is at a real risk for injury in a scarred andtight field where structures have been displaced.Even when the nerve has been identified initially,it can still be compressed and compromised by anybone graft used. A solution to this problem is use ofthe newer prostheses, which carry a ’’condyle’’ at theend of the reconstruction plate.Whenever possible, the authors make an effort to

reconstruct the inferior alveolar nerve, as suggestedby Urken et al [20]. Both the fibular and radiusosteocutaneous flaps can be raised as sensory flaps.The lateral sural nerve in the fibular flap and thelateral antebrachial cutaneous nerves may be co-apted to a recipient nerve, usually the lingualnerve. The lateral antebrachial cutaneous nerve ofthe radial forearm flap should be used, if possible,when a recipient nerve is available [22,28,29].

Reconstructive options

In the early years of head and neck reconstruction,most surgical defects were closed by primarily sutur-ing tongue to cheek or lip and allowing the mandi-ble to ’’drift’’ laterally. This method was acceptablefor lateral floor of mouth, alveolus, cheek, ortongue defects that were relatively small, but itbecame unmanageable for larger defects and was acomplete disaster for midline anterior tumors. Theend result was the typical ’’Andy Gump’’ deformity.Accordingly, attempts were made to use all sorts ofprosthetic materials to stabilize the mandible. Vari-ous local mucosal or tongue flaps were used on theinside and cheek, and cervical flaps were used onthe outside. As reconstructive surgeons gained ex-pertise, skin was brought in from distant sites toreconstruct intraoral or external defects, using fore-head flaps, scalp flaps, deltopectoral flaps, and tra-pezius flaps. The deltopectoral and pectoralis majormyocutaneous flaps markedly changed the out-comes of major head and neck surgical cancer pro-cedures. Although earlier attempts to attach a pieceof clavicle, sternum, or rib to these cutaneous ormyocutaneous flaps met with only limited success,free nonvascularized rib occasionally would helpstabilize a mandible and even became revascular-ized in some patients.The development and increased use of progres-

sively more complicated and reliable mandibularplates provided good long-term stabilization forlateral defects, but these plates were rarely satisfac-tory for anterior defects. It was only with the devel-opment of free microvascular bone and bone/softtissue flaps that the modern era of head and neckreconstruction for oromandibular defects began.One of the earliest flaps used in this regard was

the dorsalis pedis [30], which provided excellent

thin skin for intraoral reconstruction as well assome bone, although this was certainly not as reli-able as the bone flaps now available to us. Simi-larly, vascularized rib has been used intermittentlyto reconstruct the mandible and is still occasionallyused [31,32]. However, both of these osteocuta-neous flaps have significant donor-site morbidityand as a result are rarely used today.A number of well-vascularized osteocutaneous

flaps are now available for oromandibular recon-struction; their advantages and disadvantages arediscussed in detail in the next section.

Commonly used free flaps

For lining defects of the floor of the mouth, anumber of flaps are reported to yield satisfactoryresults. The most common flaps used for this pur-pose in the authors’ center are the radial forearmfasciocutaneous [22,33] and the lateral arm flaps[34]. Other flaps used for the same purpose includethe dorsalis pedis [30,35] and the scapular flaps[36]. Others have recently used the anterolateralthigh flap [37,38].The most intricate and difficult part of reconstruc-

tion is the approach to the oromandibular defectafter composite resection of the mandible. The fourflaps most commonly used for this purpose arediscussed in the following sections.

Fibula flap

In the last decade, the free fibula and radial forearmosteocutaneous flaps have probably been the flapsmost commonly used in oromandibular reconstruc-tion. The advantage of the fibula is the long bicor-tical bone that can reconstruct the entire mandibleif necessary. In contrast to other bone flaps thatprovide limited length, usually about 14 cm (radius,iliac crest, scapula), the fibula has enough bonestock to support osseointegrated dental implants,and there is generally scant morbidity to the donorextremity. It allows extirpative and harvesting teamsto work simultaneously [39–41]. The reconstructivesurgeon is well away from the extirpative surgeon,allowing maximum exposure for resection. Thisarrangement contrasts with the harvesting of theradial flap, where the field is ’’congested’’ with sur-geons and assistants. Although some surgeonsrecommend angiography, the authors no longerperform routine angiography unless the clinicalexamination casts doubt on the adequacy of theleg circulation or suggests that a peroneal arteriamagna exists. This condition is reported to occurin between 0.2% and 5% of the general popula-tion; it is associated with hypoplasia or absence ofboth the anterior and posterior tibial arteries, with

368 Thoma et al

the peroneal artery providing the sole circulationof the foot [42].When the authors are able to palpate a dorsalis

pedis and a posterior artery pulse, and the ankle/brachial index is acceptable, they are satisfied ofthe adequacy of the leg circulation. The harvestingof the flap uses the lateral approach, with the pa-tient in supine position [43]. Before surgery, theperforator vessels are mapped with the Doppler.These perforators are usually found in the middleand distal third of the leg [44,45]. In most cases,only a small portion of the fibula is used in thereconstruction; most of the harvested length isdiscarded, allowing a subperiosteal ’’pruning’’ tolengthen the vascular pedicle, which the authors

Fig. 4. (A) Central oromandibular defect after resection o56-year-old man. (B) Osteocutaneous fibula flap used.small reconstruction plates (skin paddle 5 × 8 cm; bone 1the fibula skin paddle is generally thicker than that provid

generally find necessary. The distal 7 to 8 cm mea-sured from the lateral malleolus is left intact tosupport the ankle joint. The authors routinely iden-tify the peroneal nerve at the neck of the fibula;they decompress it, and the resection of the fibulatakes place just below the neck. In this way theyprotect the peroneal nerve under direct vision at alltimes. In the rare event that the whole fibula isharvested superiorly, it is important to reattachthe lateral collateral ligament to the tibial condyle.To avoid postoperative compression of the pedi-

cle in cases of an osteocutaneous flap, it is prefera-ble to use the contralateral extremity for dissection.The authors’ usual technique is to keep the vascu-lar pedicle on the lingual side and place it poste-

f SCC of floor of mouth invading the mandible in a(C ) “Carpentry” of bone graft on side table using0 cm). (D) Fibula inset. (E ) Postoperative appearance;ed by the forearm flap.

Fig. 5. (A) Preoperative appearance of large bone tumor of the mandible in a 15-year-old male. (B) CTT appear-ance of the tumor. (C ) Radiograph after reconstruction with fibula bone graft (no skin paddle was necessary;bone graft 12 cm). (D) Postoperative appearance.


riorly. These maneuvers make it less likely that thevascular pedicle will be compressed.The most common use of the flap in the authors’

center is as an osteocutaneous flap [Fig. 4]. Lesscommonly, it is used as a vascularized bone graftwithout a skin paddle [Fig. 5]. The cutaneous flapmay be neurotized using a branch of the sural nerve(lateral cutaneous sural nerve). Other variationsof the flap described include the double barrel ar-rangement to increase the height of the recon-structed mandible [27]. Multiple skin paddles maybe designed on separate perforators [46]. The use ofparts of the soleus muscle or flexor hallucis longusto fill soft tissue defects has been described [27,47].A problematic aspect of this otherwise excellent

flap is the reported failure of the skin paddle inthe presence of viable bone [48,49]. Even in morerecently reported series, this failure remains a prob-lem. In a series of 61 patients in whom an osteo-cutaneous flap was raised, Sieg et al [27] experiencedseven total skin paddle necroses in the presenceof viable bone. The authors have also experienced

such failure, which may result from compression ofthe perforators, absence of perforators, shearing ofthe fasciocutaneous blood supply, or poor designof the cutaneous paddle. In centers where the fibulaand radial forearm flaps are used with similar fre-quency, total flap failure is reported to be higherwith the fibula than with the radial forearm osteo-cutaneous flap [50].

Radial forearm osteocutaneous flap

This flap is a true competitor to the fibula flap. It iscertainly superior to the fibula osteocutaneous flapin terms of skin paddle survival, but the bone qual-ity is inferior. As much as 14 cm of unicortical bonehas been harvested [22]. The skin paddle is thin andoften superior to the fibula’s thicker skin paddle.Likewise, the skin paddle has better flexibility inrelation to the bone than does the equivalent fib-ula flap. The radial forearm may be used as a sen-sory flap when one harvests the lateral antebrachialcutaneous flap with it. Its vascular pedicle is by farthe best of all the common flaps used in head and

370 Thoma et al

neck reconstruction. As much as 20 cm of pediclemay be easily harvested, making this flap ideal forreconstructions in vessel-depleted necks, especiallyin repeat surgical cases.Strong evidence suggests that flaps requiring vein

grafts have higher rates of flap loss (18.4%) thando flaps that do not require vein grafts (2.9%;P <0.0001) [51]. The radial forearm flap vesselsare of large caliber: 2 to 5 mm. This flap has dualvenous drainage (deep and superficial), and bothsystems may be used. If the anastomotic vein isprotected in the anticubital fossa, the two systemsmay be combined to provide a type I venous drain-age pattern with two large-caliber veins for anas-tomosis: the median basilic and median cephalicveins [52].In contrast to the fibula, in which the skin pad-

dle is the tenuous component of the flap, it is thebone component of the radial forearm flap that istenuous. Nonetheless, multiple osteotomies can beperformed on the radial bone to achieve the ap-propriate contour [53]. Caution needs to be exer-cised in harvesting this flap as one approaches thebone cortex from underneath the brachioradialistendon. The soft tissue composed of part of theflexor pollicis longus needs to remain attached tothe bone. The dissection should be parallel to theundersurface of the brachioradialis to protect thesmall periosteal vessels through which the bonederives its vascular supply [22,53]. The major draw-back of this flap is the risk of fracture of the do-nor radius, which runs at about 12%. This risk maybe minimized by limiting the harvesting of theradius to 40% of the diameter of the bone, althoughthe authors often harvest 50% of the diameter.Some precautions for avoiding donor radius frac-ture include instructing the patients preoperativelyto avoid putting weight on the hand, placing aprotective plaster splint, and controlling deliriumtremens after the surgery. Intraoperatively, oneneeds to be careful with the osteotomy on the radialside of the radius at the level of the insertion ofthe pronator teres, where the appearance that oneis osteotomizing superficially may cause one to gotoo deep [22].The fear of donor fracture has kept some investi-

gators from using this flap. Some recommend theuse of prophylactic plating of the radius [54]. Arecent cost-utility analysis comparing plating ofthe donor radius with no plating found prophylac-tic plating to be a non–cost-effective approach [55].As a result of this analysis, the authors usually onlyplate the radius in cases of fractures or when theybelieve the remaining radius is too thin. A review ofthe function of the hand in patients who fracturedthe radius did not identify any limitations, in con-trast to some early papers.

Skin graft failure and exposed tendons occa-sioned early criticisms of the radial forearm flap.These complications should never occur if one paysattention to details of skin grafting and splints theentire hand and forearm for 1 week [22]. Devascu-larization of the hand should never happen in thepresence of an intact palmar arch, as ascertained bythe Allen’s test. However, caution should be exer-cised in identifying an anomalous superficial ulnarartery, which occurs in 3% of the population. Inthis anomaly, the ulnar artery is located under thesuperficial fascia along the volar forearm. As aresult, it can easily be mistaken for a superficialvein. Normally, in the distal forearm, the ulnarartery and the venae comitantes are found underthe flexor carpi ulnaris. In the case of this anomaly,no vessels are found under the flexor carpi ulnaris.Failure to recognize this congenital anomaly candevascularize the whole hand. This problem canbe prevented by not exsanguinating the forearmbut simply elevating the flap before inflating thetourniquet [56].Osteotomies may be performed on the radial

bone graft to reconstruct central mandibular defectsjust as well as with the fibula [see Fig. 3]. Althoughosseointegrated implants are possible [57], theauthors believe that the bone quality is not as robustas that of the fibula, where shorter osteotomizedsegments are tolerated and dental restoration withosteointegrated implants is more reliable.

Scapula flap

This flap, subtended by the circumflex scapularvessels, may have a longer pedicle if the dissectionextends deeper into the triangular space and usesthe subscapular vessels [58]. The subscapular ves-sels are larger, and the vein is less friable at thislevel. One large or two venae comitantes may beanastomosed. Two variations of the osteocutane-ous flap have been described. The first variationuses the lateral edge of the scapula with thickerbone but with a shorter vascular pedicle [59,60].The second variation uses the medial edge of thescapula, in which the bone component is awayfrom the pedicle, effectively allowing the anasto-mosis to be performed on the opposite neck [61].Another advantage of the medial flap is that itminimizes the dissection around the teres majorand minor and the glenohumeral joint, therebyavoiding prolonged stiffness to the shoulder. Themajor disadvantage of the medial scapular flap ascompared with its lateral equivalent is the thick-ness of the bone. Twelve to 14 cm of bone maybe harvested with either variation of the scapulaflap [61]. Two key advantages of this flap comparedwith the fibula and radius flaps are the superiorcolor match of the skin paddle when used for

Fig. 6. (A) Malignant fistula from recurrent SCC arising from submandibular gland and invading the mandible in a70-year-old man. (B) Resected mandible. (C ) Medial scapular osteocutaneous flap used (skin paddle 17 × 8 cm;bone graft 11 cm). (D) Intraoperative view of reconstruction. (E ) Postoperative appearance showing excellentcolor match (frontal view). (F ) Profile view of same patient.


coverage [Fig. 6] and the possibility of harvestinglarge and multiple skin paddles from the posteriorthorax—ideally suited to the patient with a massivesoft tissue defect or a through and through defect[61]. Even in flaps as large as 39 × 10 cm, the donorsite can be closed directly [62].The major drawback of this flap (both lateral and

medial varieties) is the necessity of repositioningthe patient. Rarely is it possible for the ablative andreconstructive teams to work simultaneously. As aresult, it is necessary to change the patient’s positiontwice, which adds considerable time to the surgery.To minimize this time, the authors compromise by

estimating the defect ahead of time and harvestingthe flap before the cancer is resected. They takemore bone and soft tissue than they expect to use,close the defect, and let the ablative team take over.The flap is kept on ice on a side table until hourslater, when the reconstructing team returns to insetit. The prolonged cold ischemia has not been aproblem in the authors’ patients.

Iliac crest

This free flap based on the deep circumflex arteryand vein was first introduced by Ian Taylor in 1979[63,64]. Since then, it has been used successfully by

Table 1: Comparison of select qualities of the four osteocutaneous flaps in oromandibularreconstruction

Fibula Radius Scapula Iliac crest

Reliability of skin paddle (survival) ++ ++++ ++++ +++Skin paddle pliability (intraoral) +++ ++++ ++ +Color match (external coverage) ++ ++ ++++ +Bone quality +++ ++ ++ ++++Osseointegration of dental implants +++ ++ ++ ++++Vessel quality +++ ++++ +++ +++Vascular pedicle length +++ ++++ ++ (++++) ++Sensory innervation ++++ ++++ + +Donor-site morbidity ++ ++ ++ (+++) +Donor-site cosmesis +++ +++ +++ +++

++++ denotes best; + denotes worst; (+) denotes the medial scapular osteocutaneous flap.

372 Thoma et al

a number of surgeons for the reconstruction of thecomposite oromandibular defect [13,14,65]. A largevolume of unicortical or bicortical bone, up to 14 to16 cm in length, may be harvested. This flap canbe used to reconstruct various mandibular defects,and multiple osteotomies may be performed toachieve the necessary contour. Urken et al [66]used the internal oblique muscle with this bonegraft to resurface soft tissue and intraoral mucosaldefects. This flap is probably the best available flapfor accepting osseointegrated implants [66]. Itsadvantages are the volume of the bone, the height,which is appropriate for accepting dental implants,and the large soft tissue mass that can be harvestedwith it. Unfortunately, this flap carries with it sig-nificant donor-site morbidity, including severepostoperative pain, contour irregularity, and iatro-genic hernia [67]. In the authors’ experience, theskin paddle of the iliac crest flap is usually toobulky and unsuitable for reconstruction of the lin-ing and bone defects. For external coverage andbone defects or for through and through defects,the color match is inferior to other flaps. The iliaccrest flap’s vascular pedicle is also much shorterthan those of competitor flaps, such as the radialforearm, scapula, and fibula. It is still useful forisolated bone defects in which dental implants aredesired, particularly for central mandibular defects.Although other flaps have been used less fre-

quently (eg, humerus bone attached to the lateralarm flap), the authors have included here what theyconsider to be the four most useful osteocutaneousflaps. Table 1 presents some key distinguishingqualities of these four flaps, based on the authors’experience of using all of them intermittently over aperiod of more than 20 years. The authors of thisarticle anticipate some understandable disagree-ment with this assessment; they recommend read-ing the article in this issue that addresses, amongother matters, the question of bias and hierarchy ofevidence in clinical decision making. One aspect of

oromandibular reconstruction that appears to gen-erate heated controversy at national meetings is theosseointegration of any one of the bone graftsdescribed here. In the authors’ view, this debate isvastly overrated, because the proportion of patientswho undergo this step in most reported series issmall indeed. The patients undergoing oroman-dibular resection generally belong to a compara-tively low socioeconomic class and cannot affordthe added expense of osseointegration of dentalimplants. Certainly, we need to make progress inthis area by lobbying the local and federal govern-ments to provide this aspect of care. However, untilosseointegration becomes a common practice, weshould keep our minds open to the other qualitiesof the competing flaps.

Future predictions and summary

Our understanding of the reconstructive needs ofthe oromandibular defect is now clear. We need toreturn the anatomy as close to its original state aspossible to improve function and cosmesis. Recon-structive surgeons’ efforts in the last 2 decades haveyielded remarkable improvements in both areas.We need to acknowledge, however, that our resultswill continue to be suboptimal as long as we usedissimilar tissues from other parts of the body forthe reconstruction. We are presently at the dawnof a new era of reconstruction of the oroman-dibular defect. Within a few months of this writ-ing, the composite tissue allotransplantation ofthe human face, supported by immunosuppressivedrugs, will become a reality. Soon afterward, thetransplantation of the mandible and parts of thefloor of the mouth and tongue will become feasi-ble. Future head and neck microsurgeons will havean easier job. Banked parts of human faces willbe at their disposal. An entire article in this issuehas been devoted to this exciting prospect.


At national and international microsurgical meet-ings and in published articles, authors tout thebenefits of the flaps with which they are familiarand criticize those with which they are not. Becausesurgeons in general are not familiar with the prin-ciples of evidence-based microsurgery, the editorsof this issue have included an article that addressesthe methodologic issues of comparing free flapsand techniques (known and still undiscovered) inoromandibular reconstruction. The authors antici-pate that these principles will help surgeons decidewhich flap to use based on the best evidence avail-able, rather than on ‘‘expert‘‘ pronouncements—including their own!

References

[1] Public Health Agency of Canada. Cancer surveil-lance on-line page. Available at: http://dsol-smed.phac-aspc.gc.ca/dsol-smed/cancer/index_e.html.Accessed November 30, 2004.

[2] Yousif NJ, Matloub HS, Sanger JR, et al. Soft-tissue reconstruction of the oral cavity. Clin PlastSurg 1994;21(1):15–23.

[3] Sankaranarayanan R, Masuyer E, Swaminathan R,et al. Head and neck cancer: a global perspectiveon epidemiology and prognosis. Anticancer Res1998;18(6B):4779–86.

[4] Hoffman HT, Karnell LH, Funk GF, et al. TheNational Cancer Data Base report on cancer ofthe head and neck. Arch Otolaryngol Head NeckSurg 1998;124(9):951–62.

[5] Inagi K, Takahashi H, Okamoto M, et al.Treatment effects in patients with squamous cellcarcinoma of the oral cavity. Acta OtolaryngolSuppl 2002;547:25–9.

[6] Funk GF, Karnell LH, Robinson RA, et al. Pre-sentation, treatment, and outcome of oral cavitycancer: a National Cancer Data Base report. HeadNeck 2002;24(2):165–80.

[7] Jemal A, Tiwari R, Murray T, et al. Cancer Sta-tistics, 2004. CA Cancer J Clin 2004;54(1):8–29.

[8] Mork J. Forty years of monitoring head and neckcancer in Norway—no good news. AnticancerRes 1998;18(5B):3705–8.

[9] Freije J, Kumar JV. Prevention of cancers of oralcavity and pharynx in New York State. N Y StateDent J 2001;67(3):26–30.

[10] Shiboski CH, Shiboski SC, Silverman Jr S. Trendsin oral cancer rates in the United States, 1973–1996. Community Dent Oral Epidemiol 2000;28(4):249–56.

[11] Tankere F, Golmard JL, Barry B, et al. Prognosticvalue of mandibular involvement in oral cavitycancers. Rev Laryngol Otol Rhinol (Bord) 2002;123(1):7–12.

[12] Pugliano FA, Piccirillo JF, Zequeira MR, et al.Clinical-severity staging system for oral cavitycancer: five-year survival rates. Otolaryngol HeadNeck Surg 1999;120(1):38–45.

[13] David DJ, Tan E, Katsaros J, et al. Mandibular

reconstruction with vascularized iliac crest: a10 year experience. Plast Reconstr Surg 1988;82(5):792–803.

[14] Jewer DD, Boyd JB, Manktelow RT, et al.Orofacial and mandibular reconstruction withthe iliac crest free flap: a review of 60 cases and anew method of classification. Plast Reconstr Surg1989;84(3):391–403.

[15] Urken ML, Weinberg H, Vickery C, et al.Oromandibular reconstruction using microvas-cular composite free flaps. Report of 71 cases anda new classification scheme for bony, soft-tissueand neurologic defects. Arch Otolaryngol HeadNeck Surg 1991;117(7):733–44.

[16] Boyd JB, Gullane PJ, Rotstein LE, et al. Classifica-tion of the mandibular defects. Plast ReconstrSurg 1993;92(7):1266–75.

[17] Markowitz BL, Calcaterra TC. Preoperative as-sessment and surgical planning for patients un-dergoing immediate composite reconstructionof oromandibular defects. Clin Plast Surg 1994;21(1):9–14.

[18] Preda L, Dore R, Benazzo M, et al. The roleof computed tomography in the preoperativeassessment and follow-up of oromandibularreconstruction with microvascular osteomyocu-taneous free flaps. Dentomaxillofac Radiol 1999;28(6):338–43.

[19] Komisar A. The functional result of mandibu-lar reconstruction. Laryngoscope 1990;100(4):364–74.

[20] Urken ML, Buchbinder D, Weinberg H, et al.Functional evaluation following microvascularreconstruction of the oral cancer patient: a com-parative study of reconstructed and nonrecon-structed patients. Laryngoscope 1991;101(9):935–50.

[21] Curtis DA, Plesh O, Miller AJ, et al. A compari-son of masticatory function in patients with orwithout reconstruction of the mandible. HeadNeck 1997;19(4):287–96.

[22] Thoma A, Khadaroo R, Grigenas O, et al.Oromandibular reconstruction with the radial-forearm osteocutaneous flap: experience with60 consecutive cases. Plast Reconstr Surg 1999;104(2):368–78.

[23] Wagner JD, Coleman III JJ, Weisberger E, et al.Predictive factors for functional recovery afterfree tissue transfer oromandibular reconstruc-tion. Am J Surg 1998;176(5):430–5.

[24] Boyd JB, Mulholland RS, Davidson J, et al. Thefree flap and plate in oromandibular reconstruc-tion: long-term review and indications. PlastReconstr Surg 1995;95(6):1018–28.

[25] Wei FC, Celik N, Yang WG, et al. Complicationsafter reconstruction by plate and soft-tissue freeflap in compositemandibular defects and second-ary salvage reconstruction with osteocutaneousflap. Plast Reconstr Surg 2003;112(1):37–42.

[26] Iconomou TG, Zuker RM, Phillips JH. Mandibu-lar reconstruction in children using the vascu-larized fibula. J Reconstr Microsurg 1999;15(2):83–90.

374 Thoma et al

[27] Sieg P, Zieron JO, Bierwolf S, et al. Defect-relatedvariations in mandibular reconstruction usingfibula grafts. A review of 96 cases. Br J OralMaxillofac Surg 2002;40(4):322–9.

[28] Urken ML. The restoration or preservation ofsensation in the oral cavity following ablativesurgery. Arch Otolaryngol Head Neck Surg 1995;121:607–12.

[29] Boyd B, Mulholland S, Gullane P, et al.Reinnervated lateral antebrachial cutaneous neu-rosome flaps in oral reconstruction: are wemaking sense? Plast Reconstr Surg 1994;93(7):1350–9.

[30] Robinson DW. Microsurgical transfer of thedorsalis pedis neurovascular flap. Br J Plast Surg1976;29(3):209–13.

[31] Daniel RK. Mandibular reconstruction with freetissue transfers. Ann Plast Surg 1978;1(4):346–71.

[32] Thoma A, Heddle S, Archibald S, et al. The freevascularized anterior rib graft. Plast ReconstrSurg 1988;82(2):291–8.

[33] Soutar DS, Scheker LR, Tanner NS, et al. Theradial forearm flap: a versatile method for intra-oral reconstruction. Br J Plast Surg 1983;36(1):1–8.

[34] Katsaros J, Schusterman M, Beppu M, et al. Thelateral upper arm flap: anatomy and clinical ap-plications. Ann Plast Surg 1984;12(6):489–500.

[35] Zucker RM, Manktelow RT. The dorsalis pedisfree flap: technique of elevation, foot closure,and flap application. Plast Reconstr Surg 1986;77(1):93–104.

[36] Barwick WJ, Goodkind DJ, Serafin D. The freescapular flap. Plast Reconstr Surg 1982;69(5):779–87.

[37] Koshima I. Free anterolateral thigh flap forreconstruction of head and neck defects follow-ing cancer ablation. Plast Reconstr Surg 2000;105(7):2358–60.

[38] Wei FC, Jain V, Celik N, et al. Have we found anideal soft tissue flap? An experience with 672 an-terolateral thigh flaps. Plast Reconstr Surg 2002;109(7):2219–26.

[39] Hidalgo DA, Rekow A. A review of 60 consecu-tive fibula free flap mandible reconstructions.Plast Reconstr Surg 1995;96(3):585–96.

[40] Shpitzer T, Neligan PC, Gullane PJ, et al. The freeiliac crest and fibula flaps in vascularized oroman-dibular reconstruction: comparison and long-term evaluation. Head Neck 1999;21(7):639–47.

[41] Babovic S, Johnson CH, Finical SJ. Free fibuladonor-site morbidity: the Mayo experience with100 consecutive harvests. J Reconstr Microsurg2000;16(2):107–10.

[42] Kim D, Orron DE, Skillman JJ. Surgical signifi-cance of popliteal artery variants: a unified an-giographic classification. Ann Surg 1989;210(6):776–81.

[43] Gilbert A. Vascularized transfer of the fibulashaft. Int J Microsurg 1979;1:100–2.

[44] Wei FC, Seah CS, Tsai YC, et al. Fibulaosteoseptocutaneous flap for reconstruction of

composite mandibular defects. Plast ReconstrSurg 1994;93(2):294–304.

[45] Jones NF, Monstrey S, Gambier BA. Reliability ofthe fibular osteocutaneous flap for mandibularreconstruction: anatomical and surgical confir-mation. Plast Reconstr Surg 1996;97(4):707–16.

[46] Yang KC, Leung JK, Chen JS. Double-paddleperoneal tissue transfer for oromandibular re-construction. Plast Reconstr Surg 2000;106(1):47–55.

[47] Baudet J, Panconi B, Caix P, et al. The compositefibula and soleus free transfer. Int J Microsurg1982;4:10–26.

[48] Schusterman MA, Reece GP, Miller MJ, et al.The osteocutaneous free fibula flap: is the skinpaddle reliable? Plast Reconstr Surg 1992;90:787–93.

[49] Winters HA, de Jongh GJ. Reliability of theproximal skin paddle of the osteocutaneous freefibula flap: a prospective clinical study. PlastReconstr Surg 1999;103(3):846–9.

[50] Thoma A, Mongiardi C, Veltri K. Comparison ofthe free fibula and radial forearm osteocuta-neous flaps in oromandibular reconstruction.Presented at the American Society for Recon-structive Microsurgery Annual Meeting. PalmSprings (CA), January 19, 2004.

[51] Kroll SS, Schusterman MA, Reece GP, et al.Choice of flap and incidence of free flap suc-cess. Plast Reconstr Surg 1996;98(3):459–63.

[52] Thoma A, Archibald S, Jackson S, et al. Surgicalpatterns of venous drainage of the free forearmflap in head and neck reconstruction. PlastReconstr Surg 1994;93(1):54–9.

[53] Thoma A, Allen M, Tadeson BH, et al. The fateof the osteotomized free radial forearm osteocu-taneous flap in mandible reconstruction. J Re-constr Microsurg 1995;11(3):215–9.

[54] Bowers KW, Edmonds JL, Girod DA, et al.Osteocutaneous radial forearm free flaps: thenecessity of internal fixation of the donor-sitedefect to prevent pathological fracture. J BoneJoint Surg 2000;82A(5):694–704.

[55] Rockwell GM, Thoma A. Should the donorradius be plated prophylactically after harvestof a radial-forearm osteocutaneous flap? J Re-constr Microsurg 2004;20(4):297–306.

[56] Thoma A, Young JE. The superficial ulnar artery‘‘trap‘‘ and the free forearm flap. Ann Plast Surg1992;28(4):370–2.

[57] Mounsey RA, Boyd JB. Mandibular reconstruc-tion with osseointegrated implants into the freevascularized radius. Plast Reconstr Surg 1994;94(3):457–64.

[58] Serafin D. Atlas of microsurgical composite tis-sue transplantation. Philadelphia: WB Saun-ders; 1996.

[59] Teot L, Bosse JP, Moufarrege R, et al. The scapularcrest pedicled bone graft. Int J Microsurg 1981;3:257–62.

[60] Swartz WM, Banis JC, Newton ED, et al. The os-teocutaneous scapular flap for mandibular and


maxillary reconstruction. Plast Reconstr Surg1986;77(4):530–45.

[61] Thoma A, Archibald S, Payk I, et al. The freescapular osteofasciocutaneous flap for head andneck reconstruction. Br J Plast Surg 1991;44(7):477–82.

[62] Thoma A, Heddle S. The extended free scapularflap. Br J Plast Surg 1990;43(6):709–12.

[63] Taylor GI, Townsend P, Corlett R. Superiorityof the deep circumflex iliac vessels as the supplyfor the free groin flaps. Plast Reconstr Surg1979;64(5):595–604.

[64] Taylor GI, Townsend P, Corlett R. Superiorityof the deep circumflex iliac vessels as the supply

for the free groin flaps: clinical work. PlastReconstr Surg 1979;64(6):745–59.

[65] Moscoso JF, Urken ML. The iliac crest compositeflap for oromandibular reconstruction. Otolar-yngol Clin North Am 1994;27(6):1097–117.

[66] Urken ML, Vickery C, Weinberg H, et al. Theinternal oblique–iliac crest osseomyocutaneousfree flap in oromandibular reconstruction. Re-port of 20 cases. Arch Otolaryngol Head NeckSurg 1989;115(3):339–49.

[67] Forrest C, Boyd B, Manktelow R, et al. The freevascularized iliac crest tissue transfer: donor sitecomplications associated with eighty-two cases.Br J Plast Surg 1992;45(2):89–93.

oromandibular reconstruction after cancer resection

Documents