Our system of classifying tongue defects. A indicates left-sided hemiglossectomy; B, left three-quarter glossectomy; and C, total oral glossectomy defects. Base of the tongue (BOT) defects were in the zone outlined in grey, anterior to the larynx.
This patient underwent a right-sided hemiglossectomy. Reconstruction of the defect was accomplished by using a longitudinal fold of a free flap from the forearm.
A, The fold-and-roll technique was used to reconstruct three-quarter oral tongue defects. In the posterior (shaded) position, the flap is folded on itself. In the anterior (stippled) position, the flap is asymmetrically de-epithelialized, rolled on itself in the direction of the arrow, and inset. B, Final healed result of the fold-and-roll technique at 8 months postoperatively (preoperative radiation therapy only). The native tongue remnant is atrophied from a previous anastomosis of cranial nerves XII to VII.
Multidimensional technique. A, The flap is laid in transversely, ready for bilateral posterior-to-anterior advancement suturing. B, Suturing the redundant anterior flap edges to each other vertically in the midline creates anterior neotongue height. The midline of the remaining superior free margin is anchored to the vertical suture line to form a 3-point junction. Arrows indicate the direction of tissue displacement.
Multidimensional technique. Diagram (A) and photograph (B) show fashioning of right- and left-sided, anterolaterally directed cones that are de-epithelialized (stippled area) and sutured together in the midline. A neotongue tip is thus created, trapping dermis for bulk and (relative) stiffening. Curved arrows indicate arc of rotation of flap corners, to be sutured together in the midline following de-epithelialization.
Healed neotongue 9 months postoperatively, using the multidimensional fold technique after resection of the entire oral tongue (region C, depicted in Figure 1).
Neotongue reconstruction for a case of total oral glossectomy (region C of Figure 1). The multidimensional folding technique was used, allowing the patient to manipulate the neotongue tip for contact with the upper lip.
Healed result 2 months after planar reconstruction of the floor of the mouth only, absent a neotongue.
Speech intelligibility scores in 30 patients after free-flap reconstruction of the tongue. The asterisk denotes patients who were observed to have no neotongue and therefore no flap-to-palate tissue contact. Scoring is explained in the "Evaluations" subsection of the "Methods" section.
Functional outcome swallowing scale scores for the 41 patients available for swallowing evaluation. Scoring is explained in the "Evaluations" subsection of the "Methods" section.
Haughey BH, Taylor SM, Fuller D. Fasciocutaneous Flap Reconstruction of the Tongue and Floor of MouthOutcomes and Techniques. Arch Otolaryngol Head Neck Surg. 2002;128(12):1388-1395. doi:10.1001/archotol.128.12.1388
To quantify functional and other outcomes after major resection and fasciocutaneous free-flap reconstruction of the tongue and floor of mouth, and to describe reconstructive technique.
A hypothesis-generating, retrospective cohort study of 43 patients who underwent, at minimum, a hemiglossectomy and resection of the floor of the mouth for oral cancer followed by fasciocutaneous free-flap reconstruction.
A tertiary academic medical center in the midwestern United States.
Main Outcome Measures
Speech intelligibility, swallowing, interval to decannulation, length of stay, free-flap success rates, patient survival, and complications.
Thirty patients underwent oral tongue reconstructions, and 13, tongue base reconstructions. Median intelligibility scores were greater among patients in the tongue base group (98% intelligibility) than in the oral-tongue group (76% intelligibility) (P<.001). Of the 38 patients undergoing swallowing evaluation, 32 (85%) were able to feed entirely by mouth, most with mild to moderate dysphagia. All patients underwent decannulation (mean interval, 13.7 days). The mean length of hospital stay was 11 days, and free flaps in 42 patients (98%) survived. Twenty-eight patients (65%) were still alive by the end of the study, yielding a mean survival time of 27.4 months. Seven patients (16%) had severe medical and 3 (7%) had major surgical complications.
The folding techniques used in this study for reconstruction of the tongue with fasciocutaneous free flaps were associated with recovery of adequate speech and swallowing in most patients.
Then will the mute tongue shout for joy. Isaiah 35:6
Then will the mute tongue shout for joy. Isaiah 35:6
RESTORATION OF tongue function is a "cause celebre," lauded even by the ancient Israelite prophet Isaiah. Adequate speech and swallowing are dependent on tongue shape, size, and mobility. These functions can be profoundly disturbed by resection of tumors that involve the oral tongue and/or the base of the tongue.1,2 Glossectomy is effective in achieving local control of tongue malignancy.3 Reconstruction of the tongue after such resections is designed to assist patients achieve an appropriate level of oral function and remains a major challenge for the reconstructive surgeon. Numerous techniques have been used for reconstructing large defects (hemiglossectomy or greater) of the human tongue, including split-thickness skin grafts and locoregional flaps. Techniques of this type fail to provide the tissue bulk and pliability likely to benefit functional recovery. As a result, free-tissue transfer has become standard for tongue reconstruction. Various flaps have been used for this purpose; however, the fasciocutaneous flap has evolved as the "workhorse" for oral reconstruction.4- 6 Sensory reconstruction7 has also long been postulated in oral cavity reconstruction as a reasonable goal with probable benefit for functional recovery. Sensation recovery has been documented using fasciocutaneous flaps in oral reconstruction,8,9 but a causal link between sensory recovery in the neotongue and function recovery has not been proved.10,11 The primary goal of this study was to determine postoperative swallowing, speech, and airway results achieved in patients undergoing glossectomy with reconstructive designs that harnesses different flap-folding techniques according to the location and size of the defect.
In this study, we reviewed data from 252 resections and free-tissue transfers that we performed at Barnes-Jewish Hospital, St Louis, Mo, from March 1, 1990, through May 31, 1999. The database contained demographic data and data from the pathologist's report for each patient, surgery, and morbidity. We chose 43 of these records for study according to the following entry criteria: a minumum of a hemiglossectomy centered in the oral tongue, or a base of the tongue resection; partial or total resection of the floor of the mouth; fasciocutaneous free-flap reconstruction of oral soft tissue; any histologic type of malignancy; patients treated with curative intent; and American Society of Anesthesiologists classes I through III. All patients included in this study underwent resection of at least half of the oral tongue or two thirds of the tongue base, followed by free-flap reconstruction of their defect. The series did not include any patients undergoing total glossectomy, for whom we prefer a motor-innervated musculocutaneous flap reconstruction.12 We included 5 patients in whom we did not finally establish vertical height of the reconstruction; specifically, the floor of the mouth only was ultimately reconstructed, and therefore tissue did not extend above the plane of the mandibular alveolus. This tissue distribution arose in 2 cases because we used an osteocutaneous flap for segmental reconstruction of the mandible and floor of the mouth. We used a pectoralis major myocutaneous flap for wound closure in 2 additional cases with orocutaneous fistulas, and in the fifth case, a flap failure healed by secondary intention.
We obtained fasciocutaneous flaps from the forearm in 40 of the 43 patients and from the lateral thigh in 1 patient. The other 2 patients underwent segmental mandibulectomies and required fibular and scapular osteocutaneous flaps, with the fasciocutaneous paddle used for intraoral soft tissue reconstruction. A vascularized portion of the brachioradialis in the case of forearm flaps, or a free-fat graft, was included in the fasciocutaneous flap when necessary to supplement the bulk of the free-tissue transfer. We estimated the required size of a flap by measuring the tongue defect and performed model reconstructions with supple sterile templates immediately after resection. Twenty-eight (65%) of the 43 free flaps were reinnervated according to availability of a recipient nerve. The lingual nerve was used in 24 cases; the internal branch of the superior laryngeal nerve, in 2 cases; and the inferior alveolar and hypoglossal nerves, in 1 case each. We used the hypoglossal nerve in the latter case because the stump was the only one available in the wound, and it is thought to contain afferent sensory (spinal roots C2 and C3) fibers.13
We varied the technique for folding the flap according to the extent and location of the tissue defect (Figure 1). Our goal was to reconstruct the neotongue so that the vertical height would lie sufficiently cephalad, above the mandibular-alveolar plane, to achieve contact with the palate and, if possible, anterior contact with the lips.
Resections were classified in the "oral tongue" category if they were centered in the oral tongue, with no more than 1 cm of tongue base included.
For patients who underwent hemiglossectomy (region A in Figure 1), we used a longitudinal-fold reconstruction technique (Figure 2). A small amount of overcorrection (10%-15%) was allotted for the reconstruction, as the flap tends to diminish slightly in volume during the course of 1 year when folded in this manner.
For patients who had lost three quarters of the oral tongue (region B in Figure 1), we used a "fold-and-roll" technique. The goal was to endow dorsal bulk posteriorly and a new tongue tip anteriorly. First, we sutured and folded the flap longitudinally into the posterior extent of the defect in the native tongue, similar to the hemiglossectomy technique. The tissue was also tacked medially onto itself and/or the deep tongue muscle before it was folded laterally to fashion the floor of mouth. Then we created the tip by de-epithelializing the anteromedial region of the flap, maintaining the dermis, and rolling this component in a plane oblique to its long axis (Figure 3A). The remaining flap was inset to itself and to the anterior alveolar ridge, forming the tip. The photograph in Figure 3B shows a healed result 8 months postoperatively. Varying amounts of free-fat graft from the arm or the abdomen were added to the core of the folded flap to maintain the bulk of the neotongue in 3 patients who had thin native donor tissue to achieve the overcorrection mentioned above.
For patients who lost the entire oral tongue (region C in Figure 1), we used a multidimensional fold to reconstruct the tip and body of the oral tongue. Our goal in these cases was to reconstruct an entire new (albeit passive) tongue tip and body of oral tongue. We sutured the flap across the anterior edge of the tongue base remnant with its long axis in a transverse orientation. At the lateral floor of mouth border, the flap was sutured by means of an advancement procedure to the alveolar ridges from posterior to anterior (Figure 4A). Anteriorly, each side of the flap then met in the midline. Redundant flap edge at the anterior aspect was then elevated cephalad with a skin hook, and the unattached margins of the flap were sutured vertically in the midline (Figure 4B). This maneuver recreated the glossomandibular sulcus and established the height required for tongue-to-palate contact. The final steps consisted of depression of the midline anterior free edge of the flap inferiorly (Figure 4B), creating a 3-point junction; suture of the right and left lateral corners onto themselves to form anterolaterally directed cones; de-epithelialization of these cones; and swinging the cones anteriorly and suturing them together in the midline (Figure 5). The goal was to trap a longitudinal core of dermis for bulk and stiffening of the neotongue tip. Figure 6 shows a healed result of this technique, and Figure 7 shows how, in another case, the flap can be protruded to touch the upper lip. A free-fat graft was also added in these cases for central bulk when necessary, and the anterior belly of the digastric muscle, when available, was reattached to the mandible beneath the flap for antigravity support.
Resections were classified in the "tongue base" category if they were centered in the tongue base and did not extend into the floor of mouth or oral tongue for more than 2 cm. Such defects were reconstructed using a simple planar inset of the forearm flap between the oral tongue superiorly and the hyoid inferiorly. We sutured the base of the flap to the thyroid cartilage perichondrium if a hyoidectomy was performed. In the presence of a contiguous pharynx defect, the flap was folded 90° from the plane of the tongue base to provide lateral pharyngeal wall lining. At least 1 lingual artery and 1 hypoglossal nerve were preserved in all tongue resections. No patient required a formal supraglottic laryngectomy for oncologic reasons or a total laryngectomy later for aspiration. The epiglottis or lingual epiglottic mucosa, however, was excised as a margin when necessary, leaving denuded cartilage that we covered with a flap extension. We used sutures to suspend the larynx from the anterior mandible.
We recorded patient and flap survivals, complications, length of stay, and interval to decannulation. Swallowing was assessed by means of the Functional Outcome Swallowing Scale (FOSS).14 Scoring in this scale can be summarized as follows: 0 indicates normal; 1, minimal, fully compensated dysphagia; 2, swallow delayed up to one third, with modified food consistency; 3, decompensated dysphagia with weight loss of 10% and aspiration; 4, nonoral for 75% of nutrition; and 5, nonoral feeding. Speech intelligibility was assessed by recording a patient's recitation of 50 words from the Central Institute for the Deaf (CID) sentences.15 The recordings were then played to 5 naive individuals, all of whom had no experience interpreting the speech of patients with cancer of the head and neck (naive listeners). The scores of these 5 independent listeners were then averaged, and the resulting percentage correct score was assigned to each patient.
We included 27 men and 16 women who ranged in age from 33 to 82 years (mean age, 58 years). One patient had mucoepidermoid carcinoma, whereas the remaining 42 had squamous cell carcinoma. Twenty-two (51%) of the 43 patients had undergone previous partial glossectomy (n = 16), radiation therapy (n = 18), and/or chemotherapy (n = 5) before the surgery described in this study. Twelve patients underwent postoperative radiotherapy before speech assessment. We reconstructed the oral tongue and floor of mouth in 30 of the 43 patients studied and the base of the tongue in the other 13. Of the oral tongue resections, 4 were from region A, 19 from region B, and 7 from region C (Figure 1). Three patients undergoing region B and 2 undergoing region C resections required fat grafts, which lost no more than 10% to 15% of their bulk by 1 year. Two patients also underwent segmental and 3 underwent marginal mandibulectomies. In 38 (88%) of the 43 patients, we were able to complete the reconstruction of the tongue so that the flap's soft tissue projected vertically above the plane of the mandibular alveolus (3-dimensional reconstructions). Five patients undergoing oral resection healed with only a 2-dimensional result, ie, there was no neotongue, and soft tissue reached only to the plane of the mandibular alveolus, not cephalad toward the palate (Figure 8). These patients' speech results are denoted with an asterisk in Figure 9. Sensation recovery was not formally measured, but light touch was recovered in most patients 6 to 12 months after surgery, sometimes with variable areas of narrow or wide 2-point discrimination.
All patients underwent decannulation at a mean of 13.7 days (range, 31 [4 to 35] days) after surgery. The mean interval for the tongue-base group was 14.6 days compared with 12.8 days for the oral-tongue group. The difference between the groups was not statistically significant.
Thirty-eight patients, 30 with oral tongue and 8 with tongue base reconstructions, were available for postoperative swallowing evaluations. Swallowing was evaluated from 1 to 36 months postoperatively. Thirty-two patients (84%) were able to maintain their nutritional requirements by mouth, whereas 6 (16%) were fed by tube. Figure 10 demonstrates the distribution of FOSS scores for the whole group under study, with a median score of 2.5 and a modal score of 2.0. Table 1 indicates measures of central tendency in scores grouped by site of defect or reconstruction, which yielded no significant difference (P = .09; Wilcoxon rank sum test), despite better mean swallowing scores for the tongue base vs the oral tongue groups.
Thirty patients were available for speech evaluation, performed at an average of 5 months (range, 1-36 months) after surgery. Speech intelligibility scores are shown in Figure 8, and summary statistics for each group are found in Table 2. The improved intelligibility scores for the tongue base vs oral tongue groups were statistically significant (P<.001) in a Wilcoxon rank sum test with pair-wise comparison of the means. The observation that 5 patients scored principally at the lower end of intelligibility scores (denoted with an asterisk, Figure 8) with only floor-of-mouth level soft tissue reconstructions is of possible significance for future studies. These 5 patients had no new tissue that could reach the palate or lips (Figure 7); in the absence of a neotongue, their final healed tissue result differed considerably from the rest of the oral tongue group. Pursuant to this, if the oral tongue group is stratified by presence (mean intelligibility score, 79%) or absence (mean intelligibility score, 41%) of a neotongue, the difference between means is significant (P<.001, Wilcoxon rank sum test).
The mean length of hospital stay was 11 days (range, 5-29 days).
Twenty-eight (65%) of the 43 patients were alive at the time of this report, which yielded a mean overall survival time of 27.4 months (range, 3-68 months).
Flap loss occurred in 1 patient in the study. Complications that required further surgery developed in 2 other patients, a hematoma in one and a fistula, which required myocutaneous flap closure, in the other. Minor problems at the resection or donor site, such as wound dehiscence, developed in 9 other patients (21%). These patients underwent successful local wound care. Postoperative medical complications developed in 9 patients, which were severe in 7. These included 1 case of an arrhythmia, 1 of diabetic coma, and 1 of penetrating bleeding peptic ulcer, all of which resulted in patient death. Two cases of respiratory failure and 1 case each of congestive heart failure and stroke also occurred. Postoperative hypertension and urinary retention developed in 1 patient each.
To improve tongue reconstruction for enhanced swallowing, several techiniques have used a variety of flaps.16- 18 Using a pectoralis myocutaneous flap, 8 (62%) of 13 patients with near-total glossectomy were able to swallow.19 Salibian et al16 described 7 patients who had ulnar forearm flaps for reconstruction of subtotal defects. They found that all patients were able to feed orally. Specifically, 4 patients resumed a regular diet, whereas the remaining 3 received a soft diet. Using a free rectus flap for reconstructing 6 subtotal defects, Lyos et al18 reported that 3 patients (50%) had diets of pureed food or better. Finally, Yamamoto et al20 described 3 patients with subtotal glossectomy defects, all of whom achieved a soft diet.
Swallowing outcomes were satisfactory in our study. Most patients (84%) were able to maintain their nutritional needs by mouth. The mean score of 2.4 and a mode of 2.0 on the FOSS scale14 for 30 patients who had oral tongue reconstruction indicated that compensated dysphagia is a prevalent outcome. For the 8 patients who underwent tongue base reconstruction, a mean score of 1.6 indicated that dysphagia was also compensated in this group. The difference in scores between the oral tongue and tongue base groups was not statistically significant, but the trend is interesting considering the important sensory and motor function of the tongue base in deglutition. Also, only 1 patient of 36 in the 3-dimensional reconstruction group was dependent on nonoral feeding. This finding translates into 97% of this subgroup of patients successfully feeding by mouth. All patients underwent decannulation, indicating that any aspiration was tolerable and that the reconstruction did not compromise airway patency.
Speech production and intelligibility are often adversely affected after glossectomy.21,22 Articulation deficits are associated with the extent of resection, mobility of the tongue remnant, and method of reconstruction.4,23,24 Rentschler and Mann25 described 9 patients who had isolated anterior tongue defects and found a mean intelligibility score of 47%. A similar study by Leonard and Gillis26 of 3 patients with anterior tongue defects showed a mean score of 39%. Pauloski et al27 described 16 patients with isolated anterior tongue defects and found a mean speech intelligibility score of 61%. These patients, however, had limited resections (<20%) of the anterior tongue and had varying methods of reconstruction. A study from Japan consisting of 5 patients who had radial forearm free flaps for anterior tongue reconstruction found that intelligibility scores were 49%.28 In a reconstructive technique comparison study, McConnel et al29 described 9 patients who had small defects of the anterior tongue (0%-20% resected) or base of the tongue (0%-25% resected) who underwent reconstruction with free flaps. They found that the mean speech intelligibility score was 49% in the combined patient groups of oral tongue and tongue base reconstructions. None of these patients, however, had as great a percentage of their tongue resected as any patient in our study, so their results are difficult to compare with ours. Overall, their study showed that the best speech results for limited resections (0%-30%) of the tongue are achieved with primary closure. Our clinical experience also supports this finding. Few studies have addressed speech in isolated tongue base defects. A study by Logemann et al30 of 11 patients with greater than 1 cm of base of tongue, tonsil, and faucial arch resected showed a mean intelligibility score of 78%. These patients, however, underwent resection of the mandible on the side of the tumor and primary soft tissue closure in all cases. Our very encouraging speech results in the tongue base group suggest that the anterior tongue mobility was facilitated by the supple free flap at the base.
We found that patients who underwent 3-dimensional reconstructions of the tongue had superior speech results. Specifically, the mean intelligibility score was 98% for patients undergoing tongue base resection and 79% for those undergoing oral tongue resection. When both groups were combined, patients with 3-dimensional reconstructions had a mean speech intelligibility score of 87%. Our results therefore compare favorably with the literature to date for anterior tongue and base of tongue defects.
Limitations of our study include lack of control for some factors that may have an impact on functional results, such as the number of sessions with the speech and swallowing therapist, reinnervation, radiation, volume of tissue other than the tongue (eg, floor of mouth mucosa and muscle) removed, mandible surgery, cranial nerves resected, and route of surgical approach to the aerodigestive tract. However, McConnel et al31 found only the percentage of the tongue resected to be a determining variable for swallowing outcome in a study that included most of the above variables. Future studies could incorporate such factors in a multivariable model. We have not proved a direct casual association between the outcomes measured and the techniques used, such as would be proved by a prospective, controlled study. Clearly, many factors contribute to complex neuromuscular functions such as speech and swallowing, but other reports support the hypothesis that method of reconstruction contributes to speech and swallowing outcomes. We believe that our study adds an additional perspective because of its use of a set of previously undescribed tissue manipulation techniques.
Folded fasciocutaneous flaps survived reliably and were associated with functional recovery of patients with large (>50%) oral tongue and tongue base resection defects. The versatility, robust axial blood supply, and pliability of these flaps permit the use of longitudinal fold, fold-and-roll, or multidimensional fold techniques in oral tongue reconstruction. A planar orientation or 90° fold is useful for tongue base defects. These folding techniques are associated with satisfactory postoperative speech intelligibility and swallowing and significantly superior speech, but not significantly poorer swallowing in tongue base vs oral tongue cases (univariate analysis). Our data also raise the possibility that a functional element in oral tongue reconstruction is a neotongue that extends above the upper margin of the mandibular alveolus. Poorer speech resulted in our study if the flap was simply placed or ended up at the plane of the floor of mouth, without vertical bulk that can contact the palate and/or the lips. However, other factors such as mandibulectomy or scarring from a fistula or revision surgery could have played a role in this difference and should be controlled for in future studies. These studies should be designed as multivariable analyses and could include the hypotheses that the site of the defect (oral vs base) and final height of the neotongue between the lower alveoli and palate may be predictive variables for function, especially speech outcomes. In the interim, surgeons may wish to use the techniques of tongue reconstruction detailed in this report to optimize postoperative function in their patients.
Accepted for publication June 27, 2002.
We thank the cosurgeons for patients' resections: John Fredrickson, MD; Donald Sessions, MD; and Gerson Spector, MD.
Corresponding author and reprints: Bruce H. Haughey, MBChB, FRACS, Division of Head and Neck Surgical Oncology, Department of Otolaryngology, Washington University School of Medicine, Campus Box 8115, 660 S Euclid, St Louis, MO 63110 (e-mail: firstname.lastname@example.org).