Design of skin islands for a myocutaneous serratus anterior free flap—An anatomical study and clinical implication for pharyngeal reconstruction after laryngopharyngectomy

Objectives The main purpose of this study was to evaluate flap size and flap design of skin islands in myocutaneous serratus anterior free flaps (SAFFs) in fresh cadavers and to further investigate whether myocutaneous SAFFs are suitable flaps for pharyngeal reconstruction after laryngopharyngectomy. Methods Dissection and injection of methylene blue were performed in 20 hemithoraces of 13 fresh cadavers to evaluate flap size and location of skin islands. Based on these pre‐clinical data, we performed pharyngeal reconstruction with myocutaneous SAFF in five patients after laryngopharyngectomy. Results Perfused skin paddles were found in all specimens with a mean size of perfused skin islands of 85.6 ± 49.8 cm2. Lengths and widths of skin islands ranged from 10‐21 cm and 6‐20.5 cm respectively. Flap size did not significantly differ between males and females (P = 0.998), left compared to right hemithoraces (P = 0.468) and between paired specimens (P = 0.915). All skin islands were found within the upper 29.3%‐51.7% of hemithorax (calculated from axilla to costal arch), and between latissimus dorsi muscle posteriorly and anterior axillary line anteriorly. Accordingly, myocutaneous SAFFs were used for pharyngeal reconstruction after laryngopharyngectomy in five patients with advanced hypopharyngeal carcinomas. Three patients had uneventful courses, while one patient developed immediate intraoperative flap loss and another patient developed partial necrosis of SAFF on postoperative day 7. Conclusion Skin islands of SAFF have reliable blood supply, which allow harvest of large myocutaneous SAFFs that can be used also for pharyngeal reconstruction after laryngopharyngectomy.

its main vascular supply from the thoracodorsal artery (TDA), representing a branch of the subscapular artery (SA), which in turn originates from the axillary artery (AA). 6 However, numerous anatomical variations are possible that have been already described in literature.
Accordingly, vascular supply can directly arise from the AA, 8 the intercostal artery 9,10 or the lateral thoracic artery. 2,11 Particularly for reconstruction of the head and neck, the SAFF offers some significant advantages, including a long and reliable vascular pedicle, the thinness and pliability of the flap, ease of harvest and a low donor site morbidity. Moreover, due to its central localisation, the SAFF has a low risk for peripheral arterial vascular disease. 6 Although the exact extent and localisation of perfused skin paddles for myocutaneous SAFF have not been defined so far, the SAFF has been successfully used as a myocutaneous free flap for the reconstruction of the face, 2 the head and neck 12 or the lower limb. 13 Hence the main purpose of the study was to perform an anatomical study in order to evaluate the optimal area and size for the harvest of skin islands perfused by the thoracodorsal artery and to define applicable recommendations for clinical routine. Based on these pre-clinical data, we additionally report on our first clinical experiences of pharyngeal reconstruction using a myocutaneous SAFF after laryngopharyngectomy.

| MATERIAL AND METHODS
The anatomical part of the study was performed at the Center of

| Anatomical study
Twenty hemithoraces, including the upper limb, were dissected from 13 fresh cadavers (five female, eight male). Six specimens were excluded due to former operations on the lateral thoracic wall or chest (eg, insertion of pacemaker or thoracotomy). The specimens were obtained from voluntary donors who consented during lifetime to donate their body for research and teaching purpose to the Center for Anatomy and Cell Biology at the Medical University of Vienna. Specimens were dissected in supine position with full abduction (90°) of the upper limb. First, a vertical skin incision (~10 cm) within the anterior axillary was done. Second, the pectoralis major muscle was separated and the AA was identified at the level of its transition into the brachial artery. Next, the SA was found in the third part of the AA, which directly ends up in the TDA. After the TDA was cannulated, 40-60 mL methylene blue was injected (Figure 1). Finally, size and localisation of skin islands were evaluated.

| Assessment of skin islands
Two different methods were used for assessing and analysing skin islands. First, a coordinate system was used consisting of anatomical reference lines (posterior axillary line, anterior axillary line, midclavicular line) and landmarks (costal arch, axilla, latissimus dorsi muscle, nipple) to describe and analyse localisation of skin islands. The distance between axilla and costal arch (ACA), measured within anterior axillary line, has been used as surrogate marker for length of hemithorax. Moreover, distances between costal arch to the inferior border (CAIB) and axilla to the superior border (ASB) of stained skin islands were measured to assess the vertical position of skin flap. Additionally, the horizontal extent of stained skin island was evaluated with regard to latissimus dorsi muscle and posterior axillary line posteriorly and nipple with midclavicular line anteriorly. We used a 1.0 mm ruler to measure the longest vertical (length) and horizontal (width) diameter of stained skin islands. After all variables were measured, stained skin islands were marked and photographed. In a second step, all photos were processed with IMAGEJ software (open source Java image processing program) and size of marked and stained skin islands were finally assessed.

| Clinical implications
Based on our pre-clinical anatomical results, we used myocutaneous SAFF for pharyngeal reconstruction in five patients who underwent laryngopharyngectomy due to a stage IV hypopharyngeal squamous cell carcinoma (SCC). All patients were treated at the Department of

| Statistical methods
SPSS software (version 21; IBM SPSS Inc, Chicago, IL, USA) was used for statistical analysis of data. All data are indicated as mean ± standard deviation (SD) within result section. Unpaired students t test was used to compare means of normally distributed variables, while Chi-square test was performed to investigate the association between nominal variables. All tests were two-sided and P-values below 0.05 were considered as statistically significant.

| Ethical considerations
Ethical approval was obtained from the ethics committee of the

| Flap size
We dissected and analysed 10 left and 10 right specimens of seven paired and six unpaired hemithoraces. Perfused skin islands were found in all specimens (20 out of 20). Mean length and width of skin islands were 15.5 ± 9.9 cm (range 10-21 cm) and 10.9 ± 13.9 cm (range 6-20.5 cm). Mean area of stained skin islands was 85.6 ± 49.8 cm 2 with a range of 38.4-223.6 cm 2 (Table 1). Despite the great diversity of stained skin islands, there was no significant difference between female compared to male cadavers (85.6 ± 35.3 cm 2 vs 85.5 ± 59.1 cm 2 ; P = 0.998) or left compared to right hemithoraces (93.9 ± 44.7 cm 2 vs 77.2 ± 55.5 cm 2 ; P = 0.468; Figure 2). Additionally, area of stained skin islands did not significantly differ in left compared to right paired specimens (81.6 ± 65.8 cm 2 vs 84.0 ± 43.9 cm 2 ; P = 0.915).

| Localisation of flap
Next, we wanted to define the area with highest probability for successful harvest of skin islands. Therefore, several anatomical reference lines and landmarks were used for description and characterisation of skin flaps. Mean distance between axilla and costal arch (ACA) was 30.8 ± 2.9 cm. Distance between costal arch (CA) and inferior border (IB) of stained skin island was 12.4 ± 1.8 cm (CAIB; range 9-16 cm). On the other hand, mean distance between axilla (A) and superior border (SB) was 3.3 ± 2.7 cm (ASB; range 0-8.5 cm). In 25% of cases (five out of 20), skin islands reached superiorly to axilla. To simplify flap harvest, we described superior (ASB) and inferior border (CAIB) of stained skin islands related to ACA, which was used as marker for the length of hemithorax. With respect to ACA, superior and inferior borders were found between the upper 0%-29.3% of hemithorax, calculated from axilla, and the lower 32.8%-48.3% of hemithorax, calculated from CA respectively.

| Flap failure
Mean area of skin islands was 9 × 7.8 cm ranging from 7 × 7 cm up to 10 × 9 cm, respectively. Among our cohort, one patient developed immediate intraoperative flap loss and another patient developed partial necrosis of the skin island within postoperative course. and subsequently cervical anastomosis. Within endoscopic examination, we found a partial necrosis with defect of the skin island measuring 2 × 2 cm on postoperative day 7, which was most likely caused by the mechanic stress on the anastomosis. Therefore, it was most necessary to perform a reoperation to remove the necrotic area and to cover the defect with a PMF.

| Percutaneous fistula and swallowing
Those patients with flap failure also developed PCFs (two out of five). Especially the female patient developed PCF on postoperative day 14 and another fistula on postoperative day 28. It was necessary to cover both PCFs with PMFs. Conversely, no surgical intervention was necessary for the male patient and PCF closed spontaneously.
After spontaneous closure of PCF, the further postoperative course of the male patient was uneventful. In strong contrast, the female patient developed tracheoesophageal fistula 6 months after surgery, which was recently closed with a regional skin flap. Moreover, due to stenosis of the neopharynx and upper part of oesophagus and the recently covered tracheoesophageal fistula, she received nutrition via percutaneous endoscopy gastrostomy (PEG) tube.
Conversely, swallowing function was inconspicuous in the remaining four patients with uneventful postoperative course without signs of strictures of neopharynx and normal oral nutrition.

| DISCUSSION
The SAFF has a great versatility in reconstruction of the head and neck due to easy harvest, low donor site morbidity, constant vascular anatomy with long vascular pedicle and especially due to its pliability. 6 The SAFF has been already used for reconstruction of skull base, 14    were described for patients with ALT, which was lower compared to RFFF with 20% and 10% or gastro-omental free flap with 16% and 22%, but similar to jejunum flap with 12% and 11% respectively. 17 Despite similar PCF and stenosis rates between ALT and jejunum free flap, ALT is superior with regard to length of operation, donor side morbidity or tolerability of postoperative radiotherapy. The latter is significantly worse for jejunum free flaps compared to ALT flaps. 17,18 Consequently, ALT flap is primarily used for pharyngeal reconstruction followed by RFFF. 17 Although complication rates are higher in patients who under-   represents an additional option for pharyngeal reconstruction. Of course, further studies with larger patient numbers are warranted to elucidate the reliability and versatility of myocutaneous SAFF for pharyngeal reconstruction.

| CONCLUSION
In conclusion, SAFF has a reliable skin island of appropriate size that can be used for pharyngeal reconstruction after laryngopharyngectomy, when specific anatomic boundaries are considered for flap harvest.