Reconstruction in Salvage Surgery for Head and Neck Cancers

Abstract

Introduction: Salvage surgery is the treatment option in recurrences and second primary tumors. This paper aimed to study the options and outcomes of reconstruction and the predictors of poor reconstructive outcomes in salvage surgery for head and neck cancers. Study Design: This is a retrospective study of all patients who underwent reconstructive flap surgery as part of salvage surgery for head and neck cancers between the years 2004 and 2017. Methods: The initial treatment may be single modality radiotherapy or surgery or multimodality with combinations of surgery, radiotherapy, and chemotherapy. Any pathology that required surgical salvage was included. Any procedures done purely as reconstructive surgery were excluded. Predictor variables included demographical, clinical, and treatment factors. The outcome parameter was the occurrence of any flap-related complication or not. The complications and morbidity related to the procedures are reported. Results: Ninety-three patients underwent loco-regional flaps (LRF group), and 100 had free flaps (FF group). Pectoralis major flap was the commonest flap used in 68 patients (73.1%). Anterolateral thigh (ALT) flap was the commonest free flap and comprised 41% of the FF group. Any skin-related complication was seen in 35 patients (37.6%) and 41 (41%), respectively, in LRF and FF subsets. Any flap-related complication was seen in 16 patients (17.2%) and 29 patients (29%), respectively, in LRF and FF subsets. A summary measure “any one of the complications” was seen in 46 (49.5%) and 57 (57%), respectively, in LRF and FF subsets. Univariate and multivariate analysis for any flap-related complication identified no statistically significant predictor. Conclusions: Soft tissue flaps were preferred in salvage reconstruction, though the defects had a bony component. In the microvascular free flap reconstruction era, pectoralis major flap has shifted its role from a “workhorse flap” to a “salvage flap.” About half of the patients develop some complications. Flap-related complications are also common. In salvage surgery, it is important that an appropriate flap is selected, suitable for the setting, according to the indications, neck, and patient conditions.

Introduction

Recurrences and second primary tumors are common in head and neck cancers. Salvage surgery is the treatment option that is commonly opted in such cases. [1] Reconstruction plays a pivotal role in salvage surgery for both functional rehabilitation and esthetic appearance. Reconstruction in salvage head and neck cancer ablative surgical defects is most challenging due to the effects of previous treatments like surgery, radiotherapy, or chemo-radiotherapy. [2] Surrounding tissues, especially in the index tumor area and neck, would have compromised healing property, attributable to previous treatment causing fibrosis and damage to the blood vessels. [3] Vascularized free tissue transfer can be safely used in salvage oral cavity and oropharyngeal cancer defects following radiotherapy or concurrent chemoradiation considering the increased post-operative complications and hospital stay. [4,5,6] It would be best to know the predictors of poor outcomes before embarking on flap reconstructive surgery.

This paper aimed to study the options and outcomes of reconstruction in salvage surgery for head and neck cancers. The specific objectives were to analyze the predictors of poor reconstructive outcomes in patients who underwent flap reconstruction as part of salvage reconstructive surgery.

Methods

This is a retrospective analysis of all patients who underwent reconstructive flap surgery as part of salvage surgery for head and neck cancers between the years 2004 and 2017. Appropriate ethical and scientific review board approvals were taken. Since this is a retrospective study, informed consent from each patient was not required. The electronic medical records of the clinical, surgical, pathological, and follow-up details were reviewed to obtain the data. Salvage surgery is defined as any surgery to treat recurrences or any subsequent primary cancer. The initial treatment may be single modality radiotherapy or surgery or multimodality with combinations of surgery, radiotherapy, and chemotherapy. Any pathology that required surgical salvage was included. Any procedures done purely as reconstructive surgery were excluded.

Study Variables

Demographic characteristics, initial cancer and treatment details, type of recurrence, the details of salvage surgery including reconstruction, the details of morbidity, and mortality including those of the perioperative period and early follow, up to 6 months after the salvage surgery were collected. The cancer survival and disease outcomes were reported in an earlier paper. [7] The outcomes and predictors of salvage surgery, including the morbidity, are also being reported as a separate paper. This article concentrates on reconstructive surgery alone. The predictor variables were either continuous or nominal dichotomous. The age, functional status at the time of salvage surgery (Karnofsky’s Performance score (KPS)), and the duration in months from the date of completion of the initial treatment to salvage surgery date were the continuous variables. The dichotomous variables were age grouped as below 60 years and above or equal to 60 years, gender, smoking status (current/ previous smoker or not), alcohol habit, any medical comorbidity status, Site of the disease grouped as either larynx or hypopharynx/oral cavity, oropharynx, paranasal sinus or nasopharynx, prior surgery or not, prior radiotherapy or not, prior chemotherapy or not, a non-surgical modality of initial treatment (Radiotherapy (RT) vs chemo-radiotherapy (CTRT)), the type of recurrence or second primary (local/ regional vs local and regional), the T classification at sal- vage (rT1T2 vs r T3T4), N classification at salvage (r N0 vs N+) and the reconstructive procedure at salvage (free flap vs local/regional reconstruction. The outcome parameter was occurrence of any flap-related complication or not (including any of total flap loss, partial flap loss, or any flaprelated re-exploration procedure).

The complications and morbidity related to the procedures are reported in 2 subsets, the patients who underwent local or regional flap procedures and those undergoing any free flap procedure. The morbidity is summarized as (1) Any skin-related complication as a composite and separately for wound infection, oro-cutaneous fistula and pharyngocutaneous fistula, (2) Any flap-related complication as a composite and separately for any total flap loss, partial flap loss, and flap described re-exploration procedure, (3) Any systemic complication, (4) Post-operative mortality and a composite measure of all complications, (5) Any one of the complications.

Statistical Analysis

A descriptive summary of the variables using numbers and percentages is presented first as 2 groups. A univariate analysis using binomial logistic regression was done. The P-value, Odds Ratio (OR), and the 95% confidence interval limits are presented. Subsequently, a multivariate model was built for each outcome variable incorporating the significant variables in the initial univariate analysis. A P-value of less than .2 was used as the cut-off for this selection. In the multivariate logistic regression model, a P-value of less than .05 was considered statistically significant. Multicollinearity was seen among the predictors using variance inflation factor (a cut-off value of 10). Prior chemotherapy and the modality of initial treatment (RT vs CTRT) were seen to have multicollinearity, and hence prior chemotherapy was removed from the models.

Results

283 patients underwent salvage surgery. This paper focuses on a subgroup of 193 patients who underwent flap reconstructive surgery as part of salvage surgery. The data were analyzed in 2 groups, 93 patients who underwent local or regional flaps (LRF group) and 100 patients who underwent free flap surgery (FF group). The mean (SD) age of the patients in the LRF group was 61.4 (13.1) years, and that of the FF group was 54.6 (11.5) years. The mean (SD) Karnofky’s performance status for the LRF and FF groups were 85.4 (7.7) and 87.6 (6.2), respectively. The mean duration between the primary surgery and salvage surgery was 23.2 months in the LR group and 34.9 months in the FF group. Males were common in both the groups; 79.6% and 75% in LRF and FF groups. The majority of the patients had oral cavity primary. Squamous cell carcinoma was the predominant pathology (96.8%). 49 patients (52.1%) and 45 patients (47.9%) had initial non-surgical treatment.

Table 1. Demographics, Clinical Characteristics, and Details of Initial Treatment.

shows the demographic, initial disease, and treatment characteristics. The majority of the patients in both groups had local recurrence. Thirty-two patients (34.4%) and 39 (39%) had T4a stage (AJCC seventh Edition TNM Classification) on recurrence. In the LRF group, surgery involving primary alone was predominant; 52 patients (55.9%). But in the FF group, combined primary and neck surgery was the predominant category. Fifty-nine patients (59%) underwent such surgery.

Table 2. Details of Recurrence.

shows the details of recurrence.

Salvage Reconstruction

Pectoralis major flap was the commonest flap used in salvage reconstruction. Sixty-eight patients underwent this flap constituting 73.1% of the LR group. Anterolateral thigh (ALT) flap was the commonest free flap and comprised 41% of the FF group.

Table 3. Details of Salvage Surgery and Reconstruction.

shows the details of the flap surgery.

Table 4. Details of the Vessel Anastomosis in Free Flap Subset (n = 100).

shows the details of free flap vessel anastomosis. Ipsilateral neck vessels were available in 79 (79%) cases, though violated. Ipsilateral superior thyroid vessels and facial vessels were the commonest vessels used in salvage free flap reconstruction. If the contralateral neck was used, the facial artery and vein were available.

Table 5. Site-Wise Distribution of Flaps Used in Salvage Reconstruction.

shows the site-wise distribution of the flaps used. In oral cavity, 36 had pectoralis major flap, 28 had ALT flap, 30 had radial forearm flap (RFFF). Free fibula flap for bony reconstruction was used in 30 patients. In larynx and hypopharynx subsite, pectoralis major flap was preferably used.

Complications in the FF and LRF Groups

Table 6. Details of Complications in Salvage Reconstructive Surgery.

shows the list of complications. Any skin-related complication was seen in 35 patients (37.6%) and 41 patients (41%), respectively, in LRF and FF subsets. Any flaprelated complication was seen in 16 patients (17.2%) and 29 patients (29%), respectively, in LRF and FF subsets. Total free flap loss occurred in 7 free flaps and 3 local/regional flaps. Lost free flaps were replaced by subsequent free flaps in 4 cases and pectoralis major flap in the remaining 3. There was 2 pectoralis major flap loss, one of which was managed with a forehead flap and the other defect was corrected with a prosthetic obturator. One submental flap loss healed without any subsequent flap. Systemic complications occurred in 9 (9.7) % and 8 (8%), respectively. In terms of group comparability, free flaps were more done in the group with the oral cavity (P < .001) cancer patients and locoregional flaps more in the larynx/hypopharynx group. Patients who had prior surgical modality had more free flaps (P = .030), and patients with previous radiotherapy had more loco-regional flaps (P = .006). The mean (SD) ICU stay for the LR group and FF group were 2.9 (1.8) days and 5.6 (1.9) days, respectively. The mean (SD) hospital stay for the LR group and FF group were 14.2 (6.5) days 18.3 (10.3) days, respectively.

Univariate and Multivariate Analysis

A Univariate analysis for any flap-related complication in the group of patients who underwent flap reconstruction identified no statistically significant predictor. The variables nearing statistical significance (P < .2) were tested in the multivariate model. Multivariate analysis also could not identify a significant predictor (

Table 7. Univariate and Multivariate Analysis for any Flap-Related Complication as the Outcome in Patients with Flap Reconstruction, (n = 193).

).

Discussion

The role of salvage surgery is critical in head and neck cancers, and surgery is generally regarded as the best treatment option in recurrent resectable head and neck cancers. [1] Surgeons face an increasingly challenging job of ablative and reconstructive procedures in patients with significant effects of failed previous surgical or non-surgical treatment. Reconstruction plays a pivotal role in salvage surgery for both functional rehabilitation and esthetic appearance. 55%–66% of patients need pedicle flaps or free flaps in salvage surgery to repair surgical defects. [8,9] Data from a meta-analysis by Goodwin et al [10] showed that salvage surgery conveyed a fair chance of 5-year survival (39.4% in 1080 patients). Overall 2-year disease-free survival in the prospective observational study part of the same paper [10] was comparable with the meta-analysis average (44% vs 51%). The lower mortality rate in the prospective observational study was attributed probably due to the more reliable reconstructive techniques than meta-analysis (1.8% vs 5.2%). Surgical complications were reported in 20% of patients in the prospective study and 39% of 388 patients of meta-analysis. Wound complications were manageable, even though 99% of patients in meta-analysis and 88% in the prospective study received prior full course radiotherapy. [10] A recent systematic review and meta-analysis of pooled data reported complications rates as 33% fistulas, 24% wound infections, and 3% flap failures in salvage surgery for advanced-stage cancers who received prior radiation or chemoradiation. [11]

It would be best to know the predictors of poor outcomes before embarking on flap reconstructive surgery. In the present study, none of the factors were found as statistically significant predictors for flap-related complications, especially previous treatments like radiotherapy, surgery, or chemotherapy. In the recurrent cases, salvage surgery itself would be the major determinant of flap-related complications, and the different subsets in it might not have significant variations within them. These findings are consistent with other studies in the literature. [6,12,13,14] But, Mücke et al [15] study showed that previous neck dissections and radiotherapy were significant negative predictors for success in free flap transfer.

There are studies in literature comparing free flaps vs loco-regional flaps. A study by Chepeha et al comparing pectoralis flap vs free flap showed that minor complications rate (57% vs 21%, P < .001) and hospital stay (14 days vs 12 days, P < .006) was more in the pectoralis flap group. Gastric tube dependence was more in the pectoralis flap group than the free flap group (42% vs 17%, P < .004) in the salvage setting. [16] Quality of life analysis in oral cavity cancer patient study showed that free flap group had better speech, shoulder function and better mood status than pectoralis group. [17] Pectoralis major myocutaneous flap patients had more complications than the free gracilis flap group (12 vs 8, statistically not significant) for buttressing the repair of salvage laryngectomy defect repair, but fistula rates were similar. [18] A systematic review and meta-analysis by Jørgensen et al [19] that compared submental flap vs free flap showed that complete flap loss, debulking revisions and oncological recurrences were comparable except operating time and length of hospital stay. Operating time (6.7 vs 8.1 hours, P =.002) and hospital charges (32% lower, P = .0001) were significantly lower in the supraclavicular island flap compared to the fascio-cutaneous free flaps for head and neck reconstruction, but wound infection rate, wound dehiscence rate and length of stay were comparable. [20,21] We have not made an attempt to statistically compare the outcomes of FF group and LRF group to prove the superiority of one over the other, in this setting. There is an obvious selection bias while selecting the reconstruction in salvage surgery. Free flaps are more done in the oral cavity and loco-regional flaps in the larynx. Patients who underwent prior surgical modality had more free flaps, and patients with previous radiotherapy had more loco-regional flaps. But, this preference is probably due to the appropriateness of the flap for the indications. Pectoralis major flap is preferred to and better in the larynx in laryngopharyngectomy defects. In salvage surgery, it is important that an appropriate flap is selected, suitable for the setting, according to the indications, neck and patient conditions. Comparison between free and non-free groups may be flawed in this situation due to the unavoidable selection bias and the choice is determined by the patient and defect characteristics.

Soft tissue flaps were preferred in salvage reconstruction, though the defects had a bony component. Bony reconstruction was done only in 23 cases (12%) in the present series. The reasons for this preference are multifactorial. Laryngopharyngeal defects do not have a bony component. Complex oromandibular defects in salvage surgery are usually large with an inner lining, outer cover, and bony components. But the preference is to give the lining and cover over the bone for immediate speech and swallowing rehabilitation. This is comfortably possible with ALT flap, radial forearm flap (RFFF) or pectoralis major flap, where bipaddling can be done to attain this goal. The soft tissue component of the defect is not satisfactorily covered by the skin paddle of the fibula flap or the musculocutaneous element of the deep circumflex iliac artery (DCIA) flap. Among the free flaps, vessel length is better for ALT and RFFF than bone flaps like free fibula or DCIA. Bony reconstruction is usually done only in absolute indications like central mandibular and maxillary defects to prevent gross functional and esthetic problems. The presence of trismus due to prior surgical and radiation-induced fibrosis is an added problem. Alleviation of trismus is usually better with a soft tissue alone reconstruction.

Soft tissue flaps were preferred in salvage reconstruction, though the defects had a bony component. Bony reconstruction was done only in 23 cases (12%) in the present series. The reasons for this preference are multifactorial. Laryngopharyngeal defects do not have a bony component. Complex oromandibular defects in salvage surgery are usually large with an inner lining, outer cover, and bony components. But the preference is to give the lining and cover over the bone for immediate speech and swallowing rehabilitation. This is comfortably possible with ALT flap, radial forearm flap (RFFF) or pectoralis major flap, where bipaddling can be done to attain this goal. The soft tissue component of the defect is not satisfactorily covered by the skin paddle of the fibula flap or the musculocutaneous el- ement of the deep circumflex iliac artery (DCIA) flap. Among the free flaps, vessel length is better for ALT and RFFF than bone flaps like free fibula or DCIA. Bony re- construction is usually done only in absolute indications like central mandibular and maxillary defects to prevent gross functional and esthetic problems. The presence of trismus due to prior surgical and radiation-induced fibrosis is an added problem. Alleviation of trismus is usually better with a soft tissue alone reconstruction.

Pectoralis major was the most commonly used flap in the present study. It has many advantages in the salvage setting. In the microvascular free flap reconstruction era, this flap has shifted its role as a “workhorse flap” to a “salvage flap.” It can be used for primary reconstruction, salvage, or emergency reconstruction due to free flap failure, oro/pharyngocutaneous fistula closure and protection to the carotid artery in emergency blowouts. It can also be used as a simultaneous reconstruction along with a free flap to fill the dead space and to cover the cervical skin defects and great vessels. [22,23,24] It is available in the adjacent non-irradiated field, sturdy and better for logistic reasons in patients with multiple comorbidities. There is no need to depend on the status of the neck vessels. Also, its ability to accept skin grafts eliminates an additional necessity of a flap for skin cover. Sayles and Grant, in their systematic review and meta-analysis, have shown that prophylactic onlay flaps can prevent pharyngocutaneous fistula in total laryngectomy; the number needed to prevent one fistula was 6.05.

The authors concluded that prophylactic flaps should be offered in the salvage setting, especially in post chemoradiation patients. The most common flap used as an onlay in this study was the pectoralis major flap (151 out of 188 cases). [25] A systematic review by Paleri et al showed the clear advantage of the vascularized flap outside the radiation field to repair the salvage laryngectomy defect. Pharyngocutaneous fistula formation is 50% higher in primary closure than in flap reconstruction, and to prevent one fistula, 11 patients need vascularized flaps for reconstruction of salvage laryngectomy defect. [4]

There are reports showing the safety of free flaps in salvage surgery with a success rate of 91.3%–96.8%. [5,6,14,15,26,27,28,29] The commonest free flaps used for salvage surgery in this study are ALT flap and RFFF. ALT flap can be harvested as thin fascio-cutaneous flap to chimeric flap with fascia lata, vastus lateralis and rectus femoris muscle and skin with good pedicle length, concealed donor site scar and simultaneous two-team approach for flap harvest and resection of head and neck cancers. [30] Complex defects after salvage surgery can be reconstructed with good aesthetic, functional results and low complications by double island ALT flap. [31] ALT flap can be used for total or partial circumferential pharyngo-esophageal defect reconstruction (71% had prior surgery, radiotherapy, or both) with comparable fistula and stricture rates, minimum donor site morbidity and quick recovery rates. [32] RFFF is a thin fasciocutaneous flap with a long pedicle, able to reach the contralateral side of the neck for micro-anastomosis without any vein graft. RFFF can be used to cover the long mucosal salvage surgical defects in the oral cavity, tonsil, palate, larynx, and hypopharynx. [12]

Vessel depleted neck is a challenge in salvage surgery. Ipsilateral neck vessels were available in 79% of our cases, though the neck was violated by previous surgery. Ipsilateral superior thyroid or facial vessels were available and used. Proper pre-operative assessment and multiple backup plans for each patient based on individual characteristics would be essential. [33] Ipsilateral facial artery and vein, lingual artery and superior thyroid vessels are routinely used for microvascular reconstruction. In majority of the salvage surgical cases, these vessels would be either used in previous surgery or affected by the prior radiotherapy. Dual-phase Computed Tomography angiography was found helpful to find the recipient artery and vein. [34] Options for arterial anastomosis in vessel depleted neck are superficial temporal artery, transverse cervical artery, external carotid artery, internal mammary artery, inferior thyroid artery, thoracoacromial artery, and common carotid artery. [20,35,36,37,38] Veinous options are external jugular, superficial temporal, internal mammary, thoracoacromial veins, and cephalic vein transposition. [33,35,38,39] Corlett loop, that is, an extended length of the cephalic vein, can be used to achieve arterial and vein anastomosis when local artery and vessels are not available. [40] In extreme situations, previous flap pedicle [41,42] and extracorporeal perfusion [43] can also be used for microvascular reconstruction.

The strength of the study is the sample size. This is one of the largest studies which has looked into the morbidity related to flap reconstruction in the salvage surgery setting. The limitation from a methodological viewpoint would be the selection bias in the comparison between FF and LRF groups. But we aim to show that within the existing protocols of flap selection, either of these would be done with similar complication rates. Being a retrospective study, the inherent limitations might have also influenced the results.

Conclusions

Soft tissue flaps were preferred in salvage reconstruction, though the defects had a bony component. Pectoralis major was the most used flap in the present study. It has many advantages in the salvage setting. In the microvascular free flap reconstruction era, this flap has shifted its role from a “workhorse flap” to a “salvage flap.” The study could not identify a predictor for morbidity related to flap reconstruction. In salvage surgery, it is important that an appropriate flap is selected, suitable for the setting, according to the indications, neck, and patient conditions. Statistical comparison between free and non-free groups to prove the superiority of one over the other may be flawed in this situation due to the unavoidable selection bias. The choice is determined by the patient and defect characteristics.