Salvage surgery in post-chemoradiation laryngeal and hypopharyngeal carcinoma: outcome and review.

Our objective was to evaluate recurrence patterns of hypopharyngeal and laryngeal carcinoma after chemoradiation and options for salvage surgery, with special emphasis on elderly patients. In a retrospective study all patients who underwent chemoradiation for hypopharyngeal and laryngeal carcinoma in a tertiary care academic center from 1990 through 2010 were evaluated. Primary outcome measures were the survival and complication rates of patients undergoing salvage surgery, especially in elderly patients. Secondary outcome measures were the predictors for salvage surgery for patients with locoregional recurrence after failed chemoradiotherapy. A review of the literature was performed. Of the 136 included patients, 60 patients had recurrent locoregional disease, of whom 22 underwent salvage surgery. Fifteen patients underwent a total laryngectomy with neck dissection(s) and 7 neck dissection without primary tumour surgery. Independent predictors for salvage surgery within the group of 60 patients with recurrent disease, were age under the median of 59 years (p = 0.036) and larynx vs. hypopharynx (p = 0.002) in multivariate analyses. The complication rate was 68% (14% major and 54% minor), with fistulas in 23% of the patients. Significantly more wound related complications occurred in patients with current excessive alcohol use (p = 0.04). Five-year disease free control rate of 35%, overall survival rate of 27% and disease specific survival rate of 35% were found. For the 38 patients who were not suitable for salvage surgery, median survival was 12 months. Patients in whom the tumour was controlled had a 5-year overall survival of 70%. In patients selected for salvage surgery age was not predictive for complications and survival. In conclusion, at two years follow-up after chemoradiation 40% of the patients were diagnosed with recurrent locoregional disease. One third underwent salvage surgery with 35% 5-year disease specific survival and 14% major complications. Older patients selected for salvage surgery had a similar complication rate and survival as younger patients.

Introduction

Treatment of advanced stage squamous cell carcinoma of larynx and hypopharynx constitutes a challenging situation. Cisplatin-based chemoradiation is an established treatment for selected moderately advanced laryngeal and hypopharyngeal carcinoma, as it may be organ and function sparing -.

For recurrent disease, salvage laryngectomy or neck dissection may be available as a curative option for selected patients. However, the complication rate of salvage surgery after chemoradiation is relatively high. Wound healing problems are a well-known consequence of surgery in irradiated patients. Fistula rates of 11-58% after salvage laryngectomy are reported -. Due to further locoregional recurrence after salvage laryngectomy, distant metastases, second primaries and other causes, the 5-year overall survival is in the range of 31-57% -. The literature suffers from heterogeneity as to tumour site, previous therapy and salvage therapy.

Herein, we aim to provide insight into the recurrence pattern after chemoradiation for laryngeal and hypopharyngeal carcinoma and the options for salvage surgery. We were specifically interested in the complications after salvage surgery, with focus on age. Moreover, the outcome of patients after salvage surgery is evaluated.

Materials and methods

Patients

Sixty patients with locoregional disease after chemoradiation were identified from a database of 136 patients with laryngeal or hypopharyngeal squamous cell carcinoma treated by chemoradiation with curative intent between January 1990-April 2010. The hospital charts of these patients were retrospectively reviewed. In all patients response to chemoradiation was evaluated within or at 3 months after treatment, unless patients had died during or shortly after the chemoradiation. Resectability prior to treatment was determined by physical examination, imaging and endoscopy. Approval of the Medical Ethics Committee of the VU University Medical Center in Amsterdam was obtained. Patient and treatment characteristics are shown in Table I.

Table I.

Patient characteristics of 60 patients with locoregional disease after chemoradiation.

Variables	Number	Percentage (%)
Gender
Male	51	85
Female	9	15
T-stage (prior to chemoradiation)
T2	3	5
T3	30	50
T4	27	45
N-stage (prior to chemoradiation)
N0	15	25
N1	7	12
N2a	3	5
N2b	13	22
N2c	17	28
N3	5	8
Primary site
Hypopharynx	35	58
Larynx	25	42
Operability
Unresectable	4	7
Organ preservation approach	56	93
Chemoradiation schedule
Cisplatin IA*	2	3
Cisplatin IV*	24	40
Cisplatin/5-FU alternating**	14	24
Cisplatin/5-FU sequential ***	13	22
Cetuximab****	5	8
Cetuximab/TPF/Cisplatin or carboplatin*****	2	3

Concurrent four intra-arterial cisplatin (150 mg/m2) infusions or three intravenous cisplatin (100 mg/m2) infusions. In both schemes patients were radiatedwith 70 Gy irradiation (6-7 weeks);

cisplatin 20 mg/kg and 5-FU 200 mg/kg intravenously in week 1, 4, 7, 10; radiotherapy in week 2, 3, 5, 6, 8, 9, total dose 60 Gy;

cisplatin 100 mg/kg and 5-FU 1000 mg/kg intravenously, 4 courses; followed by 7 weeks radiotherapy, total dose 70 Gy;

weekly cetuximab in combination with 7 weeks radiotherapy, total dose 70 Gy;

2-4 courses of TPF (Docetaxel, Platinum, Fluorouracil), followed by cisplatin or carboplatin with concurrent 70 Gy radiotherapy (7 weeks), in some patients cetuximab was given during this treatment.

We defined chemoradiation as the combined use of cisplatin based chemotherapy and/or targeted therapy and radiotherapy for the primary treatment. Different schemes are used. Fourteen patients were treated according to an alternating scheme (cisplatin 20 mg/kg and 5-FU 200 mg/kg (i.v.) in week 1, 4, 7 and 10; radiotherapy in week 2, 3, 5, 6, 8 and 9, total dose 60 Gy) and 13 according to a sequential scheme (cisplatin 100 mg/kg and 5-FU 1000 mg/kg i.v., 4 courses; followed by 7 weeks radiotherapy, total dose 70 Gy). Twenty-four patients were treated with concomitant intravenous administration of 3 × 100 mg/m2 cisplatin on day 1, 22 and 43 with simultaneous radiotherapy. Two patients were treated according to the intra-arterial chemoradiation schedule consisting of four consecutive weekly selective intra-arterial infusions of cisplatin (150 mg/m2) followed by intravenous sodium thiosulphate rescue combined with simultaneous radiotherapy. All patients were irradiated daily for 6-7 weeks to a total dose of 70 Gy (2 Gy per fraction, 5-6/week). Two patients were treated according to the concomitant intravenous cisplatin protocol, in combination with cetuximab. Five patients received weekly cetuximab (loading dose 400 mg/m2, followed by weekly 250 mg/m2) with daily radiotherapy for 7 weeks to a total dose of 70 Gy. Both sides of the neck were radiated in all patients, regardless of the lymph node status.

Definitions

Postoperative complications were categorised into surgical complications (fistula, infection, necrosis, haemorrhage and chyle leakage), pneumonia and other complications (e.g. spondylodiscitis). Complications were classified as major if they required re-operation.

Statistics

Statistical analysis was performed with SPSS 15.0. Survival rates were calculated with the Kaplan-Meier method, with follow-up intervals calculated from the date of salvage surgery. Univariate analysis of survival parameters was done using the log-rank test. Univariate analyses of complication patterns were assessed by utilizing the χ2-test or the independent- samples T-test whenever applicable. Multivariate analysis of survival was performed with Cox regression. A model developed by Tan et al. with stratification factors for survival, was applied to our population.

Results

After a median follow-up of 25 months (range 0-130 months), 60 patients (44%) had presented with recurrent disease. One-third of the patients with recurrent disease (n = 22) underwent salvage surgery for local, regional or locoregional disease. This is 16% of the total group of patients initially treated by chemoradiation. Twenty-four percent of patients with laryngeal carcinoma vs. 10% of patients with hypopharyngeal carcinoma underwent salvage surgery. Two-thirds of patients (n = 38) were not suitable for salvage surgery because of distant metastases (n = 30), poor general condition of the patient (n = 3), refusal of surgery by the patient (n = 1) or unresectability of the tumour (n = 4).

Of the 6 patients with an initial unresectable tumour, 4 patients developed recurrent disease, which was not statistically different from the organ preservation (initial resectable) group. Two patients developed distant metastases and 2 patients were diagnosed with persistent unresectable local disease. Independent predictors for salvage surgery within the group of 60 patients with recurrent disease, were age younger than 59 years (p = 0.036) and larynx vs. hypopharynx (p = 0.002) in multivariate analyses. Gender, Tand N-stage were not associated with surgery for salvage.

The median interval between radiotherapy and recurrence for the 22 patients was 4 months.

The study population consisted of 19 males and 3 females with a median age of 59 years (range: 40-69 years), with primary tumours in larynx (n = 15) and hypopharynx (n = 7) (Table II).

Table II.

Patient and salvage surgery characteristics.

Patient	Gender	Age	T	N	Site	Recurrence	Larynx	ND*	Reconstruction	Subsequent reconstruction
1	F	52	4	1	Hypopharynx	Local	Laryngectomy	Bilateral	FRFF+PM	PM
2	M	57	3	2b	Larynx	Regional	-	Unilateral	PC
3	M	64	4	0	Larynx	Local	Laryngectomy	Unilateral	PM
4	M	63	4	0	Larynx	Local	Laryngectomy	Bilateral	PM	PM
5	M	62	4	0	Larynx	Local	Laryngectomy	Unilateral	PM
6	M	59	3	2c	Hypopharynx	Regional	-	Bilateral	PM
7	M	41	3	0	Hypopharynx	Local	Laryngectomy	Unilateral	PM
8	M	55	4	2c	Larynx	Local	Laryngectomy	Bilateral	PM
9	M	67	4	0	Larynx	Local	Laryngectomy	Bilateral	PM
10	M	57	3	2b	Larynx	Local	Laryngectomy	Bilateral	PM
11	M	52	4	3	Hypopharynx	Regional	-	Unilateral	PM
12	M	49	3	2a	Larynx	Regional	-	Unilateral	PC	$
13	M	54	4	2c	Larynx	Locoregional	Laryngectomy	Bilateral	PC
14	M	58	2	2b	Hypopharynx	Regional	-	Unilateral	PC
15	M	53	3	2c	Larynx	Local	Laryngectomy	Bilateral	PM
16	M	55	3	0	Larynx	Local	Laryngectomy	Bilateral	PM
17	M	61	3	2c	Hypopharynx	Regional	-	Unilateral	PM
18	F	59	3	0	Larynx	Local	Laryngectomy	Bilateral	PM
19	M	69	4	0	Larynx	Local	Laryngectomy	Bilateral	PM
20	F	67	3	1	Larynx	Local	Laryngectomy	Bilateral	PM
21	M	58	4	3	Hypopharynx	Local	Laryngectomy	Bilateral	PM
22	M	64	3	2c	Larynx	Regional	-	Unilateral	PM

1 year after ND: total salvage laryngectomy with FRFF and PM, followed by a second PM for complications. ND: neck dissection; PC: primary closed; PM: pectoralis major flap; FRFF: free radial forearm flap.

Neck dissection was performed in all patients. In 15 patients the salvage operation consisted of a total laryngectomy with unilateral (n = 2) or bilateral (n = 13) neck dissection. In 7 patients the surgery was limited to a neck dissection because the primary was controlled.

Histopathological examination of total laryngectomy with neck dissection showed negative resection margins in 11 patients (74%), close margins in 2 patients (13%) and microscopic positive margins in 2 patients (13%). Of the patients with neck dissection without laryngectomy 6 had negative resection margins (86%) and 1 microscopic positive margins (14%). No difference in histopathological results between the larynx and hypopharynx was found. Of the two patients with positive margins one was treated with postoperative radiotherapy, but he developed a local recurrence for which he received palliative chemotherapy. One of the patients with close margins developed a recurrence at the stoma and oesophagus, and underwent palliative radiotherapy and chemotherapy.

Reconstruction

The pectoralis major myocutaneous or myofascial pectoralis major (PM) flap was the most often used flap after total laryngectomy with or without pharyngectomy. Primary closure was only possible with smaller defects. With larger defects, when between one-third and three-quarters of the pharyngeal circumference has been resected, reconstruction was performed by utilising a pectoralis major myocutaneous PM flap. A circumferential pharyngeal defect not extending into the chest was free radial forearm flap (FRFF). A myofascial PM flap was also used to reinforce pharyngeal defects . The mucosal defect was closed primarily in 9 of the 15 patients (60%), reconstructed with a PM flap in 5 patients, and reconstructed with a free radial forearm flap (FRFF) in 1 patient. A PM flap to prevent wound healing problems was used in 9 of the 15 patients with a total laryngectomy and in 4 of the 7 patients with a neck dissection without laryngectomy.

Postoperative complications

No perioperative death occurred. Postoperative complications were observed in 15 (68%) of the 22 patients (Table III). Three patients experienced major complications that required re-operation. This concerned fistula in two patients and a bleeding in one patient that were closed with a (second) PM flap during re-operation. Most of the complications concerned wound healing problems (n = 13; 59%), as fistula (n = 5), wound dehiscence or wound infection (n = 7) or haemorrhage (n = 1). Other complications were pneumonia and spondylodiscitis in 2 patients.

Table III.

Postoperative complications for the total salvage surgery group, the group with an opened pharynx vs. the group with a closed pharynx.

Complications	Total		Pharynx open		Pharynx closed
	N	%	N	%	N	%
None	7	31%	4	27%	3	43%
Wound healing	13	59%	9	59%	4	57%
- Infection or dehiscence	7	31%	5	32%	2	29%
- Haemorrhage	1	5%	0	0%	1	14%
- Fistula	5	23%	4	27%	1	14%
Pneumonia	1	5%	1	7%	0	0%
Other	1	5%	1	7%	0	0%
Total	22	100%	15	100%	7	100%

Univariate analysis showed significantly more wound healing problems in patients with excessive alcohol intake (8 of 16 patients (50%) vs. none of 5 patients, p = 0.04). Furthermore, none of the following parameters were predictive for the development of postoperative complications: tobacco use or excessive alcohol intake at the time of presentation for the primary tumour, T- or N-stage, site of the primary tumour and age under the median of 59 years. No significant reduction in overall complications, wound related complications or fistula was found in our group of patients with a PM flap after neck dissection compared to patients with a primarily closed neck dissection.

Survival

Overall, 5-year disease free control rate was 35%, with 5-year locoregional and distant metastases control rates of 54% and 77%, respectively. Five-year overall survival was 27% (median 30 months) (Fig. 1), and disease specific survival was 35% after salvage surgery. For the 38 patients with residual or recurrent disease after chemoradiation who were not suitable for salvage surgery median survival was 12 months. Patients with tumour control (n = 76) had a 5-year survival of 70% (median 96 months) (Fig. 1). In uni- and multivariate analyses no significant predictors for overall survival after salvage surgery were found. Thus, age under or above the median 60 years was not a predictor factor for survival. The model of Tan et al. stratified patients with none, one or two of the following predictors for post salvage overall survival: stage IV (vs. other stages) and simultaneous (vs. local or regional) failure. When this model was applied to our population, no significant differences between the groups could be found (Fig. 2), although the group with stage IV disease showed a worse overall survival compared to patients with stage II or III disease (Fig. 3).

Fig. 1.

Five-year survival after the last treatment (chemoradiation or salvage surgery).

Fig. 2.

Survival after salvage surgery with model according to Tan et al. Comparison between patients with zero, one or two of the following presalvage predictors: stage IV vs. other stages and simultaneous locoregional vs. local or regional failure. No significant difference between the three groups was found.

Fig. 3.

Survival after salvage surgery with model according to Tan et al. Comparison between patients initial stage IV vs. initial non-stage IV disease. A trend (p = 0.05) towards a worse survival for patients with initial stage IV disease was found.

With a median length of follow-up after salvage surgery of 26 months (range 6-127 months), recurrent disease was found in 12 of the 22 patients (64%). These recurrences included local and/or regional recurrences in 8 patients and distant metastases in 4 patients. Local recurrences, regional recurrences and distant metastasis developed after a median interval of 6.5 months (range 2.5-14.3), 7.5 months (range 4.4-14,3) and 3.7 months (range 0-28.2) after salvage surgery, respectively.

Discussion

In this study, 37% of the patients with local and/or regional recurrences after chemoradiation for a laryngeal or hypopharyngeal tumour underwent salvage surgery, which is similar to rates reported by other authors, 33-66% . A larger proportion of patients with recurrent laryngeal than hypopharyngeal tumours underwent salvage surgery. This is in accordance with the report by Esteller et al. Independent predictors for salvage surgery within in the group of patients with locoregional failure, were age less than 59 year and larynx primary (vs. hypopharynx).

Fifteen patients underwent laryngectomy with neck dissection and 7 patients neck dissection only.

The rates of complications after salvage surgery are known to be high, with wound related complications and especially pharyngocutaneous fistula as a major problem. In this study 23% of patients developed a fistula. Review of the literature shows complication rates of 5-78%, with fistula in 4-73% of the patients (Table IV). Studies are difficult to compare, because of lack of homogeneity in patients (tumour site, stage) and in primary treatment (radiotherapy, chemoradiation).

Table IV.

Previous studies on complications and survival outcome in patients with salvage surgery after chemoradiotherapy for squamous cell carcinoma of the hypopharynx and larynx.

Authors	Year	N	Site	Comp	Fistula	LR	OS	DSS	Remarks
Stoeckli 7	2000	36	L	28%	14%			63% (5 y)	RT and CRT
Stoeckli 7	2000	9	H	40%	11%			20% (5 y)	RT and CRT
Leon 14	2001	28	L	21%	17%		57% (5 y)		endoresection
Weber 35	2003	75	L	~59%	~30%	74% (2 y)	69-71% (2 y)
Ganly 26	2005	38	L	53%	32%
Clark 13	2006	138	L/H	70% (salvage)	31%		31% (5 y) (salvage)		PT: none, RT, CRT
Fung 20	2007	14	L		29%				Interposition graft. ngopharyngectomy defects
Furuta 44	2008	34	L	47% wound	24%
Gil 5	2009	18	L		39%				PL/TL, RT and CRT
Patel 6	2009	17	L		24%				CRT or RT?
Relic 45	2009	16	L/H		73%		38% (3 y)		1 PL
Tsou 8	2010	48	H		58%
Paleri* 46	2011	>350	L			87% (2 y)	83% (3 y)	91% (2 y)	RT and CRT, PL
vd Putten 15	2011	120	L			70% (5 y) 79% (5 y) regional	50% (5 y)	58% (5 y)	RT and CRT, TL
Klozar 47	2012	208	L/H		34%				RT and CRT
Sewnaik 48	2012	24	L/H	92%
Patel 9	2013	359	L/H		27%				RT and CRT, primLE
Li 12	2013	100	L			70% (5 y)		55-70% (5 y)	RT and CRT, survival
Basheeth 49	2013	45	L/H	44%					Major complications, N0
Suzuki 50	2013	24	H	33%			50% (2 y)
Sayles* 10	2014	33 st	L/H		34%
Timmermans 31	2014	98	L/H		26%				RT and CRT, primLE
Omura 51	2014	42	H					40% (3 y)	RT and CRT, ICT
Powell 52	2014	45	L/H		22%
Suslu 11	2015	151	L/H		13%				RT and CRT, ICT
Sassler 30	1995	18	HN	61% major	50%				Sequential CRT
Newman 53	1997	17	HN	35%	20%
Lavertu 34	1998	26	HN	46%	4%
Goodwin 37	2000	109	HN	20%	6%			Med 21.5 months	PT: surgery, RT, CRT (17%)
Goodwin* 37	2000	1633	HN	39%			39% (5 y)		PT: surgery, RT, CRT
Agra 54	2003	124	HN	78% (CRT)					PT: surgery, RT, CRT
Gleich 55	2004	48	HN			20% (5 y)		15% (5 y)	Local recurrence
Taussky 56	2005	17	HN N	76%	24%		46% (3 y) 13% (5 y)		RT and CRT
Morgan 57	2007	38	HN	11%	5%				Local compl 23%
Encinas 58	2007	26	HN	31%					Article not available
Richey 39	2007	38	HN	24%		42% (2 y)	27-60% (1,2 y)
Tan 16	2010	38	HN	63%			43% (2 y) 37% (5 y)
Inohara 36	2010	30	HN	30%	7%		74-87% (3 y)
Esteller 18	2011	32	HN	28%	19%		34% (5 y)
Simon 59	2011	21	HN	33%	10%
Leon 60	2015	24	HN	63% 13%			26%(5 y) 70%(5 y)		CRT Bioradiotherapy
Present study		22	HN	73%	23%	58% (5 y)	27% (5 y)	36% (5 y)
Davidson 27	1999	34	N	38%	12%				37% CRT
Stenson 61	2000	69	N	25%	10%
Newkirk 62	2001	33	N						13 CRT,20 RT.
Grabenbauer 63	2003	56	N	25%	4%		44%(5 y)	55%	Planned ND
Kutler 64	2004		N	~30%					Only abstract
Brizel 65	2004	52	N	8% major		75% (4 y)	77%(4 y)		Planned ND in N2-3. survival cCR
Frank 66	2005	39	N	5% surgical
vd Putten 67	2007	61	N			79% (5 y)	36% (5 y)
Nouraei 68	2008	41	N			95% (5 y)		64% (5 y)	Survival hemineck
Vedrine 69	2008	28	N	14% severe
Christopoulos 70	2008	32	N	13%
Lango 71	2009	65	N	18%	5%				55%CRT
Relic 45	2009	12	N	8%
Hillel 72	2009	41	N	17%
Bremke 73	2009	25	N	24%
Goguen 28	2010	105	N	37%
Robbins 74	2012	30	N					60% (5 y)	SSND, CRT
Yirbesoglu 75	2013	44	N			71-73% (3 y)	55-64% (3 y)

Systematic review;

y: year; Com: complications, fistul: fistula, LR: locoregional control rate, OS: overall survival, DSS: disease specific survival, L: larynx, H: hypopharynx, HN: head and neck, N: neck, y: year, DFS: disease free survival, CRT: chemoradiotherapy, RT: radiotherapy, CT: chemotherapy, ND: neck dissection, PT: previous treatment, PL: partial laryngectomy, TL: total laryngectomy, primLE: primary laryngectomy, SSND: super selective neck dissection, st: studies, ICT: induction chemotherapy.

If wound healing problems are likely, pedicled PM flaps are very useful to cover important structures in the neck with well vascularised, non-irradiated tissue. In the present study, in 59% of the patients a PM flap was used for prevention of wound related complications. Unfortunately, no significant reduction in overall complications, wound related complications or fistula was found in our group of patients with a PM flap after neck dissection compared to patients with a primarily closed neck dissection. In our population only patients with considerable postradiation effects who were considered to be prone to wound healing problems underwent reconstruction with PM flap in the neck. Most studies evaluating reconstructive methods are conducted in patients undergoing salvage laryngectomy (Table V). Similar to our results, no difference in the incidence of local wound complications or fistula between the groups with and without PM flap was found by Gil et al. and Righini et al. Although it was an effective technique to prevent major complications, free vascularised tissue reinforcement did not alter the overall fistula rate as compared to when no flap was used, as reported by Fung et al. Smith et al. reported a significant reduction in fistula formation in patients with as compared to patients without PM flap reconstruction, but the percentage of patients with initial chemoradiation vs. primary surgery was not described. Sayles et al. performed a review and meta- analysis of 33 studies, and found only 10% fistula for salvage laryngectomy with onlay flap-reinforced closure compared to 28% fistula for salvage laryngectomy when no flap was used. Recently Paleri et al. described in a systematic review of nearly 600 patients a reduced risk of fistula by one-third in patients who have flap reconstruction/ reinforcement. Reconstruction of the mucosal defect using a PM flap may be associated with a higher rate of fistulae as compared to primary closure whereas a PM flap used as layer between mucosa and skin may reduce the risk of fistula formation. According to Righini et al. , fistula formation in postradiotherapy salvage surgery was reduced from 73% to 13% when a flap was used in the subgroup of patients with diabetes mellitus, a history of vascular disease or a poor nutritional status. Tsoe et al. and Withrow et al. suggest to reconstruct laryngectomy defects with an ALT (anterolateral thigh) or FRFF flap, as the incidence of fistula was low in their study.

Table V.

Comparison of pharyngocutaneous fistula in patients with salvage laryngectomy with and without flap reinforcement. In two studies, besides for reinforcement, the flap was also used to reconstruct pharyngeal defects .

Authors	Year	N	Site	Flap	Fistula	WC	Fistula No flap	WC No flap	p	Remarks
Righini 19	2005	60	larynx larynx	PMMF PMMF	23% 13%		50% 73%		0.06 0.018	Radiotherapy Subgroup: diabetes mellitus, vascular history, poor nutritional status
Fung 20	2007	41	larynx	FVT	29%	0%	30%	15%	n.s.
Patel 6	2009	17	larynx	PMF	0%		57%		0.02	(Chemo)radiotherapy
Gil 5	2009	80	larynx	PMMF	27%		24%		n.s.	PMMF 64% CRT, nonPMMF 25% CRT
Smith 21	2003	223	larynx	PMF	<1%		23%
Withrow 23	2007	37	larynx	FRFF	18%		50%			FRFF 41%CRT, nonFRFF 35%CRT
Patel 9	2013	359	larynx	PMF FVT	15% 25%		34%		0.02 0.07	(Chemo)radiotherapy
Powell 52	2014	45	larynx	FVT/PMF	0%		26%
Sayles* 10	2014	33 studies	larynx	Onlay flap	10%		28%		0.001	(Chemo)radiotherapy
Paleri* 22	2014	591	larynx							Pooled relative risk 0.63 (reduction one third compared to no flap)

Systematic review;

WC: wound complications, p: p-value, PMMF: pectoralis major myofascial flap, FVT: free vascularised tissue, PMF: pectoralis major flap, FRFF: free radial forearm flap, CRT: chemoradiotherapy.

Besides hypoalbuminaemia, neck dissection, comorbidities with diabetes mellitus or ischaemic heart disease, Tsoe et al. 8 found that reconstruction of the pharynx with primary closure had a statistically significant increased rate of fistula formation. On the contrary, Shemen et al. and Herranz et al. found an increase in complication rate when flap reconstruction was required. These patients had no history of radiation, and probably had a greater defect when a flap was required. Ganly et al. found no association between wound complications and flap reconstruction or neck dissection. The only significant independent predictor found was chemoradiation. Other suggested potential predictors for increased wound complications and fistula are: postoperative haemoglobin level lower than 12.5 g/dl, albumin level less than 40 g/l, prior tracheotomy, preoperative radio- an/or chemotherapy, concurrent neck dissection, radical neck dissection, poor nutritional status, tobacco and excessive alcohol use, poor renal and hepatic function, radiotherapy doses in excess of 70 Gy, early removal of drains (within 3 days of operation), vacuum drain duration and surgery extended to the pharynx or hypopharynx cancer -. We found more wound related complications in patients with current excessive alcohol use. This might be caused by immunosuppression due to ethanol, or alcohol-related lifestyle factors such as certain dietary deficiencies owing to unevenly composed diets .

Whether the interval between chemoradiotherapy and salvage surgery influences the risk of fistula formation remains uncertain. Increased incidence of wound complications was reported when the interval was shorter . However, Lavertu et al. and Weber et al. found no significant difference between groups with short and longer interval between chemoradiation and salvage surgery. Inohara et al. found no difference in complication incidence between salvage surgery for persistent or recurrent disease. We also did not find an association between interval and complication rate.

Comparing our results in patients with salvage laryngectomy: a) after previous radiotherapy in a previous study ; to b) after previous chemoradiotherapy in the present study, shows a worse 5-year prognosis for local disease control (58% vs. 70%) and overall survival (27% vs. 50%) in the chemoradiotherapy group. The total complication rate is 73% after chemoradiotherapy vs. 56% after radiotherapy.

The 5-year overall survival of 27% is comparable to other series, with a relatively better survival for patients with recurrent laryngeal carcinoma (compared to hypopharyngeal) or patients with a regional recurrence (Table IV). Even after adjusting for covariates, Goodwin found that a history of chemotherapy was associated with poorer survival after salvage surgery, suggesting a more aggressive tumour biology . Because of the low survival and high complication rates, the profit of salvage surgery is sometimes questioned. Salvage surgery is definitely worthwhile in a subset of patients. Reliable predictors for survival after salvage surgery are needed. Tan et al. suggested a model with stage IV tumours and concurrent local and regional failures as independent negative predictors. Esteller et al. found no significant differences in survival when analysed according to the classification of Tan. We found a worse survival for stage IV initial tumours. Other suggested potential predictors for a worse survival are: residual disease, older age, N3, positive resection margins and neck nodes with extranodal spread .

In this series age was an independent predictor for salvage surgery. Older patients were less frequently candidate for salvage surgery if recurrent tumour was diagnosed. Elderly patients with head and neck cancer generally have multiple and more severe comorbidity . Comorbidity is associated with a higher complication rate and poorer survival after major surgery -. Selection of elderly patients based on comorbidity seems to be the explanation for the similar complication rate and survival after salvage surgery. Moreover, patients with severe comorbidity would not have been treated with chemoradiation in the first place and therefore not included in this study.

In conclusion, one third of the patients with local and/or regional disease after chemoradiation underwent salvage surgery. Most of the patients not suitable for salvage surgery had distant metastases. Forty percent of the laryngectomies needed a flap reconstruction to cover mucosal defects. Patients who were at forehand more prone to wound healing failures underwent reconstruction with a PM flap in the neck to prevent wound related complications. One in four patients developed a pharyngocutaneous fistula. Only current excessive alcohol use was associated with complications. No significant independent predictors of survival were identified. The 5-year overall survival rate was 27% after salvage surgery. Older patients with recurrent laryngeal or hypopharyngeal carcinoma after chemoradiation selected for salvage surgery have a similar complication and survival rate compared to younger patients.