Outcome Following Sentinel Node Biopsy Plus Wide Local Excision Versus Wide Local Excision Only for Primary Cutaneous Melanoma

Analysis of 5840 Patients Treated at a Single Institution

Augustinus P. T. van der Ploeg, MD; Lauren E. Haydu, BSCHE, MIPH; Andrew J. Spillane, MD; Michael J. Quinn, MBBS; Robyn PM. Saw, MBMS; Kerwin F. Shannon, MBBS; Jonathan R. Stretch, MBBS, DPhil(Oxon); Roger F. Uren, MD; Richard A. Scolyer, MD; John F. Thompson, MD

Disclosures

Annals of Surgery. 2014;260(1):149-157.

Abstract and Introduction

Abstract

Objective: Worldwide, sentinel node biopsy (SNB) is now a standard staging procedure for most patients with melanomas 1 mm or more in thickness, but its therapeutic benefit is not clear, pending randomized trial results. This study sought to assess the therapeutic benefit of SNB in a large, nonrandomized patient cohort.

Methods: Patients with primary melanomas 1.00 mm or more thick or with adverse prognostic features treated with wide local excision (WLE) at a single institution between 1992 and 2008 were identified. The outcomes for those who underwent WLE plus SNB (n = 2909) were compared with the outcomes for patients in an observation (OBS) group who had WLE only (n = 2931). Median follow-up was 42 months.

Results: Melanoma-specific survival (MSS) was not significantly different for patients in the SNB and OBS groups. However, a stratified univariate analysis of MSS for different thickness subgroups indicated a significantly better MSS for SNB patients with T2 and T3 melanomas (>1.0 to 4.0 mm thick) (P = 0.011), but this was not independently significant in multivariate analysis. Compared with OBS patients, SNB patients demonstrated improved disease-free survival (DFS) (P < 0.001) and regional recurrence-free survival (P < 0.001). There was also an improvement in distant metastasis-free survival (DMFS) for SNB patients with T2 and T3 melanomas (P = 0.041).

Conclusions: In this study, the outcome for the overall cohort after WLE alone did not differ significantly from the outcome after additional SNB. However, the outcome for the subgroup of patients with melanomas more than 1.0 to 4.0 mm in thickness was improved if they had a SNB, with significantly improved disease-free and DMFS.

Introduction

Worldwide, sentinel node (SN) biopsy (SNB) is now a standard staging procedure for most patients with melanomas 1.0 mm or more in thickness.^[1–4] The American Joint Committee on Cancer included micrometastasis diagnosed by SNB in the two latest (sixth and seventh) editions of its TNM staging system.^[5,6] Survival in patients with microscopic SN metastases is considerably better than in patients with clinically evident metastases,^[7] and SN status is the most important prognostic factor for survival in patients with early-stage melanoma.^[1,8,9]

Nevertheless, the role of SNB is still being defined, and an overall survival benefit in patients having SNB, with immediate completion lymph node dissection (CNLD) if found to be SN-positive, has yet to be demonstrated in a randomized clinical trial (RCT).^[2,10,11] In the third interim analysis of the first Multicenter Selective Lymphadenectomy Trial (MSLT-I), which compared patients who had SNB and patients who had nodal observation (OBS), there was not a statistically significant survival benefit for SNB patients, but disease-free survival (DFS) was improved.^[1] However, there was substantially improved 5-year survival (72.3% vs 52.4%, P = 0.004) in patients with intermediate-thickness melanomas (1.2–3.5 mm) with nodal metastases who had an immediate completion lymphadenectomy (CLND), compared with those who had a CLND when metastatic nodal disease became clinically apparent.^[1]

Most retrospective studies comparing outcome between SNB and OBS patients have also shown a DFS benefit in favor of SNB, with no overall survival benefit.^[12–17] However, a survival benefit has been reported in specific groups stratified by thickness.^[14,17] Several investigators have analyzed the group of patients with nodal metastases only, comparing SN-positive patients undergoing CLND with OBS patients undergoing therapeutic lymph node dissection (TLND) when regional lymph node metastasis was diagnosed clinically.^[16–22] Results have been conflicting, with 4 studies showing a survival benefit for CLND patients and 3 reporting no statistical difference. Meta-analysis of 6 of these studies, however, did show an overall survival benefit for SNB patients undergoing immediate CLND compared with patients having a TLND for clinically evident lymph node disease.^[22]

The criteria for recommending SNB differ in melanoma management guidelines. The most recent guideline of the American Society of Clinical Oncology and Society of Surgical Oncology, based on critical review of all available evidence, advocates offering SNB to patients with melanomas 1.0 mm or more to 4 mm. SNB may be recommended to patients with thick melanomas (>4 mm) for staging purposes and to facilitate regional disease control. The European and Australian guidelines merely encourage discussion of SNB in patients with melanomas more than 1.0 mm and more than 1.2 mm, respectively, or in patients with thinner melanomas when 1 or more adverse prognostic features (such as ulceration, a mitotic rate ≥1/mm², or Clark level IV invasion) is present.^[23–25] The guidelines of the American Society of Clinical Oncology and Society of Surgical Oncology make a similar recommendation for patients with thin melanomas, particularly those with melanomas 0.75 to 1.0 mm in thickness and also recommend consideration of SNB in patients with thick (>4 mm) melanomas (to provide better staging and to improve disease control in the regional node field).

The aims of this study were to compare regional recurrence-free survival, distant metastasis-free survival (DMFS) and melanoma-specific survival (MSS) of SNB patients with OBS patients in a large patient cohort treated at a single institution, as well as comparing the outcomes for SNB patients undergoing early CLND with those of OBS patients undergoing a delayed TLND for recurrence.

1 2 3 4

Next Section

Abstract and Introduction
Patients and Methods
Results
Discussion
References

Table 1. Baseline Characteristics of Patients in the SNB and OBS Groups and of Patients With Nodal Metastases
	All Patients			Patients With Nodal Metastases
	SNB n (%)	Observation n (%)	P	SNB Positive* n (%)	SNB False-Negative† n (%)	Observation With Recurrence† n (%)	P
Total	2909 (49.8)	2931 (50.2%)		394 (13.5)‡	89 (18.4)§	417 (14.2)¶
Follow-up, mo
Median (IQR)	44 (20–76)	40 (18–81)	0.190	43 (22–73)	42 (25–81)	51 (27–97)	0.003
Mean ± SE	53.4 ± 0.8	54.2 ± 0.8	0.500	52.0 ± 1.8	59.2 ± 4.6	64.4 ± 2.2	<0.001
Gender
Female	1167 (40.1)	1253 (42.7)	0.041	147 (37.3)	38 (42.7)	160 (38.4)
Male	1742 (59.9)	1678 (57.3)		247 (62.7)	51 (57.3)	257 (61.6)	0.640
Age, yr
Mean ± SE	56.1 ± 0.3	60.2 ± 0.3	<0.001	53.0 ± 0.9	57.3 ± 1.7	57.7 ± 0.9	<0.001
≤50	1015 (34.9)	880 (30.0)	<0.001	170 (43.1)	32 (36.0)	148 (35.5)	0.069
>50	1894 (65.1)	2051 (70.0)		224 (56.9)	57 (64.0)	269 (64.5)
Site
Extremity	1319 (45.5)	1157 (39.7)	<0.001	165 (42.0)	41 (46.1)	171 (41.0)	0.059
Trunk	1133 (39.1)	1025 (35.2)		169 (43.0)	29 (32.6)	154 (36.9)
Head and Neck	445 (15.4)	731 (25.1)		59 (15.0)	19 (21.3)	92 (22.1)
Thickness, mm
Median	1.8	1.5	<0.001	3	2.4	2.2	<0.001
Mean ± SE	2.47 ± 0.03	2.33 ± 0.04	0.015	3.41 ± 0.11	3.00 ± 0.22	2.94 ± 0.13	0.016
≤1.00/T1	328 (11.3)	811 (27.7)	<0.001	20 (5.1)	5 (5.6)	53 (12.7)	<0.001
>1.00–≤2.00/T2	1328 (45.7)	1055 (36.0)		107 (27.2)	31 (34.8)	134 (32.1)
>2.00–≤4.00/T3	840 (28.9)	671 (22.9)		149 (37.8)	36 (40.4)	156 (37.4)
>4.00/T4	413 (14.2)	394 (13.4)		118 (29.9)	17 (19.1)	74 (17.7)
Mitotic rate,/mm²
Median (IQR)	3 (2–7)	2 (1–6)	<0.001	5 (3–9)	6 (3–12)	5 (2–9)	0.037
Mean ± SD	5.18 ± 0.11	4.67 ± 0.12	0.003	6.74 ± 0.30	8.16 ± 0.74	6.60 ± 0.33	0.120
Clark level
II–III	791 (27.7)	833 (29.2)	0.161	72 (18.5)	20 (22.7)	95 (23.2)	0.503
IV	1825 (64.0)	1759 (61.6)		265 (68.1)	58 (65.9)	270 (65.9)
V	237 (8.3)	263 (9.2)		52 (13.4)	10 (11.4)	45 (11.0)
Ulceration
Absent	1863 (71.2)	1734 (71.0)	0.922	217 (59.0)	46 (56.8)	210 (57.4)	0.883
Present	755 (28.8)	707 (29.0)		151 (41.0)	35 (43.2)	156 (42.6)
Melanoma subtype
SSM	863 (39.4)	879 (43.1)	<0.001	102 (32.4)	16 (21.3)	116 (38.4)	0.002
NM	800 (36.5)	595 (29.2)		147 (46.7)	29 (38.7)	118 (39.1)
Other	529 (24.1)	567 (27.8)		66 (21.0)	30 (40.0)	68 (22.5)
Positive nodes
Mean ± SE				1.69 ± 0.06	2.57 ± 0.32	2.92 ± 0.17	<0.001
1				243 (61.7)	51 (57.3)	181 (47.0)	<0.001
2–3				91 (23.1)	12 (13.5)	79 (20.5)
>3				60 (15.2)	26 (29.2)	125 (32.5)
First recurrence type
Local	116 (20.9)	86 (12.7)	<0.001
In transit	95 (17.1)	51 (7.5)
Nodal	111 (20.0)	367 (54.3)
Distant	234 (42.1)	172 (25.4)

*Early CLND was performed in 394 sentinel node-positive patients.
† Delayed lymphadenectomy was performed in 89 false-negative SNB patients and 417 patients with nodal observation.
‡394 of 2909 (13.5%) SNB patients had positive SNs.
§89 of 2515 (3.5%) patients with a negative SNB had their first recurrence in a lymph node in the SNB field. False negativity rate = false-negative
SNB/(false-negative SNB + true positive SNB) = 89/483 = 18.4%
¶ 417 of 2931 patients (12.5%) with nodal observation had regional lymph node recurrence.
NM indicates nodular melanoma; SE, standard error of the mean; SSM, superficial spreading melanoma.

Table 2. First Recurrence Data in the SNB and OBS Groups According to Primary Tumor Thickness Criteria for SNB*
Group	None, N (%)	Local, N (%)	In Transit, N (%)	Regional Node, N (%)	Distant, N (%)	Total, N (%)
OBS, mm (n = 1471)
<1†	152 (76.4)	6 (3.0)	0 (0.0)	29 (14.6)	12 (6.0)	199 (13.5)
1	107 (78.7)	4 (2.9)	0 (0.0)	18 (13.2)	7 (5.1)	136 (9.2)
>1	536 (47.2)	76 (6.7)	51 (4.5)	320 (28.2)	153 (13.5)	1136 (77.2)
SNB, mm (n = 1447)
<1†	40 (81.6)	3 (6.1)	0 (0.0)	2 (4.1)	4 (8.2)	49 (3.4)
1	56 (83.6)	3 (4.5)	0 (0.0)	3 (4.5)	5 (7.5)	67 (4.6)
>1	795 (59.7)	110 (8.3)	95 (7.1)	106 (8.0)	225 (16.9)	1331 (92.0)

*Including patients with at least 5 years of follow-up or a recurrence event within 5 years. In this subgroup of patients, the median Breslow thickness in the OBS group (1.7 mm) was significantly lower than that in the SNB group (2.0 mm) P < 0.001). Similarly, the median mitotic rate in the OBS group (3/mm²) was significantly lower than that of the SNB group (4/mm²) (P < 0.001). Percentage of ulcerated cases did not differ significantly between the OBS and SNB groups (31% and 30%, respectively).
†Clark IV/V or ulceration or mitoses present.

Table 3. Multivariate Analysis of DFS and MSS in All Patients (A) and the SNB Group (B)
Factor	Value	HR	95% CI	P	HR	95% CI	P
A			DFS (n = 4671)			MSS (n = 4473)
Management*	Observation	1.40	1.23–1.58	<0.001	1.04	0.88–1.22	0.642
Gender	Male	1.17	1.02–1.34	0.021	1.50	1.24–1.80	<0.001
Age	Years	1.01	1.00–1.01	0.005	1.01	1.00–1.01	0.003
Breslow thickness	Mm	1.11	1.09–1.14	<0.001	1.15	1.11–1.18	<0.001
Mitotic rate	/mm²	1.03	1.02–1.04	<0.001	1.02	1.01–1.04	<0.001
Clark level (ref: II–III)	IV–V	1.52	1.30–1.78	<0.001	1.37	1.11–1.68	0.003
Ulceration	Present	1.73	1.52– 1.98	<0.001	1.82	1.53–2.16	<0.001
Primary site	Trunk	0.97	0.84–1.13	0.707	1.19	0.98–1.45	0.072
(ref: extremity)	Head and neck	1.15	0.98–1.36	0.093	1.31	1.05–1.64	0.017
B			DFS (n = 2479)			MSS (n = 2352)
Sentinel node status	Positive	3.04	2.50–3.70	<0.001	2.97	2.34–3.77	<0.001
Gender	Male	0.96	0.82–1.21	0.957	1.21	0.94–1.57	0.148
Age	Years	1.02	1.01–1.03	<0.001	1.02	1.01–1.02	<0.001
Breslow thickness mm	mm	1.13	1.09–1.18	<0.001	1.16	1.11–1.22	<0.001
Mitotic rate	/mm²	1.03	1.02–1.04	<0.001	1.03	1.01–1.04	0.001
Clark level (ref: II–III)	IV–V	1.32	1.05–1.67	0.018	1.22	0.92–1.63	0.168
Ulceration	Present	1.50	1.24–1.82	<0.001	1.98	1.55–2.52	<0.001
Primary site	Trunk	0.91	0.74–1.12	0.359	1.35	1.04–1.75	0.026
(ref: extremity)	Head and Neck	1.16	0.90–1.50	0.264	1.39	1.00–1.94	0.053

*SNB vs OBS.

Table 4. Published Series Reporting OS or MSS and DFS for SNB and OBS Patients, Respectively, After Immediate Completion Lymph Node Dissection After a Positive Sentinel-Node Biopsy (CLND) and Delayed TLND After Regional Lymph Node Recurrence
Author	Group of Patients	No.Patients	Median Follow-up, mo	Thickness, mm	P* (DFS)	P* (MSS/OS†)	HR MSS/OS†	5-yr MSS/OS† (%)
Mohrle et al¹²	OBS	2617	58					NA
	SNB	271	35	Any	NA	0.37	0.80	NA
Gutzmer et al¹³	OBS	377	59.7					NA
	SNB	296	35.5	Any	0.0064	0.317	0.98	NA
Morton et al¹	OBS	500	48.4					86.6
	SNB	769	48.4	1.20–3.50	0.009	0.58	0.92	87.1
Koskivuo et al¹⁵	OBS	616	74					85.2
	SNB	305	16	Any	0.414	0.656	0.70	87.8
	OBS	324	74					NA
	SNB	159	16	>1.00–4.00	0.499	0.646	NA	NA
Starz and Balda¹⁴	OBS	61	115					±93
	SNB	87	74	0.76–1.00	0.01	0.03	NA	100
Leiter et al¹⁶	OBS	440	57.6					81.5
	SNB	439	54.3	>1.00	0.003	0.09	0.74	85.5
Satzger et al¹⁷	OBS	377	64.0					80.3
	SNB	296	72.5	>1.00	0.001	0.049	NA	84.8
Present study OBS	2931	40						85.8
	SNB	2909	44	Any	<0.001	0.561	0.96	85.0
	OBS	1726	40					85.3
	SNB	2168	44	>1.00–4.00	<0.001	0.011^‡	0.79	86.8
	OBS	2120	40					82.2
	SNB	2581	44	>1.00	<0.001	0.012^‡	0.82	83.6
Morton et al¹⁸	CLND	287						73
	TLND	287	12–360	Any	NA	<0.001	2.0	51
Kretschmer et al¹⁹	CLND	314	32					62.5
	TLND	623	123	Any	NA	0.002	1.82	50.2
Morton et al¹	CLND	122	48.4					72.3
	TLND	78	48.4	1.20–3.50	NA	0.004	1.95	52.4
Van Akkooi et al²⁰	CLND	64	37					±70
	TLND	124	56	Any	NA	0.115	1.60	±57
Nowecki et al²¹	CLND	258	35					52.5
	TLND	286	37	Any	NA	0.04	1.24	39.5
	CLND	111	35					57.2
	TLND	100	37	>1.00–4.00	NA	0.0006	NA	37.9
Pasquali et al²²	CLND	100	50					68.9
	TLND	90	56	Any	NA	0.17	1.07	50.4
	CLND	58	50					71
	TLND	58	56	>1.00–4.00	NA	0.49	NA	57.8
Leiter et al¹⁶	CLND§	72	47.0					52.9
	TLND	72	48.0	>1.00	NA	0.196¶	2.2	42.0
Satzger et al¹⁷	CLND	77	72.5					NA
	TLND	147	64.0	>1.00	NA	0.006	NA	NA
Present study	CLND	394	44					64.1
	TLND	417	40	Any	<0.001	0.146	1.18	60.5
	CLND	256	44					68.3
	TLND	290	40	>1.00–4.00	<0.001	0.147	1.24	62.7
	CLND	374	44					63.1
	TLND	364	40	>1.00	<0.001	0.149	1.19	57.9

*Results of univariate analysis.
†When assessed in paper, melanoma-specific survival results are shown earlier. Otherwise, overall survival results are shown.
‡Not significant after adjusting for prognostic factors in multivariate analysis.
§Patients with recurrence after a negative SNB followed by a delayed TLND were added to the CLND group.
¶P = 0.009 in multivariate analysis.
CLND indicates early completion lymphadenectomy after a positive SNB; NA, not applicable/not reported; TLND, delayed therapeutic lymphadenectomy when melanoma recurred in the regional lymph node in follow-up.