Prognostic Significance of Sentinel Lymph Node Mapping in Merkel Cell Carcinoma: Systematic Review and Meta-Analysis of Prognostic Studies

Table 1

Characteristics of the included studies.

First author	Publication year	SLN mapping method	Total number of patients/mean age (year)/mean tumor size (cm)/male gender (%)/head and neck location (%)/number of patients underwent regional lymph node dissection in SLN+/SLN− patients (% of total patients)	Inclusion of patients in the study at a common point of the disease course	Duration of follow- up	Method of outcome evaluation (death or recurrence)/Blind outcome evaluation to SLN results	Adjustment for important confounding variables	Variable used for prognostication/major findings
Tarantola [3]	2013	n/a	114 (34 with SLN mapping, 80 with other regional treatments)/70.1/1.37/70/46.3/n-a/n-a	Yes, all were included at clinical stages x and II of disease	3.3 years (mean)	Death of all causes/n-a	Yes	SLN mapping versus other nodal treatments/OS in the patients underwent SLN mapping was higher, HR: 1.04 [0.51–2.15], 𝑃=0.91
Kachare [14]	2014	n/a	1193 (474 with SLN mapping, 719 with nodal observation)/75.9/n-a/58.8/n-a/104 (21.9%)/n-a	Yes, all were included at clinical stages I and II of disease	Median of 21 months (0–83 months)	Death of disease/n-a	Yes	SLN mapping versus nodal observation, in addition prognostic significance of SLN pathological status was evaluated/DFS was higher in patients underwent SLN mapping, HR: 1.43 [1.01–2.05], 𝑃=0.04; DFS was also higher in SLN+ as compared to SLN− patients. This part of study was duplicate of Fritsch et al. [16] study and excluded form final analysis
Bajetta [17]	2009	n/a	63 (21 with SLN mapping and 42 with nodal observation)/n-a/n-a/45/18/8 (38%)/0 (0%)	Yes, all were included at clinical stages I and II of disease	Median of 65 months	Death of disease/n-a	Yes	SLN mapping versus nodal observation/operative nodal staging with SLN biopsy, HR 3.44 [1.17–10]; 𝑃=0.023) predicted better DFS
Sattler [18]	2013	n/a	47 (19 with SLN mapping and 27 with nodal observation)/70.32/n-a/26.3/31.6/n-a/n-a	Yes, all were included at clinical stages I and II of disease	Median of 20 months (2–234 months)	Death of disease, and death of all causes/n-a	Yes	SLN mapping versus nodal observation/SLN m aping predicted better DFS (HR 1.38 [0.45–4.20]) and OS 1.39 [0.25–7.76]) as compared to nodal observation.; HR was calculated from the survival curves according to Parmar method
Fields [15]	2011	Combined radiotracer and blue dye	153 (45 SLN+, and 108 SLN−)/69/n-a/59/21.5/45 (29.4%)/0 (0%)	Yes, all were included at clinical stages I and II of disease	Median of 41 months	Death of disease/n-a	Yes	SLN pathologic status/SLN− status predicted better OS (HR: 1.86 [0.22–15.49], 𝑃=0.56), and DFS (HR: 1.69 [0.37–7.65], 𝑃=0.49); HR was calculated from the survival curves according to Parmar method
Kouzmina [19]	2013	Radiotracer in all, blue dye in 16	28 (9 SLN+ and 19 SLN−)/n-a/n-a/39.3/39.3/8 (28.5%)/0 (0%)	Yes, all were included at clinical stages I and II of disease	Mean of 3.6 years	Death of all causes/n-a	No	SLN pathologic status/SLN− status predicted better OS (HR: 4.82 [0.79–29.34], 𝑃=0.08); HR was calculated from the survival curve according to Parmar method
Fritsch [16]	2014	n/a	721 (186 SLN+ and 535 SLN−)/n-a/n-a/61.7/24/n-a/n-a; 12 patients were excluded from survival analysis due to less than 1 months follow-up	Yes, all were included at clinical stages I and II of disease	Median of 34 months	Death of disease/n-a	Yes	SLN pathologic status/SLN− status predicted better DFS in head and neck (HR: 2.22 [0.84–5.86], 𝑃=0.1), and other parts of the body (HR: 3.01 [1.77–5.11], 𝑃=0.00005); The second 𝑃 value was calculated by CMA software
Cases in the literature with enough prognostic data	During the period of 1996–2013	Radiotracer and/or blue dye	172 (65 SLN+, and 107 SLN−)/68.6/1.84/45.3/40.1/30 (17.4%)/5 (2.9%)	Yes, all were included at clinical stages I and II of disease	Mean of 27.5 months (1–120 months)	Death of disease or death of all causes/n-a	Yes (refer to Table 2)	SLN pathologic status/SLN− status predicted better DFS in head and neck (HR: 4.09 [1.41–11.88], 𝑃=0.009), and other parts of the body (HR: 1.65 [0.7–3.9], 𝑃=0.15); SLN− status predicted better OS in head and neck (HR: 14.06 [1.63–121.28], 𝑃=0.01), and other parts of the body (HR: 3.79 [0.96–14.6], 𝑃=0.055).; All analyses were performed by SPSS version 11.5