Radiotherapy and Oncology 134 (2019) 211–219

Contents lists available at ScienceDirect

Radiotherapy and Oncology

journal homepage: www.thegreenjournal.com

Systematic Review

Adjuvant radiotherapy for Merkel cell carcinoma: A systematic review

and meta-analysis
Fausto Petrelli a,⇑, Antonio Ghidini b, Martina Torchio c, Natalie Prinzi c, Francesca Trevisan d,
Pierpaolo Dallera e, Agostina De Stefani d, Alessandro Russo e, Elisabetta Vitali d, Lorenza Bruschieri d,
Antonio Costanzo e, Silvia Seghezzi f, Michele Ghidini g, Antonio Varricchio e, Mary Cabiddu a, Sandro Barni a,
Filippo de Braud c, Sara Pusceddu c
a
Medical Oncology Unit, ASST Bergamo Ovest, Treviglio; b Medical Oncology Unit, Casa di Cura Igea, Milano; c Medical Oncology Unit, Fondazione IRCCS Istituto Nazionale Tumori di
Milano; d Radiotherapy Unit; e Surgical Oncology Unit; f Nuclear Medicine Unit, ASST Bergamo Ovest, Treviglio; and g Medical Oncology Unit, ASST Cremona, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Merkel cell carcinoma (MCC) is a rare and aggressive cutaneous malignancy with a high propensity for
Received 13 November 2018 local recurrence and regional and distant metastases. The main treatment is surgery with narrow excision
Received in revised form 17 February 2019 margins and draining nodes, plus or minus adjuvant radiotherapy (RT) on the surgical bed and/or lymph
Accepted 18 February 2019
nodes. We performed a systematic review and meta-analysis of the benefits of adjuvant RT in MCC treat-
Available online 28 February 2019
ment. PubMed, EMBASE, and the Cochrane Library were systematically searched to identify relevant
studies published before September 2018. Prospective trials and retrospective series comparing adjuvant
Keywords:
RT vs. no RT in resected primary MCCs were included. Primary endpoint was to evaluate the outcomes of
Merkel cell carcinoma
Adjuvant radiotherapy
MCC patients who received adjuvant RT in term of overall survival (OS) and disease-free survival (DFS).
Meta-analysis Hazard ratios (HRs) for OS and DFS were aggregated according to a fixed or random effect model.
Secondary endpoints were local, locoregional, and distant DFS.
A total of 17,179 MCCs across 29 studies were analysed. There was a significant difference in OS
between the RT and no RT arms (HR = 0.81, 95%CI 0.75–0.86, P < 0.001). There was also a significant dif-
ference in DFS in favour of adjuvant RT (HR = 0.45, 95%CI 0.32–0.62, P < 0.001). Adjuvant RT improved
locoregional DFS and local DFS but not distant DFS (HR = 0.3, 95%CI 0.22–0.42; HR = 0.21, 95%CI 0.14–
0.33, and HR = 0.79, 95%CI 0.49–1.14, respectively). Meta-regression analysis showed that high
Newcastle–Ottawa scale scores, stage I–II MCCs, shorter follow-up durations, size >2 cm, and being of
a younger age were associated with increased OS. This systematic review and meta-analysis suggests a
survival and DFS benefit for postoperative radiation of MCCs. Intermediate stage MCCs derive the max-
imum benefit with local and regional relapses reduced by 80% and 70%, respectively. Conversely, distant
metastases were not significantly prevented.
Ó 2019 Elsevier B.V. All rights reserved. Radiotherapy and Oncology 134 (2019) 211–219

Merkel cell carcinoma (MCC) is a rare and aggressive cutaneous Considering the biological aggressiveness of MCCs, there is an
malignancy with a high propensity for local recurrence and regio- unmet need regarding the optimal diagnostic and therapeutic
nal and distant metastases. Primarily affecting the elderly (average approach. This need has become more and more obvious over
age of presentation is 69 years), the incidence of MCC is slightly the past years due to the realisation that MCCs have an extremely
higher in men and occurs most commonly in sun-exposed areas high rate of regional lymphatic involvement, i.e., from 15% to 35%
of the body (50% head and neck, 40% extremities) [1]. MCCs are of cases. Even in low-risk cases, occult nodal positivity has been
considered rare tumours that have almost tripled in incidence over shown to occur in 15–20% of cases [3]. Unfortunately, the rarity
the last 25 years; from a rate of 0.15 cases per 100,000 in 1986 to of this tumour has made it difficult to conduct prospective trials.
0.44 cases per 100,000 in 2001. This increase is probably due to There is particular concern regarding adjuvant treatments (e.g.,
ameliorated diagnostic tools and developed knowledge on the link radiotherapy [RT], and chemotherapy) due to the lack of prospec-
between immunosuppression and MCC [2]. tive trials. MCCs are highly radiosensitive tumours and published
studies have suggested that there is an association between the
use of adjuvant RT and a reduced risk of loco-regional recurrence.
⇑ Corresponding author at: Piazzale Ospedale 1, 24047 Treviglio, BG, Italy. Among the positive studies, a limited effect of adjuvant RT on the
E-mail address: [email protected] (F. Petrelli).

https://doi.org/10.1016/j.radonc.2019.02.015
0167-8140/Ó 2019 Elsevier B.V. All rights reserved.
212 Adjuvant RT for Merkel cell carcinoma

disease’s local recurrence has been reported for specific patient Statistical analysis
subpopulations with a high risk of recurrence. At present, no trials
The HRs for DFS, OS, L-DFS, LR-DFS, and D-DFS with the relative
have shown a significant effect of adjuvant RT on overall survival
95% CIs were extracted from each study. Summary HRs were calcu-
(OS), while a small number of trials have observed a significant
lated using random- or fixed-effects models depending on the
increase in disease-free survival (DFS). In a previous systematic
heterogeneity of the included studies. We also calculated the rela-
review, the survival benefit of RT was observed at both the one-
tive risk (RR) for all recurrences (excluding deaths) and the CIs of
and three-year timepoints [4].
events in patients assigned to the RT arm compared to the control
High-risk subpopulations have been differently defined in stud-
arm (surgery alone) for each study. Statistical heterogeneity
ies according to site of primary tumour (most retrospective analy-
between the trials included in the meta-analysis was assessed
ses were conducted on head and neck primary MCCs), stage of
using the Chi [2] test and inconsistency was quantified with the
disease (I versus IIa and IIb), type of excision margins (close versus
I2 statistic [10].
extended), and type of lymph node resection (sentinel node biopsy
The assumption of homogeneity was considered invalid for P
versus complete loco-regional nodal dissection) and node involve-
values lower than 0.1. When substantial heterogeneity was not
ment [5]. The same studies have observed a ‘low-risk patient sub-
observed, the pooled estimate based on the fixed-effects model
population’ in which adjuvant RT seemed to have a lower/non-
was reported using the inverse variance method. When substantial
significant effect on reducing the risk of local recurrence (for exam-
heterogeneity was observed, the pooled estimate based on the
ple those with stage I MCC and with negative sentinel node
random-effects model was reported using the DerSimonian et al.
biopsy). This subpopulation could potentially avoid adjuvant RT
method [11], which considers both within- and between-study
[6].
variations. A two-tailed P value lower than 0.05 was considered
Most investigations were limited by small sample sizes, a lack
significant.
of patient follow-up data, and a lack of matched comparison
Potential publication bias was determined by visually evaluat-
groups [5]. RT at both the primary MCC site and the regional lymph
ing the asymmetry of the funnel plot. Begg’s and Egger’s tests were
node bed is becoming widely adopted, however, even without
also used to quantitatively detect publication bias [12,13]. The
compelling supportive data and without a consensus on RT sched-
trimmed values using Duval and Tweedie’s method are presented
ule and volume [7]. This article aims to analyse the literature on
in the results section only if the two-tailed P value was lower than
the current MCC primary site therapeutic approach and explore
0.1 in the Egger’s test. A leave-one-out sensitivity analysis was per-
the controversial role of adjuvant RT in MCC treatment.
formed by iteratively removing 1 study at a time to confirm that
our findings were not driven by any single study.
Material and methods The statistical analyses were performed using the RevMan soft-
ware for meta-analysis (v.5.3).
Selection of studies and inclusion criteria
We searched PubMed, EMBASE, and the Cochrane Library for Results
relevant studies published before 31 August 2018. The entry terms
for the search were (‘‘Merkel cell carcinoma”[All Fields] OR ‘‘Merkel Initially, 1023 potentially relevant articles were retrieved, leav-
carcinoma”[All Fields]) AND (‘‘radiotherapy”[Subheading] OR ing a total of 701 articles after the elimination of duplicates. A total
‘‘radiotherapy”[All Fields] OR ‘‘radiotherapy”[MeSH Terms]). The fol- of 665 articles comprising irrelevant papers, reviews, case reports,
lowing inclusion criteria were adopted: randomised/prospective meta-analyses, and studies with less than 20 patients were
trials or retrospective studies reporting data with the addition of excluded based on the titles and abstracts (Fig. 1). The remaining
adjuvant RT in at least 20 non-metastatic MCC patients. Principal 36 articles were then further reviewed based on the inclusion cri-
exclusion criteria were overlapping publications, lack of relevant teria described above to assess their suitability. Seven overlapping
outcome data, and no surgical treatment in the control arm. If series or studies that did not provide HRs or survival curves were
more than one publication was found for the same trial, the most excluded.
recent, complete, and updated version was included in the final Finally, 29 studies were included in the final meta-analysis
analysis. The quality of the randomised studies was assessed using [1,2,5–7,14–34,3,35–37]. Among these, only one was a randomised
the Jadad 5-item scale, which considers randomisation, double trial. The 29 eligible studies involved a combined total of 17,179
blinding, and withdrawals. The final score ranged from 0 to 5 [8]. patients, among which 7819 underwent RT on the primary tumour
The assessment of the risk of bias in retrospective studies was con- and/or lymph-nodes. The enroled number of patients per study
ducted independently by two authors using the Newcastle–Ottawa ranged from 24 to 6908. All included studies that were conducted
scale (NOS) [9]. abroad were published between 1990 and 2018. Most studies
included stage I–II MCCs (78%), while 18% included stage III, 3%
unknown, and 1% stage IV. Fifty-five percent of MCCs were located
Data extraction and endpoints
in the head and neck regions. The characteristics of the included
The data extraction was conducted independently by three co- studies and main outcomes are provided in Tables 1 and 2. Data
authors (FP, AG, and AC) according to the preferred reporting items regarding toxicities were only provided in two studies, therefore,
for systematic review and meta-analysis (PRISMA) statement. Any they were not analysed.
discrepancies were resolved by consensus between these three
authors. The data extracted for each trial were: first author’s name,
Effect of adjuvant RT on OS
year of publication; study type; number of evaluable patients;
median age and follow up; stage, site, and size of primary tumour; Data on OS were available in 20 studies and data on disease-
type of surgery; margin status; dose and type of RT; toxicity; and specific survival were available in one publication. There was a sig-
the hazard ratios (HRs) for DFS and OS with the relative 95% con- nificant difference in OS between the RT and no RT arms
fidence intervals (CIs). OS and DFS were the primary endpoints (HR = 0.81, 95%CI 0.75–0.86, P < 0.001; I2 = 0%, P = 0.47; fixed effect
evaluated in the experimental (surgery + RT) and control arms model; Fig. 2). The results were similar when a random effect
based on the HRs and relative 95% CIs. Local, locoregional, and dis- model was used (HR = 0.81, 95%CI 0.76–0.87; P < 0.001). After
tant DFS (L-DFS, LR-DFS, and D-DFS) were the secondary endpoints. exclusion of the 5 largest, registry-based studies, the final result
 F. Petrelli et al. / Radiotherapy and Oncology 134 (2019) 211–219 213

Potentially relevant publications and

screened for retrieval Duplicated excluded n = 322
n =1023

RCTs retrieved for more detailed

evaluation
n = 701

Potentially appropriate studies to be Publications excluded from meta-

included in the meta-analysis analysis n = 665
n = 701 List reasons: they were review,
letters, commentary, preclinical
studies, radiobiologic endpoints
evaluated, provided data of
adjuvant chemotherapy or were
Studies included in meta-analysis mixed series of skin tumors
n =36

Studies n=7 did not reported survival of

RT treated patients or were
overlapping series
Studies with usable information,
n = 21

Fig. 1. Flow diagram of included studies.

was similar to the main analysis (HR = 0.78, 95%CI 0.64–0.92; the trim and fill method. To further explore the potential sources
P = 0.004; I2 = 0%). of the heterogeneity observed in this analysis, we excluded each
study sequentially to determine its effect on the main summary
estimate. HRs ranged from 0.792 (excluding the trial of Mojica
Effect of adjuvant RT on DFS
et al.[27]) to 0.814 (excluding the trial by Jabbour et al.[23]). The
Data on DFS were available in 15 studies and, among them, data meta-regression analysis showed that high NOS scores, stage I–II
on relapse-free survival were provided in three of these studies. MCCs, shorter follow-up durations, size >2 cm, and being of a
There was a significant difference in DFS in favour of adjuvant RT younger age were associated with increased OS.
(HR = 0.45, 95%CI 0.32–0.62, P < 0.001; I2 = 69%, P < 0.001; Fig. 3).

Discussion
Effect of adjuvant RT on locoregional, local, and distant DFS
Data for regional, local, and distant control were available in Our meta-analysis of studies where adjuvant RT was delivered
five, seven, and two studies, respectively. Adjuvant RT improved after surgery for MCC suggests a possible survival benefit and a
LR-DFS and L-DFS but not D-DFS (HR = 0.3, 95%CI 0.22–0.42; major contribution to locoregional control, with the exception of
HR = 0.21, 95%CI 0.14–0.33, and HR = 0.79, 95%CI 0.49–1.14, distant metastases for postoperative radiation. From regression
respectively) despite risk of distant metastases analysis derives analysis, it seems that optimal primary treatment with radical sur-
from two publications and this flaws the power of the result. gery and adjuvant RT to the bed and/or draining lymph nodes is
appropriate for stage I–II MCCs that are at least 2 cm in size. The
adjuvant treatment for MCC remains controversial because there
Sensitivity analysis
is a lack of data on which to base treatment algorithms, recom-
To evaluate the robustness of the association results, we per- mendations, and guidelines. A lower number of retrospective stud-
formed a leave-one-out sensitivity analysis by iteratively removing ies have collected data on the optimal treatment of MCC that is
one study at a time and recalculating the summary HR. The sum- either in the advanced or initial stages.
mary HRs remained stable (range 0.79–0.81), indicating that our Although patients are generally treated with surgery as first-
results were not driven by any single study reporting OS data. line approach, there is still some controversy regarding the extent
of surgical intervention. It is important to note that these observa-
tions came from melanoma and other cutaneous tumour surgical
Publication bias
recommendations (i.e. Mohs surgery and sentinel node biopsy).
The analysis of OS was found to have a publication bias accord- More doubtful is the role of adjuvant RT in improving the clinical
ing to the Egger test (p = 0.07) and to the visual inspection of the outcomes of MCC patients because the evidence is derived from
funnel plot (Fig. 4). These values remained unchanged after using retrospective series or small prospective cohorts. Furthermore,
214 Adjuvant RT for Merkel cell carcinoma

Table 1
Characteristics of included studies.

Author/year Type of study N° of Median Median Site of primary tumour Size (cm) % Stage (%) Type of surgery (%) R+ %
pts follow age (%)
up (months)
H&N Trunk Extremities
Asgari/2014 Retrospective 218 75.6 77 45 3.7 51.3 <2 (35%) I–II (57.8) – Surgery ± SLNB ± LN dissection 6.9
>2 (11) III (25.7)
Unknown IV (16.1)
(53.7)
Balakrishnan/2013 Retrospective 54 22 76 100 0 0 NR I–II (96) – WLE (72) 21
III (4) – Other (11)
Bhatia/2016 Retrospective 6908 NR 76 44 12 44 <2 (39) I–II (70) – WLE ± SNB (stage I-II) or LN dis- 12
>2 (61) III (30) section (only in stage III)
Boyer/2002 Retrospective 45 27.5 72 69 2 28.5 1.3 (median) I (100) – Mohs surgery or WLE NR
Chen/2015 Retrospective 4815 NR NR 100 0 0 <2 (65) I–II (27) – WLE (92) or Mohs surgery (8) 21
>2 (35) III (73)
Clark/2007 Retrospective 110 21.6 70 100 0 0 NR I–II (52) – WLE (93) + LN dissection (45) 35
III (17)
Unknown
(56)
Eich/2002 Retrospective 31 22 70 42 16 42 NR I–II (97) – WLE (100) 0
III (3) – WLE + LN dissection (55)
Fields/2011 Retrospective 364 43.2 70 36 17 38 <2 (75) I–II (74) – WLE (29) 7
>2 (25) III (26) – WLE + SNB (33)
– WLE + LN dissection (38)
Ghadjar/2011 Retrospective 180 60 73 50 18 22 NR I–II (100) – WLE (93) or biopsy (7) 35
Gillenwater/2001 Retrospective 66 NR 68.4 100 0 0 <2 (67) cN+ (11) – WLE ± neck dissection + RT (51) NR
2–5 (15%) cN0 (77) – WLE ± neck dissection (39)
>5 (6) cNX (9) – RT (9)
NR (9)
Howle/2012 Retrospective 136 38 75 47 7 29 1.1 (median) I (66) – Surgery alone (39) 66
II (34) – Surgery + CT (1)
– Surgery + RT (51)
– Surgery + CTRT (4)
– RT alone (6)
Jabbour/2007 Retrospective 82 23 72 54 1 35 <2 (79) I (61) – WLE (89) + LN dissection (35) 30
2–5 (13) II (35) – Adj RT (82)
>5 (5) – Adj CTRT (11)
Jouary/2011 Randomized 83 57.7 70.9 43.4 6 50.6 NR I (100) – WLE + RT on tumour bed (100) 0
Kang/2012 Retrospective 62 44 74 51 5 27 15 (5–60) I–II (68) – WLE or excisional biopsy (80) 0
III (32)
Kim/2013 Retrospective 747 NR 76 40 13 47 <2 (61) I–II (53) – Local tumour destruction or WLE NR
>2 (39) III (47) (100) + SLNB or nodal dissection
(22.8)
Meuwissen/1995 Retrospective 80 20 74 100 0 0 <2 (66) I–II (67) – Radical surgery (90) + LN dissec- NR
2–5 (16) III (33) tion (11)
NR (16) – Incomplete surgery (9) + LN dis-
section (15)
– CTRT (1)
Mojika/2007 Retrospective 1187 40 74 NR NR NR 2 (median) I–II (55) – WLE (82) NR
III (31) – Radical surgery + LN dissection
IV (6) (10)
Morrison/1990 Retrospective 54 NR 70 70 14 16 <2 (78) I–II (83) – WLE (91) + LN dissection (5) NR
>2 (22) III (17) – RT alone (9)
Pectasides/2008 Retrospective 24 24 68 33 17 50 NR I–II (88) – WLE (67) 0
III (12) – WLE + LN dissection (21)
Poulsen/2010 Retrospective 60 56 76 0 0 100 NR I (30) – Surgery alone (28) 10
II (20) – Surgery + RT (48)
III (48) – Surg + CTRT (10)
– CTRT (5)
– RT alone (7)
– Other (2)
Rastrelli/2018 Retrospective 90 29 69 20 21 59 <2 (21) I–II (42) – WLE ± SLNB ± LN dissection NR
>2 (79) III (52)
IV (6)
Reichgelt/2011 Retrospective 808 NR 76 44.2 9.7 39.6 <2 (41) I (55) – Surgical excision, with or without NR
>2 (21) II (31) SNLB
NA (14) III (6) – Surgical excision, with ELND
– Surgical excision, with RLND
Senchenkov/2007 Retrospective 38 NR 66.6 0 0 100 NR I (71) – WLE (84) NR
II (29) – Mohs micrographic surgery (15)
Servy/2016 Retrospective 87 39 70.7 31 11.5 57.5 <2 (61) I (100) – WLE + SLNB (100) NR
2 (39)
 F. Petrelli et al. / Radiotherapy and Oncology 134 (2019) 211–219 215

Table 1 (continued)

Author/year Type of study N° of Median Median Site of primary tumour Size (cm) % Stage (%) Type of surgery (%) R+ %
pts follow age (%)
up (months)
H&N Trunk Extremities
Sexton/2014 Retrospective 42 54 70.5 42 26 32 <1 (100) I (100) – WLE (100) + LN dissection or NR
SLNB (20)
Strom/2016 Retrospective 171 33 74 41.5 10.5 47.9 1.5 (median) pN0 (59.1) – WLE (98.2), with (87.7) or with- 9.4
2 (42.1) pN1 (31) out SNLB
>2 (18.1) cN1 (9.9) – EB (1.8)
NR (39.8) – LN dissection (12.3)
– CT (15.4)
– Adj RT (76)
Takagishi/2016 Retrospective 46 44.4 66.5 100 0 0 <1 (85) I (100) – WLE + SLNB (100) 0
>1 (15)
Tarantola/2012 Retrospective 240 NR 70.1 46.3 6.7 38 1.37 (median) I (31.3) – EB (4.6) NR
II (17.1) – WLE (78.8)
III (23.3) – Mohs micrographic surgery
IV (2.1) (13.8)
NR (26.3) – CT (12.5)
– Adj RT (42.1)
Van Veenendaal/ Retrospective 351 28 74 49 11 31 17 (median) I–II (42) – Surgical excision ± LN dissection 14
2018 III (32)
IV (3)
Unknown
(23)

WLE, wide local excision; EB, excisional biopsy; NR, not reported; RT, radiotherapy; CT, chemotherapy; SNLB, sentinel lymph node biopsy; ELND, elective lymph node
dissection; RLND, radical lymph node dissection.

the majority of published studies are hampered by a marked positive MCC (stage III) are often conflicting [15,19,6,40,42]. Evi-
heterogeneity of patient characteristics, e.g. mixed stages of dis- dence from the National Cancer Database [15] for 6908 patients
ease; differences in surgical procedures performed prior to RT has shown that adjuvant RT improves OS compared to surgery in
(biopsies, debulking, whole local excision, marginal status, lymph stage I and II patients, but not in stage III patients. This was also
node sentinel evaluation); differences in or lack of information confirmed in the multicentric retrospective study of Servy et al.
about performance status, comorbidity, immunosuppression con- [7], which evaluated the role of sentinel lymph node status and
dition, and tissue immunohistochemistry positivity for MCC poly- adjuvant RT. They concluded that survival in stage III MCC could
omavirus; and adjuvant RT performed after surgery (radiation be driven by the presence of subclinical distant metastases, which
doses and volume, timing of RT, and whether RT was administered are often present in these patients, rather than adjuvant therapeu-
to only the primary tumour site, in the regional lymph nodes, or tic approaches [19]. Conversely, the Moffitt Cancer Center’s retro-
included in both sites). spective study of 171 patients showed that postoperative RT
Due to a lack of prospective comparative data, the international improves locoregional control and disease-specific survival in
guidelines currently include adjuvant RT as a treatment ‘option’ for patients with positive lymph nodes, but not in patients with nega-
all stages of the disease (I––III) to be performed within four to six tive lymph nodes [6].
weeks of surgery, although it is still unclear which subgroups of Our study has potential limitations, however. First, the use of a
patients with MCC can really reap the greatest benefits from treat- meta-analysis for observational studies is controversial and the
ment (NCCN 2019, EORTC 2015). In general, for large N0 MCCs, heterogeneity of the studies’ designs and patient populations
consensus panels and international guidelines (EADO/EORTC, may have affected the pooled estimation. All studies were not ran-
NCCN) agree on the possible role of adjuvant RT after surgical exci- domized in their nature, therefore, there could be a potential
sion for increasing local control in the primary tumour bed, imbalance in patient characteristics (performance status and
whereas there is more uncertainty for patients in III (any tumour comorbidities), with more fit patients and with a more advanced
size with nodes positivity) except for cases with multiple affected disease that potentially could have received adjuvant RT. Although
lymph nodes of extracapsular extension [38,39]. Ideally, prospec- randomised controlled trials provide the most reliable evidence,
tive studies are needed to identify which subgroups of patients such studies are currently lacking for MCC and a meta-analysis of
in stage I–III could be recommended adjuvant RT treatment. retrospective or non-randomised studies might be appropriate to
To date, the only prospective randomised trial for MCC was con- assess treatment efficacy for such rare disease. Despite the retro-
ducted by Jouary et al. in stage I patients treated with wide local spective nature of most studies, the results were not driven by
excision and RT to tumour bed was randomized to adjuvant regio- any single study reporting OS data and heterogeneity was absent.
nal radiotherapy or to observation alone [24]. The trial was closed Second, the type of RT has changed over the years with modern
prematurely due to a drop in recruitment related to the introduc- application of RT delivery (e.g., IMRT) that now represents the
tion of the lymph node sentinel evaluation in the management of standard of care. Furthermore, in the paper included, acute toxici-
MCC. No significant improvement in OS or PFS was demonstrated ties and long-term follow up of irradiated patients was not known.
after RT, however, a significant decrease in the risk of local recur- Finally, the contribution of adjuvant chemotherapy is not well
rence when compared with the group under exclusive observation known and is not evaluated in the present meta-analysis. It is likely
(0% vs. 16.7%, P = 0.007) was highlighted. that in stage III MCC, RT is not enough to delay the progression of
It is still not clear whether adjuvant RT improves DFS and OS in the disease and increase survival without the addition of systemic
patients with node-negative disease rates or if it reduces the risk of therapy. Despite these limitations, our analysis suggests the possi-
local recurrence [15,7,6,33,40,41]. Even the results of retrospective ble use of postoperative radiation for MCC, at least for stages I–II,
studies regarding the clinical benefits of RT in patients with node- which are the most represented stages in this meta-analysis. Sur-
Table 2

216
Radiotherapy delivery and outcomes.

Author/year N° received RT/Type RT/target of RT (%) Dose of RT (Gy) OS DFS Locoregional-DFS Local-DFS Distant-DFS Type of Quality of
RT vs no RT RT vs no RT RT vs no RT RT vs no RT RT vs no RT analysis paper
Asgari/2014 71/NR/NR NR HR = 0.9 (0.6–1.3) NR 0.3 (0.1–0.6) NR 1.5 (0.8–3.1) MVA 7
Balakrishnan/ 46/NR/regional RT (NR) 50 HR = 0.37 (0.13–1.07) DSS 0.63 (0.23–1.71) TTR NR NR NR UVA 7
2013
Bhatia/2016 2976/NR/NR NR HR = 0.71 (0.64–0.80) for NR NR NR NR MVA 6
stage I
HR = 0.77 (0.66–0.89) for
stage II
HR = 0.98 (0.86–1.12) for
stage III
Boyer/2002 20/NR/T and N 45–50 (T), 45–60 HR = 0.95 (0.26–3.51) DSS NR NR NR NR UVA 7
(N)
Chen/2015 2330/NR/NR NR HR = 0.8 (0.7–0.92) NR NR NR NR MVA 6
Clark/2007 66/NR/T (55) and N (60) 50 HR = 0.52 (0.25–1.07) HR = 0.64 (0.37– NR NR NR MVA 8
1.12)
Eich/2002 16/photons or electrons/T (56) or T + N (44) 55 NR NR NR HR = 0.32 (0.11–0.93) NR UVA 7
SFLD
Fields/2011 160/external beam/T (47) or N (53) 50 NR NR HR = 0.48 (0.24– HR = 0.53 (0.13–2.19) NR UVA 8
0.97)
Ghadjar/2011 131/NR/T (75) and N (21) 50–60 HR = 0.95 (0.57–1.58) HR = 0.25 (0.16– HR = 0.27 (0.16– HR = 0.18 (0.07–0.43) HR = 0.41 (0.23– MVA 8

Adjuvant RT for Merkel cell carcinoma

0.41) 0.49) 0.75)
Gillenwater/2001 26/NR/T + N + surgical bed (100) 46–66 HR = 1.05 (0.54–2.05) NR NR NR NR UVA 6
Howle/2012 69/NR/NR 50 HR = 0.94 (0.55–1.6) HR = 0.27 (0.14– NR NR NR UVA 8
0.51)
Jabbour/2007 48/Photons or electrons (89)/T (75) + N (64) 50* HR = 0.39 (0.18–0.82) HR = 0.39 (0.2–0.75) NR NR NR MVA 8
Jouari/2011 39/photon (T) and photon + electron (N)/T 50 HR = 1.23 (0.39–4.54) NR NR NR NR UVA -
(100) + N (50)
Kang/2012 43/electron (T) and megavoltage (N)/T (69) + N 46–50 NR HR = 0.33 (0.15– HR = 0.16 (0.05– HR = 0.05 (0.01–0.3)° NR UVA 8
(69) 0.74) 0.47)^
Kim/2013 404/NR/NR NR HR = 0.78 (0.61–0.99) NR NR NR NR UVA 5
Meuwissen/1995 51/electrons (T) and electrons or photons (N)/ 50 or 45–60 (R2 NR HR = 0.2 (0.11–0.37) NR NR NR UVA 7
T + N (100) only)
Mojika/2007 477/NR/NR NR HR = 0.85** (0.75–0.96) NR NR NR NR MVA 6
Morrison/1990 12/electrons/NR NR HR = 0.69 (0.3–1.57) HR = 0.57 (0.27– NR NR NR UVA 5
1.22) RFS
Pectasides/2008 11/NR/NR NR HR = 0.53 (0.13–2.14) HR = 0.46 (0.16– NR NR NR UVA 6
1.33)
Poulsen/2010 42/NR/T (50) + N (59) 48–49 NR HR = 0.51 (0.05– NR NR NR MVA 8
4.76) RFS
Rastrelli/2018 32/NR/NR NR HR = 0.87 (0.45–1.7) HR = 2.72 (1.28– NR NR NR UVA 7
5.77) RFS
Reichgelt/2011 259/NR/NR NR HR = 0.82 (0.65–1.04) NR NR NR NR MVA 6
Senchenkov/2007 21/NR/tumour bed and regional N 46–61 HR = 0.52 (0.16–1.66) NR NR HR = 0.29 (0.1–0.85) NR MVA 8
Servy/2016 75/NR/tumour bed and regional N 50 HR = 0.16 (0.05–0.51) HR = 0.19 (0.07– NR NR NR MVA 8
0.50)
Sexton/2014 18/NR/T (24) + N (20) NR NR HR = 0.68 (0.29– NR NR NR UVA 8
1.58)
Strom/2016 130/NR/T + N 50 (40–66) HR = 0.53 (0.21–0.93) HR = 0.42 (026–0.7) HR = 0.28 (0.14– HR = 0.18 (0.07–0.46) NR MVA 8
0.56)
Takagishi/2016 23/photons and/or electrons/T (100) + N (26) 50 (16–66) HR = 0.9 (0.23–3.5) NR NR HR = 0.15 (0.03–0.76) NR UVA 8
Tarantola/2012 101/NR/NR NR HR = 0.82 (0.44–1.53) NR NR NR NR UVA 8
Van Veenendaal/ 122/NR/T (65) and N (35) 60 HR = 1.1 (0.6–1.8) HR = 0.54 (0.3–0.9) NR NR NR MVA 8
2018

TTR, time to recurrence; Mo, months; SFLD, survival-free of local disease; RFS, relapse-free survival; NR, data not available.
*
Median dose.
°
Stage I–II that received local RT.
^
Patients having regional RT.
**
Out of 603 patients with data available.
 F. Petrelli et al. / Radiotherapy and Oncology 134 (2019) 211–219 217

Fig. 2. Forrest plot for overall survival analysis.

Fig. 3. Forrest plot for disease-free survival analysis.

vival data are robust and the results did not change even after gist, radiotherapist, and surgeon in reference centres dedicated to
exclusion of largest registry-based series. the treatment of neuroendocrine tumours. In reality, however,
In summary, it should be specified that due to the rarity of the the choice of treatment is more often defined by a single specialist.
disease, the therapeutic approach for MCC must always be decided This issue further worsens the heterogeneity of the data that are
by a multidisciplinary group comprising a dermatologist, oncolo- often evaluated in retrospective studies.
218 Adjuvant RT for Merkel cell carcinoma

0,0

0,2

Standard Error
0,4

0,6

0,8

-2,0 -1,5 -1,0 -0,5 0,0 0,5 1,0 1,5 2,0

Log hazard ratio

Fig. 4. Funnel plot for publication bias in overall survival analysis.

We are cautious about recommended adjuvant RT for all and end results database analysis. JAMA Dermatol 2013;149:831–8. https://
doi.org/10.1001/jamadermatol.2013.409.
patients with MCC and we believe that this treatment should be
[6] Strom T, Carr M, Zager JS, et al. Radiation therapy is associated with improved
discussed by a multidisciplinary expert team and aimed at patients outcomes in Merkel cell carcinoma. Ann Surg Oncol 2016;23:3572–8. https://
with a significantly high risk of local or lymph node recurrence. doi.org/10.1245/s10434-016-5293-1.
Therefore, we suggest several prognostic factors should be consid- [7] Servy A, Maubec E, Sugier PE, et al. Merkel cell carcinoma: value of sentinel
lymph-node status and adjuvant radiation therapy. Ann Oncol 2016;27:914–9.
ered, such as the size of the primary tumour (>2 cm, stage II); the https://doi.org/10.1093/annonc/mdw035.
pathological positivity of lymph nodes or sentinel lymph nodes [8] Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of
(stage III) or the absence of a correct assessment of the pathological randomized clinical trials: is blinding necessary? Control Clin Trials
1996;17:1–12. https://doi.org/10.1016/0197-2456(95)00134-4.
lymph node status; an immunosuppression condition, such as HIV, [9] Wells GA, Shea B, O’connell D, et al. The Newcastle-Ottawa Scale (NOS) for
lymphoproliferative diseases, and organ transplantation; and the assessing the quality of nonrandomised studies in meta-analyses. http://www.
presence of lymphatic and vascular invasion on pathological exam- ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed 13 Aug 2014.
Ottawa Hosp Res Inst. 2014.
ination and/or positive state of margins after surgical resection. [10] Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in
We believe that RT may not benefit pathologically node- meta-analyses. BMJ Br Med J 2003;327:557–60. https://doi.org/10.1136/
negative patients with very small primary tumours (<2 cm, stage bmj.327.7414.557.
[11] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
I). Indeed, the absence of a clear and well-defined standard adju- 1986;7:177–88. https://doi.org/10.1016/0197-2456(86)90046-2.
vant algorithm compared to a personalised approach highlights [12] Begg CB, Mazumdar M. Operating characteristics of a bank correlation test for
the need to perform prospective studies and randomised trials publication bias. Biometrics 1994;50:1088–101. https://doi.org/10.2307/
2533446.
with a homogeneous cohort of patients regarding stages (lymph
[13] Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a
nodes or SLNB positive versus negative), histopathologic character- simple, graphical test measures of funnel plot asymmetry. BMJ
istics (tumour size, marginal status, vascular and lymphatic inva- 1997;315:629–34. https://doi.org/10.1136/bmj.315.7109.629.
sion, IHC stain for Merkel cell polyomavirus), and treatments [14] Balakrishnan V, Berry S, Stew B, Sizeland A. Benefits of combined modality
treatment of Merkel cell carcinoma of the head and neck: single institution
(extension of the resection and fields of radiation treatment to experience. J Laryngol Otol 2013;127:908–16. https://doi.org/10.1017/
unify the therapeutic approach). Unfortunately, there are no S0022215113001862.
prospective clinical trials currently evaluating the role of adjuvant [15] Bhatia S, Storer BE, Iyer JG, et al. Adjuvant radiation therapy and chemotherapy
in Merkel cell carcinoma: survival analyses of 6908 cases from the National
RT in MCC treatment. Cancer Data Base. J Natl Cancer Inst 2016;108:1–9. https://doi.org/10.1093/
Despite data about adjuvant RT for MCC derive from retrospec- jnci/djw042.
tive or observational studies, this meta-analysis suggests that it [16] Chen MM, Roman SA, Sosa JA, Judson BL. The role of adjuvant therapy in the
management of head and neck Merkel cell carcinoma. JAMA Otolaryngol Neck
may be offered to patients’ resected MCC. Surg 2015;141:137. https://doi.org/10.1001/jamaoto.2014.3052.
[17] Leoncini E, Ricciardi W, Cadoni G, et al. Adult height and head and neck cancer:
a pooled analysis within the INHANCE Consortium. Head Neck 2014;36:1391.
Conflict of interest https://doi.org/10.1002/HED.
[18] Eich HT, Eich D, Staar S, et al. Role of postoperative radiotherapy in the
management of Merkel cell carcinoma. Am J Clin Oncol Clin Trials
None. 2002;25:50–6. https://doi.org/10.1097/00000421-200202000-00011.
[19] Fields RC, Busam KJ, Chou JF, et al. Recurrence after complete resection and
selective use of adjuvant therapy for stage I through III Merkel cell carcinoma.
References Cancer 2012;118:3311–20. https://doi.org/10.1002/cncr.26626.
[20] Ghadjar P, Kaanders JH, Poortmans P, et al. The essential role of radiotherapy in
the treatment of Merkel cell carcinoma: a study from the rare cancer network.
[1] Foote M, Veness M, Zarate D, Poulsen M. Merkel cell carcinoma: the prognostic Int J Radiat Oncol Biol Phys 2011;81:583–91. https://doi.org/10.1016/j.
implications of an occult primary in stage IIIB (nodal) disease. J Am Acad ijrobp.2011.05.028.
Dermatol 2012;67:395–9. https://doi.org/10.1016/j.jaad.2011.09.009. [21] Takes RP, Balm AJ, Loftus BM, Baris G, Hilgers FJ, Gregor RT. Merkel cell
[2] Boyer JD, Zitelli JA, Brodland DG, D’Angelo G. Local control of primary Merkel carcinoma of the head and neck. Clin Otolaryngol Allied Sci 1994;19:222–9.
cell carcinoma: review of 45 cases treated with Mohs micrographic surgery https://doi.org/10.1002/hed.21787.
with and without adjuvant radiation. J Am Acad Dermatol 2002;47:885–92. [22] Howle JR, Hughes TM, Gebski V, Veness MJ. Merkel cell carcinoma: an
https://doi.org/10.1067/mjd.2002.125083. Australian perspective and the importance of addressing the regional lymph
[3] Sexton KW, Poteet SP, Hill JB, et al. Adjuvant radiation therapy increases nodes in clinically node-negative patients. J Am Acad Dermatol
disease-free survival in stage IB Merkel cell carcinoma. Ann Plast Surg 2012;67:33–40. https://doi.org/10.1016/j.jaad.2011.07.029.
2014;73:531–4. https://doi.org/10.1097/SAP.0b013e31827f4c58. [23] Jabbour J, Cumming R, Scolyer RA, Hruby G, Thompson JF, Lee S. Merkel cell
[4] Hasan S, Liu L, Triplet J, Li Z, Mansur D. The role of postoperative radiation and carcinoma: assessing the effect of wide local excision, lymph node dissection,
chemoradiation in Merkel cell carcinoma: a systematic review of the literature. and radiotherapy on recurrence and survival in early-stage disease – results
Front Oncol 2013;3:1–10. https://doi.org/10.3389/fonc.2013.00276. from a review of 82 consecutive cases diagnosed between 1992 and 2004. Ann
[5] Kim JA, Choi AH. Effect of radiation therapy on survival in patients with Surg Oncol 2007;14:1943–52. https://doi.org/10.1245/s10434-006-9327-y.
resected merkel cell carcinoma: a propensity score surveillance, epidemiology,
 F. Petrelli et al. / Radiotherapy and Oncology 134 (2019) 211–219 219

[24] Jouary T, Leyral C, Dreno B, et al. Adjuvant prophylactic regional radiotherapy carcinomas of the head and neck. Adv Radiat Oncol 2016;1:244–51. https://
versus observation in stage I merkel cell carcinoma: a multicentric prospective doi.org/10.1016/j.adro.2016.10.003.
randomized study. Ann Oncol 2012;23:1074–80. https://doi.org/10.1093/ [34] Tarantola TI, Vallow LA, Halyard MY, et al. Prognostic factors in Merkel cell
annonc/mdr318. carcinoma: analysis of 240 cases. J Am Acad Dermatol 2013;68:425–32.
[25] Kang SH, Haydu LE, Goh RY, Fogarty GB. Radiotherapy is associated with https://doi.org/10.1016/j.jaad.2012.09.036.
significant improvement in local and regional control in Merkel cell [35] Rastrelli M, Ferrazzi B, Cavallin F, et al. Prognostic factors in Merkel cell
carcinoma. Radiat Oncol 2012;7:171. https://doi.org/10.1186/1748-717X-7- carcinoma: a retrospective single-center study in 90 patients. Cancers (Basel)
171. 2018;10:350. https://doi.org/10.3390/cancers10100350.
[26] Meeuwissen JA, Bourne RG, Kezarsley JH. The importance of postoperative [36] van Veenendaal LM, van Akkooi ACJ, Verhoef C, et al. Merkel cell carcinoma:
radiation therapy in the treatment of Merkel cell carcinoma. Pergamon l Clin clinical outcome and prognostic factors in 351 patients. J Surg Oncol December
Orig Contrib Int J Radiat Oncol Biol Phys 1995;31:325–31. https://doi.org/ 2017;2018:1768–75. https://doi.org/10.1002/jso.25090.
10.1016/0360-3016(94)E0145-A. [37] Wazen RM, Kuroda S, Nishio C, Sellin K, Brunski JB, Nanci A. NIH Public Access.
[27] Mojica P, Smith D, Ellenhorn JDI. Adjuvant radiation therapy is associated with 2014;8(9):1385-1395. doi: 10.2217/nnm.12.167.Gene.
improved survival in merkel cell carcinoma of the skin. J Clin Oncol [38] Lebbe C, Becker JC, Grob JJ, et al. Diagnosis and treatment of Merkel Cell
2007;25:1043–7. https://doi.org/10.1200/JCO.2006.07.9319. Carcinoma. European consensus-based interdisciplinary guideline. Eur J
[28] Morrison WH, Peters LJ, Silva EG, Wendt CD, Ang KK, Goepfert H. The essential Cancer 2015;51:2396–403. https://doi.org/10.1016/j.ejca.2015.06.131.
role of radiation therapy in securing locoregional control of Merkel cell [39] Tseng YD, Apisarnthanarax S, Liao JJ, Bhatia S, Nghiem PT, Parvathaneni U.
carcinoma. Int J Radiat Oncol Biol Phys 1990;19:583–91. Factors influencing radiation treatment recommendations in early-stage
[29] Pectasides D, Papaxoinis G, Pectasides E, et al. Merkel cell carcinoma of the Merkel cell carcinoma: a survey of US-based radiation oncologists. Expert
skin: a retrospective study of 24 cases by the Hellenic Cooperative Oncology Rev Anticancer Ther 2017;17:281–7. https://doi.org/10.1080/
Group. Oncology 2008;72:211–8. https://doi.org/10.1159/000112944. 14737140.2017.1285233.
[30] Poulsen M, Round C, Keller J, Tripcony L, Veness M. Factors influencing relapse- [40] Han AY, Patel PB, Anderson M, Diaz MFP, Chin R, St. John MA. Adjuvant
free survival in Merkel cell carcinoma of the lower limb-a review of 60 cases. radiation therapy improves patient survival in early-stage merkel cell
Int J Radiat Oncol Biol Phys 2010;76:393–7. https://doi.org/10.1016/j. carcinoma: a 15-year single-institution study. Laryngoscope
ijrobp.2009.02.014. 2018;128:1862–6. https://doi.org/10.1002/lary.27031.
[31] Reichgelt BA, Visser O. Epidemiology and survival of Merkel cell carcinoma in [41] Harrington C, Kwan W. Radiotherapy and conservative surgery in the
the Netherlands. A population-based study of 808 cases in 1993–2007. Eur J locoregional management of merkel cell carcinoma: the British Columbia
Cancer 2011;47:579–85. https://doi.org/10.1016/j.ejca.2010.11.002. Cancer Agency Experience. Ann Surg Oncol 2016;23:573–8. https://doi.org/
[32] Setoyama T, Natsugoe S, Okumura H, et al. a-Catenin is a significant prognostic 10.1245/s10434-015-4812-9.
factor than E-cadherin in esophogeal squamous cell carcinoma. J Surg Oncol [42] Hoeller U, Mueller T, Schubert T, et al. Regional nodal relapse in surgically
2007;95:148–55. https://doi.org/10.1002/jso. staged Merkel cell carcinoma. Strahlenther Onkol 2015;191:51–8. https://doi.
[33] Takagishi SR, Marx TE, Lewis C, et al. Postoperative radiation therapy is org/10.1007/s00066-014-0756-4.
associated with a reduced risk of local recurrence among low risk Merkel cell