Author

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

Hur, 2019

Literature search up to October 13, 2017 and updated on May 4, 2018

A: Cao (2017)

B: Ledderose (2015)

C: Lee (2015)

D: Lombardi (2016)

E: Lund (2012)

F: Meng (2014)

G: Miglani (2017)

H: Swegal (2014)

I: Won (2015)

Origin of the study

A: Yangzhou, China

B: Munich, Germany

C: Seoul, South Korea

D: Varese and Brescia, Italy

E: London, UK

F: Shanghai, China

G: Phoenix, Arizona, US

H: Cleveland, Ohio, US

I: South Korea

Study design:

SR and meta-analysis of 8 retrospective cohort studies and 1 prospective cohort study

Setting and Country:

Caruso Department of Otolaryngology – Head and Neck Surgery, Keck School of Medicine, University of Southern California, Los Angeles, California.

Source of funding:

The author(s) received no financial support for the research, authorship, and/or publication of this article.

conflicts of interest:

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Inclusion criteria SR:

- Participants: patients diagnosed with sinonasal mucosal melanoma (SNM);

- Intervention: ER of SNM;

- Comparison: open resection (OR) or combined resection (endoscopic and open) of SNM;

- Outcomes: overall survival or disease-free survival

Exclusion criteria SR:

- Studies which did not report survival rates or no comparison was performed between endoscopic and open approaches.

7 studies included

Important patient characteristics at baseline:

Sample size (N)

(ER= endoscopic resection; OR=open resection)

A:

ER: N = 15

OR: N = 18

Total: 33

B:

ER: N = 12
OR: N = 10
Total: 22

C:
ER: N = 16
OR: N = 15
Total: 31

D:
ER: N = 37
OR: N = 7
ER + OR: N = 14
Total: 58

E:

ER: N = 31
OR: N = 86
Total: 117

F:
ER: N = 41
OR: N = 28
Total: 69

G:
ER: N = 9
OR: N = 13
Total: 22

H:
ER: N = 12
OR: N = 13
Total: 25

I:
ER: N = 59
OR: N = 63
ER + OR: N = 11
Total: 133

Groups comparable at baseline?

Not reported

Describe intervention:

A: Endoscopic resection

B: Endoscopic resection

C: Endoscopic resection

D: Endoscopic resection

E: Endoscopic resection

F: Endoscopic resection

G: Endoscopic resection

H: Endoscopic resection

I: Endoscopic resection

Describe control:

A: Open resection

B: Open resection

C: Open resection

D: Open and endoscopic resection

E: Open resection

F: Open resection

G: Open resection

H: Open resection

I: Open and endoscopic resection

End-point of follow-up:

A: Median ER = 0.8 years, Median OR = 0.6 years

B: Not reported in study

C: Median = 9.0 months

D: Median = 30 months

E: Median = 37.5 months

F: Mean = 34 months

G: Median = 25 months

H: Median ER = 1.4 years, Median OR = 2.0 years

I: Mean = 40.9 months

For how many participants were no complete outcome data available?

(intervention/control)

Not reported.

Overal survival

Defined as 5-year survival

Effect measure: Hazard ratio (95% CI):

A: HR= 1.43 (95% CI= 0.09 to 23.64) favoring OR

B: HR= 0.96 (95% CI= 0.09 to 10.71) favoring ER

E: HR= 0.52 (95% CI= 0.30 to 0.92) favoring ER

F: HR= 1.21 (95% CI= 0.61 to 2.40) favoring OR

G: HR= 0.06 (95% CI= 0.00 to 31.88) favoring ER

H: HR= 1.01 (95% CI= 0.15 to 6.76) favoring OR

I: HR= 0.58 (95% CI= 0.34 to 0.99) favoring ER

Pooled effect (random effects model):

HR= 0.68 (95% CI= 0.49 to 0.95) favoring ER

Heterogeneity (I²): 0.0%; p=0.550

Disease-free survival

Definition of disease-free survival is not explicitly reported in the study. Since the study reported 5-year overall survival it is assumed that the study also used 5-year disease-free survival as an outcome of the study.

A: HR= 1.36 (95% CI= 0.35 to 5.38) favoring OR

B: HR= 0.38 (95% CI= 0.03 to 5.34) favoring ER

C: HR= 0.35 (95% CI= 0.10 to 1.20) favoring ER

G: HR= 0.49 (95% CI= 0.06 to 3.74) favoring ER

H: HR= 0.60 (95% CI= 0.06 to 5.72) favoring ER

Pooled effect (random effects model):

HR= 0.59 (95% CI= 0.28 to 1.25) favoring ER

Heterogeneity (I²): 0.0%; p=0.668

Author’s conclusion:

Based on the available literature, an endoscopic approach for SNM resection has improved OS and similar DFS compared to an open approach. Further multicenter trials or meta-analyses are necessary to assess oncologic outcomes, postoperative morbidity, and cost between the 2 approaches.

Lu, 2019

SR and meta-analysis of retrospective cohort studies

Literature search up to April 2019

A: Hagemann (2019)

B: Huang (2018)

C: Fu (2017)

D: Farquhar (2016)

E: Mortuaire (2016)

F: Naunheim (2016)

G: Bhayani (2014)

H: Guo (2014)

I: Saedi (2014)

J: Arnold (2012)

Origin of the study

A: Mainz, Germany

B: Guangdong, China

C: Toronto, Canada

D: Philadelphia, US

E: Lille, France

F: Boston, US

G: Houston, US

H: Shanghai, China

I: Tehran, Iran

J: Berne, Switzerland

Study design:

SR and meta-analysis of 10 retrospective observational cohort studies.

Setting and Country:

Prince of Wales Clinical School, University of New South Wales,

Sydney, Australia

Source of funding

The author(s) received no financial support for the research,

authorship, and/or publication of this article.

Conflicts of interest:

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of

this article.

Inclusion criteria SR:

- Histopathological primary malignancies involving the sinonasal cavity;

- In comparative cohorts managed by both ER and OR approaches;

- With reported at least one surgical outcome;

- In patients > 18 years.

Exclusion criteria SR:

- Mucosal melanoma (MM) histology as that represents a distinctly different tymor type;

- Tumors that are not deemed malignant, for example, angiofibroma and papilloma;

- Reuccrent pathologies;

- Hybrid cranioendoscopic approaches to resection;

- Case series/reports with no comparative cohort.

10 studies included

Important patient characteristics at baseline:

Sample size, N (%) (ER= endoscopic resection; OR= open resection)

A:

ER: N = 123 (55%)

OR: N = 102 (45%)

B:

ER: N= 27 (57%)

OR: N= 20 (43%)

C:

ER: N = 15 (17%)

OR: N = 72 (83%)

D:
ER: N = 82 (66%)

OR: N = 42 (34%)

E:

ER: N = 20 (47%)

OR: N = 23 (53%)

F:

ER: N = 10 (18%)

OR: N = 45 (82%)

G:

ER: N = 14 (27%)
OR: N = 38 (73%)

H:

ER: N = 8 (33%)

OR: N = 16 (67%)

I:

ER: N = 72 (45%)

OR: N = 88 (55%)

J:

ER: N = 28 (34%)

OR: N = 55 (66%)

Total:

ER: N = 399 (44%)

OR: N = 501 (56%)

Median age

A: Not reported

B: Not reported

C: 56 years

D: 54 years

E: 70 years

F: 58 years

G: 57 years

H: 35 years

I:47 years

J: 63 years

Sex:

A: 60% male

B: Not reported

C: 57% male

D: 53% male

E: 98% male

F: 69% male

G: 73% male

H: 42% male

I:70% male

J: 67% male

Radiation therapy (%) / Chemotherapy (%)

A: 58% / 20%

B: Not reported

C: 77% / 15%

D: 73% / 34%

E: Not reported

F: 82% / 62%

G: Not reported

H: Not reported

I: 38% / 27%

J: 65% / 23%

Groups comparable at baseline?

Not reported

Describe intervention:

A: Endoscopic resection

B: Endoscopic resection

C: Endoscopic resection

D: Endoscopic resection

E: Endoscopic resection

F: Endo- and cranioendoscopic resection

G: Endoscopic resection

H: Endoscopic resection

I: Endoscopic resection

J: Endoscopic resection

Describe control:

A: Open resection

B: Open resection

C: Open resection

D: Open resection

E: Open resection

F: Open resection

G: Open resection

H: Open resection

I: Open resection

J: Open resection

Median follow-up:

A: 45.5 months

B: 73.6 months

C: Not reported

D: 40.1 months

E: 79.2 months

F: 27.4 months

G: 55.3 months

H: 67 months

I: 21 months

J: 36.3 months

Overall survival

A:

ER: 5-year OS rate = 76%

OR: 5-year OS rate = 59%

P=0.36

B:

ER: mean OS = 80 months

OR: mean OS = 65 months

P=0.14

D:

ER: 3-year OS rate = 91%

OR: 3-year OS rate = 76%

P=0.02

I:

ER: median OS = 24 months

OR: median OS = 28 months

P=0.13

The data were not robust enough for meta-analysis due to the different scales and timeframes

Complications

Effect measure: RR (95% CI)

A: RR= 0.74 (95% CI= 0.50 to 1.10) favoring ER

C: RR= 2.40 (95% CI= 0.67 to 8.54) favoring OR

D: RR= 0.55 (95% CI= 0.29 to 1.03) favouring ER

E: RR= 0.23 (95% CI= 0.01 to 4.50) favoring ER

F: RR= 0.26 (95% CI= 0.04 to 1.76) favoring ER

J: RR= 0.18 (95% CI= 0.01 to 3.07) favoring ER

Pooled effect (random effects model):

RR= 0.68 (95% CI= 0.42 to 1.10) favoring ER

Heterogeneity (I²): 27.2%; p=0.231

Length of stay

Effect measure: RR (95% CI)

A: RR= -4.75 (95% CI= -5.27 to -4.23) favoring ER

B: RR= -3.50 (95% CI= -5.57 to -1.43) favoring ER

C: RR= -0.70 (95% CI= -1.30 to -0.11) favoring ER

D: RR= -2.90 (95% CI= -3.95 to -1.85) favoring ER

E: RR= -2.60 (95% CI= -3.45 to -1.75) favoring ER

Pooled effect (random effects model):

RR= -2.87 (95% CI= -4.68 to -1.05) favoring ER

Heterogeneity (I²): 96.0%; p=0.000

Recurrence; RR (95% CI)

Effect measure: RR (95% CI)

A: RR= 0.91 (95% CI= 0.69 to 1.19) favoring ER

B: RR= 0.12 (95% CI= 0.02 to 0.91) favoring ER

D: RR= 0.63 (95% CI= 0.34 to 1.18) favoring ER

E: RR= 0.57 (95% CI= 0.20 to 1.63) favoring ER

H: RR= 0.81 (95% CI= 0.36 to 1.84) favoring ER

I: RR= 1.24 (95% CI= 1.03 to 1.50) favoring OR

Pooled effect (random effects model):

RR= 0.84 (95% CI= 0.58 to 1.21) favoring ER

Heterogeneity (I²): 67.6%; p=0.009

Conclusion

Current pooled evidence suggested that when compared

to OR, ER is a comparable surgical approach for SNMs

based on direct data in the literature in terms of many surgical outcomes, tempering superiority assumptions implied by indirect comparisons. Outcomes which will improve our understanding of approach significance in the future include long-term OS outcomes as well as

longer term recurrence control. It is likely that particular

patients and presentations will benefit more from one of the approaches versus the other; however, greater

research is required to identify these potential surgical

subgroups and overcome the selection bias concerns present in the current literature.

Author

Study characteristics

Patient characteristics ²

Intervention (I)

Comparison / control (C) ³

Follow-up

Outcome measures

Kılıç, 2018

Type of study:

Observational study.

Setting and country:

Department of Otolaryngology–Head and Neck Surgery, Rutgers New

Jersey Medical School, Newark, NJ, USA

Funding

None provided.

Conflicts of interest:

None provided.

Inclusion criteria:

- Cases of SNSCC diagnosed between January 1, 2010 and December 31, 2014;

- Cases that received definitive primary site surgery, with a known surgical approach.

Exclusion criteria:

- Cases diagnosed before 2010;

- Cases that received preoperative chemotherapy and/or radiotherapy;

- Cases that received postoperative chemotherapy and/or radiotherapy 180 days after surgery;

- Cases coded as “stage 0” disease and those with unknown overall stage;

- Cases with any N stage other than N0;

- M1/stage IVC cases;

- Cases with unknown vital status or follow-up time.

N total at baseline: 1438

Intervention: N = 353

Control: N = 1130

Age <60; N (%):

I: N = 135 (38.2%)

C: N = 389 (34.4%)

Total: N = 524 (35.3%)

Age 60-70; N (%):

I: N = 99 (28%)

C: N = 376 (33.3%)

Total: N = 475 (32%)

Age >70; N (%):

I: N = 119 (33.7%)

C: N = 365 (32.3%)

Total: N = 484 (32.6%)

P=0.168

Sex:

I: N = 218 (61.8%) M

C: N = 745 (65.9%) M

Total: N = 963 (64.9%)

P=0.152

Groups comparable at baseline?

The 2 groups did not differ in any demographic characteristics, but they did differ significantly in facility

type, anatomic subsite, AJCC clinical stage group, tumor

size, and histologic grade. Mortality at 30 days and 90 days did not significantly differ between the 2 groups (p = 0.901 and p = 0.307, respectively), suggesting that there is no significant difference in perioperative mortality.

Describe intervention:

Endoscopic surgical approach

Describe control:

Open surgical approach

Length of follow-up:

Not reported.

Loss-to-follow-up:

Not reported.

Incomplete outcome data:

Not reported.

5-year overall survival (BEFORE PROPENSITY SCORE-MATCHING); % (95% CI)

I: 46.0% (95% CI= 33.2 to 58.8)

C: 56.5% (95% CI= 51.3 to 61.6)

p=0.953

5-year overall survival (AFTER PROPENSITY SCORE-MATCHED COHORT); % (95% CI)

I: 50.8% (95% CI= 37.7 to 63.5)

C: 56.53% (95% CI= 46.5 to 66.5)

P=0.850

Length of stay (days) (BEFORE PROPENSITY SCORE MATCHING)

I: 2.50 days

C: 4.67 days

P<0.0001

Length of stay (days) (AFTER PROPENSITY SCORE MATCHING)

I: 2.54 days

C: 4.69 days

P<0.0001

Surgical approach; Univariate HR (95% CI)

I: HR= 1.00 (REF)

C: HR= 1.01 (95% CI= 0.81 to 1.26)

P=0.953

Surgical approach; Multivariate HR (95% CI)

I: HR= 1.00 (REF)

C: HR= 1.01 (95% CI= 0.77 to 1.34)

P=0.932

Complications

Not reported due to a limited number of variables related to complications.

Quality of life

Not reported.

Study

First author, year

Appropriate and clearly focused question?¹

Yes/no/unclear

Comprehensive and systematic literature search?

Yes/no/unclear

Description of included and excluded studies?³

Yes/no/unclear

Description of relevant characteristics of included studies?⁴

Yes/no/unclear

Appropriate adjustment for potential confounders in observational studies?5

Yes/no/unclear/notapplicable

Assessment of scientific quality of included studies?⁶

Yes/no/unclear

Enough similarities between studies to make combining them reasonable?⁷

Yes/no/unclear

Potential risk of publication bias taken into account?⁸

Yes/no/unclear

Potential conflicts of interest reported?⁹

Yes/no/unclear

Hur (2019)

Yes

Aim
This study aims to systematically review the current literature to compare the survival outcomes of endoscopic versus nonendoscopic surgical resection in patients with SNM.

Inclusion criteria
Studies were included if they met the following criteria: (1) participants: patients diagnosed with SNM; (2) intervention: ER of SNM; (3) comparison: open resection (OR) or combined resection (endoscopic and open) of SNM; (4) outcomes: OS or DFS.

Exclusion criteria
Exclusion criteria included studies which did not report survival rates or no comparison was performed between endoscopic and open approaches.

Yes

Period
Until October 13, 2017and updated on May 4, 2018

Databases
Pubmed, Embase, Cochrane library and Web of Science

Yes

We obtained full-text articles for 56 titles in our review. Forty-seven articles were excluded either because they did not compare an endoscopic versus an open approach for resection of SNM or they did not report OS or DFS.

Yes

The studies gives a brief description of the sample size, follow-up period, design, interventions, disease stage etc. of the included studies.

Unclear

Yes

The Newcastle-Ottawa Quality Assessment Scale was used to review the quality of the included studies.

Yes

Seven out of 9 studies had sufficient data to be included in the meta-analysis of OS. The 2 studies excluded either did not report OS or had missing data such as the absence of a Kaplan–Meier survival curve.

No

Publication bias may have prevented smaller case series or publications with negative results to be published, leading our data set to have an inflated survival rate.

Yes

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Lu (2019)

Yes

Aim
the aim of this meta-analysis was to critically evaluate ER versus OR by pooling the current literature to identify and compare clinical outcomes between the 2 approaches in treating all primary SNMs based on comparative studies only.

Inclusion criteria
(1) histopathological primary malignancies involving the sinonasal cavity, (2) in comparative cohorts managed by both ER and OR approaches, (3) with reported at least one surgical outcome, and (4) in patients>18 years. Surgical outcomes were positive margins, complications, and length of stay (LOS).

Exclusion criteria
(1) mucosal melanoma (MM) histology as that represents a distinctly different tumor type, (2) tumors that are not deemed malignant, for example, angiofibroma and papilloma, (3) recurrent pathologies, (4) hybrid cranioendoscopic approaches to resection, and (5) case series/reports with no comparative cohort.

Yes
Period
Up to April 2019.

Databases
Ovid Embase, Pubmed, Scopus and Cochrane

No

Reasons for exclusion of the full-texts were not reported.

After removal of 479 duplicate studies, inclusion and exclusion criteria were applied to titles and abstracts of the 521 articles. This yielded 26 studies that underwent full-text analysis. Ten individual retrospective cohort studies20–29 fulfilled the selection criteria and were included for quantitative analysis (Table 1).

Yes

Overall, all included studies reported the survival prognosis of 900 SNM patients, with median age ranging from 35 to 70 years, and 42% to 98% were male. ER and OR were utilized in 399 (44%) and 501 (56%) cases, respectively (Table 1). Median follow-up of all patients ranged from 21 to 74 months across all studies. Where reported, radiation therapy and chemotherapy were used in management in 38% to 82% and 15% to 62% of cases, respectively. When studies reported outcomes for multiple SNM histologies, SCC was the most common.

Yes

Meta-regression was performed to analyze potential effect modification by mean age, proportion male, median follow-up, radiation therapy use, and chemotherapy use on overall outcomes with 5 studies. These results were summarized by the slope, which identified the direction of the modification by the variable. All P values were 2-sided with significance set at P<.05.

Yes

According to the Newcastle Ottawa Scale criteria, all studies were identified to be of good quality.

No

Three studies reported the incidence of positive margins. Although not robust enough for meta-analysis due to their heterogenous values, we describe their individual findings in turn.

Yes

Publication and small-study biases were assessed through the generation of a funnel plot and assessedfor asymmetry. Egger’s linear regression test and Begg’s correlation test were used to investigate suspect asymmetry.

Yes

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Study reference

(first author, year of publication)

Bias due to a non-representative or ill-defined sample of patients?¹

(unlikely/likely/unclear)

Bias due to insufficiently long, or incomplete follow-up, or differences in follow-up between treatment groups?²

(unlikely/likely/unclear)

Bias due to ill-defined or inadequately measured outcome ?³

(unlikely/likely/unclear)

Bias due to inadequate adjustment for all important prognostic factors?⁴

(unlikely/likely/unclear)

Kılıç (2018)

Unlikely

The study applied specific eligibility criteria

Unlikely

Patients with unknown follow-up time were excluded.

Unlikely

The study investigates hard (objective) outcome measures (5 year overall survival).

Unlikely

Multivariate binary logistic regression was used to identify factors associated with endoscopic surgical approach. Variables with p ≤ 0.1 on univariate analysis were

included in the multivariate model. p<0.05 was considered significant for all other tests.