Predictors of survival in head and neck mucosal melanoma

Oral Oncology 73 (2017) 36–42

Contents lists available at ScienceDirect

Oral Oncology journal homepage: www.elsevier.com/locate/oraloncology

Predictors of survival in head and neck mucosal melanoma Matthew Q. Schmidt a,b,1, John David b,c,1, Emi J. Yoshida b,c, Kevin Scher c, Alain Mita c, Stephen L. Shiao b,c, Allen S. Ho c,d, Zachary S. Zumsteg b,c,⇑ a

Creighton University School of Medicine, Phoenix Regional Campus, 525 W. Earll Dr., Phoenix, AZ 85013, USA Department of Radiation Oncology, Cedars-Sinai Medical Center, 8700 Beverly Blvd., Los Angeles, CA 90048, USA c Samuel Oschin Comprehensive Cancer Institute, Cedars-Sinai Medical Center, 8700 Beverly Blvd., Los Angeles, CA 90048, USA d Department of Surgery, Cedars-Sinai Medical Center, 8700 Beverly Blvd., Los Angeles, CA 90048, USA b

a r t i c l e

i n f o

Article history: Received 29 March 2017 Received in revised form 26 July 2017 Accepted 3 August 2017

Keywords: Melanoma Head and neck cancer Database Demographics Survival analysis Nasopharynx cancer Paranasal sinus cancer Oral cancer Nasal cancer

a b s t r a c t Objectives: To evaluate hospital-based data of head and neck mucosal melanoma patients in order to identify predictors of survival. Materials and methods: The National Cancer Data Base was used to identify 1368 patients with head and neck mucosal melanoma diagnosed between the years of 2004 and 2012. The Kaplan-Meier method was utilized to estimate overall survival, and multivariate Cox regression analyses were performed to assess the impact of covariates on survival after adjusting for confounding variables. Results: Median follow-up was 55.2 months. Median survival of all patients was 29.3 months, and the 5-year survival was 27.4%. After adjusting for other prognostic factors in multivariate analysis, paranasal sinus location [hazard ratio (HR) = 1.54, 95% Confidence Interval (CI) = 1.30–1.82, P < 0.001)] and the use of radiotherapy alone for definitive local treatment (HR = 2.27, 95% CI = 1.72–2.98, P < 0.001) were associated with worse survival. Similar results were seen in the subgroup of patients with complete clinical staging information. In terms of patterns of care, the use of combined surgery and radiotherapy as the primary local treatment modality has significant increased from 2004 and 2012 (P = 0.03). Conclusion: Outcomes in mucosal melanoma of the head and neck remain suboptimal, despite increased use of multimodality local therapy, likely due to the high risk of distant metastases. Mucosal melanomas arising from the paranasal sinuses have particularly poor prognosis. Novel therapeutic paradigms for head and neck mucosal melanoma, incorporating systemic therapies to decrease the risk of distant relapse, should be pursued in clinical trials. Ó 2017 Elsevier Ltd. All rights reserved.

Introduction Mucosal melanomas are among the most lethal of head and neck cancers. Although they can arise from any mucosa-lined surface, mucosal melanomas most commonly occur in the sinonasal regions of the head and neck [1]. In comparison to cutaneous, acral, and ocular melanomas, mucosal melanoma exhibits an aggressive clinical course with poor prognosis [2–5]. Diagnosis is often delayed given few and nonspecific presenting symptoms mimicking more common benign processes [6]. Furthermore, head and neck mucosal melanomas (HNMMs) have relatively high rates of ⇑ Corresponding author at: Department of Radiation Oncology, Cedars-Sinai Medical Center, 8700 Beverly Blvd., Los Angeles, CA 90048, USA. E-mail addresses: [email protected] (M.Q. Schmidt), John. [email protected] (J. David), [email protected] (E.J. Yoshida), Kevin.Scher@cshs. org (K. Scher), [email protected] (A. Mita), [email protected] (S.L. Shiao), [email protected] (A.S. Ho), [email protected] (Z.S. Zumsteg). 1 Contributed equally. http://dx.doi.org/10.1016/j.oraloncology.2017.08.002 1368-8375/Ó 2017 Elsevier Ltd. All rights reserved.

local, regional, and distant metastases despite aggressive multimodality therapy [7]. Due to the rarity of the disease, HNMM literature mainly consists of case reports and retrospective, single institution studies with limited patient numbers and heterogeneous anatomic sites. Thus, there is no clear consensus on treatment guidelines. Primary resection is a current standard therapy for resectable stage III-IVA disease, but the role of adjuvant or definitive radiotherapy is less certain [5,8]. Numerous studies have demonstrated improved locoregional control with adjuvant radiotherapy [9–17], but generally with no impact on overall survival. Thus, the optimal treatment paradigm in HNMM remains to be defined. Given the limited knowledge about prognostic factors and optimal therapeutic paradigms, we sought to use the National Cancer Data Base to analyze the impact of clinicopathologic, demographic, and treatment-related factors in a relatively large cohort of HNMM patients.

37

M.Q. Schmidt et al. / Oral Oncology 73 (2017) 36–42

Materials and methods We retrospectively reviewed patients with head and neck mucosal melanoma (HNMM) using the National Cancer Data Base, a hospital-based cancer database sponsored by the Commission on Cancer of the American College of Surgeons and the American Cancer Society. The database includes approximately 70% of newlydiagnosed cancer patients from approximately 1600 accredited cancer care programs in the United States and Puerto Rico. This study was deemed to be exempt from full institutional review board review by Cedars-Sinai Medical Center. The Consolidated Standards of Reporting Trials (CONSORT) diagram details the study inclusion criteria (Supplementary Fig. 1). Patients with non-mucosal lip cancers (N = 8589) and salivary primaries (N = 28,206) were excluded given that these are not mucosal sites. For other head and neck sites, International Classification of Diseases for Oncology 3rd edition morphologic codes 8720– 8790 were used to extract 2009 patients diagnosed with invasive mucosal melanoma of all non-cutaneous, mucosal head and neck

sites between the years 2004 and 2012. Patients were staged according to the American Joint Committee on Cancer 7th edition. We excluded patients with unknown details regarding follow up (N = 233), chemotherapy (N = 60), or the timing of radiotherapy with respect to surgery (N = 11). Patients who received nonexternal beam radiotherapy (N = 6), intraoperative radiotherapy (N = 1), or underwent a non-excisional surgical procedure were excluded (N = 2). Patients that presented with distant metastases (N = 214) or patients without any definitive local therapy (N = 114) were also excluded from this study. The remaining 1368 patients were included in our analysis. Tumor and nodal clinical staging information was only available for patients diagnosed in 2010 or later. Patients with full clinical staging information were analyzed in a separate subgroup analysis. Kaplan-Meier calculations were used to estimate survival. Survival curves were compared using log-rank tests. Median follow-up was calculated with the reverse Kaplan-Meier method. Univariate and multivariate Cox regression analysis was used to calculate hazards ratios (HR) and 95% confidence intervals (CI).

Table 1 Baseline demographic and clinical characteristics of patients with head and neck mucosal melanomas, stratified by anatomic site. RT, radiotherapy. N/A, not applicable. Patient demographics Variable

Nasal cavity

Paranasal Sinuses

Other Sites

Total

Total patients (N) Age [Median (range)]

749 73 (29–90)

293 71 (24–90)

326 66 (21–90)

1368 71 (21–90)

N (%)

N (%)

Sex Male Female

335 (44.7) 414 (55.3)

136 (46.4) 157 (53.6)

Race White Black Other

681 (90.9) 30 (4.0) 38 (5.1)

280 (95.6) 7 (2.4) 6 (2.0)

Year of diagnosis 2004–2006 2007–2009 2010–2012

226 (30.2) 264 (35.2) 259 (34.6)

93 (31.7) 104 (35.5) 96 (32.8)

Academic center No Yes Unknown

285 (38.1) 455 (60.7) 9 (1.2)

100 (34.1) 186 (63.5) 7 (2.4)

Charlson/Deyo Score 0 1 2

637 (85.0) 97 (13.0) 15 (2.0)

255 (87.0) 28 (9.6) 10 (3.4)

Clinical T stage cT3 cT4a cT4b cTX

136 (18.2) 50 (6.7) 24 (3.2) 539 (72.0)

24 (8.2) 27 (9.2) 11 (3.8) 231 (78.8)

Clinical N Stage cN0 cN1 cNX

215 (28.7) 13 (1.7) 521 (69.6)

59 (20.1) 3 (1.0) 231 (78.8)

668 (89.2) 81 (10.8)

268 (91.5) 25 (8.5)

695 (92.8) 54 (7.2)

249 (85.0) 44 (15.0)

320 (42.7) 40 (5.3) 389 (51.9)

90 (30.7) 37 (12.6) 166 (56.7)

Neck dissection No Yes Chemotherapy No Yes Definitive treatment Surgery Alone RT Alone Surgery and RT

p*

N (%)

0.6

p**

180 (55.2) 146 (44.8) 0.04

651 (47.6) 717 (52.4) 0.02

279 (85.6) 26 (8.0) 21 (6.4) 0.8

1240 (90.6) 63 (4.6) 65 (4.8) 0.04

93 (28.5) 95 (29.1) 138 (42.3) 0.3

412 (30.1) 463 (33.8) 493 (36.0) 0.02

95 (29.1) 213 (65.3) 18 (5.5) 0.1

482 (35.2) 855 (62.5) 34 (2.4) 0.3

271 (83.1) 43 (13.2) 12 (3.7) 0.001

1163 (85.0) 168 (12.3) 37 (2.7) 0.01

73 (22.4) 35 (10.7) 5 (1.5) 213 (65.3) 0.8

233 (17.0) 112 (8.2) 40 (2.9) 983 (71.9) <0.001

74 (22.7) 43 (13.1) 209 (64.1) 0.3

348 (25.4) 59 (4.3) 961 (70.2) <0.001

200 (61.3) 126 (38.6) <0.001

1136 (83.0) 232 (17.0) 0.05

291 (89.2) 35 (10.7) <0.001

Bold values identify statistically significant covariates (P <= 0.05). * Difference between frequencies in nasal cavity and paranasal sinuses, Chi-Squared analysis. ** Difference between frequencies in nasal cavity and other anatomic sites, Chi-Squared analysis.

N (%)

0.002

1235 (90.3) 133 (9.7) <0.001

172 (52.8) 24 (7.4) 130 (39.9)

566 (41.4) 98 (7.2) 704 (51.5)

38

M.Q. Schmidt et al. / Oral Oncology 73 (2017) 36–42

Chi-squared analyses were performed to compare frequencies between subgroups of categorical variables, and trends in patterns of care were analyzed with linear regression. A two-sided p-value of 0.05 was considered statistically significant. Statistical analyses were performed with IBM SPSS Software version 23 (Armonk, New York). Results A total of 1368 patients with head and neck mucosal melanoma (HNMM) diagnosed between the years 2004 and 2012 and meeting the inclusion criteria of the study were identified (Supplementary Fig. 1). Median follow up for the cohort was 55.2 months. Demographic data from the cohort is shown in Table 1. Overall survival at 1, 3, and 5 years was 77.9%, 42.4%, and 27.4%, respectively (Supplementary Fig. 2). Median survival for the entire cohort was 29.0 months. Of note, excluded patients presenting with distant metastatic disease or those not receiving any local treatment had worse prognosis than patients meeting inclusion criteria, with median survival of 13.6 months and 10.7 months, respectively. For definitive local treatment, 566 (41.4%) patients underwent surgery alone, 98 (7.2%) underwent radiotherapy alone, and 704 (51.4%) underwent both surgery and radiotherapy. Between the years 2004 and 2012, we observed a statistically significant increase in the rate of combined modality therapy with both surgery and radiotherapy from 43.7% to 53.1% (P = 0.03) (Fig. 1). For patients undergoing surgery alone, radiotherapy alone, or combined surgery and radiotherapy, 5-year overall survival was 31.0%, 8.1%, and 26.9% (overall P < 0.001), respectively (Fig. 2A). Radiotherapy alone was also associated with decreased survival in comparison to patients undergoing surgery in both univariate analysis (hazard ratio [HR] = 2.49, 95% confidence interval [CI]

Fig. 1. Rates of use of surgery, radiotherapy (RT) and surgery with radiotherapy between the years 2004 and 2012 in non-metastatic head and neck mucosal melanoma. P-values represent the comparison of the linear regression line and a line with slope equal to zero for each treatment modality.

= 1.96–3.17, P < 0.001) and multivariate analysis (HR = 2.27, 95% CI = 1.72–2.98, P < 0.001, Table 2). By contrast, there was no significant difference in survival comparing surgery alone to combined surgery and radiotherapy in univariate (HR = 1.06, 95% CI = 0.92– 1.22, P = 0.4) or multivariate analysis (HR = 0.99, 95% CI = 0.85– 1.15, P = 0.9). In terms of anatomic site of origin, paranasal sinus mucosal melanoma had significantly worse overall survival in comparison to nasal cavity mucosal melanoma in both univariate (HR = 1.61, 95% CI = 1.37–1.89, P < 0.001) and multivariate analysis (HR = 1.54, 95% CI = 1.30–1.82, P < 0.001) (Table 2). Non-sinonasal tumor location (73.9% oral cavity, 13.8% nasopharynx, 10.1% oropharynx, 2.1% other head and neck sites) was associated with better survival than nasal cavity mucosal melanoma in univariate (HR = 0.75, 95% CI = 0.63–0.89, P < 0.001) and multivariable analyses (HR = 0.71, 95% CI = 0.58–0.86, P < 0.001). 5-year overall survival for nasal cavity, paranasal sinuses, and other head and neck locations was 25.9%, 18.4%, and 38.7% (overall P < 0.001), respectively (Fig. 2B). Other variables found to be independent predictors of worse survival on multivariate analysis for the overall cohort include older age (HR = 1.02 per year, 95% CI = 1.02–1.03, P < 0.001), clinical tumor stage cT4b (HR = 2.09, 95% CI = 1.34–3.25, P < 0.001), Charlson comorbidity score of 2 or more (HR = 1.58, 95% CI = 1.06–2.33), and clinical nodal stage cN1 (HR = 2.22, 95% CI = 1.46–3.38, P < 0.001) and cNx (HR = 1.98, 95% CI = 1.24–3.15, P = 0.004) compared to stage cN0 patients (Table 2). Additionally, women had superior survival (HR = 0.82, 95% CI = 0.72–0.94, P = 0.005) in comparison to men. Given that patients diagnosed prior to 2010 lacked clinical staging information, a subgroup analysis was performed in patients diagnosed in 2010 or later and with complete clinical tumor and nodal staging (N = 372). The median follow-up for these patients was 32.1 months. In this subgroup, 3-year overall survival for patients undergoing surgery alone, radiotherapy alone, and combined surgery and radiotherapy was 45.3%, 18.7%, and 44.7% (overall P = 0.01), respectively (Supplemental Fig. 3A). As in the full cohort, there was significantly worse survival for patients undergoing radiotherapy alone in univariate analysis (HR = 2.37, HR = 1.38–4.07, P = 0.002, Table 3) and multivariate analysis (HR = 1.88, 95% CI = 1.01–3.51, P = 0.05) in comparison to those treated with surgery alone. Of note, only 23 patients in this subgroup were treated with radiotherapy alone. There was no difference in survival between patients undergoing surgery with or without radiotherapy in univariate and multivariate analysis. Additionally, although survival in patients with paranasal sinus mucosal melanoma remained significantly worse (HR = 2.17, 95% CI = 1.45–3.24, P < 0.001) than those with nasal cavity primaries in multivariate analysis, there was no significant difference in survival between nasal cavity and non-sinonasal mucosal melanoma, in contrast to what was observed in the full cohort (Table 3). Threeyear overall survival was 46.4%, 23.4%, and 52.4% (overall P < 0.001) for patients in this subgroup with mucosal melanoma of the nasal cavity, paranasal sinuses, and other anatomic sites, respectively (Supplemental Fig. 3B). Additional independent predictors of decreased survival in this subgroup include increasing age (HR = 1.03, 95% CI = 1.01–1.04, P < 0.001), black race (HR = 1.92, 95% CI = 1.06–3.47, P = 0.03), stage cT4b disease (HR = 2.46, 95% CI = 1.50–4.03, P < 0.001), and cN1 nodal stage (HR = 2.13, 95% CI = 1.22–3.71, P < 0.001).

Discussion In this study, we found the 5-year survival for non-metastatic head and neck mucosal melanoma (HNMM) to be 27.4%,

M.Q. Schmidt et al. / Oral Oncology 73 (2017) 36–42

confirming the relatively poor prognosis of this disease reported in previous studies [5,14,18,19]. When adjusting for stage, anatomic site, and other prognostic factors in multivariable analyses, we observed significantly improved overall survival for patients undergoing surgery with or without radiotherapy in comparison to patients treated with non-surgical approaches in the overall cohort. Although non-surgical treatment was only associated with borderline significantly worse survival in a multivariate analysis of the subgroup of patients diagnosed after 2010 with complete staging information (HR = 1.88, 95% CI = 1.01–3.51, P = 0.05), this was most likely due to a lack of statistical power. Only 23 of 372

39

patients in this group did not undergo surgery, and follow-up was relatively limited given the more recent years of diagnosis, leading to wide confidence intervals. In sum, these data support surgery as a primary management strategy in patients with resectable disease, although it is possible that the non-surgical cohort had more advanced disease characteristics and aggressive biologic features that could not be completely accounted for in our statistical analyses. It is also notable that there was no difference in survival comparing surgery alone to surgery with radiotherapy in either the overall cohort or in the subgroup with complete staging informa-

Fig. 2. Overall survival for patients with non-metastatic head and neck mucosal melanoma, stratified by (A) local treatment modality and (B) anatomic site in the entire cohort.

40

M.Q. Schmidt et al. / Oral Oncology 73 (2017) 36–42

Table 2 Univariate and multivariate Cox regression analysis of overall survival in the entire cohort of patients with non-metastatic head and neck mucosal melanoma. CI, confidence interval HR, hazard ratio. RT, radiotherapy. Univariate Analysis Variable

Multivariate Analysis

HR (95% CI)

p

HR (95% CI)

p

1.02 (1.02–1.03)

<0.001

1.02 (1.02–1.03)

<0.001

0.96 (0.84–1.10)

0.6

0.82 (0.72–0.94)

0.005

Race White Black Other

1.00 (reference) 1.37 (1.01–1.84) 1.08 (0.80–1.47)

– 0.04 0.6

1.00 (reference) 1.19 (0.86–1.63) 1.24 (0.90–1.71)

– 0.3 0.2

Academic center Yes vs No (reference)

0.87 (0.76–1.00)

0.05

0.93 (0.81–1.07)

0.3

Comorbidities Charlson/Deyo 0 Charlson/Deyo 1 Charlson/Deyo 2

1.00 (reference) 1.01 (0.82–1.23) 1.50 (1.02–2.21)

– 1.0 0.38

1.00 (reference) 1.06 (0.86–1.30) 1.58 (1.06–2.33)

– 0.6 0.02

Anatomic site Nasal Cavity Paranasal Sinuses Other

1.00 (reference) 1.61 (1.37–1.89) 0.75 (0.63–0.89)

– <0.001 0.00

1.00 (reference) 1.54 (1.30–1.82) 0.71 (0.58–0.86)

– <0.001 <0.001

Clinical T stage cT3 cT4a cT4b cTX

1.00 (reference) 1.54 (1.12–2.13) 2.51 (1.67–3.77) 1.33 (1.07–1.64)

– 0.01 <0.001 0.01

1.00 (reference) 1.37 (0.98–1.91) 2.09 (1.34–3.25) 0.73 (0.45–1.20)

– 0.06 0.001 0.2

Clinical N Stage cN0 cN1 cNX

1.00 (reference) 1.64 (1.13–2.36) 1.27 (1.06–1.52)

– 0.01 0.01

1.00 (reference) 2.22 (1.46–3.38) 1.98 (1.24–3.15)

– <0.001 0.004

Neck dissection Yes vs No (reference)

0.86 (0.71–1.03)

0.10

1.13 (0.91–1.40)

0.3

Chemotherapy Yes vs No (reference)

1.39 (1.12–1.72)

0.004

1.18 (0.933–1.49)

0.2

Definitive treatment Surgery Alone RT Alone Surgery and RT

1.00 (reference) 2.49 (1.96–3.17) 1.06 (0.92–1.22)

– <0.001 0.4

1.00 (reference) 2.27 (1.72–2.98) 0.99 (0.85–1.15)

– <0.001 0.9

Age Sex Female vs. Male (reference)

Bold values identify statistically significant covariates (P <= 0.05).

tion, consistent with prior reports [5,7,17]. Although the National Cancer Data Base does not report patterns of recurrence, the lack of survival benefit from multimodality local therapy, despite multiple studies showing that radiotherapy improves local control [7,9,17], is likely because distant metastasis occurs at very high rates in HNMM and is the primary driver of survival. Thus, improving survival by escalating the intensity of local therapy is unlikely, at least until better systemic control is achievable, and novel therapeutic paradigms designed to abrogate the risk of distant relapse should be pursued in clinical trials. In particular, immune checkpoint antibodies, such as PD-1 inhibition, have shown some activity in this disease, with a response rate of 23% in metastatic mucosal melanoma in one retrospective series [20]. Thus, clinical trials investigating the benefit of immune checkpoint antibodies in the adjuvant setting may be one promising approach for improving outcomes. For patients with specific molecular alterations, trials testing adjuvant therapy with molecularly targeted inhibitors is another potential strategy. Although prospective trials in this disease are challenging to perform given the rarity of the disease, further improvement in the dismal outcomes observed at unlikely without them. Despite the primacy of distant relapse in terms of survival, radiotherapy can have an important role in mucosal melanoma, and we observed a significant increase in the use of combined sur-

gery and post-operative radiotherapy from 2004 and 2012. Achieving locoregional control in HNMM is an important objective given the morbidity associated with disease progression and salvage local therapy in this intricate anatomic region containing numerous critical nerve, vascular, and functional pathways. Surgery is the primary local treatment modality for resectable HNMMs, but anatomic constraints make achieving en-bloc negative margin resection difficult without excessive morbidity. Multiple studies, have demonstrated that adjuvant radiotherapy improves local control in HNMM [7,9,17], but not overall survival, consistent with our results. Despite the lack of survival benefit, we believe postoperative radiotherapy is an important component of treatment for patients with localized disease in order to decrease the morbidities, costs, and psychological impacts associated with disease recurrence and local salvage treatments. Moreover, the impact of adjuvant radiotherapy on survival may emerge as improved systemic therapies capable of decreasing the risk of distant metastasis are incorporated in the multimodality therapeutic paradigm of HNMM. We observed significantly worse survival with mucosal melanomas arising from the paranasal sinuses compared to other anatomic locations, even after adjusting for other covariates, in both the overall cohort and in patients with full clinical staging information. There has been conflicting data regarding the inde-

41

M.Q. Schmidt et al. / Oral Oncology 73 (2017) 36–42

Table 3 Univariate and multivariate Cox regression analysis for overall survival in the subset of patients with head and neck mucosal melanoma and complete clinical staging data. CI, confidence interval HR, hazard ratio. RT, radiotherapy. Univariate analysis Variable

Multivariate analysis

HR (95% CI)

p

HR (95% CI)

p

1.02 (1.01–1.04)

<0.001

1.03 (1.01–1.04)

<0.001

1.08 (0.81–1.46)

0.6

1.08 (0.79–1.48)

0.7

Race White Black Other

1.00 (reference) 1.89 (1.11–3.21) 0.96 (0.47–1.95)

– 0.02 0.9

1.00 (reference) 1.92 (1.06–3.47) 1.06 (0.51–2.20)

– 0.03 0.9

Academic Center Yes vs No (reference)

0.85 (0.63–1.15)

0.3

1.02 (0.73–1.41)

0.9

Comorbidities Charlson/Deyo 0 Charlson/Deyo 1 Charlson/Deyo 2

1.00 (reference) 1.11 (0.75–1.68) 1.02 (0.45–2.30)

– 0.6 1.0

1.00 (reference) 1.35 (0.87–2.08) 0.82 (0.35–1.91)

– 0.2 0.7

Anatomic Site Nasal Cavity Paranasal Sinuses Other

1.00 (reference) 1.90 (1.32–2.73) 0.90 (0.63–1.30)

– 0.00 0.6

1.00 (reference) 2.17 (1.45–3.24) 0.96 (0.61–1.52)

– <0.001 0.9

Clinical T Stage cT3 cT4a cT4b

1.00 (reference) 1.61 (1.16–2.25) 2.82 (1.83–4.37)

– 0.005 <0.001

1.00 (reference) 1.32 (0.92–1.89) 2.46 (1.50–4.03)

– 0.1 <0.001

Clinical N Stage N1 vs N0 (reference)

1.53 (1.05–2.22)

0.03

2.13 (1.22–3.71)

0.008

Chemotherapy Yes vs No (reference)

1.30 (0.82–2.08)

0.3

1.15 (0.68–1.95)

0.5

Definitive Treatment Surgery Alone RT Alone Surgery and RT

1.00 (reference) 2.20 (1.30–3.72) 1.03 (0.75–1.41)

– 0.003 0.9

1.00 (reference) 1.88 (1.01–3.51) 0.99 (0.70–1.42)

– 0.05 0.7

Age Sex Female vs. Male (reference)

Bold values identify statistically significant covariates (P <= 0.05).

pendent impact of anatomic site on outcome [5]. For example, although multiple older studies have demonstrated no impact of anatomic site on survival [12,15,21], several more recent studies have shown worse outcomes with paranasal mucosal melanoma compared to other HNMM sites [7,22,23]. Our results strongly support anatomic site as an important predictor of survival in HNMM. There are several plausible explanations that may underlie the inferior outcomes observed with paranasal sinus primaries. Given the anatomic location of these tumors, they tend be detected relatively late and present with extensive primary tumors, as do virtually all histologies arising from this site. Although our multivariate models adjust for stage, it is possible that paranasal HNMMs are more extensive even within the same stage category, as the current mucosal melanoma staging definitions are broad and may not be sensitive enough to capture the full spectrum of heterogeneity of primary tumor extent. Additionally, obtaining a gross total resection is generally more challenging in the paranasal sinuses than in other locations of the head and neck, such as the nasal cavity and oral cavity, given their deep location and proximity to other critical structures. Lastly, biologic differences in HNMMs arising in different anatomic locations may in part drive their outcomes. Although relatively little is known about the pathogenesis and risk factors for HNMM, it has been shown that there are different gene alterations implicated among various types of melanomas, such as higher rates of KIT alterations in HNMMs, BRAF mutations in cutaneous malignant melanomas, and GNAQ and GNA11 mutations in uveal melanoma [5,24,25]. There is some evidence that the genetic landscape may vary between individual HNMM sites [5,24,26], and

these differences in biology may manifest as different clinical courses. Multiple limitations of this analysis warrant additional discussion. First, this is retrospective observational study, with all the biases inherent to any such study. For instance, we recognize that patients in the surgery group may have more favorable disease and are more likely to have better outcomes independent of surgery itself. Additionally, the National Cancer Data Base only captures overall survival, but has no information regarding cancer-specific outcomes such as locoregional control, distant recurrence, or cause-specific survival. We were also limited by a large proportion of patients with incomplete clinical staging, due to the fact that a universal staging system for HNMM was not developed until 2010. To overcome this, we performed multivariable analyses both in the overall cohort and in patients with complete clinical staging information, and have focused our conclusion on findings that are consistent in both analyses. Finally, the National Cancer Data Base does not collect data regarding certain factors that are critical determinants of survival, including smoking and body mass index. Conclusion Despite aggressive treatment, survival for HNMM remains poor. There has been increased use of surgery combined with radiotherapy in the definitive treatment of local disease between the years 2004 and 2012. Patients undergoing surgery with or without radiation have improved survival in comparison to patients treated with radiation alone. Although there is no difference in survival

42

M.Q. Schmidt et al. / Oral Oncology 73 (2017) 36–42

with or without post-operative radiotherapy, likely due to high distant relapse rates in this disease, multimodality therapy to achieve locoregional control is reasonable if it can be achieved with acceptable morbidity. Paranasal sinus anatomic location portends worse prognosis even after adjusting for other prognostic factors. Clinical trials in HNMM that investigate novel therapeutic approaches, focusing on treatments that may abrogate the risk of distant recurrence, are imperative.

[8] [9]

[10]

[11]

Funding [12]

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Conflict of interest The authors have no conflicts of interest related to this work. ZSZ is on the external advisory board for the Scripps Proton Therapy Center and is a consultant for EMD Serono. All other authors have no disclosures. Appendix A. Supplementary material

[13]

[14] [15]

[16]

[17]

[18]

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.oraloncology. 2017.08.002.

[19]

References

[20]

[1] Jethanamest D, Vila PM, Sikora AG, Morris LG. Predictors of survival in mucosal melanoma of the head and neck. Ann Surg Oncol 2011;18(10):2748–56. [2] Bishop KD, Olszewski AJ. Epidemiology and survival outcomes of ocular and mucosal melanomas: a population-based analysis. Int J Cancer 2014;134 (12):2961–71. [3] Furney SJ, Turajlic S, Stamp G, Nohadani M, Carlisle A, Thomas JM, et al. Genome sequencing of mucosal melanomas reveals that they are driven by distinct mechanisms from cutaneous melanoma. J Pathol 2013;230(3):261–9. [4] Mendenhall WM, Amdur RJ, Hinerman RW, Werning JW, Villaret DB, Mendenhall NP. Head and neck mucosal melanoma. Am J Clin Oncol 2005;28 (6):626–30. [5] Lazarev S, Gupta V, Hu K, Harrison LB, Bakst R. Mucosal melanoma of the head and neck: a systematic review of the literature. Int J Radiat Oncol Biol Phys 2014;90(5):1108–18. [6] Gasparyan A, Amiri F, Safdieh J, Reid V, Cirincione E, Shah D. Malignant mucosal melanoma of the paranasal sinuses: two case presentations. World J Clin Oncol 2011;2(10):344–7. [7] Samstein RM, Carvajal RD, Postow MA, Callahan MK, Shoushtari AN, Patel SG, et al. Localized sinonasal mucosal melanoma: Outcomes and associations with

[21]

[22]

[23]

[24] [25]

[26]

stage, radiotherapy, and positron emission tomography response. Head Neck 2016;38(9):1310–7. Gavriel H, McArthur G, Sizeland A, Henderson M. Review: mucosal melanoma of the head and neck. Melanoma Res 2011;21(4):257–66. Benlyazid A, Thariat J, Temam S, Malard O, Florescu C, Choussy O, et al. Postoperative radiotherapy in head and neck mucosal melanoma: a GETTEC study. Arch Otolaryngol Head Neck Surg 2010;136(12):1219–25. Medina JE, Ferlito A, Pellitteri PK, Shaha AR, Khafif A, Devaney KO, et al. Current management of mucosal melanoma of the head and neck. J Surg Oncol 2003;83(2):116–22. Temam S, Mamelle G, Marandas P, Wibault P, Avril MF, Janot F, et al. Postoperative radiotherapy for primary mucosal melanoma of the head and neck. Cancer 2005;103(2):313–9. Owens JM, Roberts DB, Myers JN. The role of postoperative adjuvant radiation therapy in the treatment of mucosal melanomas of the head and neck region. Arch Otolaryngol Head Neck Surg 2003;129(8):864–8. Meleti M, Leemans CR, de Bree R, Vescovi P, Sesenna E, van der Waal I. Head and neck mucosal melanoma: experience with 42 patients, with emphasis on the role of postoperative radiotherapy. Head Neck 2008;30(12):1543–51. Gal TJ, Silver N, Huang B. Demographics and treatment trends in sinonasal mucosal melanoma. Laryngoscope 2011;121(9):2026–33. Nandapalan V, Roland NJ, Helliwell TR, Williams EM, Hamilton JW, Jones AS. Mucosal melanoma of the head and neck. Clin Otolaryngol Allied Sci 1998;23 (2):107–16. Ang KK, Peters LJ, Weber RS, Morrison WH, Frankenthaler RA, Garden AS, et al. Postoperative radiotherapy for cutaneous melanoma of the head and neck region. Int J Radiat Oncol Biol Phys 1994;30(4):795–8. Jarrom D, Paleri V, Kerawala C, Roques T, Bhide S, Newman L, et al. Mucosal melanoma of the upper airways tract mucosal melanoma: a systematic review with meta-analyses of treatment. Head Neck 2017;39(4):819–25. Cheng YF, Lai CC, Ho CY, Shu CH, Lin CZ. Toward a better understanding of sinonasal mucosal melanoma: clinical review of 23 cases. J Chin Med Assoc 2007;70(1):24–9. Lee SP, Shimizu KT, Tran LM, Juillard G, Calcaterra TC. Mucosal melanoma of the head and neck: the impact of local control on survival. Laryngoscope 1994;104(2):121–6. Shoushtari AN, Munhoz RR, Kuk D, Ott PA, Johnson DB, Tsai KK, et al. The efficacy of anti-PD-1 agents in acral and mucosal melanoma. Cancer 2016;122 (21):3354–62. Lund VJ, Howard DJ, Harding L, Wei WI. Management options and survival in malignant melanoma of the sinonasal mucosa. Laryngoscope 1999;109(2 Pt 1):208–11. Khan MN, Kanumuri VV, Raikundalia MD, Vazquez A, Govindaraj S, Baredes S, et al. Sinonasal melanoma: survival and prognostic implications based on site of involvement. Int Forum Allergy Rhinol 2014;4(2):151–5. Roth TN, Gengler C, Huber GF, Holzmann D. Outcome of sinonasal melanoma: clinical experience and review of the literature. Head Neck 2010;32 (10):1385–92. Curtin JA, Busam K, Pinkel D, Bastian BC. Somatic activation of KIT in distinct subtypes of melanoma. J Clin Oncol 2006;24(26):4340–6. Van Raamsdonk CD, Griewank KG, Crosby MB, Garrido MC, Vemula S, Wiesner T, et al. Mutations in GNA11 in uveal melanoma. N Engl J Med 2010;363 (23):2191–9. Rivera RS, Nagatsuka H, Gunduz M, Cengiz B, Gunduz E, Siar CH, et al. C-kit protein expression correlated with activating mutations in KIT gene in oral mucosal melanoma. Virchows Arch 2008;452(1):27–32.