Lymph Node Positive Cutaneous Non-Melanoma Skin Cancer: A Poor Prognosis Disease In Need of Treatment Intensification

Abstract

Locoregionally advanced non-melanoma skin cancer (NMSC) has an aggressive clinical course characterized by high rates of treatment failure and poor survival compared to localized skin cancers. Our goal is to investigate multi-modality therapy for lymph node (LN) positive NMSC. LN positive NMSC patients who underwent surgery and adjuvant therapy were retrospectively reviewed from 2000–2012 from a single tertiary cancer center. Median follow-up was 1.8 years (range, 0.5–8.5). Overall survival (OS) and progression free survival (PFS) were calculated using the Kaplan-Meier method. Chi-squared test and logistic regression was used to determine the association of locoregional control (LRC) and the following variables: evidence of extracapsular extension, number of LN positive, largest involved LN, presence of a positive margin, and use of concurrent chemoradiation (CRT). A total of 46 patients were evaluated. Thirteen (28%) received adjuvant CRT. CRT patients were younger (p<0.0001) and had a significantly greater number of positive lymph nodes (p=0.016) than patients who received adjuvant radiation alone. At 5 years, LRC was 76%, PFS was 65%, and OS was 49%. Univariate analysis demonstrated that CRT (p = 0.006), largest LN measurement (p=0.039), and ≥3 involved LN (p=0.001) predicted local recurrence. CRT (p=0.035, OR 0.2 [95% CI 0.05–0.9]) and ≥3 involved LN (p = 0.017, OR 0.07 [95% CI 0.01–0.62]) remained significant on multivariate analysis. CRT was well tolerated. No grade ≥3 toxicities observed except one asymptomatic grade 4 thrombocytopenia. Patients with LN positive NMSC do poorly. Patient selection for adjuvant therapy intensification needs clarification.

Introduction

Non-melanoma skin cancer (NMSC) is the most common malignancy in the United States, with more than 2,000,000 new cases diagnosed annually.⁽¹⁾ While most patients have an excellent prognosis⁽²⁾, a distinct minority of patients either present with regional metastases⁽³⁾ or develop lymphadenopathy as a component of recurrent disease.⁽⁴⁾ Locoregionally advanced NMSC has an aggressive clinical course characterized by high rates of treatment failure and poor survival compared to localized skin cancers.⁽⁵⁾

Although there are no formal randomized trials investigating the role of adjuvant therapy for resected high risk NMSC, the current dogma suggests consideration of adjuvant radiotherapy for the following: positive margin at the primary site, pathologically-involved lymph nodes, perineural invasion (particularly if symptomatic), and recurrent disease.⁽⁶⁾ Despite aggressive therapy (i.e., surgical excision and adjuvant radiotherapy), disease free survival for NMSC treated with multi-modality therapy ranges from 34 – 78%.^(7–11) Because most cases occur in the head and neck⁽⁴⁾ and have histologic features similar to mucosal head and neck squamous cell cancers (HNSCC)⁽¹²⁾, it is hypothesized that the survival benefit of concurrent systemic therapy for mucosal HNSCC⁽¹³⁾ can be extended to NMSC with similar prognostic features. An ongoing randomized phase III trial (TROG 05.01) tests this proposition. Until currently accruing trials are reported, practitioners must often decide whether to intensify therapy without concrete evidence of its safety and efficacy.

At our institution, we recommend primary surgery followed by risk adapted adjuvant therapy for NMSC patients. The purpose of this analysis is to investigate the outcome of patients treated with multi-modality therapy for NMSC with pathologically involved lymph nodes.

Patients and Methods

Patients who received multimodality treatment for NMSC from 2000 to 2012 at our institution were retrospectively identified. Inclusion criteria for this analysis were patients who had histologically confirmed cutaneous squamous cell carcinoma and clinical evidence of regional lymph node metastasis that were treated with a lymph node dissection and/or parotidectomy followed by adjuvant radiation. Patients were excluded if they had less than 6 months of follow-up or were treated with palliative intent. In patients with nodal disease and no obvious cutaneous lesion, a thorough evaluation was undertaken to attempt to identify a cutaneous or mucosal primary. All patients were staged using the American Joint Committee on Cancer (AJCC) 7th edition staging system.⁽¹⁴⁾ Patient demographics, relevant clinical history, tumor information, treatment, and follow-up were abstracted from the relevant medical records in accordance with an Institutional Review Board approved protocol and the Health Insurance Portability and Accountability Act (HIPAA). All toxicity was graded based on CTCAE v4.0.

The primary endpoints for analysis included locoregional control (LRC), progression free survival (PFS), and overall survival (OS). This is a difficult disease to study due to the heterogeneous population, varied treatment over time, and the often subjective clinical determination that progressive NMSC diagnosed after treatment of initial localized disease represents treatment failure or new primary tumor. The sites of relapse were categorized as follows: 1) local if it occurred in surgical bed 2) regional if tumor was detected in nearby lymph nodes or surrounding structures; and 3) distant if metastases were seen in a non-adjacent organ or lymph node. LRC was defined as lack of local or regional relapse. PFS was calculated from the date of last radiation treatment until the date of clinical or radiographic relapse. OS was calculated from the end date of radiation to the date of death or last follow-up. We explored associations between the outcomes of interest (LRC, PFS, and OS), and the following variables: evidence of nodal extracapsular extension (ECE), number of lymph nodes (LN) involved, largest pathologically (or clinically if patients were treated with induction therapy) involved lymph node, presence of a positive margin, and use of concurrent chemotherapy. We determined the association between LRC and these variables using Chi-squared tests in univariate analysis, and using logistic regression in multivariate analysis. PFS and OS were calculated using the Kaplan-Meier survival curves. Cox regression analysis was used to perform univariate and multivariate analysis of survival outcomes. A p-value of <0.05 was considered statistically significant.

Results

From 2000 to 2012, 108 patients with NMSC who received multimodality treatment at our facility were identified. Of this cohort, 46 were clinically node positive at presentation and had greater than 6 months of follow-up. Patient and tumor characteristics are listed in Table 1. As expected, the majority of tumors were located in the head and neck region (n = 40, 88%). All patients were managed with an operation. Patients presenting with parotid tumors and no clinical evidence of neck lymph node involvement were often managed without a neck dissection, since adjuvant radiation was anticipated.

Table 1:

Patient Characteristics

	RT only	CHEMORT	TOTAL	p-value
Patient Age	78 (62–96)	71 (39–78)	76 (39–96)	<0.001
Patient Gender
F	6 (18%)	2 (15%)	8 (17)	0.6
M	27 (82%)	11 (85%)	38 (83)
Immunosuppreasion
yes	3 (9%)	1 (8%)	4 (9%)	0.69
no	30 (91%)	12 (92%)	42 (91%)
Site
lip	4 (12%)	0 (0%)	4 (9%)	0.41
ear	3 (9%)	1 (8%)	4 (9%)
non lip/ear H&N	23 (70%)	9 (69%)	32 (70%)
trunk	2 (6%)	3 (23%)	5 (11%)
extremity	1 (3%)	0 (0%)	1 (2%)
Number of Lymph Nodes Dissected	18 (2–77)	28 (2–90)	19 (2–90)	0.07
Number of Positive Lymph Nodes	1 (0–8)	5 (1–84)	2 (0–84)	0.016
Perineural Invasion
yes	8 (24%)	7 (54%)	15 (33%)	0.08
no	25 (76%)	6 (46%)	31 (67%)
Positive margin
yes	6 (18%)	3 (23%)	9 (20%)	0.5
no	27 (82%)	10 (77%)	37 (80%)
Recurrence
yes	28 (85%)	12 (92%)	40 (87%)	0.49
no	5 (15%)	1 (8%)	6 (13%)
Extracapsular Extension
yes	17 (52%)	9 (69%)	26 (57%)	0.4
no	5 (15%)	0 (0%)	5 (11%)
unknown	11 (33%)	4 (31%)	15 (33%)
Largest lymph node size (cm)*	2 (0.5–6.5)	2.5 (0.5–5.5)	2 (0.5–6.5)	0.59
Differentiation
poorly or undifferentiated	14 (42%)	8 (62%)	22 (49%)	0.14
moderately or well	8 (24%)	0 (0%)	8 (17%)
unknown	11 (33%)	5 (38%)	16 (35%)

^*Measured lymph nodes only

A total of 13 patients (28%) received concurrent systemic therapy (cetuximab [4], cisplatin [7], carboplatin [1], and paclitaxel [1]) with adjuvant radiation (Table 1). Three of these patients were recommended to have concurrent chemoradiation (CRT) as a result of gross tumor progression in the surgical bed prior to commencing adjuvant radiation. Patients who received chemotherapy had higher risk disease, as demonstrated by a significantly greater number of positive lymph nodes (p=0.016) and were also significantly younger with a median age of 71 (range, 39–78) vs. 78 (range, 62–96) for those who did not (p<0.001). Other factors such as margin status, extracapsular extension, or recurrent disease were evenly distributed between the two groups (Table 1).

At a median follow-up of 1.8 years (range, 0.5 – 8.5), 15 (33%) patients developed recurrence. The location of the initial recurrence is located in Figure 1. Of the patients who failed regionally, only 1 patient failed in the contralateral neck. At 5 years, LRC was 76%, PFS was 65%, and OS was 49%. The PFS of patients who received RT alone was significantly better than those that received chemoradiation (76% v 38%, p = 0.014, Figure 2) although the OS difference was not significant (p = 0.25, Figure 3) with a median overall survival for the entire cohort of 4.2 years.

Figure 1:

Location of first documented failure

Adjuvant Chemoradiation

Adjuvant Radiation

Figure 2:

Progression Free Survival

Figure 3:

Overall Survival

On univariate analysis, type of adjuvant therapy (radiation alone vs. CRT; p = 0.006), mean largest lymph node size (p = 0.039) and the number of positive neck nodes (<3 vs. ≥3; p = 0.001) were all statistically significant predictors for LRC (Table 2). The type of adjuvant therapy (p = 0.048, OR 0.15 [95% CI 0.02–0.99]) and 3+ positive lymph nodes (p = 0.017, OR 0.07 [95% CI 0.01–0.62]) remained statistically significant on multivariate analysis (Table 2). Univariate analysis for PFS showed that type of adjuvant therapy (p = 0.021) and number of positive neck nodes (p = 0.003) were statistically significant. On multivariate analysis, number of positive lymph nodes remained significantly associated with PFS (p = 0.036, OR 3.05 [95%CI 1.36–6.89]). For OS, only number of positive neck nodes (p = 0.007) was statistically significant on univariate analysis and remained positive on multivariate analysis (p = 0.025, HR 2.70 [95% CI 1.13–6.43], Table 3).

Table 2:

Locoregional Control

	Locoregional control (35)	Locoregional failure (11)	Univariate p- value	Multivariate p- value	Odds ratio (95% CI)
Positive ECE	19 (54%)	7 (64%)	0.73	0.38	2.70 (0.29–24.90)
3+ LNs positive	9 (26%)	9 (82%)	0.001	0.017	0.07 (0.01–0.62)
Positive margin	8 (23%)	1 (8%)	0.3	0.22	5.03 (0.39–65.30)
Largest lymph node (cm) - mean (SD)	2.9 (1.7)	4.2 (2.1)	0.039	0.5	0.84 (0.59–1.41)
Chemotherapy	6 (17%)	7 (64%)	0.006	0.048	0.15 (0.02–0.99)

Table 3:

Progression Free Survival and Overall Survival

	Progression free survival				Overall survival
	Univariable		Adjusted		Univariable		Adjusted
	HR (95%CI)	p-val	HR (95%CI)	p-val	HR (95%CI)	p-val	HR (95%CI)	p-val
ECE	2.48(0.78–7.94)	0.13	1.37(0.39–4.87)	0.63	1.36(0.62–2.96)	0.44	0.98(0.40–2.37)	0.96
3+ positive LNs	5.33(1.76–16.20)	0.003	3.65(1.09–12.31)	0.036	3.05(1.36–6.89)	0.007	2.70(1.13–6.43)	0.025
Largest LN (per 1cm increase)	1.25(0.99–1.56)	0.059	1.08(0.81–1.43)	0.6	1.16(0.98–1.37)	0.087	1.09(0.89–1.31)	0.44
Positive margin	0.98(0.21–4.46)	0.98	0.91(0.19–4.26)	0.9	2.82(0.99–8.01)	0.052	2.90(0.98–8.60)	0.055
Chemotherapy	3.48(1.21–9.98)	0.021	2.23(0.73–6.83)	0.16	1.63(0.70–3.75)	0.26	1.15(0.49–2.75)	0.74

Chemotherapy was well tolerated. All patients who received cetuximab developed an acneiform rash and three out of four patients were sufficiently symptomatic to require treatment with minocycline or other systemic antibiotics. Five of the nine patients who received non-cetuximab chemotherapy experienced a dose reduction or received less than the planned dose due to neutropenia, tinnitus, decrease in creatinine clearance, severe dehydration, and/or failure to thrive, although these did not seem to interfere with RT delivery (median days to complete radiation was 44 and 42 days, in the CRT and radiation group respectively).

Discussion

In patients with locally advanced cutaneous NMSC, multiple retrospective analyses support multimodality therapy; survival seems to be increased in those treated with surgery plus radiotherapy as opposed to surgery alone.^{(10, 11, 15)} Reported 5-year disease specific survival rates for patients treated with adjuvant radiation therapy alone range from 34%−78%.^(7–11) Although death due to competing causes is not uncommon in this population⁽¹⁶⁾, advanced NMSC is still a significant cause of mortality for these patients. Better outcomes are needed for this aggressive disease.

Extrapolating from the results of prospective trials demonstrating a survival benefit of adjuvant CRT for mucosal HNSCC, as well as the evidence of the sensitivity of metastatic NMSC to cetuximab⁽¹⁷⁾, adjuvant CRT is selectively administered to select patients with lymph node positive cutaneous NMSC at our institution. In general, chemoradiation therapy was well tolerated. There were no grade 3 or higher toxicities (except for one case of asymptomatic grade 4 thrombocytopenia), nor was there evidence of increased treatment breaks. Thus, the significantly worse PFS of patients who received tri-modality therapy was not expected. It seems likely that the worse outcome of patients treated with CRT in our series was due to selection bias - patients recommended adjuvant chemoradiation were determined to be at the highest risk of recurrence by a multidisciplinary team. Several aspects of this selection bias merit comment. First, contrary to other series, we included patients with gross local recurrence after complete surgical resection and prior to the initiation of adjuvant therapy (median delay to commence adjuvant radiation 48 days). This is an underreported event that may occur more often than appreciated ^{(18, 19)} and is reflective of the challenges of managing poor prognosis tumors with primary surgery. Second, a majority of our patients were referred for recurrent (87%) tumors that had failed previous treatment, most commonly to the primary site. Finally, all of our patients were treated for lymph node positive disease, a relatively infrequent ⁽²⁰⁾ occurrence in cutaneous carcinoma.

In contrast to our analysis, a recent single institution retrospective analysis demonstrated improved recurrence free survival in patients with cutaneous squamous cell carcinoma of the head and neck treated with concurrent chemoradiation compared with adjuvant radiation alone. ⁽²¹⁾ In this study, patients treated with CRT trended to have worse baseline characteristics primarily at the primary site, with more poorly differentiated tumors, perineural invasion, and higher T-stage among patients treated with CRT. In addition, 15% of patients treated with CRT were N0. Our series had no T4 patients (those patients are typically managed with definitive chemoradiation at our institution) and our indications were based primarily on regional involvement which was supplemented by indications at the primary site. As a consequence, seven of the 8 (88%) patients in our CRT group with disease recurrences had distant metastasis as part of their recurrence; this is markedly higher than the 5% reported in the recent analysis ⁽²¹⁾ and the 7–13% previously seen in other retrospective studies of lymph node positive NMSC ^{(10, 11)} treated with surgery and radiation. Overall survival in our series was similar to that of the Tanvetyanon study (50 and 40.3 months respectively) likely due to competing risk factors for death in an elderly patient population.

Our results with tri-modality therapy demonstrate that treatment of high risk NMSC remains challenging. It is possible that there are patients with poor prognostic features who should be spared the toxicity of adjuvant radiation given their extremely high risk for distant disease. It is now our practice to perform post-operative imaging prior to initiation of adjuvant radiation. This is done in addition to all pre-operative imaging, in order to assess for distant metastases which may have manifested in the interval between scans at diagnosis and radiation treatment planning. Patients with evidence of early metastatic disease might therefore be spared the toxicity of additional locoregional treatment and proceed with systemic therapy only.

Patients in our series were recommended either targeted therapy (cetuximab) or cytotoxic chemotherapy (most commonly cisplatin). No randomized, prospective evidence is available at this time for what constitutes the ideal regimen for lymph node positive NMSC. Retrospective studies in treatment of mucosal HNSCC have shown similar late toxicity in patients treated with radiation and concurrent cetuximab vs concurrent cisplatin⁽²²⁾ and current prospective study of efficacy and toxicity of these two regimens are being evaluated in RTOG 1016. The addition of cetuximab to radiotherapy was shown to be well tolerated in the treatment of mucosal HNSCC with no significant differences in global health status/quality of life scores compared with radiation alone.⁽²³⁾ A recently reported phase I trial⁽²⁴⁾ of concurrent erlotinib demonstrated a similar PFS to this report, with a 2-year DFS/PFS of 60%. Other investigations have also studied novel potential targets, such as combined inhibition of epidermal growth factor receptor and insulin-like growth factor-1 receptor, which may warrant additional investigation in NMSC.^{(25, 26)}

This analysis is limited by its retrospective nature, small heterogeneous patient population, and extended time frame. Nonetheless, it demonstrates valuable information – lymph node positive NMSC is poorly controlled by conventional therapy (76% PFS of patients treated with surgery + RT alone) and those selected for intensification do even worse (38% PFS of patients treated with surgery + chemoradiation). Prudent, tolerable strategies for intensification are needed. In addition, this analysis supports the contention that indications for adjuvant therapy for high risk NMSC has risk factors similar to that of high risk mucosal HNSCC, including number of positive nodes, presence of ECE, and size of the largest involved node. With current population trends suggesting an increase of numbers of elderly patients in the future⁽²⁷⁾ and thus probably more patients with advanced NMSC, an effective regimen for these patients is needed.

Lymph node positive NMSC has a poor outcome despite aggressive therapy. The risk factors for treatment failure seem to be similar to that of mucosal head and neck squamous cell carcinoma. It is unknown whether the addition of systemic therapy to surgery and radiation for treatment of advanced NMSC is efficacious, but prospective study is warranted in view of the poor disease control currently achieved with the current standard of care.