The authors analyzed a new clinical staging system and its correlation with pathologic findings and patient survival. Patients were eligible for inclusion in this longitudinal retrospective cohort study if they had cutaneous squamous cell carcinoma on the head or neck, underwent surgery and had a minimum 3 year follow-up. The primary study variable was using a new clinical staging system. Secondary variables included the parotid as a predictor of metastatic spread to the lymphatic nodes in the neck and primary lesion histopathologic traits. The outcome variable was patient survival. Associations between variables were assessed using Fisher’s exact test, Mann–Whitney test, Kaplan–Meier method and Mantel log-rank test. p < 0.05 was considered significant. The sample comprised 103 patients. Regional metastatic disease was found in 24 patients. Histopathological analysis showed a higher frequency of neck metastatic disease if the parotid was positive for metastases ( p = 0.022). An extended staging system showed significant correlation between survival rate and substages ( p = 0.0105). Perineural invasion was a negative prognostic factor ( p = 0.0151). The results of this study suggest that combining curative parotidectomy and elective neck dissection could be beneficial in high risk patients. Both neck and parotid metastases should be included in the clinical and histological N classification.
Cutaneous squamous cell carcinoma (CSCC) comprises nearly 20% of all non-melanoma malignant skin tumours and represents a significant global health issue. Most lesions (80–90%) occur in sun-exposed regions of the head and neck. Although early stage cure rates are excellent, in some patients (5–10%) regional metastatic disease develops, lowering the survival rate by 50%. The rate of distant metastatic disease remains low, even in those patients with regional metastases. The inability to identify high risk lesions at the time of initial treatment is aggravated by a lack of consensus on the appropriate management of regional metastases. Adequate surgical excision of the primary CSCC and an immediate curative lymphadenectomy remain essential in achieving control over disease progression and patient survival. There is no consensus on the indications for an elective neck dissection and extent of surgery, while further benefit can be expected from sentinel lymph node biopsy, selective neck dissection in high risk patients and adjuvant radiotherapy. There are only limited data on the relationship between clinical staging of advanced regional disease and patient survival rates. The parotid gland is the first site for metastatic spread of head and neck CSCC, and parotid involvement poses a significant diagnostic and therapeutic challenge. It is therefore important to examine the relationship between patterns of metastatic spread in the parotid and the neck when planning surgical treatment.
Materials and methods
This retrospective longitudinal cohort study and its protocol were approved by the University of Zagreb School of Medicine Bioethics Board, adhering to the Helsinki Declaration of 1983. All patients with head and neck CSCC who were surgically treated from 1 September 1983 to 31 October 2007 were eligible for inclusion in a comprehensive oncological database (tum-2). Patients were eligible for inclusion if: they had a histopathologically confirmed CSCC on the head or neck; they underwent primary surgical treatment in the authors’ institution; the primary tumour measured >2 cm in its highest diameter; and they had a minimum follow-up of 3 years. The patients underwent extensive examinations and were treated according to standard clinical practices approved by a multidisciplinary oncological team. Clinical staging was carried out according to the current AJC staging system, but was revised using a new staging system (O’Brien’s PN classification) shown in Table 1 , which separates patients into groups according to the extent of parotid disease and presence or absence of clinical neck disease. The patients were included consecutively, and follow-up lasted for a minimum of 3 years, until 1 May 2010 or the patient’s death. Demographic information, prior history of skin malignancy, location, lesion size, previous therapy, surgical procedure and reconstruction method details, histopathologic findings (margin status, number of positive lymph nodes, number of examined lymph nodes, extracapsular spread and/or perineural invasion) and evidence of regional metastatic disease was collected. Histologic margins were defined as ‘clear’ if there was a definitive healthy tissue margin of >5 mm around the cancer tissue, while ‘near’ margins were defined with <5 mm of healthy tissue. Patients with a ‘near’ or ‘positive’ margin were evaluated by a multidisciplinary oncology team and underwent re-excision if necessary. The total number of surgically excised CSCC was well over a thousand patients, but due to strict study protocol criteria, the following patients were not eligible: patients with primary tumours <2 cm at the widest diameter (there were no patients in the oncological database with tumours of this size who developed regional metastatic disease); incomplete documentation; incomplete follow-up or day clinic patients; or localization of CSCC on the lip, which is known to have clinical and histological characteristics and a poorer prognosis than other head and neck CSCC. Patterns of metastatic involvement of cervical nodes were analyzed and correlated with primary lesion localizations. These primary lesions were divided into 5 anatomic sublocalizations: (1) frontonasal; (2) auriculotemporal; (3) mental, submental and submandibular; (4) orbitozygomatic; (5) buccal and preauricular region ( Fig. 1 ). Localizations of regional metastatic disease were divided into the following anatomic regions: the parotid gland and neck regions I–VI, according to the recommendations proposed by the American Head and Neck Society and the American Academy of Otolaryngology-Head and Neck Surgery.
|P0||No clinical disease in the parotid|
|P1||Metastatic node up to 3 cm in diameter|
|P2||Metastatic node >3 cm and up to 6 cm in diameter or multiple nodes|
|P3||Metastatic node >6 cm in diameter or disease involving the facial nerve or skull base|
|N0||No clinical disease|
|N1||Single ipsilateral neck node up to 3 cm in diameter|
|N2||Single node >3 cm in diameter or multiple nodes or|
Statistical analysis was performed using MedCalc software (Version 11.2.1 © 1993–2010. MedCalc Software bvba Software, Broekstraat 52, 9030 Mariakerke, Belgium), using standard descriptive statistics and frequency tabulation as indicated. The data for the n = 24 cohort were expressed as ratios due to n < 100. Associations between variables were assessed using Fisher’s exact test, the Mann–Whitney test and Kaplan–Meier method with Mantel log-rank test applied to endpoints to establish survival or recurrence probabilities. All tests of statistical significance were performed using a two-sided 5% type I error rate.
This study included 103 patients treated surgically from 1 September 1983 to 31 October 2007, while maintaining strict exclusion criteria. Of the patients eligible for inclusion, 56 were women and 47 were men. Average age was 74.5 years (69 years in men; 79 years in women) while ages ranged from 47 to 92 years. The highest incidence of head and neck CSCC was in the 80–98 year age group, followed by the 70–79 year age group. The average age of primary lesion occurrence was significantly greater in women ( p < 0.0012). Of the 103 patients, a cohort of 24 patients with clinically apparent regional metastatic disease and neck dissection was formed. Data from the total cohort population ( n = 103) shows that the most frequent localization of the primary lesion was in the buccal and preauricular sublocalization (56.3%), followed by the frontonasal sublocalization (29.12%). The incomplete excision rate was 6.7%, and there was a significant positive statistical correlation with higher T stage in the cohort with regional metastatic disease ( p = 0.0018). Average lesion size ranged from 2.1 to 13.5 cm, with a median of 4.3 cm ( n = 103). There were 26 recurrences of the disease both in the primary and metastatic disease localizations (25.2%), including 15 patients with primary localization recurrence, 5 with recurring regional neck disease, 1 with recurring parotid disease after parotidectomy, and 5 with new primary lesions in the follow-up interval. Perineural invasion was recorded in 5 patients, while 2 patients had vascular invasion, all in the patient cohort with regional metastatic disease ( n = 24). Three patients with extensive primary lesions (8.1, 11.3 and 13.5 cm in diameter) were not submitted to skull base procedures on review by a multidisciplinary oncological team. The regional metastatic disease rate was 23.3% (24/103) and the patients were singled out to form a separate cohort group. This group comprised 12 men and 12 women with the highest incidence of CSCC in the 70–79 year age group. The most common localization of primary lesions was the auriculotemporal area (9/24), followed by the buccal and preauricular area (6/25). There were 10 patients with clinical N1 stage and 8 with clinical metastatic disease in the parotid gland only. There were 6 patients with clinical N2 stage. All 24 patients underwent curative neck dissection, whereas 18 patients underwent total parotidectomy with preservation of the facial nerve, and 1 patient underwent total parotidectomy sacrificing the facial nerve, and 3 patients underwent superficial parotidectomy with preservation of the facial nerve. In the 2 remaining patients the parotid was not surgically treated.
There were 8 patients with clinical parotid disease and 16 patients with clinical neck disease, while there were no patients with a combined clinical finding. Histopathological analysis discovered 9 additional histopathologically positive metastases in the parotid gland that were clinically unrecognized. There was a statistically significant higher frequency of neck metastatic disease if the parotid gland is also positive for metastases ( p = 0.022). A positive histopathological finding in the parotid gland is related with statistical significance to a higher frequency of neck metastases in all 5 regions ( p = 0.044), and a negative finding is related to a significantly higher frequency of metastatic spread in regions II and III ( p = 0.033). In the cohort with regional metastatic disease, primary lesions localized in the temporoauricular, buccal and preauricular areas have a statistically significant higher rate of metastatic spread into neck regions IV and V than primary lesions localized in the frontonasal, orbital and zygomatic, mental and submental areas ( p = 0.0299).
The survival rate calculated by Kaplan–Meier’s method and using the standard AJC staging system is displayed in Fig. 2 . According to the most recent staging system, the N1 stage includes patients with a positive clinical finding in the parotid, neck or both. When an extended staging system was applied to the same group of patients, a statistically significant correlation between survival rate and substages of the PN classification is observed ( Fig. 3 ) ( χ 2 = 9.1072, df = 2, p = 0.0105). Cumulative disease-specific survival was highest in N1 stage, followed by P and N2 stages. Figure 4 shows a significant correlation between survival and P stage with cumulative disease-specific survival decreasing significantly with increasing P stage ( χ 2 = 12.9407, df = 3, p = 0.0048).
In Fig. 5 , a significant correlation between histopathological findings and cumulative disease-specific survival was identified, with highest survival in the group that had histopathologically verified parotid disease only, and significantly lower survival rates in stages N1 and N2 ( χ 2 = 12.7359, df = 3, p = 0.0052). Further data analysis yielded a significant correlation between survival rate in patients with isolated neck or parotid metastatic disease and those with metastatic disease in both the parotid and neck ( Fig. 6 ) ( χ 2 = 10.2545, df = 2, p = 0.0059).