An evaluation of high-stage cutaneous squamous cell carcinoma outcomes by sex

Dear Editor, Data from the World Health Organization indicate that global overall age‐adjusted cancer incidence and mortality rates are 1·2‐ and 1·5‐fold higher in men than in women, respectively.1 Nonmelanoma skin cancer is the most common cancer in white populations, yet the prognostic significance of sex remains unclear.2 A possible relationship may exist given that women with melanoma have a lower risk of mortality than men.3,4 Most metastases and deaths from nonmelanoma skin cancer occur in patients diagnosed with high‐stage cutaneous squamous cell carcinoma (HSCSCC), particularly Brigham and Women's Hospital (BWH) stage T2b (two to three risk factors) and T3 tumours (four risk factors or bone invasion). BWH risk factors include a tumour diameter ≥ 2 cm, perineural invasion diameter ≥ 0·1 mm, poorly differentiated histology and depth beyond fat.5,6 BWH stages T2b and T3 cutaneous squamous cell carcinoma therefore represent a clinically relevant subtype for the identification of sex differences in prognosis. The objective of this study was to examine the association between sex and poor outcomes in a cohort of patients diagnosed with HSCSCC. Patients diagnosed with BWH stages T2b and T3 cutaneous squamous cell carcinoma from 2000 to 2013 were analysed. The study cohort included both patients diagnosed at BWH from 2000 to 2009 (as previously reported),5 and those diagnosed from 2010 to 2013 (not previously published). The detailed methodology used for data collection has been previously described.5 In brief, all pathology reports with a diagnosis of cutaneous squamous cell carcinoma from 1 January 2000 to 31 December 2013 were analysed. Cases of noncutaneous, in situ and recurrent cutaneous squamous cell carcinoma were excluded. The patients' medical records were reviewed to determine whether outcomes of interest had occurred, including local recurrence, nodal metastasis, distant metastasis and death from cutaneous squamous cell carcinoma. Other data, such as the characteristics of the patients and of the primary tumour, were also collected. The HSCSCCs were also staged based on the American Joint Committee on Cancer (AJCC) tumour staging system.7 However, a tumour's AJCC T stage did not affect the inclusion of the case in this study's cohort as BWH T staging has been shown to define the subset of cutaneous squamous cell carcinoma cases with a significant risk of poor outcomes with greater accuracy.6 Fine and Gray competing risk analysis was used to determine univariate and multivariate associations of risk factors with the development of any poor outcome.8 Causes other than cutaneous squamous cell carcinoma were treated as a competing risk of death. The following factors were assessed via univariate analysis: age, AJCC T stage, BWH T stage, sex, immunosuppression status, tumour diameter, tumour depth, tumour differentiation, tumour location and perineural invasion. Multivariate models were built through forward stepwise variable additions followed by backward elimination. In this form of model building modelling begins with the variable with the largest effect estimate on univariate modelling. Other variables are added based on the next variable with the largest effect estimate and are retained in the model if the Wald test comparing the smaller model with the larger model is significant at P ≤ 0·05 or if the P‐value of the χ2 test comparing the two models is borderline (> 0·05 to 0·10) and if the addition of the variable changes the hazard ratios by at least 10%. Models were corrected for intrapatient correlations using the robust variance estimate as the analysis was case based, rather than patient based. Analyses were performed using Stata statistical software version 12·0 (StataCorp, College Station, TX, U.S.A.). The Partners Human Research Committee approved this study. A total of 100 patients (24% women) with 110 HSCSCCs (22% women) were included. There was no significant difference between the men and women in terms of their age, immunosuppression, tumour location and BWH T stage (T2b vs. T3) at primary tumour presentation. The risk of local recurrence and any poor outcome (nodal metastasis, distant metastasis or death from cutaneous squamous cell carcinoma) was significantly higher in men than in women (local recurrence: 27% vs. 4%, P = 0·02; any poor outcome: 41% vs. 13%, P = 0·01). Although the risk of nodal metastasis and death from cutaneous squamous cell carcinoma was threefold higher in men than in women, these trends did not reach statistical significance (nodal metastasis: 28% vs. 8%, P = 0·06; death from cutaneous squamous cell carcinoma: 20% vs. 8%, P = 0·24) (Table 1). In men, tumours on the scalp accounted for 17% of nodal metastasis, 30% of deaths from cutaneous squamous cell carcinomas and 29% of any poor outcome. Baseline characteristics of the study population AJCC, American Joint Committee on Cancer; BWH, Brigham and Women's Hospital; aT2b, 2–3 risk factors; T3, four risk factors or bone invasion; bincludes local recurrence, nodal metastasis or death from cutaneous squamous cell carcinoma. Baseline characteristics of the study population AJCC, American Joint Committee on Cancer; BWH, Brigham and Women's Hospital; aT2b, 2–3 risk factors; T3, four risk factors or bone invasion; bincludes local recurrence, nodal metastasis or death from cutaneous squamous cell carcinoma. Figure 1 depicts the cumulative incidence curves of any poor outcome stratified by sex. The 5‐year cumulative incidence of any poor outcome was 13% [95% confidence interval (CI) 3·1–30)] in women and 41% (95% CI: 31·1–52·2) in men (Gray's test: P = 0·01). Similarly, on multivariate analysis, independent predictors of disease recurrence (any poor outcome) were male sex [subhazard ratio (SHR): 3·9 (95% CI 1·2–12·5), P = 0·02] and BWH T stage [SHR: 2·5 (95% CI 1·1–5·3), P = 0·02]. In a separate multivariate analysis utilizing the AJCC staging system as opposed to the BWH staging system, male sex [SHR: 3·9 (95% CI 1·2–12·1), P = 0·02] and AJCC T stage [SHR: 3·7 (95% CI 1·8–7·6), P < 0·001] were independent predictors of any poor outcome. Cumulative incidence curves of any poor outcome stratified by sex. To our knowledge, this is the first study to examine sex differences in HSCSCC outcomes. Despite the low number of high‐stage tumours in women in this cohort, the data herein show that men have a 3·9‐fold higher risk of disease recurrence, corresponding to a nearly 30% higher 5‐year cumulative incidence. The reason for such a marked sex difference in HSCSCC outcomes is unknown. Studies have shown that biological responses to ultraviolet B damage differ between sexes.9,10 In mice exposed to equal doses of ultraviolet B, male mice developed tumours earlier and developed larger and more advanced tumours than female mice.10 Such studies may help explain why cutaneous squamous cell carcinoma, including aggressive high‐stage disease, is more common in men than women. Men may have higher ultraviolet exposure to the scalp due to baldness. However, the bulk of poor outcomes in men arose at other locations so male‐pattern baldness cannot explain the sex effect seen. A more pronounced response to ultraviolet as a carcinogen does not necessarily explain why men with high‐stage tumours with similar profiles to those of women are more prone to poor outcomes than women. Ultraviolet‐induced genetic damage is considered to be an early event in cutaneous squamous cell carcinoma carcinogenesis, with numerous other mutational events occurring afterwards to produce high‐stage disease.11,12 All tumours in this study were representative of such high‐stage disease, which has a broad mutational spectrum beyond that of ultraviolet‐induced changes. Sex differences in both ultraviolet and nonultraviolet‐induced mutational events merit further study. This study is limited by being a single‐institution cohort study. However, controlling for tumour stage via BWH and AJCC criteria and confining the study to high‐stage disease may help the results to be more broadly generalizable. The study is further limited by the relative rarity of HSCSCC overall, which is consistent with other large cohort studies that show most cutaneous squamous cell carcinomas are low stage. Nevertheless, the present study constitutes the only evidence available on the impact of sex on outcomes of high‐stage cutaneous squamous cell carcinoma. Larger studies are required to validate the findings herein and examine sex differences in potentially related cofactors, such as occupational exposure and cultural factors impacting upon ultraviolet exposure and access to care. Further research evaluating sex differences in cutaneous squamous cell carcinoma outcomes is necessary and may lead to the creation of more targeted prevention and treatment approaches. Conflicts of interest: none declared.