Squamous cell carcinoma of the skin (SCCS) is one of the most common cancers around the word 12, with an increasing incidence at a rapid rate among Caucasians in Europe, America and Australia 3456. It affects typically white skin populations and usually in body areas under sun exposure. In majority of cases the progression is indolent, and is easily cured by simple excision or radiotherapy.

Different from head and neck skin squamous cell carcinomas that present a higher incidence of lymph node addressed by several studies 789, trunk and extremities tumors usually do not present metastasis and the biological behavior is less aggressive 1011. Although it may be uncommon for a practicing dermatologist to observe a metastasis, locally advanced tumors, more frequent observed in referral hospitals, may present higher local recurrence and tissue invasion, regional lymphatic and even distant metastasis, consisting in a high risk group 1112. Most of those cases have a late progress to lymph node metastasis, with advanced local and regional disease. Nevertheless, at this point, disease control is difficult and survival is poor 10111213.

The frequency of lymph node metastasis is debatable and the reported rates vary considerable among investigators, from 0.1 to 28%, with a resulting mortality from 50–75%. Also, the inclusion of head and neck tumors in those series is common 1314. Few studies have addressed specifically the SCCS of the trunk and extremities 10111516. Some have looked at several prognostic factors, but most of them are heterogeneous or include also initial stage lesions and head and neck tumors 10111213141516.

The present study investigates a consecutive series of only localized advanced SCCS of the trunk and extremities, in order to define prognostic factors to lymph node metastasis and survival.

Patients with locally advanced SCCS (T3 and T4 classification – AJCC) 17were identified from the Hospital do Cancer de Barretos database – Barretos – Brazil. Patients with tumors from head and neck or genital origin were excluded. Patients with previous cancer diagnosis with exception to cutaneous basal cell carcinoma were also excluded due the difficulty to diagnosis the possible origin of metastasis or death of cancer. A total of 57 consecutive patients admitted and treated from October 1987 to November 2005 were evaluated. Pathologic diagnosis of SCC was confirmed in all cases. All of them were classified according to the 2002 American Joint Committee on Cancer staging system and have T3 (tumor >5 cm) or T4 (invasive of deep extra-dermal structures) classification. Institutional Review Board approval was obtained and all information retrospectively collected from medical records.

Several clinical variables were examined. These include age, gender, ethnicity, preceding chronic skin lesions(burns, scars, varicosae ulcers and others) at the site of the tumor, patient origin (rural or urban), anatomic site and treatment. Patterns of lymph node metastasis, recurrence and survival outcomes were also recorded.

We classified the patients related to lymph node metastasis as: N0 – patients with no evidence of lymph node metastasis at presentation; N0f – patients with no evidence of lymph node metastasis at any time; N1 – patients with lymph node metastasis at presentation; N1f – patients with lymph node metastasis during follow up.

We considered lymph node metastasis at presentation (N1) or during follow up (N1f) as endpoint. To analyze the association between clinical variables and lymph node metastasis the chi square, Fisher exact and T-test was used. Overall survival (OS) was also studied, and curves were constructed using the Kaplan-Meier method and compared using the log-rank test. All tests were two-sided, and a p-value of ≤ 0.05 was considered statistically significant.

We included 57 patients in this study. They were 34 men and 23 women; mean age was 61.9 years (range 30–91 years). 48 were Caucasians and 48 came from urban areas. Preceding chronic skin lesions at the site of SCC occurred in 15 cases. The tumors were sited at: the trunk, 14 patients; upper extremity (shoulder, arm and forearm), 7; lower extremity (thigh and leg) 16; hand 12 and foot 8. The T classification was T3 in 36 cases and T4 in 21 cases. 11 patients initially presented with lymph node metastasis (N1).

Surgery was performed in 46 patients; exclusive radiation therapy in 11. Eight patients received adjuvant radiation therapy. Four of them on the area of lymph node metastasis after lymphadenectomy, at discretion of the attending physician. Major concerns were coalescent tumor and more than 4 lymph node metastasis. Other 4 patients received neoadjuvant radiation on the site of primary tumor, with intention of local response and posterior limb sparing surgery, what was possible in one patient. Lymph node dissection was performed in 21 patients, 11 at the time of primary surgery and 10 during the follow up period. No elective lymph node dissection or sentinel node dissection was performed (Table 1). From 46 patients submitted to surgery, 38 had local control (82.6%) and 8 (17.4%) local recurrence. From 11 patients who received radiation therapy, 9 achieved local control. Recurrence occurred in 2 (18.2%). There was no difference between the treatment modalities (p = 1.0).

From all patients, 22 presented nodal disease. 11 of them presented lymph node metastasis at diagnosis (N1), another 11 cases presented lymph node disease during follow up (N1f) and the median time for it be diagnosed was 11.5 months (ranging from 1.6 to 33.8 months). The distribution of the lymph node metastasis in all cases is presented in Table 1. 21 of those patients underwent lymph node dissection, and 6 of them recurred at the lymph node basin.

Of 11 patients who presented lymph node metastasis during follow up, 7 (63.6%) were T3 and 4 (36.4%) were T4. Nine were submitted to lymphadenectomy, 01 was treated with radiotherapy and 01 refused treatment. The treatment of choice of the primary tumor was radiotherapy in 2 (18.2%). Five (45.5%) had surgery with local reconstruction and 4 (36.4%) amputation. All patients achieved local control of the primary tumor.

Considering the other recurrences, local ones occurred in 6 cases and 4 patients presented distant metastasis: one patient had lung metastasis only, 2 had lung and bone metastasis and the other one had disseminated metastatic disease (lung, bone, liver and skin). The untreated patient with lymph node metastasis had advanced local disease and refused any treatment.

At the time of last follow up, 19 patients were alive without evidence of disease, 17 died due to disease and 12 died from other causes. Nine patients were lost of follow up. Actuarial 5-years overall survival was 40.6%.

The only statistically significant clinical variable associated to increased number of lymph node metastasis was the presence of previous lesions on the skin. 9 of 22 patients with lymph node metastasis over the time presented previous lesions, versus 6 who did not (p = 0.047). Lesions located in upper extremity presented more lymph node metastasis, but did not reach statistical significance. 5 of 7 cases presented metastasis versus 17 of 50 from other locations (p = 0.095). Regarding the N0 patients, 11 presented lymph node metastasis during the follow up period (N1f) and again previous lesion was the only statistically significant factor: 5 of 11 patients with preceding chronic skin lesion at the site of SCC presented lymph node metastasis, versus 6 of 35 (p = 0.050). The mean time of presentation of lymph node metastasis in the 11 N1f group was 12.7 months (median, 11.5 months). Yet, the mean follow up of patients that never presented lymph node metastasis over the time (N0f) was 30.6 months (median, 22.7 months) (Table 2).

Comparative survival curves demonstrate differences in actuarial overall survival among patients with T3 and T4 classification (5 years overall survival 48.7% vs. 24.2%, p = 0.018) (Figure 1). Among N0 and N1, there was difference, but it was not statistically significant (43.3% vs. 34.1%, p = 0.070) (Figure 2). The treatment of choice of the primary tumor did not show differences in overall survival. Actuarial 5-years survival was 42.15% in the surgery group versus 40.91% (p = 0,659) in the radiation group. Overall survival estimative for N0 patients did not show difference regarding the presentation of lymph node metastasis over the time (N1f vs. N0f): (33.9% vs. 46.2%, p = 0.948) (Figure 3) (Table 3). Despite the risk for development of lymph node metastasis, comparative survival did not show difference between patients with previous skin lesions or not (37.7% vs. 42.2%, p = 0.507) (Figure 4).

Local advanced tumors are at risk of lymph node metastasis. Increased risk is associated to previous lesions at tumor site, and T4 classification tumors have worse prognosis. Lymph node recurrences in N0 patients, once treated, did not affect survival. For these patients, we propose close follow up and prompt treatment of lymph node metastasis. These results do not support indication for elective lymphadenectomy or sentinel node mapping. Further prospective studies must address this issue.

The authors declare that they have no competing interests.

VLV: Conception, design, analysis of data; TS: Acquisition and analysis of data; NMP: Acquisition and analysis of data; ALC: Interpretation of data, critical review, design. All authors read and approved the final manuscript.