Abstract
IMPORTANCE: Cutaneous squamous cell carcinoma (cSCC) has metastatic potential, but the prognostic value of current staging systems in nonselected patients is uncertain. OBJECTIVE: To assess the effect of risk factors for metastasis and to evaluate validity and usefulness of 4 staging systems for cSCC using population-based data. DESIGN, SETTING, AND PARTICIPANTS: This was a nationwide, population-based, nested case-control study. The Cancer Registry of Norway, which receives compulsory data on all cancers in the Norwegian population of approximately 5.2 million inhabitants. All patients diagnosed as having a primary invasive cSCC in the period 2000 to 2004 (n = 6721) were identified. Of these, 112 patients were diagnosed with metastasis within 5 years. As control patients, 112 patients with cSCC without metastases, matched for sex and age at diagnosis, were identified by random. Clinical data and biopsy specimens of primary cSCC were collected for all 224 patients. The biopsies were reexamined histologically by an experienced pathologist using well-established criteria for cSCC, yielding 103 patients with metastasis (cases) and 81 cSCC without metastasis (controls). MAIN OUTCOMES AND MEASURES: The ability of 4 cSCC staging systems (ie, the American Joint Committee on Cancer, 7th edition [AJCC 7] staging system, the staging system used by Breuninger et al, the Brigham and Women's Hospital [BWH] staging system, and the AJCC, 8th edition [AJCC 8] staging systems) to identify patients who developed metastasis with 5 years of follow-up. External validation was performed by logistic regression, discrimination (sensitivity, specificity, proportion correctly classified, concordance index), and calibration (R2, Hosmer-Lemeshow test, plots) statistics. RESULTS: Of 6721 patients; 3674 (54.7%) were men and 3047 (45.3%) were women, with a mean age at diagnosis of 78 years. Of the 103 patients with cSCC diagnosed with metastasis within 5 years, 60 [58.3%] were men, and mean [SD] age 72.7 [13.5] years. Of the 81 patients with cSCC without metastasis, 51 [63.0%] were men, and mean [SD] age 74.6 [11.7] 15 years. The staging systems distinguished poorly to moderately between patients who developed metastasis and those who did not. The ability to cluster patients with similar outcomes within the same staging category was low, particularly when using the AJCC 7 system. Using the AJCC 7 system, the risk of metastasis for T2 patients was 3-fold, compared with T1 patients (OR, 2.96; 95% CI, 1.43-6.15). In the system used by Breuninger et al, high-risk categories for diameter and tumor thickness and the BWH system's T2b category collected relatively homogeneous groups. In the systems used by Breuninger et al and Brigham-Women's Hospital, risk of metastasis was significantly elevated with increasing stage or risk category. Using the system by Breuninger et al, the risk of metastasis was less than 3-fold for tumors in the high-risk category of the combined variable (OR, 2.72; 95% CI, 1.29-5.74). The BWH system gave ORs for metastasis at 4.6 (95% CI, 2.23-9.49) and 21.31 (95% CI, 6.07-74.88) for the T2a and T2b categories, respectively. CONCLUSIONS AND RELEVANCE: Using population-based data, 4 current staging systems for cSCC were unsatisfactory in identifying nonselected cSCC patients at high risk for metastasis. The system used by Breuninger gave the best results.