cSCC staging systems poorly determine metastasis risk

BY MADHU RAJARAMAN in Oncology on March 15th, 2018


Current staging systems for cutaneous squamous cell carcinoma (cSCC) poorly discerned between patients with and without metastases, according to researchers.

In a population-based case-control study of 6,721 cSCC patients, the American Joint Committee on Cancer, 7th edition (AJCC 7) staging system had the lowest rate of correctly classified cases (61.8%), followed by the AJCC 8 (68.2%), the Brigham and Women’s Hospital (BWH) system (72.3%), and the Breuninger system (76.2%). The Breuninger system performed best, with sensitivity of 77.3%, specificity of 75%, correctly classified tumor rate of 76.2%, and concordance index (C-index) of 0.81, reported Ingrid Roscher, MD, of the department of dermatology at Oslo University Hospital, and her coauthors. The report was published in JAMA Dermatology .

Investigators used data from the Cancer Registry of Norway to identify 6,721 patients with a first cSCC diagnosis between Jan. 1, 2000, and Dec. 31, 2004. Metastasis status was split into one of two categories: no metastases (local disease only) and metastasis (regional lymph node or distant metastasis).

Within 5 years follow-up, 112 patients had developed metastasis, and 112 patients without metastasis were selected at random as controls. Tumor tissue was collected for all 224 patients and checked by a pathologist. Tumors were classified under all four staging systems, and a chi-squared test was performed to compare patients with and without metastasis. Relative risk of metastasis was calculated via logistic regression analyses.

Sensitivity, specificity, and correctly identified cases were used to evaluate performance of the staging systems, and C-index was used to measure discriminatory ability, wrote Dr. Roscher and her colleagues.

AJCC 7 had a sensitivity of 85.6%, specificity of 33.3%, and C-index value of 0.59; compared with AJCC scores of 67.1%, 69.6%, and 0.70 for sensitivity, specificity, and C-index, respectively. BWH had a sensitivity of 68.9%, specificity of 76.5%, and C-index of 0.75.

Within the AJCC 7 system, 85.6% of patients with metastasis and 66.7% without metastasis fell into the T2 category, and no patients were grouped into T3 or T4 (P = .003). Under the BWH system, patients without metastasis fell mostly into the T1 category, while those with metastasis were about equally distributed among T1, T2a, and T2b (P less than .001). With the Breuninger system, more patients with metastasis than without metastasis fell into the high-risk categories for tumor diameter and depth (P less than .001). Lastly, under AJCC 8, 10% of all patients fell into the T2 category, while less than 20% of patients without metastasis and more than 50% of patients with metastasis fell into the T3 category (P less than .001).

Risk of metastasis for T2 patients was greater than for T1 patients under the AJCC 7 system (odds ratio = 2.96; 95% confidence interval, 1.43-6.15). Under BWH, OR for metastasis were 4.6 (95% CI, 2.23-9.49) for T2a patients and 21.31 (95% CI, 6.07-74.88) for T2b patients. Under the Breuninger system, tumor diameter greater than 2 cm (OR = 5.92; 95% CI, 2.18-16.07) and depth of invasion greater than 6 mm (OR = 9.00; 95% CI, 3.51-32.31) increased risk of metastasis. The AJCC 8 system showed increased metastasis risk for T2 (OR = 2.00; 95% CI, 0.62-6.44) and T3 (OR = 6.14; 95% CI, 0.41-1.09) patients, the authors reported.

The results suggest that the current staging systems “distinguished poorly to moderately between patients who developed metastases and those who did not,” Dr. Roscher and her coauthors wrote. Moreover, “the poorest results were found for the AJCC 7 system, which is most widely used,” they added.

“Our findings indicate a need for a more reliable, easy-to-perform, and clinically useful staging system than those presently available,” they concluded.

Oslo University Hospital and the Cancer Registry of Norway funded the study. No other disclosures were reported.

SOURCE: Roscher et al. JAMA Dermatol. 2018 March 7 doi: 10.1001/jamadermatol.2017.6428 .


Back to top