Measurement of Tumor Thickness in Cutaneous Squamous Cell Carcinomas: Do the Different Methods Provide Better Prognostic Data?

Abstract

Cutaneous squamous cell carcinoma (cSCC) is the second most common cause of nonmelanoma skin cancers. Although it has a relatively low mortality rate, it may be locally destructive and potentially metastasize. Tumor thickness of the primary lesion is one important parameter associated with biologic behavior. Such measurement is currently performed in different ways depending on the anatomic location and subspecialty (eg, skin vs. head and neck vs. gynecologic pathology). Furthermore, the new The American Joint Committee on Cancer eighth edition has changed the previously recommended method of measurement of cSCC of head and neck from a modified Breslow thickness to measuring from the granular layer of adjacent, normal-appearing skin to the deepest invasive tumor cell. This study evaluated the clinical significance on patient outcome by measuring tumor thickness using 4 common, currently available methods (measurement from: A. uninvolved dermoepidermal junction; B. top of granular cell layer of the epidermis overlying the tumor, that is, similar to Breslow thickness; C. dermoepidermal junction with in situ cSCC; D. top of granular layer of uninvolved skin) in 85 specimens from nongenital areas of 78 patients with cSCC. Thirty-five percent of them were from the head and neck area. Measurements were performed in millimeters using the digital ruler of image analysis software (Olympus cellSens Standard) on whole-slide scanned images. Associations between recurrence-free survival (RFS) and each method were assessed. When thickness was considered as a continuous measure, there was no statistically significant association between any of the 4 measurement techniques and RFS. When using the currently recommended 6.0-mm cutoff, methods B and C were significantly associated with RFS. Similarly, when optimal cutoff values were selected, all 4 methods were significantly associated with RFS in univariable analysis. However, in a multivariable model that included the techniques and location of lesion, only method B, using the optimal cutoff value of 8.7 mm, was independently associated with RFS. In summary, in our series of cSCC, measurement of thickness using a Breslow method (method B) was significantly associated with RFS using the optimal cutoff and the currently recommended 6.0 mm in univariable analyses and the optimal cutoff in a multivariable assessment. Therefore, our data indicate that measurement of tumor thickness in a manner similar to Breslow thickness may be used to help predict recurrence in patients with cSCC.