Comparison of clinical and histopathological evaluations of basal cell carcinoma thickness

DEAR EDITOR, Basal cell carcinoma (BCC) can be classified into a number of different clinical and histopathological types. The clinical diagnosis is not always consistent with the histopathological diagnosis because BCC does not have a definite clinical appearance in all cases. Accordingly, the extent of tumour infiltration may not be easily measurable as it can have subclinical extensions and deep irregular delineation against normal tissue. This may have implications for treatment, as tumour thickness is regarded as an important predictor of response to minimally invasive topical therapies. Therefore, the clinician needs reliable information on BCC thickness to identify those tumours most likely to respond. At present, biopsy to provide a sample for histopathological evaluation of the tumour thickness is considered the ‘gold standard’. Even so, BCC is treated in many cases without histopathological investigation of its thickness. We report on the comparison of corresponding clinical and histopathological measurements of BCC thickness to evaluate whether the clinical assessment of tumour thickness is accurate. In addition, we evaluated the agreement between these measurements in different BCC subtypes. This prospective study was approved by the regional committee for medical research ethics and performed at the outpatient clinic at the Department of Dermatology, St. Olavs University Hospital, Trondheim, Norway. Part of the study sample was included in a previous report that compared measurements of BCC thickness from punch biopsy and excision specimens. Consecutive patients ˃ 18 years old, neither pregnant nor breast feeding, with primary histopathologically verified BCC gave written informed consent before study entry. To ensure sufficient material for both punch and excision biopsy, only lesions ≥ 9 mm were included. Two consultants and an experienced registrar in dermatology performed the clinical examinations of 12, 17 and 24 tumours, respectively. Tumour size was defined clinically as the mean value of its extreme length and width. Tumour thickness was obtained by clinical and histopathological investigations. First, the clinical thickness (in mm) was estimated by inspection and palpation of the tumour. Second, a 3-mm disposable biopsy punch (Kai Industries Co. Ltd, Gifu, Japan) was used to obtain a small tissue sample. Third, and on the same day, an elliptical resection of the whole tumour was made. Details of the processing of tumour specimens have been described previously. A single hospital pathologist measured the BCC thickness from below the stratum corneum to the bottom of the tumour nest on haematoxylin, eosin and saffron-stained slides using an ocular micrometre to a precision of 0 1 mm. The greatest measurement from each punch biopsy specimen was defined as the punch tumour thickness. The greatest measurement from the corresponding punch biopsy and excision specimens of each tumour was defined as the ‘maximum tumour thickness’ (MTT), since the thickest part could reside within either specimen. The tumours were subclassified histopathologically into three categories: superficial, nodular or aggressive (morpheaform, infiltrative, basosquamous) growth types. IBM SPSS Statistics (v.21; IBM Corp., Armonk, NY, U.S.A.) was used, and the data were normally distributed. Possible relationships between the differences and mean thickness measurements were analysed by Bland–Altman plots with a regression approach for nonuniform differences. Simple regression analyses were used to calculate the slope of the regression line between the means and differences and the SD of the residuals (res-SD). The 95% limits of agreement were then calculated as 1 96 9 res-SD with respect to the regression line. P < 0 05 was considered statistically significant. In total, 159 tumour thickness measurements from 53 BCCs of 46 patients (18 women, mean age 71 years and 28 men, mean age 73 years) were included. Tumours were located to: head/neck (n = 20), trunk (n = 27) and extremities (n = 6) with a mean size of 18 mm (range 10–30). Histopathologically, 14were superficial, 24 nodular and 15 of aggressive growth types. The mean (SD) thickness values by the clinical, punch biopsy and MTT evaluations were 1 5 (0 9), 1 7 (1 3) and 2 0 mm (1 5), respectively. The largest mean difference (0 5 mm) was found between clinical and MTT values. Figure 1 shows the Bland–Altman plot with 95% limits of agreement for the difference between the MTT and the clinical thickness. The slope of the regression line was significant (slope 0 49, P < 0 001). The 95% limits of agreement were 1 3 mm (res-SD 0 67). Figure 2 shows the Bland–Altman plot between the punch biopsy thickness and the clinical thickness. The slope of the regression line was significant (slope 0 34, P = 0 004). The 95% limits of agreement were 1 6 mm (res-SD = 0 81). As the regression lines show, there was a systematic relationship between differences and tumour thicknesses. For tumours of approximately 1 6 mm thickness, the clinical and histological results agreed. Thinner BCCs were evaluated as thicker by clinical evaluation than by the corresponding histological examination, whereas the relationship was the opposite for thicker tumours. The scatter plots (Figs 1 and 2) also present the values of thickness for different BCC subtypes. By