Assessing the accuracy of computed tomography in detecting bony invasion and thickness of squamous cell carcinoma of the scalp

Abstract

Objectives The aim of this study was to ascertain the accuracy of computed tomography (CT) in assessing the presence of bony involvement and thickness of squamous cell carcinoma (SCC) of the scalp. Methods A single-centre retrospective chart review was carried out. Inclusion criteria were scalp SCC, CT between January 2008 and 2018, and the availability of a reference test. Reference tests were either histology, surgical notes or clinical notes. Tabular assessment of accuracy was performed and Student’s t-test, Mann–Whitney U test and Fisher exact test were used in univariable analysis. Accuracy of thickness measurement was calculated using the limits of agreement method, and linear regression was used to examine trend. Results Thirty-nine patients were included. Most patients were male (74.4%), white (97.4%), not immunosuppressed (66.6%) and had poorly differentiated tumours (33.3%). The most common tumour sites were the vertex (28.2%) and temporal region (23.1%). Sensitivity of CT in detecting presence or absence of bony invasion of scalp SCC was 76.9% (95% CI 46.2–94.9%) and specificity was 96.2% (95% CI 80.4-99.9%). Overall accuracy was 89.7% (95% CI 75.8–97.1%), positive predictive value was 90.1% (95% CI 58.7–99.8%) and negative predictive value was 89.3% (95% CI 71.8–97.7%). No significant differences were found comparing patients with an accurate or inaccurate CT scan. Thickness on CT was found to be consistent with histological thickness at the 95% confidence level. Conclusions CT is accurate at assessing the presence of bony involvement and thickness of scalp SCC. This study was limited somewhat by small sample size.