Positive predictive value of clinical diagnosis of head and neck non-melanoma skin malignancies. How accurate are we?

Abstract

ᅟ: The purpose of the current study was to determine the accuracy of the clinical diagnosis in non-melanoma skin malignancy as confirmed by histopathological examination of the specimen. Positive predictive value (PPV) was selected as a reliable measure of diagnostic accuracy.

Methods: Clinical diagnosis was made in the outpatient clinic by a consultant oral and maxillofacial surgeon. We reviewed the electronic charts of 210 patients with non-melanoma skin malignant tumours and calculated the positive predictive value of the initial clinical diagnosis. Histological confirmation for each lesion was compared with the provisional clinical diagnosis made in the clinic.

Results: Of the 147 lesions provisionally diagnosed as basal cell carcinomas, 133 lesions were histologically confirmed to be basal cell carcinomas (BCC) (PPV 90.4 %). Of the 63 lesions provisionally thought to be squamous cell carcinomas, 47 lesions were histologically confirmed as squamous cell carcinomas (SCC) (PPV 74.6 %). The difference between the PPVs for the two types of malignancy in our study was not significant (p = 0.39). Statistics between our results for BCC and SCC and those reported from two other cancer institutes revealed no significant difference (p = 0.58 and 0.07).

Conclusions: The present study supports that relying on clinical diagnosis with the purpose to formalise a treatment plan for head and neck non-melanoma skin cancer is safe and efficient. This is more reliable in cases of basal cell carcinoma in comparison to suspected squamous cell carcinomas. Although positive predictive value represents a reliable measure of diagnostic accuracy, it is increased when populations with higher prevalence of the disease are studied.

Keywords: Accuracy; Clinical diagnosis; Head and neck; Skin malignancy.