Line-field confocal optical coherence tomography of basal cell carcinoma: A retrospective study on diagnostic performance

Introduction: Line-field confocal optical coherence tomography (LC-OCT) represents one of the newest non-invasive in vivo skin imaging techniques. Previous studies described morphologic criteria of basal cell carcinoma (BCC) under LC-OCT examination. Preliminary data on LC-OCT diagnostic performance for BCC have recently been published but showed only a modest improvement compared to dermoscopy, possibly due to study limitations.

Objectives: The main objective of this study was to find diagnostic performance parameters of LC-OCT for differentiating BCC from its clinical/dermoscopic mimickers and for discriminating among BCC subtypes. An additional objective was to suggest a simple, user-friendly diagnostic algorithm based on the most powerful LC-OCT features in the field of BCC and its imitators.

Materials and methods: Equivocal BCC lesions imaged with an LC-OCT device prior to biopsy/excision were included. Three observers blinded for histopathological diagnosis retrospectively formulated clinical, dermoscopic and LC-OCT diagnoses and evaluated LC-OCT features for each study lesion.

Results: A total of 303 lesions (173 BCCs and 130 non-BCCs) from 225 patients were included. For the differentiation of BCC from BCC imitators, the use of LC-OCT increased the diagnostic accuracy compared to clinical examination by 24% and compared to dermoscopy by 12%. For the discrimination of sBCC from other BCC subtypes, LC-OCT increased the diagnostic accuracy compared to clinical examination by 18% and compared to dermoscopy by 12%. The presence of lobule with millefeuille pattern was a significant feature for BCC diagnosis. Lobule shape and location allowed BCC subtype discrimination.

Conclusions: The accuracy of BCC diagnosis can be increased by at least 12% with the use of LC-OCT compared to clinical and dermoscopic examinations alone, both in terms of BCC differentiation from clinical/dermoscopic imitators and in terms of BCC-subtype discrimination. A diagnostic algorithm based on significant features for BCC diagnosis is proposed, for which further validation is required.