Optical coherence tomography in the assessment of cutaneous cancer margins of the face: An immediate ex vivo study.

Background: The assessment of cutaneous cancer margins intra-operatively or in the immediate postoperative phase can guide the operator into achieving clear margins. Achieving clear (tumour-free) margins following surgery is an essential factor that can reduce morbidity and disfigurement. The aim of present study was to determine the accuracy of optical coherence tomography in assessing cutaneous cancer margins of the face.

Materials And Methods: The excised tissue specimens that were examined, in this study, were acquired from 70 patients with 70 facial cancer lesions, with no nodal disease. Forty lesions were basal cell carcinomas (150 margins; 27 tumour positive) and the remaining thirty were cutaneous squamous cell carcinomas (112 margins; 22 tumour positive). These 70 resected lesions were subjected to optical coherence tomography (OCT) in the immediate ex vivo phase to assess each specimen's four margins status (anterior, posterior, medial and lateral). Two reviewers, blinded to the diagnosis, carried out the assessment of the acquired OCT images and measured the mean thickness. Intra- and inter-reviewer agreement was also calculated.

Results: On OCT, tumour-involved margins displayed sudden change in thickness associated with architectural changes. BCC-involved margins showed homogenous oval nests with dark rim and dark cysts or an empty space below dermo-epidermal junction depending on the sub-type. In the case of cutaneous SCC-involved margins, the DEJ had lost its integrity with/out the presence of small bright clusters in the papillary dermis and damage to the superficial epidermal layers. The mean thickness of the whole epidermal layer of the tumour-free margin was 128 μm, while for the BCC-involved margin 640 μm, and for the cutaneous SCC-involved margin 810 μm. The sensitivity of using OCT in examining BCC-involved margins was 88.9-92.6 % and the specificity was 96.8-98.4 %. For cutaneous SCC-involved margins, the sensitivity was 81.8-91.0 % and the specificity 85.6-91.1 %. There was "strong" inter-reviewer agreement on the BCC-involved margins, while the agreement was "moderate" for the cutaneous SCC-involved margins.

Conclusion: OCT provides good accuracy in identifying cutaneous cancer margins. This can potentially be used to guide and monitor resection in real-time. Tumour thickness could be measured due to the thin skin of the face, but may be more difficult to measure accurately in thick tumours and/or thick skin areas.