Melanoma and Nevi Subtype Histopathological Characterization with Optical Coherence Tomography

Abstract

Background: Melanoma incidence has continued to rise in the latest decades, and the forecast is not optimistic. Non-invasive diagnostic imaging techniques such as optical coherence tomography (OCT) are largely studied; however, there is still no agreement on its use for the diagnosis of melanoma. For dermatologists, the differentiation of non-invasive (junctional nevus, compound nevus, intradermal nevus, and melanoma in-situ) versus invasive (superficial spreading melanoma and nodular melanoma) lesions is the key issue in their daily routine.

Methods: This work performs a comparative analysis of OCT images using haematoxylin-eosin (HE) and anatomopathological features identified by a pathologist. Then, optical and textural properties are extracted from OCT images with the aim to identify subtle features that could potentially maximize the usefulness of the imaging technique in the identification of the lesion's potential invasiveness.

Results: Preliminary features reveal differences discriminating melanoma in-situ from superficial spreading melanoma and also between melanoma and nevus subtypes that pose a promising baseline for further research.

Conclusions: Answering the final goal of diagnosing non-invasive versus invasive lesions with OCT does not seem feasible in the short term, but the obtained results demonstrate a step forward to achieve this.

Keywords: CADx; HE; OCT; histopathology; melanoma; optical biopsy; optical properties; skin cancer; textural properties.

Figure A1

Complete HE slide of compound nevus Case 1 (20× magnification).

Figure A2

Complete HE slide of compound nevus Case 2 (20× magnification).

Figure A3

Complete HE slide of intradermal nevus Case 1 (20× magnification).

Figure A4

Complete HE slide of intradermal nevus Case 2 (10× magnification).

Figure A5

Complete HE slide of melanoma in-situ Case 1 (10× magnification).

Figure A6

Complete HE slide of melanoma in-situ Case 2 (10× magnification).

Figure A7

Complete HE slide of superficial spreading melanoma Case 1 (10× magnification).

Figure A8

Complete HE slide of superficial spreading melanoma Case 2 (20× magnification).

Figure 1

Example of dermatoscopy and OCT images from the equipment. White line in dermatoscopy image represents the scanning path (9 mm) of the OCT image (superficial spreading melanoma Case 2).

Figure 2

Clinical procedure followed in the hospitals for each database sample acquisition (superficial spreading melanoma Case 2).

Figure 3

Progression of skin lesions: from normal skin to non-invasive benign nevus and melanoma in-situ and finally invasive melanoma.

Figure 4

Automatic OCT lesion delimitation process: (A) zoomed image of lesion with lateral delimitation based on the colour of the dermatoscopic image; (B) delimitation transfer to OCT image with dotted red lines indicating delimitation; (C,D) original dermatoscopic and OCT images presented in Figure 1 (superficial spreading melanoma Case 2).

Figure 5

ROI extraction process: red bars indicate the automatic lesion delimitation as illustrated in Figure 4; purple bars delimitate the ROI at the centre of the lesion that is ROI considered for further properties extraction; green bar delimitates the healthy adjacent ROI starting on the left of the image (alternatively on the right) (superficial spreading melanoma Case 2).

Figure 6

Depth analysis comparison for optimum size selection (superficial spreading melanoma Case 2).

Figure 7

Comparison of quartile µ_t values for selected cases’ lesion tissues versus healthy adjacent tissues in the lesion centre. Y axis illustrates the µ_t value and its corresponding IQR interquartile ratio, where the middle line represents the median. X axis orders the different case studies grouped by colour as per diagnosis.

Figure 8

Contrast textural feature comparisons for case studies’ lesions with respect to adjacent healthy tissue (left: 0–25 pixels in depth; centre: 25–50 pixels in depth; right: 50–75 pixels in depth).

Figure 9

Dissimilarity textural feature comparisons for case studies’ lesions with respect to adjacent healthy tissue (left: 0–25 depth; centre: 25–50 depth; right: 50–75 depth).

Figure 10

Energy textural feature comparisons for case studies’ lesions with respect to adjacent healthy tissue (left: 0–25 pixels in depth; centre: 25–50 pixels in depth; right: 50–75 pixels in depth).

Figure 11

Homogeneity textural feature comparisons for case studies’ lesions with respect to adjacent healthy tissue (left: 0–25 pixels in depth; centre: 25–50 pixels in depth; right: 50–75 pixels in depth).

Figure 12

Correlation textural feature comparisons for case studies’ lesions with respect to adjacent healthy tissue (left: 0–25 pixels in depth; centre: 25–50 pixels in depth; right: 50–75 pixels in depth).