Title of article
Irregularity index: A new border irregularity measure for cutaneous melanocytic lesions
Author/Authors
Tim K. Lee، نويسنده , , David I. McLean، نويسنده , , M. Stella Atkins، نويسنده ,
Issue Information
روزنامه با شماره پیاپی سال 2003
Pages
18
From page
47
To page
64
Abstract
One of the important clinical features that differentiates benign melanocytic nevi from malignant melanomas is the irregularity of the lesion border. There are two types of border irregularities: texture irregularities, the small variations along the border, and structure irregularities, the global indentations and protrusions. Texture irregularities are subject to noise, whereas structure irregularities may suggest excessive cell growth or regression of a melanoma. We have designed a new algorithm for measuring the structure irregularities in the border. Our algorithm first locates all the local and global indentations and protrusions and organizes them in a hierarchical structure. Then an area-based index, called the irregularity index, is computed for each indentation and protrusion along the border. From the individual irregularity indices, two important new measures, the most significant irregularity index and the overall irregularity index are derived. These two new indices provide a measure of the degree of irregularity along the lesion border. A double-blinded test was performed to examine the effectiveness of these two new indices. Fourteen experienced dermatologists were asked to evaluate the borders of 40 pigmented lesions. The clinical evaluation result was then compared with the two new indices and other published shape measurements. The user study showed that both of the new indices vastly outperformed the other shape descriptors. Moreover, our algorithm captured the knowledge of expert dermatologists in analysing malignancy of a lesion based on its shape alone, indicating that the new measures may be useful for diagnosing melanomas.
Keywords
melanoma , Border irregularity , Shape descriptor , Scale space filtering
Journal title
Medical Image Analysis
Serial Year
2003
Journal title
Medical Image Analysis
Record number
449780
Link To Document