Irregularity index: A new border irregularity measure for cutaneous melanocytic lesions

被引:76
作者
Lee, TK [1 ]
McLean, DI
Atkins, MS
机构
[1] British Columbia Canc Agcy, Canc Control Res Program, Vancouver, BC V5Z 4E6, Canada
[2] Simon Fraser Univ, Sch Comp Sci, Burnaby, BC V5A 1S6, Canada
[3] Vancouver Gen Hosp, Div Dermatol, Vancouver, BC V5Z 4E8, Canada
[4] Ctr Hlth Sci, Div Dermatol, Vancouver, BC V5Z 4E8, Canada
[5] Univ British Columbia, Div Dermatol, Vancouver, BC V5Z 4E8, Canada
[6] British Columbia Canc Agcy, Div Dermatol, Vancouver, BC V5Z 4E6, Canada
关键词
melanoma; border irregularity; shape descriptor; scale space filtering;
D O I
10.1016/S1361-8415(02)00090-7
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
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. (C) 2002 Elsevier Science B.V. All rights reserved.
引用
收藏
页码:47 / 64
页数:18
相关论文
共 55 条
[1]  
[Anonymous], THESIS U BRIT COLUMB
[2]  
Armstrong B. K., 1996, CANC EPIDEMIOLOGY PR, V2nd, P1282
[3]  
ASADA H, 1983, IEEE T PATTERN ANAL, V8, P2
[4]   SOME INFORMATIONAL ASPECTS OF VISUAL PERCEPTION [J].
ATTNEAVE, F .
PSYCHOLOGICAL REVIEW, 1954, 61 (03) :183-193
[5]  
BALCH CM, 1989, CANCER PRINCIPLES PR, P1499
[6]   APPLICATION OF AN ARTIFICIAL NEURAL-NETWORK IN EPILUMINESCENCE MICROSCOPY PATTERN-ANALYSIS OF PIGMENTED SKIN-LESIONS - A PILOT-STUDY [J].
BINDER, M ;
STEINER, A ;
SCHWARZ, M ;
KNOLLMAYER, S ;
WOLFF, K ;
PEHAMBERGER, H .
BRITISH JOURNAL OF DERMATOLOGY, 1994, 130 (04) :460-465
[7]  
Bono A, 1999, CANCER, V85, P72, DOI 10.1002/(SICI)1097-0142(19990101)85:1<72::AID-CNCR10>3.0.CO
[8]  
2-Q
[9]   CHARACTERIZATION OF MAMMOGRAPHIC PARENCHYMAL PATTERN BY FRACTAL DIMENSION [J].
CALDWELL, CB ;
STAPLETON, SJ ;
HOLDSWORTH, DW ;
JONG, RA ;
WEISER, WJ ;
COOKE, G ;
YAFFE, MJ .
PHYSICS IN MEDICINE AND BIOLOGY, 1990, 35 (02) :235-247
[10]   A POSSIBLE NEW TOOL FOR CLINICAL-DIAGNOSIS OF MELANOMA - THE COMPUTER [J].
CASCINELLI, N ;
FERRARIO, M ;
TONELLI, T ;
LEO, E .
JOURNAL OF THE AMERICAN ACADEMY OF DERMATOLOGY, 1987, 16 (02) :361-367