Multivariate analysis of Monte Carlo generated images for diagnosis of dysplastic lesions

Jun Q. Lu; Yuanming Feng; Rosa E. Cuenca; Kai Li; Yalin Ti; Kenneth M. Jacobs; Shawn B. Jackson; Ron R. Allison; Claudio H. Sibata; Gordon H. Downie; Xin-Hua Hu

doi:10.1117/12.589649

1 April 2005 Multivariate analysis of Monte Carlo generated images for diagnosis of dysplastic lesions

Jun Q. Lu, Yuanming Feng, Rosa E. Cuenca, Kai Li, Yalin Ti, Kenneth M. Jacobs, Shawn B. Jackson, Ron R. Allison, Claudio H. Sibata, Gordon H. Downie, Xin-Hua Hu

Author Affiliations +

Proceedings Volume 5692, Advanced Biomedical and Clinical Diagnostic Systems III; (2005) https://doi.org/10.1117/12.589649
Event: SPIE BiOS, 2005, San Jose, CA, United States

Abstract

Early detection of malignant melanoma is critical to improve the survival rates of patients with this aggressive malignancy. We constructed an imaging system employing two liquid-crystal tunable filters to acquire in vivo spectral images of dysplastic lesions from patients at 31 wavelengths from 500 to 950nm. These reflectance images were analyzed in search of optical signatures for quantitative characterization of dysplastic nevi and malignant melanoma. A principal component analysis (PCA) algorithm was developed to examine the spectral imaging data in the component space and an index of spreading of clustering pixels (SCP) was defined to measure the degree of clustering in the distribution of image pixel scores in a component space. We found that SCP of differential polarimetric images correlate strongly with the degree of dysplasia for 4 lesions. However, many questions remain unanswered on the relations between PCA results and the spatial and spectral characteristics of the image data because of limited spectral image data from the patients. To fully improve our understanding on the multivariate analysis of spectral imaging data, we have developed a parallel Monte Carlo code to efficiently generate reflectance images from given distribution of optical parameters in a skin lesion phantom. With this tool, we have investigated numerically the dependence of score distribution and SCP in the component sub-spaces on lesion size and position. These numerical results provide a foundation for our future study to identify optical signature of dysplastic lesion and melanoma in the skin.

Citation Download Citation

Jun Q. Lu, Yuanming Feng, Rosa E. Cuenca, Kai Li, Yalin Ti, Kenneth M. Jacobs, Shawn B. Jackson, Ron R. Allison, Claudio H. Sibata, Gordon H. Downie, and Xin-Hua Hu "Multivariate analysis of Monte Carlo generated images for diagnosis of dysplastic lesions", Proc. SPIE 5692, Advanced Biomedical and Clinical Diagnostic Systems III, (1 April 2005); https://doi.org/10.1117/12.589649

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available