Validation of IR-spectroscopic brain tumor classification

C. Beleites; G. Steiner; S. Sobottka; G. Schackert; R. Salzer

doi:10.1117/12.647496

22 February 2006 Validation of IR-spectroscopic brain tumor classification

C. Beleites, G. Steiner, S. Sobottka, G. Schackert, R. Salzer

Proceedings Volume 6078, Photonic Therapeutics and Diagnostics II; 60782V (2006) https://doi.org/10.1117/12.647496
Event: SPIE BiOS, 2006, San Jose, California, United States

Abstract

As a molecular probe of tissue composition, infrared spectroscopic imaging serves as an adjunct to histopathology in detecting and diagnosing disease. In the past it was demonstrated that the IR spectra of brain tumors can be discriminated from one another according to their grade of malignancy. Although classification success rates up to 93% were observed one problem consists in the variation of the models depending on the number of samples used for the development of the classification model. In order to open the path for clinical trials the classification has to be validated. A series of classification models were built using a k-fold cross validation scheme and the classification predictions from the various models were combined to provide an aggregated prediction. The validation highlights instabilities in the models, error rates, sensitivity as well as specificity of the classification and allows the determination of confidence intervals. Better classification models could be achieved by an aggregated prediction. The validation shows that brain tumors can be classified by infrared spectroscopy and the grade of malignancy corresponds reasonably to the histopathological assignment.

Citation Download Citation

C. Beleites, G. Steiner, S. Sobottka, G. Schackert, and R. Salzer "Validation of IR-spectroscopic brain tumor classification", Proc. SPIE 6078, Photonic Therapeutics and Diagnostics II, 60782V (22 February 2006); https://doi.org/10.1117/12.647496

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