Abstract
Vessels supplying malignant tumors are abnormally shaped. This paper describes a blinded study that assessed tumor malignancy by analyzing vessel shape within MR images of 21 brain tumors prior to surgical resection. The program's assessment of malignancy was then compared to the final histological diagnosis. All tumors were classified correctly as benign or malignant. Of importance, malignancy-associated vessel abnormalities extend outside apparent tumor margins, thus allowing classification of even small or hemorrhagic tumors.
Original language | English (US) |
---|---|
Pages (from-to) | 645-653 |
Number of pages | 9 |
Journal | Lecture Notes in Computer Science |
Volume | 3217 |
Issue number | 1 PART 2 |
DOIs | |
State | Published - 2004 |
Event | Medical Image Computing and Computer-Assisted Intervention, MICCAI 2004 - 7th International Conference, Proceedings - Saint-Malo, France Duration: Sep 26 2004 → Sep 29 2004 |
ASJC Scopus subject areas
- Theoretical Computer Science
- General Computer Science