High-resolution image analysis was employed in the analysis of round (
discoid) erythrocytes from hematologically normal (AA) individuals, AA
individuals with nonspecific anemia, individuals with sickle cell tra
it (AS), individuals with SC disease (SC), and individuals with sickle
cell anemia (SS). The shape feature Form Factor (4 pi Area/Perimeter(
2)) was used to select round cells and to exclude sickle and other abn
ormal cells. Textural features extracted from round cells of SS and SC
patients were found to differ from those derived from cells of normal
andanemic AA individuals. Two textural features, Standard Deviation o
f Run Length Matrix Counts and Rotation Moment of the Cooccurrence Mat
rix, discriminated between patient mean values from AA samples and tho
se from SS samples. The ability of textural features to separate round
cells into classes based on genotype suggests that high resolution im
age analysis may be an effective tool in the study and monitoring of s
ickle cell disease. (C) 1994 Wiley-Liss, Inc.