Avtor/Urednik		Kayser, Klaus; Kayser, Gian; Hufnagl, Peter; Gabius, Hans-Joachim
Naslov		Plane-dependent (relational) morphometry in human lung carcinoma: a suitable technique for delineation of structure-associated stereological features
Tip		članek
Vir		Acta Stereol
Vol. in št.		Letnik 17, št. 1
Leto izdaje		1998
Obseg		str. 25-35
Jezik		eng
Abstrakt		Vascularization- and cell proliferation-associated features convey potentially valuable characteristics about human malignancies for the pathologists. Pathologists are obliged to develop suitable techniques and approaches to exploit the preserved information to its complete extent. Using stereological tools, commonly only one reference unit is chosen for the estimation of particle (volume) densities Vv, surface densities Av, or numericaldensities Nv (numbers of structures by volume). However, an image can be divided into multiple different compartments which might be used to serve as independent reference spaces. The technique described in this article is based upon the segmentation of certain "basic structures" and the use of their center of gravity as reference center for the establishment of a structure-associated reference plane (space). The surface of the segmented structure serves as inner reference boundary. The boundary will then be artificially expended with a defined perpendicular diameter as defined by the shortest inner perpendicular diameter (center of gravity to surface), and a (new) outer reference boundary is created. This computed space serves for reference space for particles, structures etc. located within the new reference plane. By expansion of the inner and outer reference boundary volume several stereological features can be assessed with respect to "distance to surfaces", features of or relations between the basic structures. The general principle and the derived mathematical equation will be explained. In addition, the results of first measurements performed on vascularization in human lung carcinomas are presented.
Deskriptorji		LUNG NEOPLASMS IMAGE PROCESSING, COMPUTER-ASSISTED