Automated data sampling in sections from selected liver diseases
| Autor: | H. Simon, Wenzelides K, Kranz D, Voss K |
|---|---|
| Rok vydání: | 1980 |
| Předmět: |
Adult
Male Pathology medicine.medical_specialty Biopsy Biology Pathology and Forensic Medicine Automated data medicine Humans Sampling (medicine) Diagnosis Computer-Assisted Liver damage Alcoholic liver damage Feulgen stain Liver Diseases Alcoholic Electronic Data Processing Liver Diseases Liver cell Chronic persistent hepatitis Cell Biology Middle Aged Hepatitis B Liver Chronic Disease Female Chemical and Drug Induced Liver Injury Differential diagnosis Contraceptives Oral |
| Zdroj: | Pathology - Research and Practice. 170:388-401 |
| ISSN: | 0344-0338 |
| DOI: | 10.1016/s0344-0338(80)80043-4 |
| Popis: | Summary Automatic collection and evaluation of data obtained from histological sections will provide unprecedented information for morphological research and practice, which had been largely unvailable in the past because too much effort was involved. Possible applications and efficacy of this system for automatic analysis of microscopic images (AMPA) were tested in selected liver diseases from material received, and comparison was made with intact livers. Included in the pathological processes checked against controls were cases of chronic persistent hepatitis B and liver damage of various severity caused by contraceptives and alcohol. Morphometric and densitometric evaluations were based on sections 4 [,m in thickness which had been stained with Feulgen and Feulgen light-green. The results obtained from this evaluation were used as points of departure from which to differentiate the above liver diseases by classification programms using 30 parameters for each specimen. The rates of classification in training were 100 per cent for clinically intact livers and alcoholic liver damage and 99.3 per cent for chronic persistent hepatitis (Typ B) and hepatosis caused by contraceptives. Automatic assessment by the above system was made for 168 slides, with 90 per cent correctly identified and ten per cent attributed to diseases different from those originally diagnosed. Only the following parameters were relevant to the automatic system for dependable differentiation between diseases: ZGES (total number of objects), ZLEB (total number of liver cell nuclei), ZSTE (total number of Kupffer cell nuclei), VHSL (1000 x ZSTE/ZLEB), ZSIN (total number of sinusoidal sections), ZVAK (total number of vacuolar sections), and VGFL (vacuolar area in each visual field). These features differ to some extent from those commonly used in routine diagnosis, but they are likely to provide new approaches to more information form histological slide specimens. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|