Premium
Bone Histomorphometry using interactive image analysis. A methodological study with application on the human temporomandibular joint
Author(s) -
Flygare Lennart,
Hosoki Hidehiko,
Rohlin Madeleine,
Petersson Arne
Publication year - 1997
Publication title -
european journal of oral sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.802
H-Index - 93
eISSN - 1600-0722
pISSN - 0909-8836
DOI - 10.1111/j.1600-0722.1997.tb00182.x
Subject(s) - temporomandibular joint , computer science , joint (building) , dentistry , bone histomorphometry , orthodontics , computer vision , medicine , pathology , osteoporosis , trabecular bone , engineering , architectural engineering
The aim was to develop a reproducible method for bone histomorphometry with the aid of a computerized image analysis program, and to examine the variation when assessing the total and the trabecular bone volume. Histologic sections of 18 temporomandibular joint autopsy specimens were read interactively using a cursor. The two parameters total bone volume and trabecular bone volume, of the condyle and the temporal component respectively, were estimated 2 × by 1 observer using 3 different threshold settings: an automatic, a semiautomatic and a manual technique. The threshold was based on the gray‐scale distribution of the image. 2 observers read the same sections with the aid of the semi‐automatic technique. The intra‐observer variation expressed as coefficient of variation ranged between 1.9% and 7.1% for the different parameters, when the automatic threshold setting technique was employed, and between 2.8% and 8.7% with the semi‐automatic technique. The manual technique resulted in a high intra‐observer variation with a coefficient of variation between 5.2% and 19.9%. There was a systematic difference between the estimates of the 2 observers. In general, intra‐ and inter‐observer variation was higher in the temporal component than in the condyle. The inter‐section variation was moderate, the coefficient of variation ranging from 3.8% to 11.1%. The automatic and semi‐automatic techniques resulted in comparable intra‐observer variation, with a lower bias in the estimates of the semi‐automatic technique. By letting one observer apply the semi‐automatic technique, it was possible to achieve fast and reproducible analysis of the total and trabecular bone volume.