Open AccessAssessing agreement between preclinical magnetic resonance imaging and histology: An evaluation of their image qualities and quantitative results
Author(s) -
Cindy Elschner,
Philippe Korn,
Maria Hauptstock,
Matthias C. Schulz,
Ursula Range,
Diana Jünger,
Ulrich Scheler
Publication year - 2017
Publication title -
plos one
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.99
H-Index - 332
ISSN - 1932-6203
DOI - 10.1371/journal.pone.0179249
Subject(s) - concordance , magnetic resonance imaging , biomedical engineering , modalities , histology , biomaterial , medicine , computer science , nuclear medicine , medical physics , radiology , materials science , pathology , social science , sociology
One consequence of demographic change is the increasing demand for biocompatible materials for use in implants and prostheses. This is accompanied by a growing number of experimental animals because the interactions between new biomaterials and its host tissue have to be investigated. To evaluate novel materials and engineered tissues the use of non-destructive imaging modalities have been identified as a strategic priority. This provides the opportunity for studying interactions repeatedly with individual animals, along with the advantages of reduced biological variability and decreased number of laboratory animals. However, histological techniques are still the golden standard in preclinical biomaterial research. The present article demonstrates a detailed method comparison between histology and magnetic resonance imaging. This includes the presentation of their image qualities as well as the detailed statistical analysis for assessing agreement between quantitative measures. Exemplarily, the bony ingrowth of tissue engineered bone substitutes for treatment of a cleft-like maxillary bone defect has been evaluated. By using a graphical concordance analysis the mean difference between MRI results and histomorphometrical measures has been examined. The analysis revealed a slightly but significant bias in the case of the bone volume( b i a s H i s t o − M R I : B o n e v o l u m e = 2.40 % , p < 0.005 )and a clearly significant deviation for the remaining defect width( b i a s H i s t o − M R I : D e f e c t w i d t h = − 6.73 % , p ≪ 0.005 ) .But the study although showed a considerable effect of the analyzed section position to the quantitative result. It could be proven, that the bias of the data sets was less originated due to the imaging modalities, but mainly on the evaluation of different slice positions. The article demonstrated that method comparisons not always need the use of an independent animal study, additionally.