Semi‐quantitative methods yield greater inter‐ and intraobserver agreement than subjective methods for interpreting^99mtechnetium‐hydroxymethylene‐diphosphonate uptake in equine thoracic processi spinosi

Abstract

Scintigraphy is a standard diagnostic method for evaluating horses with back pain due to suspected thoracic processus spinosus pathology. Lesion detection is based on subjective or semi-quantitative assessments of increased uptake. This retrospective, analytical study is aimed to compare semi-quantitative and subjective methods in the evaluation of scintigraphic images of the processi spinosi in the equine thoracic spine. Scintigraphic images of 20 Warmblood horses, presented for assessment of orthopedic conditions between 2014 and 2016, were included in the study. Randomized, blinded image evaluation was performed by 11 veterinarians using subjective and semi-quantitative methods. Subjective grading was performed for the analysis of red-green-blue and grayscale scintigraphic images, which were presented in full-size or as masked images. For the semi-quantitative assessment, observers placed regions of interest over each processus spinosus. The uptake ratio of each processus spinosus in comparison to a reference region of interest was determined. Subsequently, a modified semi-quantitative calculation was developed whereby only the highest counts-per-pixel for a specified number of pixels was processed. Inter- and intraobserver agreement was calculated using intraclass correlation coefficients. Inter- and intraobserver intraclass correlation coefficients were 41.65% and 71.39%, respectively, for the subjective image assessment. Additionally, a correlation between intraobserver agreement, experience, and grayscale images was identified. The inter- and intraobserver agreement was significantly increased when using semi-quantitative analysis (97.35% and 98.36%, respectively) or the modified semi-quantitative calculation (98.61% and 98.82%, respectively). The proposed modified semi-quantitative technique showed a higher inter- and intraobserver agreement when compared to other methods, which makes it a useful tool for the analysis of scintigraphic images. The association of the findings from this study with clinical and radiological examinations requires further investigation.

Thoracic processi spinosi findings agree among subjective, semiquantitative, and modified semiquantitative scintigraphic image evaluation methods and partially agree with clinical findings in horses with and without thoracolumbar pain

“…The association between scintigraphic and radiographic evaluation methods of processi spinosi of the equine thoracic spine as well as the clinical examination of horses with and without thoracolumbar pain was investigated in this study. In accordance with the authors’ hypothesis, an agreement between the recently described modified semiquantitative scintigraphic image evaluation technique and subjective and semiquantitative scintigraphic evaluation methods was identified.…”

“…Scintigraphic images were analysed on a computer for scintigraphic image post‐processing (Dell Optiplex 3010 with Windows XP Professional Service Pack 3 system; Dell GmbH, Frankfurt am Main, Germany) using region of interest analysis (Scintron VII v145, Medical Imaging Electronics, Seth, Germany). The regions of interest were placed over the processi spinosi as previously described, the background region of interest was modified compared to the previous protocol and adapted in size and position (Figure ) . For this purpose, a rectangular region of interest was placed over the 15th thoracic vertebral body (region of interest 1).…”

“…The modified semiquantitative analysis includes the calculation of radiopharmaceutical uptake ratios by assessment of the highest counts per pixel per region of interest . Fifty interpolated pixels were selected per region of interest.…”

“…The two most commonly applied techniques for the interpretation of scintigraphic images include the subjective grading of findings based on color scales and the analysis of semiquantitative regions of interest . The inter‐ and intraobserver agreement in subjective scintigraphic image evaluation is higher when experienced veterinary surgeons assess scintigraphic images presented in grayscale rather than in color scale . A modified semiquantitative scintigraphic image analysis has recently been proposed and showed an increased inter‐ and intraobserver agreement when compared to subjective and semiquantitative scintigraphic image analysis .…”

“…The inter‐ and intraobserver agreement in subjective scintigraphic image evaluation is higher when experienced veterinary surgeons assess scintigraphic images presented in grayscale rather than in color scale . A modified semiquantitative scintigraphic image analysis has recently been proposed and showed an increased inter‐ and intraobserver agreement when compared to subjective and semiquantitative scintigraphic image analysis . Further investigation is warranted to compare the modified semiquantitative scintigraphic image evaluation technique with the established subjective and semiquantitative scintigraphic evaluation techniques …”