This investigation sought to establish SC SPECT/CT image parameters that best correlate to clinical liver function by optimizing the automatic thresholding of SC SPECT and correlating relative ratios to composite and individual clinical parameters of liver function.
The images were converted to binary, allowing for automatic thresholding of hues and gray scales to black and white.
Golgi distribution was quantified in ImageJ by automatic thresholding of Golgi (either GM130 or TGN46) staining and measuring the % area covered for each cell.
Next, the thresholds of intensity of the HP1 α images were determined using the automatic threshold setting of the ImageJ program (between 23 and 39) and the number of HP1 α signals per nucleus (appropriately 100 nuclei in each sample) was counted.
Using the ImageJ version 1.41o software, the percentage of WGA-labelled tissue area of the total tissue area in each section of diaphragm or EDL muscle was determined in the captured images converted to binary images by applying the automatic threshold function of the software.
The accuracy, sensitivity (percentage correctly scored as sleep) and specificity (percentage correctly scored as awake) were calculated for actigraphy using high, medium, low and automatic threshold sensitivity settings of the actigraphy software.
In brief, after background subtraction and application of an automatic threshold, the plugin automatically processes the data, detects the different motility patterns and calculates the gliding speed.
This was not seen with automatic thresholding for TbN and BV/TV by in vivo μCT or ex vivo at 31 µm or with TbSp by in vivo or ex vivo μCT at any voxel size; this result demonstrates the importance of using fixed optimal thresholding instead of automatic thresholding.
In this setting, the application of the mixture model for automatic threshold calibration allowed the reduction of the falsely predicted functions down to 4percentt.
This observer dependence is not likely to be alleviated by adoption of one of the many existing automatic thresholding algorithms, many of which produce thresholded images that are equally, or even more, variable than those of the experts.
