Masks loading from ome.tiff but a significant number of cells missing from class listed in CSV

[miantgta]

I have generated a segmentation mask as an OME.TIFF in 32-bit space with objectIDs represented by different gray values along with a CSV file with headers "CellID" and "State" and imported it into a story I am building with Minerva Author. As far as I can tell all of the (>400,000) objects that are represented in the OME.TIFF are present under the "all-cells" classification. However, when I look at the "State" with the largest number of objects a large number of the cells are missing.

I have queried the gray values from a sampling of the objects that are missing from their respective "State" in Minerva and found that they are properly annotated in the CSV file. Furthermore, objects of similar "CellID" in the CSV file (far down the list) with different "State" classifications are present in Minerva Author when I load their classes.

I've included two screenshots that demonstrate the masks are loaded in Minerva Author, but they are missing from their proper "State" classification--there is also a side-by-side with the same "State" classification showing the cells properly classified in QuPath.

In this screenshot I am showing the same "State" associated with the same "CellID" in Minerva (left) and QuPath(right). There are two cells missing from the "ring" in the middle-left of each FOV.

In this screenshot I am showing that the "all-cells" group in Minerva includes the cells that are missing from their respective "State" classification (left) which are also properly classified under the same "State" classification in QuPath (right). Note: the purple cell and the slightly darker green cell in the middle-left ring

I have also noticed that many of these masked cells disappear and reappear as I zoom in and out of different pyramidal levels in the image.

Please help!