Quantitative thresholds for the MS-SSIM metric and Dice score were chosen based on visual inspection. Example lesions that, are considered dissimilar (below the threshold, top row, and at the threshold, middle row) and similar (above the threshold, bottom row) are plotted. In both (a) and (b), The left column shows four axial slices through a generated lesion volume, while the right column shows the same for a real lesion.