The image display is incorrect due to incorrect pixel processing

Hello, I encountered a DX image that did not display correctly. After analysis, it was found that bits allocated=16, bits Stored=14, and the first two bits of the 16 bit pixel were both 1.
   This resulted in the 0xF3F8 calculation of the original pixel data obtaining 62456 instead of 13304 (0x33F8). When I use 13304, it displays correctly. Is this a bug and can it be fixed?