Hi, I'm using dcm2niix to convert DICOM images but with some exams I get the "No valid DICOM files were found" error. 
You can find in attachment one of those images.
Can you explain why I have this kind of error?

These are not valid DICOM files. All DICOM elements have must have a even length. The tag 0008,0005 claims to contain 2573 bytes of data, which is not valid for DICOM (and this tag should only contain a few bytes, e.g. "ISO_IR 100". No valid DICOM tool can read these corrupted files. For example, try dcmdump:
  >dcmdump IM11
  W: DcmItem: Length of element (0008,0005) is odd
  W: DcmItem: Non-standard VR ' ' (03\01) encountered while parsing element (000f,0028), assuming 2 byte length field
  W: DcmItem: Length of element (000f,0028) is odd

The data clearly is not direct from the scanner, but has been manipulated by a PACS or anonymization tool.  I notice that the ImplementationVersionName (0002,0013) reports "AW4_7_14_081_HEL" - perhaps that is the source of the data corruption. If you trace the provenance of the image, you will identify the source of the error.

0002,0000 4@140 UL 210
0002,0001 2@156 OB
0002,0002 26@166 UI 1.2.840.10008.5.1.4.1.1.2
0002,0003 54@200 UI 1.2.840.113619.2.452.3.839543800.288.1588743345.84.109
0002,0010 20@262 UI 1.2.840.10008.1.2.1
0002,0012 20@290 UI 1.2.840.113619.6.350
0002,0013 16@318 SH AW4_7_14_081_HEL
0002,0016 12@342 AE NOCDicomSCP
0008,0000 4@362 UL 822
0008,0005 2573@374 CS <2573 bytes>

We get the DICOM due to a reconstruction console, can you suggest us a work around?

If the reconstruction console is a commercial device, the manufacturer has an obligation to support their DICOM conformance statement. These are not valid DICOM images, and are unacceptable for any usage. This is not a limitation of my software, but the tool that created the bogus tags.