Issue with loss of trailing zeros in formatted numbers during PDF parsing

### Bug
When parsing PDFs containing numbers with thousand separators (e.g., “16.000”), Docling incorrectly transforms these numbers by removing trailing zeros or altering the values (for example, “16.000” becomes “16”).
I have tried adjusting pipeline settings, such as disabling cell matching, but the issue persists. This behavior impacts the accuracy of extracted numerical data.

### Steps to reproduce
Load a PDF containing numbers formatted with dots as thousand separators (e.g., “16.000”).

Parse the PDF with Docling using default settings or with cell matching disabled.

Observe that the numbers are altered, (but not always) losing zeros and becoming incorrect values (e.g., “16.000” becomes “16”).

### Docling version
Docling version: 2.36.1
...

### Python version
Python 3.13.4


...