Multiline/splitted cells in needleman wunsch algorithm

[srg9000]

How are cases where pdf looses the order of reading because of cell splitting or content of cell overflowing to next row handled?
For example:

| ---------- | ---------- |
| Client's | Offer |
| Offer | from xyz |

Here, by parsing the pdf, words/characters will be picked up in row first order, does needleman wunch based approach consider vertical overflow of word or does top row and bottom row get converted to 2 cells as text will remain same ("Client's Offer" and "Offer from xyz"?