Is using Transaction XID or LSN the more accurate and precise way to perform PITR?

I am practising Point in Time Recovery (PITR) on PostgreSQL 13 against accidental data deletion. Since PostgreSQL offers multiple parameters to perform PITR, I am bit confused choosing between Transaction XID and LSN. If an accidental DELETE statement removed a large number of rows, and I obtained the corresponding XID and starting LSN from the WAL dump, would performing PITR yield the same result regardless of whether I use the XID or LSN? Is there a risk of losing other committed transactions if I restore using the XID?