Desai and coauthors report important observations considering the lacking prognostic value of COO in relapsed (refractory) DLBCL. The question is very relevant and the answer is clear. The paper suffers from the poor to very poor written English that I am not going to edit (please either use a professional editing company or let a native-speaker thoroughly rephrase it in a readable way using e.g. correct tenses and comma punctuation!).

Major open issues:
1. What was the physician's choice therapy (line 3, page 7)? How many patients have been adequately treated by R-CHOP (or at least mini-R-CHOP or DA-R-EPOCH or other rough R-CHOP-equivalent)? The authors should try their best to adjust results for given treatment (at least as far as possible).
2. Have BAP FISH probes been used to determine rearrangements? Please refer to a paper describing the method!
3. How did the authors handle DHL, since according to WH02008 and 2016 they represent either *I-DLBCL/BL with MYC and ... rearrangements" or "HG-BCL with MYC and ... rearrangement" and NOT DLBCL!? This must be clearly stated, and the DHL must be listed and analyzed as a separate entity and not as DLBCL!
4. The unexpected high proportion of GCB cases must be discussed; I assume that this is to a prat due to the fact that DHL have been lumped to DLBCL, yet - though I acknowledge that they mostly represent "GCB" with MCL and BCL2 rearrangement - they should be listed as a separate entity. Other potential biases should be sought for and discussed.
5. I miss p-values in the whole 2nd paragraph on page 9, lines 29-51.
6. I miss information on whether nodal and/or extranodal DLBCL have been studied; the effect of N/EN must be addressed.
7. The discussion could be shortened by 20% and become much more focused.

Minor:
Genes must always be given in italics upper case