MSK-CHORD overall survival

13 views
Skip to first unread message

Andi L.

unread,
Mar 20, 2026, 1:49:38 PM (2 days ago) Mar 20
to cBioPortal for Cancer Genomics Discussion Group
Hi all,

I'm wondering how to interpret the OS_MONTHS column of data_clinical_patients.txt of MSK-CHORD. Let me quote from the comments on top of the file:

DESCRIPTION: Overall survival from date of first sequenced sample to date of death or last contact. If patient died before first sequencing, overall survival in months is set to '0'. If patient is alive and last contact date is before date of sequencing,overall survival in months is set to 'NA'.

So is it correct to interpret these numbers relative to day 0, the day of sequencing results as used by the other timeline files? 

Why I'm unsure: firstly, the cBioPortal docs define OS_MONTHS as "Overall survival in months since initial diagnosis". Further, the associated Nature publication is generally not very explicit about this but does mention OS from "time of diagnosis to time of death" with "left truncation at time of sequencing". Both of these seem contradictory to the file header and the general setup of MSK-CHORD with day 0 being the sequencing results. 

I would greatly appreciate a clarification so I can continue to use this rich resource.

Best,
Andi


Benjamin Gross

unread,
Mar 20, 2026, 2:14:22 PM (2 days ago) Mar 20
to Andi L., cBioPortal for Cancer Genomics Discussion Group
Hi Andi,

I can see the confusion. The definition of OS_MONTHS in the cBioPortal documentation is meant as an example when describing the file format of the clinical patient data file. Each study is free to choose what “time 0” is as its timeline anchor. The Cancer Genome Atlas, for example, uses the date of initial diagnosis as “time 0.” While certain plots in the MSK-CHORD publication may use “time of diagnosis” as the anchor, the curated study in cBioPortal uses the date of the first sequenced sample as the anchor.

I hope this clarifies any confusion.

Best,
-Benjamin

--
You received this message because you are subscribed to the Google Groups "cBioPortal for Cancer Genomics Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cbioportal+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cbioportal/55c9fd2c-9b74-4c5b-896c-bfb48390f461n%40googlegroups.com.

Reply all
Reply to author
Forward
0 new messages