Survival issue (error?)

[jal...@gmail.com]

Hello,

In several datasets (e.g. Breast METABRIC data, see attached screenshot) the disease-free survival is longer than the overall survival. Biologically thinking about it, disease progression happens before (or at the time of) death, not after death. So disease-free survival for each patient should be shorter than (or equal to) overall survival, not longer. Practically speaking, the DFS curve is generally shifted to the left of the OS curve. Or am I missing something obvious?

Could it be that the column headings are switched for some datasets? Another example is the Lung Squamous Cell Carcinoma (TCGA, Provisional) data.

The only other thing that I can think of is whether the disease progression data is (rather disproportionately) missing for those who had a shorter survival, and the analysis disregards the missing data, causing the DFS curve to shift to the right.

Any explanation would be helpful.

Thanks!

[Hsiao-Wei Chen]

Hi,

     Thanks a lot for reporting this issue to us. After checking with the author of the METABRIC study, the DFS there is actually disease-specific survival (that's Cancer related deaths), not DFS. Therefore the values in OS months and DFS months columns are the same but the event indicators are different. We will correct the DFS headers by DSS months and DSS status soon, but the KM plot for DSS won't be there because currently our portal only supports KM plots for OS and DFS. 

    The author will provide us the actual DFS data in couples of weeks, so we will add it to the study then. Hope we answer your question. 

Thank you,

Annice 

--
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+unsubscribe@googlegroups.com.
To post to this group, send email to cbiop...@googlegroups.com.
Visit this group at https://groups.google.com/group/cbioportal.
For more options, visit https://groups.google.com/d/optout.

[huqing...@gmail.com]

Hello,

Would you please help me about this? I can not find disease free status and desease free month data in the clinical data for METABRIC dataset, can you tell me where can I find it?

Thank you.

Qingtao Hu

[JJ Gao]

Hi Qingtao,

Currently it's not available. We took it down because it was not accurate, but we are working with the METABRIC authors to get the DFS data.

-JJ

--

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 post to this group, send email to cbiop...@googlegroups.com.
Visit this group at https://groups.google.com/group/cbioportal.

To view this discussion on the web visit https://groups.google.com/d/msgid/cbioportal/a933cd95-735d-463d-9e9b-7b6e6b973b05%40googlegroups.com.

[Aidan Moriarty]

Hello,

I am noticing a similar error with the now reported RFS vs OS. Assuming their vital status data is correct, the issue is not due to disease specific vs total survival. I was working with downloaded data for my analysis and this pattern is replicated no matter what. Can you please explain?

Thank you very much for your help. 

Sincerely,
Aidan