Help in understanding metadata in TCGA skin cancer RangedSummarizedExperiment

Question

[Lexi]

Hey all,

I am currently trying to run an analysis in r using the skin dataset available in recount2 https://jhubiostatistics.shinyapps.io/recount/

While analysing the metadata of the samples, I found that information on stage of the patient (stages 1-4) as well as the type of tissue sample is available (e.g. from primary tumor, metastasis, healthy tissue).

However when I look into the staging + tissue sample I run into the following problem: Samples from patients in stages 0-2 are tagged as being from metastatic tissue, however as far as I know metastasis only occurs starting stage 3.

Here is my code in r:

rse_gene <- load(http://duffel.rail.bio/recount/v2/TCGA/rse_gene_skin.Rdata)
table(colData(rse_gene)[,"gdc_cases.samples.sample_type"], rse_gene$gdc_cases.diagnoses.tumor_stage)

                    stage 0 stage i stage ia stage ib stage ii stage iia stage iib stage iic

  Additional Metastatic       0       0        0        1        0         0         0         0
  Metastatic                  7      29       18       28       26        14        19        15
  Primary Tumor               0       1        0        1        4         4         9        49

                        stage iii stage iiia stage iiib stage iiic stage iv
  Additional Metastatic         0          0          0          0        0
  Metastatic                   39         15         35         55       21
  Primary Tumor                 2          1         12         12        3

Can anyone explain what is happening here? I would appreciate any input!