Open hongdavid94 opened 1 week ago
For example, for the TCGA breast cancer dataset, this is public data, and the clinical data is downloaded from the TCGA website. The format of the BRCA_surv.csv file is shown in the table below.
HE_entity_submitter_id | case_id | HE_entity_id | HE_file_name | case_submitter_id | project_id | age_at_index | race | gender | vital_status | days_to_death | days_to_last_follow_up | ajcc_pathologic_stage |
---|---|---|---|---|---|---|---|---|---|---|---|---|
TCGA-AR-A1AN-01Z-00-DX1 | 43029932-feb0-4658-97fa-645c1e775009 | 38849d6e-fddf-43ed-8943-eca9790a45a4 | TCGA-AR-A1AN-01Z-00-DX1.1118F9FE-6DF2-4496-B102-D3A10D332EC0.svs | TCGA-AR-A1AN | TCGA-BRCA | 46 | white | female | Alive | -- | 2920 | Stage IIA |
TCGA-AR-A5QQ-01Z-00-DX1 | 3af31fcf-ad0c-4fd9-a8e3-10f9176b5e9d | 805c9a58-9bff-497a-a72d-96eb6f154d08 | TCGA-AR-A5QQ-01Z-00-DX1.4185D6D2-498E-4E71-8CB3-EDB4739C6229.svs | TCGA-AR-A5QQ | TCGA-BRCA | 68 | white | female | Dead | 322 | -- | Stage IIIA |
Hello,
I am going through the steps and I finally processed the H&E images (breast cancer) of my data into a graph object, following the preprocessing step. now I need to predict survival but the given scripts (e.g. survival_model_prediction/MCO_surv_cli_external_test.py) are difficult to replicate because the git doesn't provide an example (MCO_csv_path = '../super-patch_graph_construction/preprocessed_MCO/surv_csv/MCO_surv.csv') file so I am having trouble replicating the prediction process with my own data. Could you share relevant files (a dummy one if privacy is an issue here) or other scripts you used to predict survival of H&E slides?
Thanks,