sksurv.datasets.load_breast_cancer

sksurv.datasets.load_breast_cancer()[source]

Load and return the breast cancer dataset

The dataset has 198 samples and 80 features. The endpoint is the presence of distance metastases, which occurred for 51 patients (25.8%).

See 1, 2 for further description.

Returns

  • x (pandas.DataFrame) – The measurements for each patient.

  • y (structured array with 2 fields) – e.tdm: boolean indicating whether the endpoint has been reached or the event time is right censored.

    t.tdm: time to distant metastasis (days)

References

1

https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE7390

2

Desmedt, C., Piette, F., Loi et al.: “Strong Time Dependence of the 76-Gene Prognostic Signature for Node-Negative Breast Cancer Patients in the TRANSBIG Multicenter Independent Validation Series.” Clin. Cancer Res. 13(11), 3207–14 (2007)