This PR revisits two of the survival notebooks, immune and tp53_telomerase, which construct models using extent_of_tumor_resection. The models included unknown resections (denoted "Unavailable" or "Not reported" in the data), but these categories are wholly uninterpretable so they should not be included.
This PR updates the specific survival models and associated results where this predictor was included.
There is one particularly notable result change that I observed - the model with quantiseq cell type + fractions CD274 + extent of tumor resection now only converges for HGG, not MB.
This PR revisits two of the survival notebooks,
immuneandtp53_telomerase, which construct models usingextent_of_tumor_resection. The models included unknown resections (denoted "Unavailable" or "Not reported" in the data), but these categories are wholly uninterpretable so they should not be included.This PR updates the specific survival models and associated results where this predictor was included.
There is one particularly notable result change that I observed - the model with
quantiseq cell type + fractions CD274 + extent of tumor resectionnow only converges for HGG, not MB.