Kaplan Meier Curve After 4.7

Requires: Python and Lifelines and Plotly packages.

This component leverages the Python Lifelines package to generate Kaplan-Meier curves and output statistics in a simple dashboard. It is designed to plot the classes within a column (so multiple curves).

The following are output:
1. A statistics table based on the output of the pairwise_logrank_test function with p-value and -log2(p).
2. A survival table with probability percent breakdown and corresponding timeline.
3. A case processing summary with counts, event counts and percentages.
4. Selected logrank test output.
5. Kaplan-Meier curves that plot the variables from the selected column.

Read the Lifeline docs here: https://lifelines.readthedocs.io/en/latest/index.html

Options

Max End Time For Timeline?

If checked the entire plot is generated.

Variable

Select the variable to consider.

Select Survival Status Column

Select the column that indicates the survival or failure of your data.

Select Time Column

Select the column that represents the survival time or time to failure.

alpha-value

This impacts the CI of the KM plot. Adjust based on your domain and specific research problem. It is passed in the KM fitter.

Plot End Time (static plot only)

Adjusts the output of the plot to stop at the given time interval. This is different that only calculating survival to that point.

Plot Start Time (static plot only)

Plot Start Input

Plot Resolution (Enter 0 For Default)

This determines the granularity of the survival curve. It specifies the number of time points at which to compute the survival probabilities. Somewhat experimental, read it in the docs here: https://lifelines.readthedocs.io/en/latest/fitters/univariate/KaplanMeierFitter.html?highlight=timeline#lifelines.fitters.kaplan_meier_fitter.KaplanMeierFitter.timeline

p value (For Flemington-Harrington weighting)

"The Flemington-Harrington test allows the%%00010most flexibility in terms of the choice of%%00010weights because the user provides the values of p and q" %%00010- Survival Analysis A Self Learning Text p.75.%%00010%%00010More Weight For Earlier Survival Times: p=1, q=0%%00010More Weight For Later Survival Times: p=0, q=1%%00010Reduce To Log Rank Test: p=0, q=0

q value (For Flemington-Harrington weighting)

"The Flemington-Harrington test allows the%%00010most flexibility in terms of the choice of%%00010weights because the user provides the values of p and q" %%00010- Survival Analysis A Self Learning Text p.75.%%00010%%00010More Weight For Earlier Survival Times: p=1, q=0%%00010More Weight For Later Survival Times: p=0, q=1%%00010Reduce To Log Rank Test: p=0, q=0

Select Logrank Test (if comparing 2+ curves)

Select the logrank test you would like to use, if comparing 2 or more curves. See the documentation here: https://lifelines.readthedocs.io/en/latest/Examples.html?highlight=logrank#logrank-test

Select Desired Weighting

Select the weighting you would like to use. If no weighting is desired select None. See the documentation here: https://lifelines.readthedocs.io/en/latest/Examples.html?highlight=logrank#logrank-test.%%00010%%00010The Wilcoxon, Tarone-Ware and Peto tests apply more weight to earlier death times. The Peto test is more robust than the Wilcoxon or Tarone-Ware tests when many observations are censored. When p > q, the Fleming-Harrington applies more weight to earlier death times whilst when p < q, it is more sensitive to late differences (for p=q=0 it reduces to the unweighted logrank test). The choice of which test to perform should be made in advance and not retrospectively to avoid introducing bias.

Survival Curve x-label

This will become the x axis label of the KM plot.

Survival Curve y-label

This will become the y axis label of the KM plot.

Plot Title

This will become the title of the KM plot.

Nodes

• Table View (JavaScript)4 ×
• Column Selection Configuration3 ×
• Double Configuration3 ×
• Integer Configuration3 ×
• String Configuration3 ×
• Show all 24 nodes

Extensions

• FeatureKNIME Base nodes
• FeatureKNIME JavaScript Views
• FeatureKNIME Quick Forms

Links

• No links available