#055 — Log-Rank (Mantel-Cox) Survival Test
Statistical Definition
Non-parametric hypothesis test comparing the Kaplan-Meier survival distributions of two or more right-censored patient cohorts.
Bioinformatics Use Case
Comparing overall survival (OS) or progression-free survival (PFS) between biomarker-high and biomarker-low cancer patients.
Input Variables
Time-to-event vector (t); binary event indicator (1=event, 0=censored); group label.
Application
The gold standard of clinical oncology: the Log-Rank test proves whether stratifying patients by your biomarker separates long-term survivors from high-risk progressors.
#056 — Cox Proportional Hazards Wald & Score Test
Statistical Definition
Semi-parametric regression model estimating hazard ratios (HR) for continuous or categorical covariates while adjusting for clinical confounders.
Bioinformatics Use Case
Multivariable survival modeling of a continuous gene expression risk score adjusted for patient age, sex, and tumor stage.
Input Variables
Time-to-event (t); event status (0/1); covariate matrix (X).
Application
Does your prognostic signature remain significant after adjusting for patient age and tumor stage? Cox Proportional Hazards delivers independent Hazard Ratios that clinicians trust.
#057 — Gehan-Breslow-Wilcoxon Survival Test
Statistical Definition
Weighted log-rank test that weights timepoints by the number of subjects at risk, giving higher sensitivity to early survival differences.
Bioinformatics Use Case
Immuno-oncology or acute toxicity trials where treatment arms separate early before survival curves converge later.
Input Variables
Time-to-event (t); event indicator (0/1); group label.
Application
In fast-acting therapies or acute toxicity studies, early survival separation matters most. Gehan-Breslow weights early events to capture rapid clinical benefit.
#058 — Tarone-Ware Survival Test
Statistical Definition
Weighted log-rank test using the square root of the risk-set size as weights, balancing early and late survival differences.
Bioinformatics Use Case
Comparing progression-free survival curves with moderate censoring across intermediate follow-up horizons.
Input Variables
Time-to-event (t); event indicator (0/1); group label.
Application
Tarone-Ware strikes the ideal mathematical balance between early-weighted Wilcoxon and late-weighted Log-Rank tests when clinical benefit unfolds mid-trial.
#059 — Peto-Peto Prentice Survival Test
Statistical Definition
Weighted survival test weighting timepoints by the pooled Kaplan-Meier survival estimate, robust against differential censoring patterns.
Bioinformatics Use Case
Comparing multi-center clinical trial cohorts where patient dropout/censoring rates differ between treatment arms.
Input Variables
Time-to-event (t); event indicator (0/1); group label.
Application
When control-arm patients drop out faster than treatment-arm patients, standard tests can distort. Peto-Peto weights by pooled survival to eliminate censoring bias.
#060 — Schoenfeld Residuals Proportional Hazards Test
Statistical Definition
Diagnostic hypothesis test checking whether Cox model residuals correlate with time, validating the proportional hazards assumption.
Bioinformatics Use Case
Verifying that a biomarker's hazard ratio remains constant across follow-up rather than crossing survival curves.
Input Variables
Fitted Cox PH model; survival times (t); event statuses (e).
Application
Regulators and peer reviewers routinely reject Cox models that violate proportional hazards. This diagnostic test certifies that your survival model holds over time.
#061 — Restricted Mean Survival Time (RMST) Difference Test
Statistical Definition
Compares the area under the Kaplan-Meier survival curve up to a clinically meaningful truncation time tau, valid even when hazards cross.
Bioinformatics Use Case
Immunotherapy trials with delayed clinical onset where survival curves separate only after 6 months of treatment.
Input Variables
Time-to-event (t); event status (0/1); group label; truncation horizon tau.
Application
When immunotherapy survival curves cross early, Hazard Ratios break down. RMST delivers an intuitive, model-free metric: exact months of life gained over the trial window.
#062 — Gray's Competing Risks Cumulative Incidence Test
Statistical Definition
Compares cumulative incidence functions (CIF) across groups in the presence of mutually exclusive competing events.
Bioinformatics Use Case
Evaluating cancer-specific relapse mortality while accounting for non-cancer treatment toxicity deaths as competing events.
Input Variables
Time-to-event (t); multi-state status (0=censored, 1=primary event, 2=competing event); group label.
Application
Treating non-relapse deaths as standard censored observations inflates relapse estimates. Gray's test isolates true disease-specific incidence in complex clinical tri