Covid-death mean-imputation (CoDMI) algorithm and survival analysis: First clinical application in rectal cancer

ABSTRACT

Background: We illustrate a clinical application of Covid-Death Mean-Imputation (CoDMI) algorithm in survival analysis. CoDMI algorithm is a new statistical tool that allows to adjust, through mean imputation based on the Kaplan-Meier model, Covid-19 death events in oncologic clinical trials, providing a complete sample of observations to which any statistical method in survival analysis can be applied. Methods: We analyzed a group of patients who received trimodal treatment - neoadjuvant chemoradiotherapy, followed by surgery and adjuvant chemotherapy - for primary locally advanced rectal cancer (LARC). Overall survival (OS) was calculated in months from the date of diagnosis to the first event, including date of last follow-up or death. Because Covid-19 death events potentially bias survival estimation, to eliminate skewed data due to Covid-19 death events the observed lifetime of Covid-19 cases was replaced by its corresponding expected lifetime in absence of the Covid-19 event using CoDMI algorithm. In a traditional Kaplan-Meier approach, patient died of Covid-19 (DoC) can be i) excluded to the cohort;ii) counted as censored (Cen);iii) considered as died of disease (DoD). CoDMI algorithm offers an additional, more satisfactory option iv) DoC events are mean-imputed as no-DoC cases at later follow-up times. With this approach, the observed lifetime of each DoC patient is considered as an “incomplete data” and is extended by an additional expected lifetime computed using the classical Kaplan-Meyer model. Results: In total 94 patient records were collected. At the time of the analysis, there were 16 DoD cases, 1 DoC patient and 77 Cen cases. The DoC patient died due to Covid-19 52 months after diagnosis. CoDMI algorithm computed the expected future lifetime (beyond the DoC time of occurrence) provided by the Kaplan-Meier estimator applied to the no-DoC observations as well as to the DoC data itself. Given the DoC event at 52 months, CoDMI algorithm (applied in its standard form DoC as virtual DoD) estimated that this patient would be died after 79.5 months of follow-up. Table summarizes the 2-year OS and the 5-year OS rates for the different treatment of DoC event. Since our sample contains only one DoC case, the effects on the estimates of the options considered differ very little. In this situation, however, one can better understand how CoDMI algorithm works. Conclusions: CoDMI algorithm leads to the “unbiased” (appropriately adjusted) OS probability in LARC patients with Covid-19 infection, compared with that provided by a naïve application of the Kaplan-Meier approach. This allows a proper interpretation of Covid-19 events in survival analysis. A user-friendly version of CoDMI is available at https//github.com/alef-innovation/codmi.