I am working on a project about predicting survival and recurrence from imaging-based biomarkers. In the project we analyze CT images of tumors and extracted texture parameters like tumor entropy. In the first exploratory part of the study we want to find cut point / threshold values for entropy to divide patients into good/bad prognosis groups. In the second part this cut points will then be validated on a validation cohort. I would like to use an outcome-oriented approach based on best p-value (minimum) from log rank-test from Kaplan-Meier analysis.

There is one stata specific presentation about the issue describing almost exactly what I want do however I am struggling putting the code together since the whole code is not shown in the presentation.
https://www.google.com/url?q=https:/...F6NuW0g5MVO6x8

The approach is, as I understand it, based on maxstat, an extension programmed for R:
http://r-addict.com/2016/11/21/Optim...ank-statistics

There is also an well described approach for SAS based on log rank:
https://support.sas.com/resources/pa...i28/261-28.pdf

My programming skills are limited and I was not able to translate this approach to Stata.

Is there a maxstat equivalent or other solution for Stata?

Maybe the “cutpt” command from: http://fmwww.bc.edu/RePEc/bocode/c/ may help however it is based on ROC.

My Stata version is 16.1.

Any help would be appreciated very much!
Thanks!