Correlation and ROC analysis with repeated measures data

Hi all,
I have a hypothetical study where participants are taking drug X per standard of care. I have 3 different timepoints (baseline, therapy, post-therapy follow-ups) at which participants have biomarker sampling and echocardiograms performed.
Exploratory objective: efficacy; Exploratory endpoints: biomarkers (creatinine, potassium), echocardiogram parameters (ejection fraction, shortening fraction, degree of atrioventricular valve regurgitation)

My questions are:
1. For the exploratory analysis, if I have to use correlations and ROC analysis to explore relationships between these different endpoints (i.e. creatinine and ejection fraction, potassium and degree of atrioventricular valve regurgitation), as well as relationship between echo parameters and mortality at say 90 days follow-up (yes/no), how do I do that, given that it's a repeated measures design? Degree of atrioventricular valve regurgitation is a categorical variable with 4 levels (none/mild/moderate/severe), all other echo parameters and biomarkers are continuous.

2. How does the ROC bit fit into all of this? I’m a bit confused because I have no experience using ROC before (or with doing analysis on repeated measures data, for that matter) and from what I’ve briefly read, ROC analysis is usually used for gauging diagnostic accuracy so not sure how it would be used here?

If this wasn’t a repeated measures study, I know I could use Pearson or Spearman rank correlation for association between biomarkers and continuous echo parameters but given that we have measurements for these at different time points, what should I use to explore these correlations? Is there such a thing as a “Pearson or Spearman for repeated measures data”? I read something about a rmcorr package in R and mixed models that could be used for this but not sure if this is the best route? And what about the relationship between the continuous biomarkers and the categorical variable with 4 levels? How do I explore the correlation between a categorical with more than 2 levels and continuous variable, for repeated measures?

If anybody has any suggestions, I'd greatly appreciate it. Thank you!