An interesting suggestion would be to analyze the ratio of plasma testosterone/cortisol. Search this article: Biochemical and Immunological Markers of Over-Training.
As mentioned above, Cadegiani’s work on the hormonal aspects of OTS is a great starting point. Bloods can help to paint a picture of the athlete’s training state but aren’t reliable (or sensitive enough) to confidently and proactively determine OTS in isolation. Combining blood measures with performance metrics and athlete feedback would help.