EMV is a statistical concept (with a defined formula) which calculates average outcome when there is uncertainty in the final outcome (more than one possible outcome), where as Decision Tree model is an enabling analysis tool to represent the available decision choices as decision tree with their monetary value and probability of occurrence. So I guess both needs to be used together in situations where we perform Quantitative Risk analysis to decide which option/decision to pursue.
One of the major benefits that I see is that it will aid in initial project selection phase when you have multiple projects on the list which has very high profits. Through EMV based Decision tree analysis, we would be able to select projects that gives highest EMV. Since it is based on expert judgment, it can be totally skewed without taking into consideration of any external factors. Another drawback is that it doesn’t take into consideration the risk tolerances of stakeholders which is a major decision factor when you decide on something. Also, if the number of alternatives increase, it would be difficult to follow the tree and validate.
I am not an expert user, but I think I understand its usage and would definitely use it if I have to do decision analysis of possible outcomes, provided, I have the monetary values and probability distribution values for each options.