There isn't really any reason why you couldn't use the complete [Symfony 2 stack or cherry pick specific [url=http://symfony.com/doc/current/components/index.html]components](http://symfony.com/) like the router for a such a task. I guess it would partially depend on the business purpose of the API. Using Symfony/components any loss in performance you will make up in maintainability, support, dependability, scale, and general code quality.
I'm always skeptical of projects that compare themselves to others on the basis of performance. Judging performance is bias unless the two applications/code bases do the same same exact thing meeting all the business requirements of a specific project. Though performance comparisons always tends to be based on the simplest business requirements that would rarely if ever fulfill the needs of a real-world project in continuous development.
To that end I recommend maintainability, support, and code quality as the driving factors behind choosing a framework, cms, or any other open rather than performance. Unless the business requirements are very out of the ordinary the major performance bottlenecks will always be interacting with other services. Attempts to add micro-optimizations in other places tend to be a waste of time and offset by future changes. You are better off implementing caching strategies when problems arise than micro-optimizing especially when it comes to software that is in continuous development.
That is why I would recommend a framework like Symfony or to cherry pick components as opposed to using some random project by a single individual that will most likely be abandoned all to soon because they move onto the "next big thing". With large projects like Symfony there is a very slim chance the project will be dropped. There is a high chance certain versions will not be supported in distant future. However, death of versions naturally happens in the open source life cycle.