Cp9 is text to speech possible on quiz results to read scores?

As RodWard​ said, you'd need to generate the audio files beforehand. If you limited each point value to be divisible by 5 or 10, you could greatly reduce the number of audio files needed and the resulting logic would be more manageable.

Text-to-Speech in Captivate is something you can do in Edit mode when developing a module for publishing.  It's not something you can set up to generate on the fly at run-time.

You could certainly generate some audio files beforehand and tie these into Conditional Advanced Actions so that according to the leaner's score a given file was played.  But that'a about all.