Dear Jean,
You aren't missing anything; it's a typo. I'm guessing in a previous version of the problem the treatment threshold was 3%, in which case it should have said you are willing to do up to 32 CTPAs in patients without a PE to diagnose one PE.
In any case, assume the 5% threshold is right, in which case you are willing to do 20 CTPAs to diagnose one PE.
Tom