This is a linkpost for https://entirelyuseless.wordpress.com/2017/03/04/the-practical-argument-for-free-will/
50% of the people have the lesion. That is a frequency. But if you pick a random person, that person either has the lesion or not. The probability, and not the frequency (which is not meaningful in the case of such an individual), that the random person has the lesion is 50%, because that is our expectation that the person has the lesion.
The parallel still holds. If you pick a random person with the lesion, he will either develop cancer or not. The probability that the random person with the lesion develops cancer is 90%. Is that not so?