Really? What about the use of computer vision for satellite imagery? Facial recognition?

The premise seems to be proven, just not extended to the medical domain yet.

I think you’re conflating detection with interpretation.

Radiologists like to distinguish between "perception" and "cognition", but the reality is that cognition is the same statistical process machines use, just without knowing the actual probability numbers.

Satellite image processing can certainly detect a hotspot then interpret it as either a small brushfire or a missile launch. Facial recognition detects my features then interprets who I am. It's all pattern matching just at different scales in different parameter spaces.