Konva's hit testing is as good as it gets
I mistakenly thought that a math-based mouse pointer hit testing approach might be faster than Konva's built-in approach. In the end I had to agree that Konva has it covered and there's no optimisation to be done. Here's the explanation...
Members only