Improve time efficiency of loop removal

[mjhyman]

The loop removal performed nominally on the datasets from the AD / CTE samples. However, the datasets from the CAA cases have much more loops. Small inefficiencies in the loop removal led to negligible increases in computation time with the AD / CTE samples, however it is resulting in non-negligible increases with the CAA cases.

I have been meeting with a member of the research computation team at BU, and they have been assisting me with this investigation. The purpose of this ticket is to track this improvement.