A tale of two pulse sequences: hard lessons learned in computing K^trans from dynamic contrast enhanced prostate MRI