In their article, Brosnahan et al1 suggest that patients with autosomal dominant polycystic kidney disease (ADPKD) exhibit different patterns of kidney function decline, but they do not propose a solution to predict individual disease progression. We designed a neural network2,3 to predict individual estimated glomerular filtration rate (eGFR) decline in patients with ADPKD. Neural network inputs were 5 consecutive eGFR values (in mL/min/1.73 m2) covering a 2-year period; neural network output was predicted eGFR at 5 years (pGFR).