Predicting acute kidney injury in cancer patients using heterogeneous and irregular data.

How can we predict the occurrence of acute kidney injury (AKI) in cancer patients based on machine learning with serum creatinine data? Given irregular and heterogeneous clinical data, how can we make the most of it for accurate AKI prediction? AKI is a common and significant complication in cancer patients, and correlates with substantial morbidity and mortality. Since no effective treatment for AKI still exists, it is important to take timely preventive measures. While several approaches have been proposed for predicting AKI, their scope and applicability are limited as they either assume regular data measured over a short hospital stay, or do not fully utilize heterogeneous data. In this paper, we provide an AKI prediction model with a greater applicability, which relaxes the constraints of existing approaches, and fully utilizes irregular and heterogeneous data for learning the model. In a cohort of 21,022 cancer patients who were registered into Korea Central Cancer Registry (KCCR) in Seoul National University Hospital between January 1, 2004 and December 31, 2013, our method achieves 0.7892 precision, 0.7506 recall, and 0.7576 F-measure in predicting whether a patient will develop AKI during the next 14 days.

Active proteases in nephrotic plasma lead to a podocin-dependent phosphorylation of VASP in podocytes via protease activated receptor-1.

Focal segmental glomerulosclerosis (FSGS) is associated with glomerular podocyte injury. Podocytes undergo dramatic changes in their actin structure, with little mechanistic insight to date into the human disease. Post-transplantation recurrence of FSGS is the archetypal form of the disease caused by unknown circulating plasma 'factors'. There is increasing indication that plasma protease activity could be central to this disease. Using clinical plasma exchange material, collected from patients in relapse and remission stages of disease, the effects of FSGS plasma on human conditionally immortalized podocytes (ciPods) were studied. We show that vasodilator stimulated phosphoprotein (VASP) is phosphorylated in response to relapse plasma from ten consecutively tested patients, and not in response to paired remission plasma or non-FSGS controls. The phosphorylation signal is absent in human podocytes carrying a pathological podocin mutation. To test for a plasma ligand, inhibition of proteases in relapse plasma leads to the loss of VASP phosphorylation. By the use of siRNA technology, we show that proteases in the plasma signal predominantly via protease activated receptor-1 (PAR1) to VASP. Mechanistically, FSGS plasma increases podocyte motility, which is dependent on VASP phosphorylation. These data suggest a specific biomarker for disease activity, as well as revealing a novel and highly specific receptor-mediated signalling pathway to the actin cytoskeleton.

Attempted treatment of factor H deficiency by liver transplantation.

Complement factor H (FH) deficiency is one of the causes of atypical hemolytic uremic syndrome (HUS). Most patients with FH deficiency associated HUS progress to end-stage renal disease despite plasma therapy. Moreover, the disease invariably recurs in the graft kidney and causes graft failure. We confirmed FH deficiency in a 30-month-old boy with recurrent HUS of 2 years duration, and attempted an auxiliary partial orthotopic liver transplantation (APOLT) to overcome the sustained intractable dependency on plasma therapy. APOLT restored the plasma FH level, without HUS recurrence, for 7 months. However, thereafter he suffered from serious infectious complications associated with immunosuppression and finally died 11 months after APOLT. In conclusion, although APOLT showed clinical and laboratory improvement for some period in this patient, the final fatal outcome suggests that liver transplantation should be cautiously applied to patients with HUS associated with FH deficiency.