Protective Effects of Riociguat Against Contrast-Induced Nephropathy: An Experimental and Machine Learning-Based...

Contrast-induced nephropathy (CIN) is an important cause of acute kidney injury following exposure to iodinated contrast media, and effective preventive strategies remain limited. This study investigated the renoprotective effects of riociguat, a soluble guanylate cyclase stimulator, in an experimental rat model of CIN and explored machine-learning-based prediction of renal injury using histopathological, biochemical, and inflammatory markers. Thirty-six female Wistar albino rats were randomized into control, riociguat, CIN model, and CIN + riociguat groups. CIN was induced by iohexol after dehydration, and riociguat was administered orally for 5 days. Renal injury was assessed by histopathological scoring, TUNEL assay, and biochemical parameters including serum creatinine, urea, tumor necrosis factor-alpha, nitric oxide, neutrophil gelatinase-associated lipocalin, and advanced oxidation protein products. Riociguat significantly decreased serum creatinine, urea, apoptotic index, and histopathological injury scores, reduced inflammatory and oxidative stress markers, and increased nitric oxide levels compared with untreated CIN animals (p < 0.05). Machine learning models (Random Forest, CatBoost, AdaBoost, and XGBoost) were applied for exploratory prediction and feature importance analysis. The apoptotic index and nitric oxide were identified as dominant predictors, indicating mechanistic relevance but limited clinical screening utility because these predictors require histological assessment. Overall, riociguat demonstrated significant renoprotective effects through anti-apoptotic, anti-inflammatory, and antioxidative mechanisms, and machine learning provided hypothesis-generating insight rather than a clinically deployable predictive model.