In vivo drug screening assays are far superior than computational models that are employed to predict drug-induced toxicity since the former often miss subclinical hepatotoxicity that can be ascertained only from in vivo studies. Hepatotoxicity still accounts for 50% of acute liver failure owing to drug toxicity, and to rely exclusively on QSAR-based techniques or related systems preclude us from determining the actual toxicity of potential hepatotoxins/drugs.
Cell culture techniques, in vivo studies in rat, mice, and assays should be regularly complemented with computational toxicity studies to determine the actual levels of subclinical toxicity related to Type A and Type B ADRs. Conclusive evidence should be drawn only after bioassays, and not based on computational predictions. This would likely minimize the incidence of hepatotoxicity during drug therapy, reduce the cost associated with hospitalization, and the incidence of overall hospitalization due to ADRs.