Drug-induced liver injury (DILI) is a major concern for public health, as well as for drug development in the pharmaceutical industry, since it can cause liver failure and lead to drug withdrawal from the market and black box warnings. Thus, it is important to identify biomarkers for early prediction to increase our understanding of mechanisms underlying DILI that will ultimately aid in the exploration of novel therapeutic strategies to prevent or manage DILI. DILI can be subdivided into 'intrinsic' and 'idiosyncratic' categories, although the validity of this classification remains controversial. Idiosyncratic DILI occurs in a minority of susceptible individuals with a prolonged latency, while intrinsic DILI results from drug-induced direct hepatotoxicity over the course of a few days. The rare occurrence of idiosyncratic DILI requires multicenter collaborative investigations and phenotype standardization. Recent progress in research on idiosyncratic DILI is based on key developments in 3 areas: (1) newly developed high-throughput genotyping across the whole genome allowing for the identification of genetic susceptibility markers, (2) new mechanistic concepts on the pathogenesis of DILI revealing a key role of drug-responsive T lymphocytes in the immunological response, and (3) broad multidisciplinary approaches using different platform "-omics" technologies that have identified novel biomarkers for the prediction of DILI. An association of a specific human leukocyte antigen (HLA) allele with DILI has been reported for several drugs. HLA-restricted T-cell immune responses have also been investigated using lymphocytes and T-cell clones isolated from patients. A microRNA, miR-122, has been discovered as a promising biomarker for the early prediction of DILI. In this review, we summarize recent advances in research on idiosyncratic DILI with an understanding of the key role of adaptive immune systems.