<b><i>Introduction:</i></b> Prediction of neurodevelopmental outcome in infants with hypoxic-ischemic encephalopathy remains an important challenge. Various studies have shown that the predictive ability of different modalities changed after the introduction of therapeutic hypothermia. This paper reviews the diagnostic test accuracy of the different modalities that are being used to predict neurodevelopmental outcomes following therapeutic hypothermia. <b><i>Methods:</i></b> A systematic literature search was performed using Embase and PubMed. Two reviewers independently included eligible studies and extracted data. The quality of the studies was assessed using the Quality in Prognosis Studies Tool. Meta-analyses were performed where possible. <b><i>Results:</i></b> Forty-seven articles and 3 conference abstracts were included, reporting on 3,072<i></i>infants of whom 39% died or had an adverse neurodevelopmental outcome. A meta-analysis could be performed using 37 articles on (amplitude-integrated) electroencephalography (EEG), conventional magnetic resonance imaging (MRI), diffusion-weighted imaging (DWI), and proton magnetic resonance spectroscopy (<sup>1</sup>H-MRS). Amplitude-integrated EEG (aEEG) at 24 and 72 h showed similar high diagnostic OR, while aEEG at 6 h and EEG performed less, both due to a low specificity. For MRI, most studies reported scoring systems in which early (<8 days) MRI performed better than late (≥8 days) MRI. Injury to the posterior limb of the internal capsule on MRI or to the thalami on DWI were strong individual predictors, as was an increased lactate/N-acetylaspartate peak on <sup>1</sup>H-MRS. <b><i>Conclusions:</i></b> In the era of therapeutic hypothermia, the different modalities remain good predictors of neurodevelopmental outcome. However, timing should be taken into account. aEEG may initially be false positive and gets more reliable after 24 h. In contrast, MRI should be used during the first week, as its predictive value decreases afterwards.