<b><i>Introduction:</i></b> The aim of this study was to evaluate prediction of neonatal mortality in fetuses with isolated left congenital diaphragmatic hernia (CDH) when the observed/expected lung-to-head ratio (O/E LHR) was estimated at two different gestational time points during pregnancy. <b><i>Methods:</i></b> Forty-four (44) fetuses with isolated left CDH were included. O/E LHR was estimated at the time of referral (first scan) and before delivery (last scan). The main outcome was neonatal death due to respiratory complications. <b><i>Results:</i></b> There were 10/44 (22.7%) perinatal deaths. The areas under (AU) the ROC curves were: first scan, 0.76, best O/E LHR cut-off 35.5% with 76% sensitivity and 70% specificity; last scan, AU-ROC 0.79, best O/E LHR cut-off 35.2%, with 79.0% sensitivity and 80% specificity. Considering an O/E LHR cut-off ≤35% to define high-risk fetuses at any examination, prediction for perinatal mortality showed: 80% sensitivity, 73.5% specificity, 47.1% positive and 92.6% negative predictive values, and 3.02 (95% CI 1.59–5.73) positive and 0.27 (95% CI 0.08–0.96) negative likelihood ratios. Prediction was similar in the two evaluations as 16/21 (76.2%) of fetuses considered at risk had an O/E LHR ≤35% in the two examinations; in the remaining 5 cases, two were identified only in the first and three only in the last scan. <b><i>Conclusion:</i></b> The O/E LHR is a good predictor of perinatal death in fetuses with left isolated CDH. Approximately 80% of fetuses at risk of perinatal death can be identified with an O/E LHR ≤35%, and 90% of them will have similar O/E LHR values at the first and at the last ultrasound examinations prior to delivery. In general, 88.6% of all CDH fetuses have a similar severity classification based on the O/E LHR at the first diagnostic ultrasound or at the ultrasound examination prior to delivery.