Preoperative high-dose methotrexate (HD-MTX) with folinic acid (leucovorin) rescue is still a mainstay in the treatment of osteosarcoma. This anticancer agent is characterized by a narrow therapeutic index and wide interpatients variability. To ensure effective and safe administration of HD-MTX, we had earlier developed an adaptive-dosing schedule with a feedback strategy. In our institute, the MTX dosage was tailored according to individual pharmacokinetics parameters, determined in real time both from two blood samples (3.5 and 4.5 h) and from Bayesian population parameters. Up to 20 g of MTX was safely administered as 8-h infusions. Low MTX elimination rate has, however, been reported in 15-20% of the patients, and forecasting the MTX elimination phase and the management of leucovorin rescue is still a challenging issue in clinical oncology. This study aims at identifying the clinical or biological covariates related to impaired MTX clearance, and at validating a new limited sampling strategy (LSS), allowing for the accurate prediction of the MTX terminal elimination phase. This retrospective study was carried out on 49 patients (30 men, 19 women; mean age, 26.7 years) treated for osteosarcoma with HD-MTX. The population and individual pharmacokinetics parameters were computed, before the identification of the relevant covariates. Different LSSs were then tested, to predict accurately when the MTX plasma concentrations would drop below 0.2 micromol/l, the threshold associated with the end of the rescue of leucovorin with alkaline hydration. Two main covariates (creatinemia clearance and alanine aminotransferase) were correlated with MTX clearance. Conversely, the impact of body surface area on MTX pharmacokinetics was weak, suggesting that dosing schedules based on body surface area were inadequate and potentially hazardous. A new LSS predicting accurately when the MTX concentration would reach 0.2 micromol/l has been validated; blood samples are stopped as soon as the MTX concentration drops to 1 micromol/l. With this LSS, our retrospective study suggests that 60% of the patients would have left the hospital earlier than they actually did owing to a better forecasting of the MTX decrease, thus improving their quality of life while improving the cost-effectiveness for the institute. HD-MTX can be administered safely using an adaptive-dosing strategy with drug monitoring. Moreover, pharmacokinetic modeling permits the accurate forecasting of the MTX elimination profile, thus allowing for a better management of the postinfusion care of cancer patients treated with particularly high doses of this drug.