Development of the LSS
The relationship between the predicted and actual AUC0–∞ values using the one-point sampling strategy with plasma busulfan concentrations (ng/mL) measured at 1, 2, 3, 4, and 6 h after the initiation of busulfan infusion (C1, C2, C3, C4, and C6, respectively) is shown (Figure 1). The most accurate predicted AUC0–∞ in all patients using the one-point sampling strategy was based on C6 (r2 = 0.789; MAPE, 11.0%). The multiple regression analysis of the predicted AUC0–∞ values based on C3, C4, and C6 using the two-point and three-point sampling strategies revealed that the one-point sampling strategy was highly precise. By the two-point sampling strategy, the AUC0–∞ predicted based on C3 and C6 exhibited the most favorable performance (r2 = 0.943; MAPE, 6.4%). The AUC0–∞ predicted based on the three-point sampling strategy using C3, C4, and C6 showed excellent predictive performance when compared with the actual AUC0–∞ (r2 = 0.955; MAPE, 5.9%) (Table 2).
The predictive performance of busulfan AUC based on body weight are shown in Table 3. In adolescent patients (body weight > 23 kg), the AUC0–∞ predicted based on C6was highly precise (r2 = 0.937; MAPE, 5.9%); however, this correlation was poor in infants and young children weighing ≤23 kg (r2 = 0.782; MAPE, 11.4%). By the two-point sampling strategy using C3 and C6, the predicted AUC0–∞ was acceptable even for infants and young children (r2 = 0.963; MAPE, 5.7%) and was as accurate as the AUC0–∞based on C6 in adolescent patients (r2 = 0.937; MAPE, 5.9%).