Fig 1.
(A) shows oxaliplatin concentration profile in the infusion tube. The x-axis represents the distance along the tube, the y-axis represents the time from the start of the pump delivery. For figures (B-G), the x-axis represents Clock time and starts at the beginning of the considered drug administration. The left column shows the difference between the intended delivery profiles and the simulated delivery profiles evaluated at the end of the tube (x = L), for irinotecan (B), oxaliplatin (D) and 5-fluorouracil (F). The right-hand column shows the cumulative percentage of drug delivered to the patient for the intended and actual profiles over time for irinotecan (C), oxaliplatin (E) and 5-fluorouracil (G).
Table 1.
Table describing the defining delivery values for CPT11, LOHP and 5-FU.
The main peak refers to the maximum flow rate from the intended delivery schedule. The spike peak rate refers to the maximum flow rate of the delivery caused by the glucose flush.
Fig 2.
Improved administration profiles.
(A) shows the drug solution delivery profile which consists of an initial bolus to fill the tube entirely, followed by the original profile. (B) shows the rinse solution delivery rate which continues drug delivery at correct rate while clearing the tube from any active substance, Original rinse peaks were kept unchanged, although there are not mandatory in this administration design. (C) shows how the flow rate along the tube is smoothly switched between the drug and the rinse and (D) shows the new drug delivery profile that will enter the patient compared to the original profile used in the OPTILIV study.
Fig 3.
Semi-physiological model of irinotecan PK.
Compartments were minimised to the most important components, Liver to accurately represent drug delivery, Blood which is measurement site and Organs to represent the rest of the body. Ci is the rate constant of clearance from compartment i. Irinotecan is bio-activated into its active metabolite SN38. Irinotecan was assumed to be delivered directly into the liver.
Fig 4.
Semi-physiological model of oxaliplatin PK.
Compartments were minimised to the most important components, Liver to accurately represent drug delivery, Blood which is measurement site and Organs to represent the rest of the body. Ci is the rate constant of clearance from compartment i. Each compartment contains a bound and unbound drug fraction and only unbound molecules can migrate between compartments. b and u are respectively the binding and unbinding rate constants of platinum to proteins. Oxaliplatin was assumed to be delivered directly into the liver in its unbound form.
Fig 5.
Semi-physiological model of 5-fluorouracil PK.
Compartments were minimised to the most important components, Liver to accurately represent drug delivery, Blood which is measurement site and Organs to represent the rest of the body. Ci is the rate constant of clearance from compartment i. 5-fluorouracil was assumed to be delivered directly into the liver.
Fig 6.
Patient data best-fit of irinotecan PK model.
Each subplot represents an individual patient dataset, fit to the model independently. (A) shows the fit of irinotecan plasma concentration, (B) shows that of SN38, the active metabolite of irinotecan.
Fig 7.
Patient data best-fit of oxaliplatin PK model.
Each subplot is an individual patient data, fit to the model independently. (A) shows plasma ultrafiltrate platinum concentrations, and (B) shows plasma total platinum concentrations. PK data for Patient 11 was missing.
Fig 8.
Patient data best-fit of 5-fluorouracil PK model.
Each subplot is an individual patient data fit to the model independently. PK data for Patient 6 and 11 was missing.
Fig 9.
Inter-patient variability in drug PK parameters.
The first line shows parameter variability across the considered patient population for irinotecan (A), oxaliplatin (B) and 5-fluorouracil (C), the colour and symbols represent the clusters each parameter set belongs to. The parameters are named with reference to the schematics of the models, the subscripts refer to the blood (B), organs (O) and liver (L). In the irinotecan parameters, additional subscripts cpt and sn refer to irinotecan and SN38 respectively. The second line shows multidimensional scaling representation of patient clustering based on their PK parameters for irinotecan (D), oxaliplatin (E) and 5-fluorouracil (F), the x refer to the cluster centroids and the points refer to patient PK parameters projected onto 2D plot.
Fig 10.
(A) Delivery profiles of irinotecan, oxaliplatin, 5-fluorouracil and glucose flushes as administered in the OPTILIV clinical trial. (B) Schematic of the Mélodie infusion pump (Axoncable, Montmirail, France) used in the OPTILIV study for hepatic artery infusion [20].