Clin Pharmacokinet DOI 10.1007/s40262-013-0092-3

ORIGINAL RESEARCH ARTICLE

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients Treated with Cytoreductive Surgery and Hyperthermic Intraperitoneal Oxaliplatin Carlos Pe´rez-Ruixo • Bele´n Valenzuela • Jose´ Esteban Peris • Pedro Bretcha-Boix • Vanesa Escudero-Ortiz • Jose´ Farre´-Alegre Juan Jose´ Pe´rez-Ruixo

•

Ó Springer International Publishing Switzerland 2013

Abstract Background and Objective Peritoneal carcinomatosis is an abdominal metastatic manifestation of a life-threatening tumour progression requiring standard palliative surgery and/or chemotherapy treatment. The aim of this study was to characterize the immediate neutrophilia response induced by cytoreductive surgery (CRS) and the myelosuppression effect of hyperthermic intraperitoneal oxaliplatin (HIO) in peritoneal carcinomatosis patients. Methods Absolute neutrophil counts (ANCs) from 45 patients treated with CRS and HIO diluted in isotonic 4 % icodextrin (cohort A), 21 patients undergoing CRS followed by HIO diluted in isotonic 5 % dextrose (cohort B) and 18 patients treated with CRS without HIO (cohort C) were used to estimate the system-related parameters [baseline ANC (Circ0), mean transit time (MTT) and feedback on proliferation (c)] and drug-specific (a) parameters of a modified Friberg’s model that accounts for the surgical stress-induced neutrophilia. The plasma oxaliplatin concentrations, Cp, were assumed to reduce the proliferation rate of the progenitor cells according to the function a 9 Cp. Model evaluation and simulations were undertaken to evaluate the effect of the dose, treatment C. Pe´rez-Ruixo  J. E. Peris Pharmacy and Pharmaceutical Technology Department, University of Valencia, Valencia, Spain B. Valenzuela (&)  P. Bretcha-Boix  V. Escudero-Ortiz  J. Farre´-Alegre Platform of Oncology, Hospital Quiro´n Torrevieja, Partida de la Loma s/n, 03184 Torrevieja, Alicante, Spain e-mail: [email protected] J. J. Pe´rez-Ruixo Pharmacokinetics and Drug Metabolism, AMGEN, Valencia, Spain

duration and carrier solution on the incidence of severe neutropenia. Results The typical values [between-subject variability, expressed in coefficient of variation values (%)] of the Circ0, MTT, c and a were estimated to be 3.58 9 109 cells/L (41.2 %), 144 h (70.9 %), 0.155 and 0.066 L/mg (134.9 %), respectively. Surgical stress induced a maximal 3.37-fold increase in the proliferation rate that was attenuated with a half-life of 10 days, and a maximal 68 % reduction in the MTT that was attenuated with a half-life of 28 days. Age, body surface area, sex, total proteins and carrier solution did not impact the model parameters. The model evaluation evidenced an accurate prediction of the incidence of neutropenia grade C2 and/or C3. Simulations indicated that (i) the neutropenia was reversible and shortlasting; and (ii) the HIO dose and treatment duration were the main determinants of the severity and duration of neutropenia. Conclusion The time course of neutropenia was well characterized by the model that was developed, which simultaneously accounts for the acute-immediate neutrophilia response induced by CRS and the HIO myelosuppressive effect produced in the bone marrow. This model suggests that higher doses than those evaluated to date could be used in peritoneal carcinomatosis patients without substantially increasing the risk of severe neutropenia.

1 Introduction Peritoneal carcinomatosis is an abdominal metastatic manifestation of life-threatening tumour progression requiring standard palliative surgery and/or chemotherapy (SPSC) treatment [1]. Relative to SPSC, cytoreductive surgery (CRS) followed by hyperthermic intraperitoneal

C. Pe´rez-Ruixo et al.

oxaliplatin (HIO) has been associated with increased survival in a retrospective analysis of patients with resectable peritoneal carcinomatosis of colorectal origin [2]. Furthermore, the efficacy of this innovative treatment has been reported in a phase II study in ovarian cancer [3] and in two phase III studies in colorectal and gastric cancers [4, 5]. Recently, a meta-analysis has shown a statistically significant survival benefit of CRS with hyperthermic intraperitoneal chemotherapy (HIPEC) and/or early postoperative intraperitoneal chemotherapy (EPIC) over SPSC in peritoneal carcinomatosis of colorectal origin [6]. These results encourage clinical development of this aggressive treatment, particularly in conditions such as peritoneal carcinomatosis of non-gynaecologic origin, where long-term survival is hardly ever seen after SPSC [7, 8]. Oxaliplatin is one of the drugs most commonly used in HIPEC because it has the ability to act at any stage of malignant cell replication [9], its intra-tumoral penetration is optimal [10] and its cytotoxicity is substantially improved by hyperthermia [11]. Several phase I doseescalation studies in peritoneal carcinomatosis patients treated with CRS have characterized HIO pharmacokinetics in both the peritoneum and plasma, [12–15] and evidenced that following 460 mg/m2 dosing, the maximum HIO concentration (Cmax) in the peritoneum (330 mg/L) [15] was 130-fold higher than the plasma Cmax (2.59 mg/L) after intravenous administration of 130 mg/m2 [16]. This finding indicates that HIO might achieve higher drug exposure in unresected tumour nodules and residual tumour cells in the peritoneal cavity, relative to intravenous oxaliplatin dosing. In addition, the short duration of HIO administration warrants minimum oxaliplatin access to the systemic circulation and, consequently, a low risk of haematological toxicity and peripheral sensory neuropathy, which are the dose-limiting toxicities after oxaliplatin intravenous dosing. In this context, pharmacokinetic and pharmacodynamic models that explain and predict the degree and duration of haematological toxicity after HIO are of particular clinical value in peritoneal carcinomatosis patients after CRS. One of the semi-mechanistic pharmacokinetic and pharmacodynamic models most frequently used to characterize neutropenia following cytotoxic chemotherapy in cancer patients [17] has recently been applied to describe neutrophil dynamics in peritoneal carcinomatosis patients treated with CRS and HIO [18]. The model provides a simplified mathematical framework to quantitatively describe the underlying physiological process of granulopoiesis through system-specific parameters that account for neutrophil production, maturation, regulation and elimination, and it contains drug-related parameters that quantify the cytotoxic mechanism of the oxaliplatin effect on proliferative bone marrow cells [18]. However, the model

does not account for the neutrophilic effect induced by surgical stress, which might be particularly important in projecting severe neutropenia after aggressive CRS and defining the maximum tolerated dose [19]. Thus, in the current study, the semi-mechanistic population pharmacokinetic and pharmacodynamic model that was previously developed [18] has been expanded to account for the effect of surgical stress on neutrophil dynamics. The expanded model is used to quantify the effect of HIO on the absolute neutrophil count (ANC) in peritoneal carcinomatosis patients after CRS, to evaluate the relationships between patient covariates and oxaliplatin model parameters, and to explore the potential maximum tolerated HIO dose in this setting.

2 Patients and Methods 2.1 Subject Eligibility Criteria and Treatment Adult patients were eligible for enrolment in this study if they had confirmation of peritoneal carcinomatosis without extra-abdominal metastasis. Other eligibility criteria included a World Health Organization performance status of 0–2 and anticipated life expectancy of at least 3 months. Patients were required to have a negative pregnancy test (only for female patients with reproductive potential) and normal hepatic and renal function, defined as bilirubin B1.5 times the upper limit of normality (9ULN), aspartate aminotransferase (AST) and alanine aminotransferase (ALT) B2.5 9 ULN, and serum creatinine B1.5 9 ULN. Acceptable bone marrow function, defined as ANC [1.5 9 109/L, haemoglobin [10 g/dL and platelets [100.0 9 109/L, was needed. Previous anticancer radiation therapy and/or chemotherapy, if given, had to have been discontinued at least 4 weeks before entry into the study, or 6 weeks in the case of pretreatment with nitrosoureas or mitomycin C. Patients with one or more of the following criteria were not selected: active infection, central nervous system metastases, peripheral neuropathy [grade 2, allogeneic transplant, prior extensive radiation therapy ([25 % of the bone marrow reserve), prior bone marrow transplantation or high dose chemotherapy with bone marrow or stem cell rescue, concurrent radiation therapy, chemotherapy, hormonal therapy, immunotherapy, participation in a clinical trial involving an investigational drug in the past 30 days or concurrent enrolment in another investigational study, and any coexisting medical condition that was likely to interfere with study procedures and/or results. Data from three single-arm observational studies, involving a total of 84 patients, were available for the analysis. Forty-five patients (53.6 %) out of 84 were treated with CRS followed by HIO diluted in isotonic 4 %

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients

icodextrin (cohort A), 21 patients (25.0 %) underwent CRS followed by HIO diluted in isotonic 5 % dextrose (cohort B) and 18 patients (21.4 %) underwent CRS but did not receive HIO (cohort C). The details of the CRS procedure and HIO treatment has been extensively described elsewhere [14, 18]. The studies were conducted in accordance with principles for human experimentation as defined in the International Conference on Harmonization for Good Clinical Practice guidelines and the principles of the Declaration of Helsinki. The studies were approved by the corresponding investigational review board, and informed consent was obtained from all subjects after they had been advised of the potential risks and benefits of CRS with or without HIO, as well as the investigational nature of the studies. Oxaliplatin pharmacokinetics following HIO administration were characterized in patients from cohorts A and B, using an intensive sampling schedule as described elsewhere [14]. Blood samples for determination of ANC were collected before surgery and, afterwards, daily until the patients completely recovered from the haematological toxicity. ANCs were determined using an automated haematology analyzer (AcT 5 diff AL; Beckman Coulter, Inc. Fullerton, CA, USA). 2.2 Software Nonlinear mixed-effects modelling using the first-order conditional estimation (FOCE) method implemented in the NONMEMÒ version 7.1.2 software package (ICON Development, Hanover, MD, USA) [20] was used to conduct the model-based analysis and simulations. Compilations were achieved using gfortran compiler 4.6.2 for Windows VistaTM 64 bits. The PsN 3.4.2 tool was used to conduct a stratified non-parametric bootstrap. Graphical and all other statistical analyses were performed using S-Plus 6.1 Professional Edition (Insightful Corporation, Seattle, WA, USA). 2.3 Pharmacokinetic and Pharmacodynamic Model Development The schematic of the pharmacokinetic and pharmacodynamic model used to describe neutrophil dynamics is based on Friberg’s model [17] and is displayed in Fig. 1. This model initially consisted of five compartments: one compartment representing the proliferative cells [Prol] such as stem cells and other progenitor cells; three transit compartments representing the cell maturation process from the bone marrow to the peripheral blood compartment [Transit], and one compartment describes the circulating blood cells [Circ]. Consistent with the mechanism of self-renewal

or mitosis, the generation of new cells in [Prol] was dependent on the number of cells in that compartment and was characterized by the first-order proliferation rate constant (kprol), which determines the rate of cell division, together with the feedback mechanism from the circulating cells. A feedback loop on the proliferation process was required in order to describe the rebound of cells compared with the baseline value of the ANC (Circ0). It was incorporated into the model as (Circ0/Circ)c, where c is the estimated parameter of the feedback system on the proliferation process and reflects the increase in the self-replication rate occurring when circulating cells are depleted. A maturation chain with three [Transit] compartments and first-order transition rate constants (ktr) allows prediction of the time taken for neutrophil progenitors to mature and reach the circulation, which is defined as the mean transition time (MTT). The MTT becomes apparent because it was assumed that the random loss of cells in the transit compartments was negligible. The differential equations governing the neutrophil dynamics model were as follows (Eqs. 1–5):   dProl Circ0 c ¼ kprol  Prol  EDrug  ktr  Prol dt Circ dTransit1 ¼ ktr  Prol  ktr  Transit1 dt dTransiti ¼ ktr  Transiti1  ktr  Transiti dt ) where i ¼ 2; . . .; n dCirc ¼ ktr  Transiti  kcirc  Circ dt nþ1 ktr ¼ MTT

ð1Þ ð2Þ

ð3Þ ð4Þ ð5Þ

where n is the number of transit compartments, Edrug represents the oxaliplatin neutropenic effect and kcirc represents the degradation rate of circulating cells. As the proliferative cells differentiate into more mature cell types, the cell concentration is maintained by cell division. At steady state, before CRS, dProl/dt = 0 and, therefore, kprol = ktr. Consequently, the amounts in the proliferative and transit compartments at baseline are Circ0 9 kcirc/ktr. During the model development, the number of transit compartments was optimized and the inclusion of a feedback loop on the maturation process in the model was also evaluated in order to describe the shorter maturation time of neutrophil progenitors after the ANC decrease, as reported recently [21, 22]. The liberation of multiple inflammatory mediators after surgery stimulate extra cell divisions in the proliferating compartment [23]. The peripheral neutrophil demand is increased after CRS

C. Pe´rez-Ruixo et al.

Fig. 1 Schematic of the neutrophil dynamics model. a slope of the linear relationship between Edrug and Cp, c feedback effect on the maturation process, Circ circulating blood cells, Circ0 baseline value of the absolute neutrophil count, CL systemic clearance, CLa peritoneal to plasma clearance, CLp intercompartmental clearance, Cp peritoneal oxaliplatin concentration, Edrug drug effect, IMmax maximum inhibitory effect of surgical stress on MTT, kcirc degradation rate of circulating cells, km first-order disappearance rate constant

of the surgical stress effect on maturation, kp first-order disappearance rate constant of the surgical stress effect on proliferation, kprol firstorder proliferation rate constant, ktr first-order transition rate constant, MTT mean transit time, Prol proliferation compartment, SPmax maximum stimulatory effect of surgical stress on kprol, t time, ts time when surgery starts, V1 central volume of distribution, V2 peripheral volume of distribution, V3 peritoneal volume of distribution

and, consequently, neutrophil mobilizations begin to occur from the bone marrow, reducing the total maturation transit time and resulting in an increase of the neutrophil circulating pool within 1–2 h after exposure to a surgical stress stimulus [24–26]. Therefore, CRS was assumed to increase the mitosis rate of proliferative cells and to reduce the maturation time of neutrophil precursors in the bone marrow. Mathematically, the dual effect of the surgical stress was modelled by a stimulatory function of kprol (Eq. 6) and an inhibitory effect on MTT (Eq. 7): 8 0 < kprol ; if t  ts h i ð6Þ kprol ¼ 0 : kprol  1 þ SPmax  eKp ðtts Þ ; if t [ ts 8 < MTT0 ; if t  ts h i MTT ¼ : MTT0  1  IMmax  eKm ðtts Þ ; if t [ ts

and IMmax are the maximum stimulatory and inhibitory effects of surgical stress on kprol and MTT, respectively; and kp and km are the first-order rate constants driving the disappearance of the surgical stress effect on the proliferation and maturation of neutrophils, respectively. Plasma oxaliplatin concentrations, Cp, were assumed to reduce the mitosis rate of proliferative cells according to the function Edrug, which was modelled using a linear function (Eq. 8):

ð7Þ 0 where kprol represents the first-order proliferation rate constant and MTT0 represents the mean transition time, at a time before the surgery, which starts at time (ts); SPmax

Edrug ¼ 1  a  Cp

ð8Þ

where a is the slope of the linear relationship between Edrug and Cp, which is derived from the empirical Bayesian estimates of the individual pharmacokinetic parameter obtained from updating the oxaliplatin population pharmacokinetic model previously developed and the individual peritoneal and plasma concentrations obtained from each patient [14]. In subjects with no oxaliplatin concentrations available (n = 9), typical values of pharmacokinetic parameters were assumed to generate the predicted time course of plasma oxaliplatin concentrations. Consequently, the development of the pharmacokinetic and pharmacodynamic model

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients

describing neutrophil dynamics was performed using a sequential process as described elsewhere [27]. During the model development process, it was tested whether power or sigmoid functions of Cp could describe Edrug better than the linear function. It was also evaluated whether Edrug should be a function of the drug concentration in a hypothetical effect compartment (Ce). Thus, the structural model parameters to be estimated were the system-related parameters (Circ0, MTT, kcirc and c), the surgery-specific parameters (SPmax, kp, IMmax and km) and the drug-related parameter (a), which were all restricted to be positive ([0). 2.4 Statistical Model The interindividual (or between-subject) variability (IIV) in the model parameters was assumed to follow a log-normal distribution and, consequently, an exponential error model was used. Residual variability in the ANC was evaluated using an additive error model after natural logarithmic transformation of the observations and model predictions.

by cohort [31], normalized prediction distribution errors (NPDEs) [32, 33], a prediction corrected visual predictive check (pcVPC) [34] and a numerical predictive check (NPC) [35]. NPCs were conducted for the incidence of neutropenia grade C2 (ANC \1.5 9 109/L) and grade C3 (ANC \1.0 9 109/L), and mean neutrophil counts at the nadir in patients with neutropenia grade C2 and grade C3, stratified by cohort. The distribution of the NPDE was explored by a frequency histogram, summary statistics stratified by cohort, and scatterplots of the NPDE versus the population prediction and time. Both the NPDE scatterplots and the pcVPC display the 5th, 50th and 95th percentiles of the observed values, and the 95 % confidence interval for the corresponding model-based predicted percentiles computed from 1000 Monte Carlo replicates obtained by simulating the design of the underlying dataset with the final model parameters. 2.8 Sensitivity Analysis

The improvement of the fit obtained for each nested model was assessed by the likelihood ratio test. In addition, the reduction in the IIV and residual variability, the precision and the correlation in parameter estimates, and the examination of diagnostic plots and shrinkage were used to evaluate each candidate model [28].

Based on the model developed, deterministic simulations were undertaken in order to explore the role of the model parameters quantifying the surgical stress effect on the ANC profile. An arbitrary value of a ±10 % change in model parameters was used to conduct the sensitivity analysis. Deterministic simulations were also undertaken in order to explore the role of the dose (or the initial peritoneal oxaliplatin concentration), HIO duration and carrier solution on the ANC time course.

2.6 Covariate Analysis

2.9 Model-Based Simulations

The covariates included in the analysis were age, body surface area (BSA), sex, total proteins and carrier solution. As the patients included in this analysis had normal liver and renal function biomarkers, and these biomarkers have not been associated with oxaliplatin pharmacokinetic parameters in this population [14], the effect of ALT, AST, total bilirubin and creatinine clearance on the systemrelated parameters of the neutrophil dynamics model were not formally tested because their effect had been proven to be negligible in previous analysis [29]. The effect of selected covariates on model parameters was explored following the forward-inclusion (p \ 0.05) and backwardelimination (p \ 0.01) process as described elsewhere [30]. Categorical covariates were incorporated into the model as index variables, whereas continuous covariates were evaluated using power equations after centering on the median.

Stochastic simulations were conducted to establish the relationship between the dose (or the initial peritoneal oxaliplatin concentration) and the incidence of severe neutropenia in order to determine the projected maximum tolerated HIO dose. For a range of initial oxaliplatin concentrations in the peritoneum spanning from 0 to 1,600 mg/ L, the daily ANC was simulated for two different HIO durations (30 and 60 min) and the incidence of neutropenia grade 4 or neutropenia grade 4 lasting at least 5 days was computed, for HIO administered with both icodextrin 4 % and dextrose 5 %. For each scenario, the oxaliplatin concentrations and daily ANC values for 2,000 virtual patients were simulated.

2.7 Model Qualification

The primary tumour types of the 84 eligible patients were colorectal (n = 27), ovarian (n = 20), gastric (n = 13), appendiceal (n = 11), endometrial (n = 3) and others (n = 10). Descriptive statistics of the patient

2.5 Model Selection Criteria

Four complementary methods were employed to evaluate the model developed: a non-parametric bootstrap stratified

3 Results

C. Pe´rez-Ruixo et al. Table 1 Patient characteristics at baseline, stratified by cohorta Patient characteristics

Cohort A (n = 45)b

Cohort B (n = 21)c

Cohort C (n = 18)d

Total (n = 84)

Age [years]

56 (12)

58 (11)

59 (11)

57 (12)

Bodyweight [kg]

68.6 (11.8)

69.8 (13.9)

69.0 (17.4)

69.0 (13.1)

Body surface area [m2]

1.7 (0.2)

1.8 (0.2)

1.8 (0.2)

1.8 (0.2)

Sex [% male]

36.2

33.3

55.6

40.7

ALT level [IU/L]

29 (6)

29 (8)

25 (8)

28 (7)

AST level [IU/L]

22 (6)

29 (13)

19 (9)

24 (10)

Total bilirubin level [lmol/L]

9.6 (3.7)

8.9 (3.9)

6.3 (3.1)

8.6 (3.8)

Total protein level [g/L]

71.5 (5.4)

74.0 (4.3)

67.5 (4.8)

71.7 (5.2)

Creatinine clearance [mL/min]e

81 (28)

80 (29)

85 (32)

81 (28)

Haemoglobin level [g/dL]

14.3 (2.7)

14.7 (1.1)

14.1 (0.4)

14.3 (2.0)

Neutrophil count [9109/L] Platelet count [9109/L]

3.76 (1.7) 235 (79)

4.19 (2.9) 268 (69)

3.82 (1.8) 245 (59)

3.88 (2.1) 246 (73)

Liver metastases [%]

10.6

9.5

27.8

14.0

Peritoneal carcinomatosis index

10.3 (11.5)

10.4 (9.5)

21.5 (6.4)

10.7 (10.9)

Complete cytoreduction [%]

78.7

95.2

0.0

66.3

Oxaliplatin dose [mg/m2]

358.0 (50.7)

410.9 (81.0)

–

376.2 (67.1)

Volume of carrier solution [L]

3.9 (0.8)

3.7 (0.7)

–

3.8 (0.8)

Duration of HIO [min]

37.3 (7.3)

32.4 (4.4)

–

35.7 (6.9)

ALT alanine aminotransferase, AST aspartate aminotransferase, CRS cytoreductive surgery, HIO hyperthermic intraperitoneal oxaliplatin, SD standard deviation a Continuous variables are expressed as mean (SD), whereas categorical variables are expressed as a percentage b

Cohort A were treated with CRS followed by HIO diluted in isotonic 4 % icodextrin solution

c

Cohort B were treated with CRS followed by HIO diluted in isotonic 5 % dextrose solution

d

Cohort C were treated with CRS but did not receive HIO

e

Creatinine clearance was calculated using the Cockcroft and Gault formula, and values higher than 150 mL/min were truncated to 150 mL/min

characteristics at baseline, stratified by cohort, are shown in Table 1. Cohort C had a higher proportion of females, incidence of liver metastasis and peritoneal carcinomatosis index, and a lower proportion of patients with complete cytoreduction. For the rest of the covariates, a similar distribution was found among the three cohorts, with no statistically significant differences. A total of 1,140 ANCs were included in the analysis dataset. The time course of the ANC displayed a ‘signature pattern’ profile consisting of an initial ANC peak, followed by the nadir and a subsequent peak, which was achieved before the return to baseline. In fact, the mean and the coefficient of variation (CV%) of the ANC before CRS, 3.88 9 109/L (53.1 %), substantially increased to 8.80 9 109/L (55.6 %) within 24 h after the CRS. The overall 2.27-fold increment in circulating neutrophil numbers observed was consistent with findings in other studies investigating acute mobilization of neutrophils from the bone marrow reserve, suggesting that the response to such neutrophilic stimuli is of a relatively uniform magnitude [36–38]. After this significant increase, ANC values started to diminish until the nadir value, 6.15 9 109/L (57.2 %), was reached around 7 days after the CRS. Afterwards, the

ANC rebounded up to 10.7 9 109/L (55.3 %) approximately 15 days after the CRS, just before the ANC started to return to the baseline value, which was achieved approximately 40 days after CRS. In addition to the neutrophilia induced by the surgical stress, the myelosuppressive effect of HIO was also evident from the incidence of neutropenia grade C2 and grade C3 in cohorts A, B and C (Table 2) and the oscillatory ANC profile that was observed. As expected, the original model proposed by Friberg et al. was not able to appropriately describe the neutrophilia induced by the surgical stress. When the effect of the surgical stress on kprol was included in the model, a substantial decrease in the minimum value of the objective function (DMOFV = -399.24) was observed. A further decrease in the MOFV was obtained by including the effect of surgical stress on the MTT (DMOFV = -395.15). The addition of kp and km, driving the disappearance of the surgical stress effect on the proliferation and maturation process, also improved the fit by decreasing the MOFV by -194.90 and -355.81 points, respectively. No substantial improvement of the fit was achieved by incorporating additional transit compartments or a feedback effect of the

0.84 (0.41–0.99) CI confidence interval

5.56 (0.00–11.1) 5.56

0.98 0.80 (0.32–0.98)

9.52 (4.76–14.3) 9.52

0.83 0.77 (0.34–0.95) 0.48 Mean neutrophil count at the nadir [9109 cells/L]

8.89 (4.44–11.1) 8.89 Incidence [%]

Neutropenia grade C3

5.56 (0.00–22.2)

1.21 (0.75–1.45) 0.98

5.56 14.3 (4.80–23.8)

1.19 (0.63–1.43) 1.07

14.3

0.77 Mean neutrophil count at the nadir [9109 cells/L]

15.6 (6.67–24.4) 11.1 Incidence [%]

Neutropenia grade C2

1.15 (0.78–1.34)

Predicted (95 % CI) Observed Observed Predicted (95 % CI) Observed

Predicted (95 % CI)

Cohort C Cohort B Cohort A Numerical predictive check

Table 2 Numerical predictive check of the incidence of neutropenia grade C2 and grade C3 and mean neutrophil counts at the nadir in patients with neutropenia grade C2 and grade C3

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients

ANC on the MTT. Similarly, no additional model improvement was achieved by adding an effect compartment or modelling Edrug as a power function of the plasma oxaliplatin concentration instead of using a linear function. Moreover, Emax or sigmoid Emax models for Edrug did not converge successfully, probably because the Emax was not achieved within the Cp range evaluated and the low incidence of severe neutropenia that was observed. Figure 2 shows the goodness-of-fit plots of the final model. The scatterplots representing the observed ANC and the population model prediction (panel a) and individual model prediction (panel b) showed a normal random scatter around the identity line and indicated the absence of significant bias. Similarly, the distribution of conditional weighted residuals [39] as a function of the population predictions (panel c) and time (panel d) did not show any trend that evidenced model inadequacy. The estimates of the final model parameters, together with the results of the non-parametric bootstrap analysis, are presented in Table 3. Between-subject variability was estimated for the model parameters MTT, Circ0, kcirc, a and kp, with shrinkage lower than 27.6 %. Furthermore, the exploratory graphical analysis of the correlation between age, BSA, sex, total proteins and the carrier solution with model parameters did not suggest any statistically significant association, and therefore they were not included in the model. Nevertheless, the carrier solution was a covariate for the volume of distribution in the peritoneum, as previously reported [14]. The population estimates for the final model were very similar to the mean of the 74 bootstrap replicates (37 %) that minimized successfully, and were contained within the 95 % confidence intervals obtained from the bootstrap analysis. The precision of the parameter estimates for the fixed effects was acceptable, with a relative standard error (RSE) lower than 21 %, except for a. In addition, the precision for the random effect was also adequate, with the RSE ranging from 18.2 to 75.3 %. The distribution of the parameter estimates across bootstrap replicates that minimized successfully (n = 74) or not (n = 126) was similar, as the difference in the mean was lower than 8 %. The results of the pcVPC are depicted in Fig. 3 (panel a), and evidence that the model developed is appropriate to describe the time course of the ANC and its variability in peritoneal carcinomatosis patients after CRS, regardless of HIO administration and the carrier solution used [34]. The overall distribution of the NPDE is presented in Fig. 3 (panel b) and approximately follows a normal distribution with a mean of 0 and a standard deviation (SD) of 1. Actually, the mean and SD of the NPDE [32] for the ANC in cohort A were 0.01 (95 % CI -0.07 to 0.08) and 0.98 (95 % CI 0.922–1.03), respectively. The mean and SD of the NPDE for the ANC in cohort B were 0.10 (95 % CI -

C. Pe´rez-Ruixo et al.

Fig. 2 Goodness-of-fit plots for the neutrophil dynamics model. Upper panels representing the observed absolute neutrophil counts as a function of the population model prediction (a), and the individual

model prediction (b). Lower panels representing conditional weighted residuals as a function of the population predictions (c) and time (d)

0.01 to 0.22) and 1.00 (95 % CI 0.92–1.10), respectively, while for cohort C they were 0.04 (95 % CI -0.08 to 0.18) and 0.94 (95 % CI 0.83–1.04), respectively. In addition, the distribution of the NPDE as a function of the population predictions (Fig. 3, panel c) and time (Fig. 3, panel d) did not show any trend that evidenced model misfit, and confirmed the model accuracy and precision in describing neutrophil dynamics and their variability in peritoneal carcinomatosis patients who underwent CRS and did or did not receive HIO. In Table 2, the results of the NPC show that the model properly described the incidence of neutropenia grade C2 and C3, and the mean ANC at the nadir in the three cohorts that were studied. Figure 4 shows the model’s ability to describe the individual time course of ANCs in peritoneal carcinomatosis patients randomly selected from the three cohorts analyzed in this study. Figure 5 shows the results of the sensitivity analysis in the absence of HIO. These simulations show that the model properly characterized the initial transient neutrophilia

observed immediately after the end of the CRS, the time to the ANC nadir and the rebound phenomena, before the ANC returned to the physiologic baseline. Each part of the ANC time course signature pattern is governed by different model parameters. The rapid ANC increase after surgery is mainly controlled by the effect of CRS on progenitor cell maturation. Indeed, due to a reduction in the maturation time, the neutrophils are rapidly mobilized to the peripheral circulation, causing an initial neutrophilia, and its magnitude is mainly determined by the parameter IMmax. Actually, the first ANC peak after the surgery increases as the IMmax increases (Fig. 5, panel a). Furthermore, the effect of CRS on maturation leads to rapid depletion of the proliferative cells, which results in an ANC decrease after the first ANC peak, reaching the nadir around 5–8 days after surgery. The contribution of the feedback effect on the depletion of proliferative cells is limited, since the proliferation rate is decreased by only 13 % right after the CRS. The ANC nadir after the first ANC peak is determined by

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients Table 3 Parameter estimates and non-parametric bootstrap analysis of the neutrophil dynamics model for peritoneal carcinomatosis patients treated with cytoreductive surgery with or without hyperthermic intraperitoneal oxaliplatina Pharmacodynamic model parameters

Original dataset

Non-parametric bootstrap (n = 74 replicates out of 200)

Estimateb

Mean (RSE %)

95 % CI

System-related parameters Circ0 [9109/L]

3.58

3.54 (6.25)

3.10–4.05

MTT [h]

144

148 (13.7)

112–204

c

0.155

0.163 (17.4)

0.133–0.253

kcirc [/h]

0.088

0.090 (20.5)

0.065–0.141

Surgical stress-related parameters SPmax 2.37

2.41 (11.0)

1.97–2.94

IMmax

0.684

0.683 (3.17)

0.643–0.723

kp [10-3/h]

3.47

3.59 (13.7)

2.72–4.68

km [10-3/h]

1.99

2.11 (20.4)

1.55–3.08

0.064 (43.2)

0.014–0.126

Drug-related parameter a [L/mg]

0.066

Interindividual variability [CV %] x Circ0

41.2

41.0 (18.2)

33.2–48.2

x MTT

70.9

70.6 (29.4)

47.2–89.2

x kcirc

104.4

105.2 (29.4)

66.6–132.7

x kp

15.5

15.7 (44.6)

10.8–21.9

xa

134.9

161.6 (75.3)

78.3–285.8

32.7 (6.90)

28.5–36.7

Residual variability [CV %] r

32.7

a slope of the linear relationship between Edrug and Cp, c feedback effect on the maturation process, CI confidence interval, Circ0 baseline value of the absolute neutrophil count, Cp peritoneal oxaliplatin concentration, CV coefficient of variation, Edrug drug effect, IMmax maximum inhibitory effect of surgical stress on MTT, kcirc degradation rate of circulating cells, km first-order disappearance rate constant of the surgical stress effect on maturation, kp first-order disappearance rate constant of the surgical stress effect on proliferation, kprol firstorder proliferation rate constant, MTT mean transit time, RSE relative standard error of the parameter estimate, SPmax maximum stimulatory effect of surgical stress on kprol a

The shrinkage values (%) for Circ0, MTT, kcirc, kp, a and r were 6.43, 22.8, 26.4, 19.2, 27.6 and 12.3, respectively [28]

b

The covariance step failed. Therefore, RSEs of pharmacodynamic parameters are not provided

the magnitude of the proliferative cell depletion, which is mainly dependent on the SPmax/IMmax ratio (Fig. 5, panels a, c). Consequently, a faster ANC decrease is observed with a higher SPmax/IMmax ratio, which may result in lower ANC nadir values (Fig. 5, panel a), depending on the magnitude of the proliferative cell depletion. The ANC nadir also increases as the SPmax increases (Fig. 5, panel c). Similarly, the second ANC peak increases as the SPmax increases and the magnitude of the proliferative cell depletion decreases (Fig. 5, panels a, c). The effect of km and kp on the first ANC peak and the subsequent ANC

nadir is negligible, since both events occurs relatively faster after CRS relative to the corresponding half-life associated with km and kp. The modest effect of km and kp on the ANC profile is mainly observed in the second ANC peak, which increases as km increases (Fig. 5, panel b) and kp decreases (Fig. 5, panel d). Figure 6 (panels a and b) shows the results of the deterministic model-based simulations evaluating the impact of the initial HIO concentration, treatment duration and carrier solution on the time course of the ANC. The simulations clearly show that the main determinants of the severity and the duration of the neutropenia are the initial HIO concentration in the peritoneum, the treatment duration and, to a lower extent, the carrier solution. For example, the ANCs on day 7 following CRS and 30 min HIO administration of 400, 800 and 1,600 mg/L with icodextrin 4 % were 4.87, 3.48 and 1.75 9 109/L, respectively. If the HIO duration was maintained for 60 min, the ANCs on day 7 were 3.76, 2.05 and 0.6 9 109/L, respectively. On the other hand, if dextrose 5 % was selected as the carrier solution, the ANCs on day 7 after CRS and 30 min HIO administration of 400, 800 and 1,600 mg/L were 4.65, 3.16 and 1.44 9 109/L, respectively. If the HIO duration was prolonged to 60 min, the ANCs on day 7 decreased to 3.50, 1.77 and 0.47 9 109/L, respectively. Given that the HIO effect on proliferative cells is proportional Cp, the effect of the HIO dose on the ANC profile illustrated in Fig. 6 is a surrogate of the drug effect on the ANC profile. In addition, deterministic simulations also suggest that the neutropenia is reversible and short-lasting. The model-based relationship between the initial HIO concentration in the peritoneum and the incidence of severe neutropenia is presented in Fig. 6 (panels c and d), stratified for the treatment duration and carrier solution. This figure suggests that a 30 min HIO administration of 1,000 mg/L diluted in icodextrin 4 % produces an approximate 20 % incidence of neutropenia grade 4 lasting more than 5 days. If the treatment duration is extended to 60 min, the initial HIO concentration should be reduced by 25 % (to 750 mg/L) to achieve a similar incidence of neutropenia grade 4 lasting more than 5 days. Moreover, if instead of using icodextrin 4 % as the carrier solution, an isotonic dextrose 5 % solution is preferred, the initial HIO concentrations should be reduced by 10 or 12 % for an HIO duration treatment of 30 or 60 min, respectively, in order to obtain a similar incidence to that obtained with icodextrin 4 % carrier solution.

4 Discussion In this study, the relationship between oxaliplatin pharmacokinetics and the time course of the ANC in patients with peritoneal carcinomatosis receiving HIO after CRS

C. Pe´rez-Ruixo et al.

Fig. 3 Prediction corrected visual predictive check (a), normalized prediction distribution error histogram (b) and scatterplots of the normalized prediction distribution error versus model prediction

(c) and time (d). The solid red lines display the 5th, 50th and 95th percentiles of the observed values, and the associated shaded areas represent the 95 % confidence intervals (CIs). SD standard deviation

was investigated by extending a semi-mechanistic model that had previously been developed [17, 18]. The extension of this model was able to properly characterize the transient increase in the ANC induced by CRS in the immediate postoperative period and the myelosuppressive effect caused by the HIO. Under normal circumstances, neutrophils follow an orderly maturation and progression from the bone marrow to the circulatory system; however, the surgical stress induces a profound systemic inflammatory process, which is associated with an acute-immediate neutrophilia response, as evidenced in the patients included in this study [19, 40]. This neutrophilia may occur by an accelerated release of cells from the marrow into the blood, an increase in proliferative cell production, a shift from the marginal pool to the circulating pool within the blood, a reduced exit of neutrophils from the blood to the tissues, or a combination of these mechanisms [41]. The neutrophil dynamics model developed here takes into account the first two pathophysiological processes, described, in a similar way to a previous model developed to quantify the effect of

granulocyte colony-stimulating factor (G-CSF) on the stimulation of neutrophil production and maturation [42]. However, the model could not account for the demargination process that neutrophils suffer when entering the peripheral circulation from areas of intravascular marginated polymorphonuclear cell pools, nor the decreased egress (outwards migration) of neutrophils from the peripheral circulation into the tissues. Shifts between these cell pools take only a few minutes [41, 43], and the ANC data collected based on daily sampling schedules did not contain enough information to properly characterize these processes. The response to the surgical stress is associated with the liberation of multiple inflammatory mediators like leukotriene B4, complement component C5a, interleukin-8, alpha tumour necrosis factor (TNFa), G-CSF and chemokine (C-X-C motif) ligand 1 (CXCL1), which stimulate extra cell divisions in the proliferating compartment [19, 23, 44]. In fact, a maximal 3.37-fold increase in the proliferative cell production rate in the bone marrow was observed. This maximal stimulus was attenuated with a half-life of 10 days. The neutrophil dynamics model also

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients

Fig. 4 Representative individual time course of absolute neutrophil counts (ANCs) and the corresponding model predictions for cohorts A, B and C. CRS cytoreductive surgery, HIO hyperthermic intraperitoneal oxaliplatin

accounted for a rapid release of neutrophil precursors from the bone marrow into the blood due to the surgical stress. The post-mitotic maturation time at baseline, MTT, was estimated to be 144 h, which was very similar to the MTT values previously reported in healthy volunteers [45] and cancer patients receiving other antineoplastic agents [17, 21, 48]. In the immediate postoperative period, there was a maximal 68.4 % reduction in the MTT, which slowly disappeared following a first-order process with a half-life of 28 days. The magnitude of the increase in the mitotic activity in the bone marrow and the reduction of the maturation time are consistent with the magnitude of the G-CSF effects on these two processes as previously reported [46–48]. In addition, since the attenuation of the surgical effect on the proliferation rate is faster than the estimated effect on the maturation time, the time course of the ANC generates an oscillatory profile, as shown in Figs. 4 and 6, due to the depletion of proliferative cells. Furthermore, the reduction in the MTT due to the inflammatory condition results in an increase in immature or band-neutrophils in the peripheral circulation [19, 49], and some authors have hypothesized

that the increase in immature neutrophils might translate in a defective neutrophil function and, therefore, the incidence of opportunistic infections could be raised [50, 51]. However, other authors such as van Dijk et al. [52] found no change in neutrophil phagocytosis after surgery. Moreover, Mollit et al. [53] reported an enhancement of neutrophil function following elective surgery in healthy children. These discrepancies suggest that further studies should be conducted in order to assess neutrophil function and its association with the host defence in the context of peritoneal carcinomatosis patients treated with HIO after CRS. With respect to other system-related parameters, the estimation of Circ0 before CRS was 3.58 9 109/L (41.2 %), which is consistent with physiologic ANC values. In peritoneal carcinomatosis patients, Valenzuela et al. [18] reported a higher Circ0 of 7.05 9 109/L (42.3 %). However, this value actually reflects the ANC after the CRS, which is not comparable with Circ0 before CRS as reported here [18]. The estimated c value of 0.155 was also similar to the values previously obtained for other anticancer drugs, such as irinotecan (0.132) and docetaxel

C. Pe´rez-Ruixo et al.

Fig. 5 Influence of surgical stress-related parameters quantifying the impact of cytoreductive surgery on maturation (a, b) and proliferation (c, d) on the neutrophil dynamics profile. IMmax maximum inhibitory effect of surgical stress on MTT, km first-order disappearance rate

constant of the surgical stress effect on maturation, kp first-order disappearance rate constant of the surgical stress effect on proliferation, kprol first-order proliferation rate constant, MTT mean transit time, SPmax maximum stimulatory effect of surgical stress on kprol

(0.161) [17]. The estimated neutrophil half-life of 7.8 h was very similar to other values reported in the literature, ranging from 6.7 to 10 h [45, 54, 55], which suggests that CRS with or without HIO does not increase the residence time of neutrophils in plasma. In addition, the interindividual and residual variabilities were moderate to large, consistent with the values reported for other drugs, when ANC data are analyzed using similar models [17, 56]. The relationship between plasma oxaliplatin concentrations Cp and the neutropenic effect was described by a linear function, and the slope was estimated to be 0.066 L/mg, which is consistent with the value recently reported for carboplatin, after accounting for the differences in protein binding and molecular weight between carboplatin and oxaliplatin [57]. However Valenzuela et al. [18] reported a higher slope, 0.182 L/mg, probably because their structural model could not explain that a portion of the ANC decrease was due to the priming effect in neutrophil precursors induced by surgical stress, rather than the HIO neutropenic effect [18]. As a consequence of the linear drug effect model, an approximate 28.4 % decrease in the

ANC at day 7 is predicted per each 400 mg/L increase in the initial HIO concentration for 30 min treatment, regardless of the carrier solution. In addition, a 22.8 % decrease in the ANC at day 7 for each 400 mg/L increase in the initial HIO concentration is predicted when the treatment is extended from 30 to 60 min, regardless of the carrier solution. The four complementary methods that were used to evaluate the predictive performance of the neutrophil dynamics model (the non-parametric bootstrap, pcVPC, NPDE and NPC) showed that the precision of the NONMEMÒ parameter estimates was acceptable and that the model developed was suitable to describe the neutrophilic and neutropenic effect caused by CRS and HIO, respectively. Although the fact that only 37 % of bootstrap replicates minimized successfully and the covariance step failed might indicate a weak convergence around the parameter estimates, these phenomena are believed to be related to the use of FOCE. In fact, the distribution of the parameters estimates across bootstrap replicates was similar between the replicates that minimized successfully and

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients

Fig. 6 Effect of initial oxaliplatin concentrations in the peritoneum, treatment duration and carrier solution on neutrophil dynamics (a, b) and the incidence of neutropenia grade 4 and neutropenia grade 4 lasting 5 days (c, d). HIO hyperthermic intraperitoneal oxaliplatin

those that did not, and the parameter estimates obtained by the stochastic approximation expectation maximization (SAEM) algorithm in NONMEMÒ were similar to those obtained with FOCE. These findings allowed us to rule out concerns about the convergence and confirmed the estimates of the parameters. Furthermore, the successful model qualification provided supportive evidence that the severity and duration of neutropenia can be predicted in cancer patients with peritoneal carcinomatosis treated with CRS and HIO and, therefore, simulations were undertaken to explore the relationship between the HIO dose and the incidence of severe neutropenia, stratified by the treatment duration and carrier solution. Deterministic simulations revealed that HIO induced a reversible and short-lasting neutropenia, which is largely dependent on the dose administered (or the initial HIO concentration in the peritoneum) and the treatment duration. Figure 6 (panels a and b) shows that increasing the initial HIO concentration in the peritoneum and/or extending the HIO duration lead to a greater fluctuation in ANC values and consequently increase the likelihood of severe neutropenia. Model-based stochastic simulations also indicated that it is possible to reduce the incidence and

severity of neutropenia by employing treatment regimens with shorter HIO durations or with icodextrin 4 % as the carrier solution, while the overall dose administered (and the initial HIO concentration in the peritoneum) remains the same. Icodextrin, an a-1-4-linked glucose polymer of 12,000 to 20,000 D diluted at 4 %, is an isotonic high molecular weight solution widely used for peritoneal dialysis and has also been employed as a carrier solution for HIO. This high molecular weight solution is able to maintain the intraperitoneal fluid volume for longer, reduces the amount of oxaliplatin absorbed from the peritoneum into the plasma, and therefore explains the slight reduction in the incidence and severity of neutropenia with respect to other carrier solutions like dextrose 5 % isotonic solution [14, 58]. Actually, for a given HIO dose and duration, the risk of neutropenia grade 4 lasting more than 5 days was 1.13-fold higher when dextrose 5 %, instead of icodextrin 4 %, was used as the carrier solution. However, for a given HIO dose and carrier solution (dextrose 5 % or icodextrin 4 %), extending the duration from 30 to 60 min was associated with a 1.35-fold increase in the risk of neutropenia grade 4 lasting more than 5 days.

C. Pe´rez-Ruixo et al.

Qualitatively, the results of the simulation exercise emphasized the importance of the HIO dose, treatment duration and carrier solution effect on the incidence of severe neutropenia. However, generalization of the quantitative results obtained from this modelling and simulation exercise should be done with caution, due to the semimechanistic nature of the neutrophil dynamics model developed, the range of doses evaluated, the limited number of patients and the specific characteristics of this complex treatment in our centre. These circumstances per se demand further model evaluation in additional studies. 5 Conclusion For the first time, the neutrophil dynamics model has been extended to simultaneously account for the immediate neutrophilia response induced by surgical stress and chemotherapy-induced myelosuppression in the bone marrow. This model, successfully applied to describe the time course of the ANC in peritoneal carcinomatosis patients treated with CRS with or without HIO, suggests that higher doses than those evaluated to date (460 mg/m2, corresponding to an initial concentration of 230 mg/L for a 2 L/m2 volume of the carrier solution) could be used without substantially increasing the risk of severe neutropenia, if the incidence of other adverse events is not increased. However, additional clinical studies should be conducted in order to identify the HIO maximum tolerated dose (or exposure). In those studies, primary prophylaxis with G-CSF should be considered to minimize the risk of infections due to bone marrow myelosuppression and to control the potential impaired function of neutrophil bandcells in the systemic circulation as a consequence of surgical stress, especially if high initial oxaliplatin concentrations and long treatment durations are evaluated [59]. Finally, the model developed in this study could be a useful tool to optimize the design of future clinical studies in this setting. Acknowledgments The authors would like to thank the patients and the medical, nursing and laboratory staff of the Hospital Quiro´n Torrevieja who participated in the present study. The authors would like to thank Dr. Ricardo Nalda for his valuable help during the simulation exercise. This work was supported by Consellerı´a de Sanidad of Comunidad Valenciana (grant GE-079/11). The authors also like to thank the peer reviewers of this manuscript for their valuable comments, which helped to improve the quality of the work and Dr. Ricardo Nalda-Molina for his comments and support at the beginning of this project. Conflict of interest Carlos Pe´rez-Ruixo, Bele´n Valenzuela, Jose´ Esteban Peris, Pedro Bretcha-Boix, Vanesa Escudero-Ortiz, Jose´ Farre´-Alegre and Juan Jose´ Pe´rez-Ruixo have indicated no potential conflicts of interest, other than those reflected in their affiliations. Disclaimer The views expressed in this article are the personal view of the authors, reflecting their scientific knowledge of this topic, and should not be understood or quoted as being made on behalf of the companies where the authors work.

References 1. Spiliotis JD. Peritoneal carcinomatosis cytoreductive surgery and HIPEC: a ray of hope for cure. Hepatogastroenterology. 2010;57: 1173–7. 2. Elias D, Lefevre JH, Chevalier J, et al. Complete cytoreductive surgery plus intraperitoneal chemohyperthermia with oxaliplatin for peritoneal carcinomatosis of colorectal origin. J Clin Oncol. 2009;27:681–5. 3. Di Giorgio A, Naticchioni E, Biacchi D, et al. Cytoreductive surgery (peritonectomy procedures) combined with hyperthermic intraperitoneal chemotherapy (HIPEC) in the treatment of diffuse peritoneal carcinomatosis from ovarian cancer. Cancer. 2008;113:315–25. 4. Verwaal VJ, van Ruth S, de Bree E, et al. Randomized trial of cytoreduction and hyperthermic intraperitoneal chemotherapy versus systemic chemotherapy and palliative surgery in patients with peritoneal carcinomatosis of colorectal cancer. J Clin Oncol. 2003;21:3737–43. 5. Yang XJ, Huang CQ, Suo T, et al. Cytoreductive surgery and hyperthermic intraperitoneal chemotherapy improves survival of patients with peritoneal carcinomatosis from gastric cancer: final results of a phase III randomized clinical trial. Ann Surg Oncol. 2011;18:1575–81. 6. Cao C, Yan TD, Black D, et al. A systematic review and metaanalysis of cytoreductive surgery with perioperative intraperitoneal chemotherapy for peritoneal carcinomatosis of colorectal origin. Ann Surg Oncol. 2009;16:2152–65. 7. Sugarbaker PH. Intraperitoneal chemotherapy and cytoreductive surgery for the prevention and treatment of peritoneal carcinomatosis and sarcomatosis. Semin Surg Oncol. 1998;14:254–61. 8. Elias D, Pocard M, Goere D. HIPEC with oxaliplatin in the treatment of peritoneal carcinomatosis of colorectal origin. Cancer Treat Res. 2007;134:303–18. 9. Alcindor T, Beauger N. Oxaliplatin: a review in the era of molecularly targeted therapy. Curr Oncol. 2011;18:18–25. 10. Elias D, El Otmany A, Bonnay M, et al. Heated intraoperative intraperitoneal oxaliplatin after complete resection of peritoneal carcinomatosis: pharmacokinetics and tissue distribution. Ann Oncol. 2002;13:267–72. 11. Atallah D, Marsaud V, Radanyi C, et al. Thermal enhancement of oxaliplatin-induced inhibition of cell proliferation and cell cycle progression in human carcinoma cell lines. Int J Hyperthermia. 2004;20:405–19. 12. Stewart JH, Shen P, Russell G, et al. A phase I trial of oxaliplatin for intraperitoneal hyperthermic chemoperfusion for the treatment of peritoneal surface dissemination from colorectal and appendiceal cancers. Ann Surg Oncol. 2008;15:2137–45. 13. Ferron G, Dattez S, Gladieff L, et al. Pharmacokinetics of heated intraperitoneal oxaliplatin. Cancer Chemother Pharmacol. 2008;62: 679–83. 14. Pe´rez-Ruixo C, Valenzuela B, Peris JE, et al. Population pharmacokinetics of hyperthermic intraperitoneal oxaliplatin in patients with peritoneal carcinomatosis after cytoreductive surgery. Cancer Chemother Pharmacol. 2013;71:693–704. 15. Elias D, Sideris L. Pharmacokinetics of heated intraoperative intraperitoneal oxaliplatin after complete resection of peritoneal carcinomatosis. Surg Oncol Clin North Am. 2003;12:755–69. 16. Massari C, Brienza S, Rotarski M, et al. Pharmacokinetics of oxaliplatin in patients with normal versus impaired renal function. Cancer Chemother Pharmacol. 2000;45:157–64. 17. Friberg LE, Henningsson A, Maas H, Nguyen L, Karlsson MO. Model of chemotherapy-induced myelosuppression with parameter consistency across drugs. J Clin Oncol. 2002;20:4713–21. 18. Valenzuela B, Nalda-Molina R, Bretcha-Boix P, et al. Pharmacokinetic and pharmacodynamic analysis of hyperthermic intraperitoneal

Neutrophil Dynamics in Peritoneal Carcinomatosis Patients

19.

20. 21.

22.

23.

24.

25.

26. 27.

28.

29.

30.

31. 32.

33.

34.

35. 36.

37. 38.

oxaliplatin-induced neutropenia in subjects with peritoneal carcinomatosis. AAPS J. 2011;13:72–82. Tabuchi Y, Shinka S, Ishida H. The effects of anesthesia and surgery on count and function of neutrophils. J Anesth. 1989;3:123–31. Beal SL, Sheiner LB, Boeckman AJ, editors. NONMEM users guides. Ellicott City: ICON Development Solutions; 1989–2006. Gonza´lez-Sales M, Valenzuela B, Pe´rez-Ruixo C, et al. Population pharmacokinetic-pharmacodynamic analysis of neutropenia in cancer patients receiving PM00104 (ZalypsisÒ). Clin Pharmacokinet. 2012;51:751–64. Quartino AL, Friberg LE, Karlsson MO. A simultaneous analysis of the time-course of leukocytes and neutrophils following docetaxel administration using a semi-mechanistic myelosuppression model. Invest New Drugs. 2012;30:833–45. Vieira SM, Lemos HP, Grespan R, et al. A crucial role for TNFalpha in mediating neutrophil influx induced by endogenously generated or exogenous chemokines, KC/CXCL1 and LIX/ CXCL5. Br J Pharmacol. 2009;158:779–89. Price TH, Chatta GS, Dale DC. Effect of recombinant granulocyte colony-stimulating factor on neutrophil kinetics in normal young and elderly humans. Blood. 1996;88:335–40. Dale DC, Liles WC, Llewellyn C, Price TH. Effects of granulocyte-macrophage colony-stimulating factor (GM-CSF) on neutrophil kinetics and function in normal human volunteers. Am J Hematol. 1998;57:7–15. Jagels MA, Hugli TE. Mechanisms and mediators of neutrophilic leukocytosis. Immunopharmacology. 1994;28:1–18. Zhang L, Beal SL, Sheiner LB. Simultaneous vs. sequential analysis for population PK/PD data I: best-case performance. J Pharmacokinet Pharmacodyn. 2003;30:387–404. Savic RM, Karlsson MO. Importance of shrinkage in empirical Bayes estimates for diagnostics: problems and solutions. AAPS J. 2009;11:558–69. Kloft C, Wallin J, Henningsson A, et al. Population pharmacokinetic-pharmacodynamic model for neutropenia with patient subgroup identification: comparison across anticancer drugs. Clin Cancer Res. 2006;12:5481–90. Mandema JW, Verotta D, Sheiner LB. Building population pharmacokinetic-pharmacodynamic models. I: models for covariate effects. J Pharmacokinet Biopharm. 1992;20:511–28. Efron B, Tibshirani R. An introduction to the bootstrap. London: Chapman and Hall/CRC Press; 1993. Brendel K, Comets E, Laffont C, et al. Metrics for external model evaluation with an application to the population pharmacokinetics of gliclazide. Pharm Res. 2006;23:2036–49. Nguyen TH, Comets E, Mentre´ F. Extension of NPDE for evaluation of nonlinear mixed effect models in presence of data below the quantification limit with applications to HIV dynamic model. J Pharmacokinet Pharmacodyn. 2012;39:499–518. Bergstrand M, Hooker AC, Wallin JE, Karlsson MO. Predictioncorrected visual predictive checks for diagnosing nonlinear mixed-effects models. AAPS J. 2011;13:143–51. Karlsson MO, Savic RM. Diagnosing model diagnostics. Clin Pharmacol Ther. 2007;82:17–20. Dale DC, Fauci AS, Guerry D IV, et al. Comparison of agents producing a neutrophilic leukocytosis in man: hydrocortisone, prednisone, endotoxin, and etiocholanolone. J Clin Invest. 1975;56:808–13. Joyce RA, Boggs DR. Visualizing the marrow granulocyte reserve. J Lab Clin Med. 1979;93:101–10. Orr Y, Wilson DP, Taylor JM, et al. A kinetic model of bone marrow neutrophil production that characterizes late phenotypic maturation. Am J Physiol Regul Integr Comp Physiol. 2007;292:R1707–16.

39. Hooker AC, Staatz CE, Karlsson MO. Conditional weighted residuals (CWRES): a model diagnostic for the FOCE method. Pharm Res. 2007;24:2187–97. 40. Kim C, Sakamoto A. Differences in the leukocyte response to incision during upper abdominal surgery with epidural versus general anesthesia. J Nippon Med Sch. 2006;73:4–9. 41. Summers C, Rankin SM, Condliffe AM, et al. Neutrophil kinetics in health and disease. Trends Immunol. 2010;31:318–24. 42. Krzyzanski W, Wiczling P, Lowe P, et al. Population modeling of filgrastim PK-PD in healthy adults following intravenous and subcutaneous administrations. J Clin Pharmacol. 2010;50:101S–12S. 43. Athens JW, Haab OP, Raab SO, et al. Leukokinetic studies. IV. The total blood, circulating and marginal granulocyte pools and the granulocyte turnover rate in normal subjects. J Clin Invest. 1961;40:989–95. 44. Gwak MS, Choi SJ, Kim JA, et al. Effects of gender on white blood cell populations and neutrophil-lymphocyte ratio following gastrectomy in patients with stomach cancer. J Korean Med Sci. 2007;22:S104–8. 45. Dancey JT, Deubelbeiss KA, Harker LA, et al. Neutrophil kinetics in man. J Clin Invest. 1976;58:705–15. 46. Sandstro¨m M, Lindman H, Nygren P, et al. Population analysis of the pharmacokinetics and the haematological toxicity of the fluorouracil-epirubicin-cyclophosphamide regimen in breast cancer patients. Cancer Chemother Pharmacol. 2006;58:143–56. 47. Panetta JC, Schaiquevich P, Santana VM, et al. Using pharmacokinetic and pharmacodynamic modeling and simulation to evaluate importance of schedule in topotecan therapy for pediatric neuroblastoma. Clin Cancer Res. 2008;14:318–25. 48. Ramon-Lopez A, Nalda-Molina R, Valenzuela B, et al. Semimechanistic model for neutropenia after high dose of chemotherapy in breast cancer patients. Pharm Res. 2009;26:1952–62. ˚ , et al. Granulocytopoi49. Fliedner TM, Cronkite EP, Killmann SA esis: II. Emergence and pattern of labeling of neutrophils granulocytes in humans. Blood. 1964;24:683–700. 50. Duignan JP, Collins PB, Johnson AH, Bouchier-Hayes D. The association of impaired neutrophil chemotaxis with postoperative surgical sepsis. Br J Surg. 1986;73:238–40. 51. Solomkin JS, Bauman MP, Nelson RD, et al. Neutrophils dysfunction during the course of intra-abdominal infection. Ann Surg. 1981;194:9–17. 52. van Dijk WC, Verbrugh HA, van Rijswijk RE, et al. Neutrophil function, serum opsonic activity, and delayed hypersensitivity in surgical patients. Surgery. 1982;92:21–9. 53. Mollitt DL, Steele RW, Marmer DJ, et al. Surgically induced immunologic alterations in the child. J Pediatr Surg. 1984;19:818–22. 54. Walker RI, Willemze R. Neutrophil kinetics and the regulation of granulopoiesis. Rev Infect Dis. 1980;2:282–92. 55. Cartwright GE, Athens JW, Wintrobe MM. The kinetics of granulopoiesis in normal man. Blood. 1964;24:780–803. 56. Hing J, Perez-Ruixo JJ, Stuyckens K, Soto-Matos A, LopezLazaro L, Zannikos P. Mechanism-based pharmacokinetic-pharmacodynamic meta-analysis of trabectedin (ET-743, Yondelis) induced neutropenia. Clin Pharmacol Ther. 2008;83:130–43. 57. Schmitt A, Gladieff L, Laffont CM, et al. Factors for hematopoietic toxicity of carboplatin: refining the targeting of carboplatin systemic exposure. J Clin Oncol. 2010;28:4568–74. 58. Mohamed F, Sugarbaker PH. Carrier solutions for intraperitoneal chemotherapy. Surg Oncol Clin N Am. 2003;12:813–24. 59. Aapro MS, Cameron DA, Pettengell R, et al. EORTC guidelines for the use of granulocyte-colony stimulating factor to reduce the incidence of chemotherapy-induced febrile neutropenia in adult patients with lymphomas and solid tumors. Eur J Cancer. 2006;42:2433–53.