Abstract
Background/Aim: Oral 5-fluorouracil (5-FU)-based prodrugs, used in cancer chemotherapeutic regimens, exhibit large inter- and intra-patient variability in plasma 5-FU concentrations, contributing to treatment failure. Although dosage determination criteria according to plasma drug concentrations are required, the relationship between pharmacokinetics and drug response after multiple oral 5-FU derivative administrations remain unknown. Materials and Methods: We evaluated the pharmacokinetics and pharmacodynamics/toxicodynamics of uracil-tegafur (UFT) after multiple administrations in colorectal cancer (CRC) model rats, and applied a pharmacometric approach to describe the time-course alterations of plasma 5-FU concentrations and tumor shrinkage. CRC was induced in rats using 1,2-dimethylhydrazine and dextran sulfate sodium. UFT (30 mg/kg as tegafur) was administered to CRC rats for 14 days. Results: Plasma 5-FU exposure levels increased with the dosing time, and large variations were observed in tumor 5-FU concentrations (32.0-125.8% with coefficient of variation). Although severe hematological toxicities were not observed, a weak correlation was observed between blood platelet count and the plasma 5-FU concentration (r=0.439, p=0.176). A simple pharmacokinetic-pharmacodynamic model was developed comprising of a small number of parameters and successfully describing plasma 5-FU levels and tumor shrinkage after multiple UFT administrations. Conclusion: A pharmacometric model approach can help establish the dose-determination criteria based on plasma 5-FU concentration of UFT-based regimens, and contribute to improvement of clinical outcomes.
Antineoplastic 5-fluorouracil (5-FU) is a key drug in colorectal cancer (CRC) therapy. Long-term infusion schedules (over 46 h) of 5-FU, such as FOLFIRI (folinic acid, 5-FU, and irinotecan) and FOLFOX (folinic acid, 5-FU, and oxaliplatin), are first-line chemotherapeutic regimens for advanced or metastatic CRC (1). However, long-term infusion procedures require catheterization and restrict the daily activities of patients. Therefore, oral prodrugs of 5-FU such as capecitabine, S-1 (tegafur, 5-chloro-2,4-dihydroxypyridine, and potassium oxonate), and uracil-tegafur (UFT) have been approved instead of long-term infusion of 5-FU.
UFT and leucovorin regimens, which comprise daily oral administration of UFT and leucovorin (28 days followed by a 7-day rest), are widely prescribed as adjuvant chemotherapy for CRC (2). Repetitive oral UFT administration offers certain therapeutic benefits; however, large inter- and intra-individual variations are observed in 5-FU plasma concentrations, which are a primary factor responsible for treatment failure (3-5). These variations in systemic exposure to 5-FU are associated with body surface area (BSA)-based dose determination (6). Multiple clinical studies have shown the advantages of pharmacokinetic (PK)-guided dose adjustments in improving clinical outcomes (6-9). Therapeutic drug monitoring (TDM) of 5-FU is strongly recommended in the intravenously continuous infusion regimen to personalize dosing and obtain adequate systemic exposure (10). However, TDM for oral 5-FU prodrugs have not been established due to the limited data of the relationship between plasma 5-FU concentrations and therapeutic responses such as antitumor efficacy and toxicity (10, 11).
A pharmacokinetic-pharmacodynamic (PK-PD) model is a mathematical model that describes the relationship between drug exposure and efficacy. Mathematical models are not only useful for developing optimal dose schedules to maximize drug efficacy with minimum toxicities, but also aid in understanding key factors determining drug efficacy. Currently, several PK models have been developed to describe 5-FU exposure after UFT administration (4, 12-14); however, these models are complex, contain many parameters, and were developed based on plasma concentration data after single-drug administration. To develop a widely applicable PK-PD model for clinical studies, a simple mathematical model composed of a small number of parameters is preferable. In a recent clinical study, multiple-cycle chemotherapy treatment using an oral prodrug of 5-FU resulted in a significant decrease of 5-FU levels in plasma; the reason being unclear (15). Few studies have developed a PK-PD model to describe PK alteration due to repetitive anticancer drug administration.
In this study, we evaluated the PK and PD/TD of 5-FU after repetitive UFT administration in a CRC rat model and developed a simple PK-PD model to identify and quantitatively describe the PK and PD/TD behavior of UFT after repetitive oral administration.
Materials and Methods
Chemicals. Tegafur was purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Uracil and 5-FU was supplied by Wako Pure Chemical Industries, Ltd. (Osaka, Japan).
Development of a rat CRC model. CRC model rats were prepared using 10-week-old male Wistar rats (Nippon SLC Co., Ltd., Hamamatsu, Japan), as described previously (16). Rats were subcutaneously administered 10% 1,2-dimethylhydrazine saline solution (40 mg/kg) three times every other day (Days 1, 3, and 5) and were provided drinking water containing 1.0% dextran sulfate sodium solution for the following one week (Day 8 to 15). The rats were housed in a temperature-controlled room under a 12-h light/dark cycle with free access to food and water. After at least 20 weeks, the rats were used in experiments after identification of a solid tumor using an endoscope (AVS AE-C1, Olympus. Co. Ltd., Tokyo, Japan). Animal experimental procedures were approved by the animal experimentation Ethics Committee of the Kyoto Pharmaceutical University (approval number: PKPD-17-001) and performed in accordance with the Kyoto Pharmaceutical University guidelines for animal experimentation.
PK and PD/TD of tegafur and 5-FU. UFT, comprising tegafur and uracil (1:4 molar ratio), was dissolved in 1% sodium carboxymethyl cellulose solution (14). CRC rats were divided in three groups based on the duration of drug administration (1, 7, and 14 days). To evaluate myelosuppression by UFT administration, blood samples (approximately 500 μl) were collected in ethylenediaminetetraacetic acid (EDTA)-treated tubes from the external left jugular vein before starting the PK study on Days 1, 7, and 14. The total number of blood cells (platelets, leukocytes, neutrophils, and lymphocytes) was determined in the Kyoto Biken Laboratories Inc. (Kyoto, Japan). To evaluate the PK, UFT (30 mg/kg as tegafur) was administered to CRC rats daily for 14 days. After UFT administration, blood (250 μl) was collected at 0.5, 1, 1.5, 2, 3, 4, 6, 8, and 24 h after drug administration on Days 1, 7, and 14. Blood samples, for determining drug plasma concentration, were centrifuged in heparinized centrifuge tubes at 14,000×g for 15 min at 4°C. Plasma samples were stored at −80°C until analysis. To evaluate the drug distribution in the tumor, all the rats were euthanized by cervical dislocation immediately after final blood sampling and tissues were perfused with phosphate-buffered saline (PBS) (pH 7.4) to remove blood, followed by collection of colon samples. The tumor tissue was harvested from the colon sample and washed with PBS. The tumor volume was calculated as follows: tumor volume (mm3)=length (mm)×width2 (mm2)/2 (17). Tumor samples were homogenized in PBS (three-fold volumes of each sample weight) using a homogenizer (PT 10-35 GT) (Kinematica AG, Lucerne, Switzerland). The homogenate samples were centrifuged at 3000×g for 15 min and the supernatant was stored at −80°C until analysis. The tegafur and 5-FU concentration in samples was measured using liquid chromatography-tandem mass spectrometry (LC-MS/MS) as described previously (14).
Non-compartmental PK analysis. Traditional non-compartmental PK analysis was conducted using the Phoenix® WinNonlin software (version 8.2) (Certara USA, Inc., Princeton, NJ, USA) before PK-PD modeling and simulation. The time (Tmax) at which plasma drug concentrations reached the maximum value (Cmax) was determined from the time-profile of plasma concentration data. The area under the plasma concentration-time curve from 0 h to the final sampling time (AUC0–24h) and to infinity (AUC0-∞) were determined using the linear trapezoidal rule. The terminal slope (λz) was determined based on the linear regression of at least three data points from the terminal portion of the plasma concentration-time curve. The elimination half-life (t1/2) was calculated as t1/2=ln2/λz. The total plasma clearance (CLtot) was calculated using the formula D/AUC0-∞, where D represents the total dosage administered. The distribution volume (Vd) was calculated using the formula CL/λz.
PK-PD modeling. The PK-PD model was developed to characterize plasma drug concentrations and tumor shrinkage after UFT administration in CRC rats using the Phoenix® WinNonlin software. The PK-PD model for myelosuppression was not used as there was no severe hepatotoxicity observed in the current in vivo experiments. Therefore, only the PK-PD model describing tumor shrinkage was used. The disadvantage of simultaneous PK-PD fitting is that unreliable PK model parameters can be obtained if the PD model structure is fragile. Therefore, the model was developed using a two-stage approach; first, PK modeling was executed separately from the PD modeling, and then a PD model was developed with fixed PK parameters. The final model was determined based on the Akaike’s Information Criteria (AIC), the coefficient of variation (CV) of parameter estimates, goodness-of-fit plots including predictions (PRED) vs. observations (OBS), conditional weighed residuals (CWRES) vs. time, and CWRES vs. PRED. The AIC were used in the modeling; a decrease in AIC by ≥2 was interpreted as a major improvement of the model. Figure 1 shows a schematic of the final model.
Schematic of the uracil-tegafur (UFT) pharmacokinetic-pharmacodynamic (PK-PD) model for describing tumor shrinkage after oral UFT administrations for 14 days. A1: Tegafur amount; A2: 5-FU amount; C1: tegafur concentration; C2: 5-FU concentration; V1: distribution volume of tegafur; V2: distribution volume of 5-FU; ka: absorption rate constant; ke: elimination rate constant; k1: rate constant; k2: measure of drug potency; X: tumor volume; 5-FU: 5-fluorouracil.
Various PK models (one-, two-, or three compartment models) were initially investigated and the above-mentioned modeling criteria was utilized to construct the simple PK model with a small number of parameters. To develop the PK model, we used the observed plasma concentration data of tegafur and 5-FU following single and multiple UFT administration in CRC rats. After the seventh administration of UFT, Day 7 showed higher plasma 5-FU levels compared to Day 1, with slightly increased plasma tegafur levels. In the source data, the prolonged Tmax value of tegafur and the slow absorption rate was observed after multiple administrations of UFT. In the initial PK model analysis using a 1-compartment model with plasma tegafur concentration data, the absorption rate constant for tegafur (ka) decreased with the number of dosages (3.87 on Day 1, 0.79 on day 7, and 0.56 on Day 14). Multiple instances of 5-FU exposure induce gastrointestinal injury, affecting the intestinal absorption of drugs and nutrients (18). Therefore, absorption inhibition was included in the final PK model (Equation 1 and 2). A conversion rate constant for the metabolism of tegafur into 5-FU (k1) and a linear elimination constant of 5-FU (ke) was incorporated in the PK model. The final PK model comprised of the following equations:
[1]
[2]
[3]
[4]
[5]
Aa, Imax, and IC50 represent the amount of tegafur in the absorption compartment, maximum inhibition constant of 5-FU, and 5-FU concentration producing 50% of the maximum inhibition. C1, A1, and V1 represent the concentration, amount, and distribution volume of tegafur, respectively. C2, A2, and V2 represent the concentration, amount, and distribution volume of 5-FU, respectively. V2 and ke values were set to 0.3 L/kg and 18.4 1/h, respectively, according to previous study (19) and results of preliminary PK model analysis.
The PD model was developed based on a previous study (17) with minor modifications. A well-defined tumor growth model (20) was initially applied to the current data to describe the time-course of tumor volume. However, in the current study, there was no tumor growth data in rats without UFT treatment. Therefore, various PD models were initially tested to construct the simple PD model comprising of a small number of parameters. The final PD model of cell death after chemotherapy treatment comprised the following equations:
[6]
[7]
[8]
X (t) and k2 represent number of tumor cells and 5-FU potency against tumor cells, respectively. W(t) and W0 represent the weight of the tumor cells and isolated tumor tissue before UFT treatment, respectively. The estimated parameters were k2 and X0.
Statistical analysis. Data are presented as the mean±standard deviation (SD). The Student’s unpaired t-test was used to compare two groups. One-way analysis of variance (ANOVA), followed by the Bonferroni adjustment, was used to compare data from multiple groups. Statistical significance was set at p<0.05. The correlation between blood cell counts and PK parameters were assessed using the Pearson’s correlation coefficient and was considered statistically significant at p<0.05.
Results
PK and PD of UFT in CRC rats. The mean plasma and tumor concentrations of tegafur and 5-FU, and tumor volume after oral administrations of UFT for 14 days in CRC rats are presented in Figure 2. The PK parameters obtained by non-compartment PK analysis are summarized in Table I. The plasma concentration of tegafur and 5-FU was altered after multiple UFT administrations in CRC rats; tegafur Cmax values on Day 7 (397.6±84.4 μmol/l) and 14 (347.4±24.5 μmol/l) were slightly higher compared to that on Day 1 (293.0±37.1 μmol/l), and the Tmax was prolonged. However, these differences were not statistically significant. No differences were observed in t1/2, Vd/F, CLtot/F, and AUC0-∞ of tegafur between Days 1, 7, and 14. The PK parameters of 5-FU were as follows: Cmax (4.4±1.5 μmol/l) and AUC0-∞ (38.3±8.1 μmol×h/l); values on Day 7 were 2.6- and 2.1-times higher compared to those on Day 1 (1.7±1.0 μmol/l and 18.0±7.4 μmol×h/l), respectively. Tmax of 5-FU on Day 14 (4.0±1.6 h) was prolonged compared to that on Day 1 (1.2±0.6 h). The intra-tumor tegafur concentration on Day 7 (0.018±0.010 μmol/g) was approximately 2-times higher than that on Day 1 (0.010±0.003 μmol/g). The mean intra-tumor 5-FU concentration tended to increase with dosages, whereas relatively large variabilities were observed (32.0-125.8 CV%). Tumor volume on Day 14 (70.1±56.6 mm3) was lower than that on Day 1 (214.7±195.0 mm3); tumor shrinkage following multiple administration of UFT was confirmed in CRC rats.
Mean plasma and tumor tegafur and 5-fluorouracil concentrations (A) and tumor volume (B) after oral uracil-tegafur (UFT) administrations for 14 days. Results are presented as the mean±S.D. (Day 1: n=3; Day 7: n=4; Day 14: n=4 rats). White, gray, and black circles represent plasma concentrations on Days 1, 7, and 14, respectively. *p<0.05 statistically significant difference compared to that on Day 1, evaluated using one-way ANOVA followed by Bonferroni test.
Non-compartmental analysis of pharmacokinetic parameters of tegafur and 5-fluorouracil (5-FU) in colorectal cancer model rats after oral uracil-tegafur (UFT) administration for 14 days.
Toxicity of UFT in CRC rats. The time-course profiles of blood cell counts and their relationship with the AUC0-24h of 5-FU after oral administrations of UFT in CRC rats are presented in Figure 3. No changes were observed in platelet, leucocyte, neutrophil, and lymphocyte counts among UFT treatments. No significant linear correlations were observed between the AUC0-24h of 5-FU and the leucocyte, neutrophil, and lymphocyte counts. There was a weak correlation between platelet counts and AUC0-24h of 5-FU (r=0.439, p=0.176).
Time-course profiles of blood cell count and their relationship with the AUC0-24h of 5-fluorouracil after oral uracil-tegafur (UFT) administrations for 14 days. Results are presented as mean±S.D. of n=5 rats. White, gray, and black circles in the four correlation figures in the lower panel represent blood cell counts and the area under the plasma drug concentration-time curve from 0 h to the final sampling time (AUC0–24h) on Days 1, 7, and 14, respectively.
PK-PD model of UFT in CRC rats. Model fitting of the mean plasma concentrations of tegafur, 5-FU, and tumor volume after UFT administration in CRC rats is presented in Figure 4. The time-course profiles of plasma concentrations were best characterized by the final PK model with the absorption inhibition factor. The slope, intercept, r, and r2 of the linear regression line in the goodness-of-fit plots showed the following values: 1.04, −4.96, 0.950, and 0.903 for tegafur, 1.00, 0.08, 0.682, and 0.466 for 5-FU, respectively. The AIC value was 350.4. On Days 7 and 14, the Cmax values of tegafur (302.4 μmol/l) and 5-FU (2.1 μmol/l) were underestimated compared to the observed mean values (tegafur: 397.6 μmol/l on Day 7 and 347.4 μmol/l on Day 14; 5-FU: 4.4 μmol/l on Day 7 and 2.3 μmol/l on Day 14). The final PK-PD model successfully described tumor shrinkage after multiple UFT administrations in CRC rats. The slope, intercept, r, and r2 of the linear regression line in the goodness-of-fit plots for the tumor volume data were 1.05, – 6.32, 0.995, and 0.990, respectively. The AIC value was 25.1. The final PK-PD parameter estimates are summarized in Table II. The CV% of each parameter estimate was <17.8%. The residual variability of plasma tegafur and 5-FU concentrations, and tumor volume was 19.0% (13.9 CV%), 45.1% (14.6 CV%) and 4.8% (40.9 CV%), respectively. These model-relating parameter results indicate that the PK-PD model adequately described the time-course alterations of tumor volume after multiple UFT administrations.
Simulated time-course profile and observed plasma tegafur and 5-fluorouracil concentrations (A), and tumor volume (B) in colorectal cancer model rats after uracil-tegafur administration for 14 days. Open circles represent mean observed data.
Pharmacokinetic-pharmacodynamic model parameters of tegafur and 5-fluorouracil in colorectal cancer model rats after oral uracil-tegafur (UFT) administration for 14 days.
PK-PD model simulation. Simulated time-course profiles were produced using the PK-PD model for the plasma concentration of 5-FU and tumor volume after different doses of UFT (Figure 5). Plasma 5-FU levels were predicted to increase with dosages. The PK-PD model can predict tumor shrinkage with increased dose concentrations.
Pharmacokinetic-pharmacodynamic model-based simulation of time-course profiles of plasma 5-fluorouracil concentration on Day 1 and the tumor volume in colorectal cancer model rats after uracil-tegafur (UFT) administration for 14 days.
Discussion
The standard first-line chemotherapy for CRC is an oxaliplatin and irinotecan-based regimen (1). However, UFT and leucovorin combination therapy is occasionally preferred for elderly or frail patients as it is safer and more tolerable than standard chemotherapy (21). The growing elderly population may present increased the demand for personalization of dose determination to avoid chemotherapy discontinuation due to severe toxicities.
Plasma 5-FU concentration-based dose determination (TDM) is recommended for improving the clinical outcome of a regimen with long-term infusion of 5-FU; however, this determination method for oral 5-FU derivates has not been established. TDM of UFT has been proposed to determine the optimal dose for patients (9, 22). However, there is little evidence available to develop the dose determination criteria for each patient as plasma 5-FU concentration data after repetitive oral administration of UFT are not well documented. Oral 5-FU derivates are generally prescribed to outpatients rather than inpatients; therefore, it is difficult and ethically challenging to collect numerous blood samples from patients with cancer to evaluate PK characteristics and their correlations with anticancer effects and myelosuppression. Therefore, in the current study, we tried to identify the PK and PD/TD behavior of UFT after repetitive treatments in CRC model rats and develop the PK-PD mathematical model to establish the dose determination criteria based on plasma drug concentrations.
The dimethylhydrazine and dextran sulfate sodium-induced CRC model rat is a well-known and widely used animal model that parallels the disease in humans in terms of disease presentation and gross and microscopic pathology (16). The CRC model animals were used to evaluate the chemotherapeutic efficacy and toxicity profile of anticancer agents (23-25). We previously evaluated PK and PD of 5-FU or capecitabine (a 5-FU derivative) in CRC model rats (17, 26, 27); therefore, the CRC model rats are considered suitable for investigating the PK and PD/TD of UFT.
Multiple 5-FU exposures lead to changes in the PK of 5-FU. We previously reported that following repeated intravenous bolus 5-FU administration to CRC model rats, the metabolism of 5-FU varied with the reduction of dihydropyrimidine dehydrogenase activity, which is a rate-limiting metabolic enzyme of 5-FU in the liver, resulting in an increase of AUC values (27). In the current study, repetitive UFT treatment increased the AUC values of 5-FU in CRC rats, which is consistent with repetitive 5-FU bolus administration. Although the elimination half-life of 5-FU after bolus administration in rats is very rapid (approximately 0.3 h) (27), the blood concentration–time profile of 5-FU after UFT administration showed moderate elimination (approximately 6.7-10.4 h); the flip-flop phenomenon (elimination rate is much larger than the metabolic rate of tegafur to 5-FU) occurred. These results suggest that multiple administrations of UFT prolong the elimination rate of 5-FU, resulting in an increase of AUC values and leading to large variations in plasma and tumor 5-FU concentrations. However, in the current PK model analysis, the ke value was estimated to not decrease with the number of UFT administrations.
UFT has a low risk of adverse events compared to an oxaliplatin-based regimen (24); however, myelosuppression is a dose limiting toxicity and, in a recent clinical study, leucopenia (11.7%), neutropenia (2.6%), anemia (44.2%), and thrombocytopenia (7.8%) were observed in patients with CRC receiving UFT and leucovorin treatment regimens (28). In the current study, severe myelosuppression was not observed in CRC rats after multiple UFT treatments. Previous studies reported the correlation of the degree of myelotoxicity with plasma 5-FU levels after intravenous administration using a semi-physiological PK-PD model (19, 29, 30). However, the relationship between the severity of hematological toxicity and plasma 5-FU exposure after oral administration of a 5-FU derivative is unclear. In the current study, a weak correlation was observed between platelet counts and the AUC of 5-FU after UFT administrations, suggesting that plasma 5-FU levels could be used to predict the severity of thrombocytopenia in UFT treatments.
The current PK-PD model developed with the absorption inhibition factor of tegafur best captured the time-course of plasma drug levels and shrinkage of colorectal tumor volume after multiple UFT administrations, and reliable PK-PD parameters. Although some pharmacometric UFT models have been reported (4, 12-14), this is the first model that comprises a small number of parameters and was developed using data obtained from repetitive UFT administrations. Regarding the effects of multiple exposure to 5-FU on the intestinal absorption of compounds, it has been reported that multiple administration of 5-FU (30 mg/kg daily, for 4 days) in rats cause small intestinal epithelial barrier failure, which increases the permeation of some compounds, including 3-O-methyl glucose and fluorescein isothiocyanate-labelled dextran (18, 31). One study suggested that 5-FU-based chemotherapy results in mucosal injury in the brush border of the small intestine, interfering with the functioning of enzymes and transporters, is responsible for hydrolysis and absorption of dietary carbohydrates (32). The current PK model analysis revealed that the absorption rate constant of tegafur decreased after repeated UFT administration; however, its permeability and bioavailability were not evaluated. Although no differences were observed in the exposure levels of tegafur after multiple administrations, further studies need to investigate the effects of chemotherapy-induced intestinal injury on drug absorption.
The current study had certain limitations. First, the current model partially underestimated the observed PK data. Further studies will investigate the mechanism of PK alteration after multiple UFT treatments and the application of these other variable parameters to the model may improve the estimation of PK data. Second, the PK and hematological data were obtained at the single dose level. Therefore, to elucidate their relationship after UFT administration, PK-toxicodynamic model analysis should be performed using plasma 5-FU concentration and blood cell count data at various dose levels. Third, the currently proposed model from animal data cannot be directly applied to clinical practice. For clinical use, a pharmacometrics model should be developed using clinical data. To improve clinical outcomes using the current results, a clinical PK study needs to be conducted to establish personalization in UFT dose determination. Finally, inter- and intra-individual variabilities of PK and PD parameters after multiple UFT treatments were not evaluated; in clinical trials, population model analysis should be performed to determine the factors that contribute to large variations in drug exposure.
Conclusion
Our results reveal the potential of a PK-PD model approach to quantitatively describe the relationship between plasma 5-FU exposure levels and tumor shrinkage after multiple UFT administrations. Our findings highlight the fact that a pharmacometrics model approach can be used to establish the dose-determination criteria based on the plasma 5-FU concentration in UFT-based regimens, thereby contributing to improvement of clinical outcomes. However, additional PK and PD/toxicodynamic studies are essential in establishing a TDM strategy.
Acknowledgements
The Authors would like to thank Professor Toshiyuki Sakaeda from Kyoto Pharmaceutical University (Kyoto, Japan) for his valuable support and research guidance. This work was supported in part by JSPS KAKENHI Grant Number 21K06720.
Footnotes
Authors’ Contributions
Conception and study design: S.K. and Y.I.; experimental work and analysis: S.K., M.T., M.O. and T.N.; interpretation of data; S.K., M.T., M.O., T.N. and Y.I.; drafting of the article: S.K.; revision of the article: S.K. and Y.I. All Authors approved the final version of the article.
Conflicts of Interest
The Authors declare that they have no competing interests in relation to this study.
- Received August 25, 2022.
- Revision received September 13, 2022.
- Accepted September 26, 2022.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
