journa l homepage: www.e lsev ier .com/ locate / lungcan
elective pulmonary artery perfusion for the treatment of primary lung cancer:mproved drug exposure of the lung
art P. van Puttea,∗, Marco Grootenboersb, Wim-Jan van Bovena, M. van Oosterhoutc,erard Pasterkampd, Gert Folkerts e, Franz Schramelb
Departments of Cardio-thoracic Surgery, Sint Antonius Hospital, Koekoekslaan 1, Nieuwegein, The NetherlandsDepartments of Pulmonology, Sint Antonius Hospital, Nieuwegein, The NetherlandsDepartments of Pathology, Sint Antonius Hospital, Nieuwegein, The NetherlandsDepartment of Experimental Cardiology, University Medical Center, Utrecht, The NetherlandsDepartment of Pharmacology and Pathophysiology, Faculty of Pharmaceutical Sciences, University of Utrecht, Utrecht, The Netherlands
r t i c l e i n f o
rticle history:eceived 23 July 2008eceived in revised form 8 September 2008ccepted 12 November 2008
Introduction: Selective pulmonary artery perfusion (SPAP) is an experimental drug infusion method forthe treatment of lung cancer that aims to achieve more effective T(umour) and lymph N(ode) down-staging. The aim of this experiment was to compare drug uptake of gemcitabine and carboplatin duringSPAP and intravenous infusion (IV).Material and methods: SPAP was performed in 12 pigs using clinically applied doses of gemcitabine(1.25 g/m2, n = 4) and carboplatin (AUC 5, n = 4) and a combination of both (n = 4). All animals under-went catheterisation of the left pulmonary artery and furthermore a left thoracotomy and lumbotomyfor tissue sampling. After 2 min of SPAP, 30 min of blood flow occlusion was performed in order to delaydrug washout from the lung. Two additional groups were infused intravenously (IV) using the same doseof gemcitabine (n = 4) and carboplatin (n = 4). Peak concentrations and area under the curve (AUC) werecompared with t-tests.Results: Significantly higher pulmonary gemcitabine peak concentrations (p ≤ 0.017) and AUC (p ≤ 0.020)were observed after SPAP gemcitabine and gemcitabine/carboplatin compared to IV while no dif-ferences were shown between serum, renal and lymph tissue. Furthermore, SPAP carboplatin andgemcitabine/carboplatin also resulted in significantly higher pulmonary carboplatin peak concentra-tions (p ≤ 0.018) compared to IV while AUC was significantly higher after SPAP gemcitabine/carboplatin(p = 0.045). SPAP carboplatin resulted in significantly higher lymph concentrations at 10 min comparedto IV carboplatin (p = 0.04).Conclusion: SPAP with gemcitabine and carboplatin resulted in significantly improved drug exposure ofthe lung compared to IV. Equivalent serum concentrations and a trend towards higher mediastinal lymphnode concentrations were achieved for carboplatin.
Summery: Selective pulmonary artery perfusion (SPAP) is an experimental drug infusion method for thetreatment of lung cancer that aims to achieve more effective T(umour) and lymph N(ode) down-staging.The aim of this experiment was to compare drug uptake of gemcitabine and carboplatin during SPAP andintravenous infusion (IV). In conclusion, SPAP with carboplatin and gemcitabine or the combination ofboth showed a superior uptake profile into the lung while systemic exposure was equivalent comparedto IV. Mediastinal lymph node concentrations are comparable with IV or tended to be higher after SPAP with carboplatin.
. Introduction
Cancer remains death cause number one after cardiovascularisease in the USA [1]. After breast and prostate cancer in women
and men, respectively, lung cancer is the most common and dead-liest cancer. Patients are treated by surgery, chemotherapy and/orradiotherapy depending on their disease stage. Seventy percent of
these patients is presented in a non-surgical stage at the moment ofinitial diagnosis and is treated by intravenous chemotherapy and orradiotherapy resulting in a 5-year survival of less than 15% [2]. Fur-thermore, (neo-)adjuvant intravenous chemotherapy results in a5-year survival benefit of only 4–14% in patients eligible for surgery
n stage III NSCLC [3]. However, the current standard treatmentor stage III NSCLC is combined chemoradiotherapy [4,5]. Two ran-omized studies concluded that surgery can be performed afterhemoradiotherapy if the mediastinum is free from tumour [6,7].
In order to improve efficacy of intravenous chemotherapy byore effective down-staging of the T (tumour) and N (lymph node)
tatus, we recently proposed a new method of drug delivery byelective pulmonary artery perfusion (SPAP) [8]. In this model, thehemotherapeutic drug is infused selectively into the affected rightr left pulmonary artery using a balloon infusion catheter result-ng in significantly higher lung levels and equivalent plasma levelsompared with intravenous infusion. We furthermore concludedhat SPAP with a physiologic pulmonary artery blood flow is opti-
al. Flow reduction of the pulmonary artery blood flow during SPAPesulted in inhomogeneous drug distribution [8].
In a subsequent study we successfully evaluated blood flowcclusion of the pulmonary artery after SPAP in order to delayhe washout of the drug from the lung into the circulation [9].ose-escalation up to 125% of the initially delivered dose resulted
n increasing lung concentrations without significantly increasinglasma levels.
Based on the pharmacokinetic profile of the drug tested, SPAPs a more efficient technique of drug delivery for the treatmentf primary lung cancer compared with intravenous infusion thatight result in more effective down-staging of the T (tumour) and(lymph node) status.In preparation of a phase 1 study, we aimed to evaluate
ediastinal lymph node exposure after SPAP with gemcitabine.urthermore, we studied pharmacokinetics of a well-known syn-rgistic combination of gemcitabine and carboplatin.
. Material and methods
.1. Animals
Twenty female Dutch Landrace pigs (mean weight: 60 ± 3.7 kg)ere used. Animals were fed with a normal diet and were treated
n accordance with the Animal Welfare Act and the “Guide forhe Care and Use of Laboratory Animals” (NIH Publication 86-23,evised 1985). The experimental protocol was approved by the ani-al experimentation committee of the Utrecht University (04/220).
.2. Anaesthesia and Euthanasia
Anaesthesia was induced with midazolam (0.5 mg/kg), atropine0.04 mg/kg) and ketamine (10 mg/kg) intramuscularly. Each ani-
al received thiopental natrium 4 mg/kg through an intravenousine. After intubation, the animals were connected to a volume-ontrolled ventilator (tidal volume of 8 ml/kg, frequency of2 breaths/min guided by capnography) maintaining positive end-xpiratory pressure of 5 cm of H2O and an inspiratory oxygenraction of 0.5. Anaesthesia was maintained by continuous infusionf midazolam (0.7 mg/kg h). Analgesia was obtained with continu-us infusion of sufentanil citrate (10 �g/kg h) and muscle relaxationith pancuronium (0.1 mg/kg h). Furthermore, a continuous infu-
ion of saline (300 ml/h) was administered during the operation.fter finishing the experiment, animals were sacrificed with pen-
obarbital natrium (200 mg/kg) intravenously.A central venous line was inserted for serum sampling during
he experiment and a catheter was introduced into the right femoralrtery for arterial blood pressure monitoring.
.3. Surgery
Initially, a balloon catheter (Balloon Wedge Pressure Catheter, 7rench, 110 cm, Arrow International, USA) was introduced through
cer 65 (2009) 208–213 209
the left internal jugular vein. The catheter was positioned intothe left pulmonary artery under blood pressure guidance. Sub-sequently, a left sided anterolateral thoracotomy was performedthrough the fifth intercostal space. The left pulmonary arterywas dissected free and the position of the balloon catheter waschecked for manually. The tip of the catheter was positioned in theleft main stem pulmonary artery just proximal of the first side-branches.
Gemcitabine and/or carboplatin were infused through thelumen of the balloon catheter (SPAP) into the left pulmonary artery,or through the central venous line (intravenous administration)using an infusion pump. Occlusion of the left pulmonary arteryafter infusion was realized by insufflating the balloon of the ballooncatheter.
Tissue samples of the lung were obtained from the left lower lobeand stored in chloroform calcium and liquid N2 for later analysis.Lymph node samples were taken from lymph node station N5 onthe left side. Furthermore, serum samples were collected from thecentral venous line and stored in tubes filled with 500 �l K2-EDTAto prevent clotting and immediately frozen into liquid nitrogen. Atthe end of the procedures, renal tissue samples were obtained inthe carboplatin groups (groups 3–5) through a left lumbotomy forconcentration analysis and histologic examination.
2.4. Experimental setting
Twenty pigs were divided into five groups (n = 4 each). Twogroups were infused intravenously with gemcitabine (IV gemci) orcarboplatin (IV carbo) during 30 min. Three other groups had SPAPduring 2 min followed by 30 min of blood flow occlusion using gem-citabine (SPAP gemci), carboplatin (SPAP carbo) or a combinationof both (SPAP gemci/carbo). All animals were sacrificed at 45 min(Fig. 1).
Serum and lung samples were obtained at 10, 20, 30 and 45 minwhile lymph node samples were taken at 10 and 20 min.
2.5. Histology
After fixation in chloroform calcium during 90 min at room tem-perature lung tissue was stored in a buffer (10 ml distilled H2O, 1 gCaCl2, 0.121 M cacodylate) at 4 ◦C until further processing. Tissuesamples for light microscopic investigations were dehydrated withisopropanolol, cleared with toluol and embedded in paraffin wax.4 �m sections were stained with haematoxylin and eosin for laterassessment.
2.6. Gemcitabine
Gemcitabine (difluorodeoxycytidine, dFdC, Ely Lilly, Indianapo-lis, USA) solutions were prepared by reconstituting non-lyophilizedpowder in saline solution. All animals were treated with gemc-itabine in a dose and volume (1250 mg/m2 body surface area, solvedin 50 ml saline) as clinically applied for the treatment of NSCLC.
2.7. Gemcitabine processing and measurement
A high-performance liquid chromatographic method has beenused and validated for the determination of gemcitabine (dFdC) inplasma, lung and kidney tissue. Before analysis of tissue and serumsamples, standard samples of blanc plasma were spiked with gem-citabine (100–100,000 ng) and extracted in the same way as the
other samples and used for a calibration curve [10]. Within-runand between-run precisions were less than 10% and average accu-racies were between 90 and 110%. Before analysis, the frozen tissueand serum samples are mixed with tetrahydrouridine in order toprevent metabolization by cytidine deaminase.
210 B.P. van Putte et al. / Lung Cancer 65 (2009) 208–213
rimental setting.
2
ru3d
2
Nlw5
2
swPc<1ws
2
mm(gui
3
a
concentrations were achieved at ten minutes for both SPAP groups(SPAP gemci: p = 0.013, SPAP gemci/carbo: p = 0.000, both comparedto IV gemci) immediately followed by a rapid dilution graph in bothgroups. IV gemci resulted in a peak concentration at 30 min that was
Table 1AUC values of gemcitabine and carboplatin for serum, lung and mediastinal lymphnode tissue. *p < 0.05, **p ≤ 0.02, ***p < 0.001.
Lung Serum Lymph node
Fig. 1. Expe
.8. Gemcitabine assay by HPLC-UV
Separation was achieved on a Chrompack Spherisorb ODS-2eversed phase column (25 m × 4.6 mm, 5 �m). The mobile phasesed was Pic B7 reagent (Waters Corporation) in 15% methanol (pH.1) with a flow rate of 1.0 ml/min. Gemcitabine is detected by UVetection at 270 nm.
.9. Carboplatin
Carboplatin (paraplatin, Bristol Myers Squibb, Woerden, Theetherlands) solutions were prepared by reconstituting non-
yophilized powder in saline solution. Like gemcitabine, all animalsere treated with cisplatin in a dose (AUC 6), solved in 50 ml glucose%, as clinically applied for the treatment of NSCLC.
.10. Analysis of carboplatin
Plasma samples are diluted six times in water and 20 �l of theolution is injected in the Atomic Absorption Spectrometer (AAS)ith pyrocoated graphitetube and cathode Pt lamp (Zeeman/3030,
erkinElmer, Shelton, USA). Quantification is performed with aalibration curve (matrix matched). The detection limit (LOQ) is10 �g/l (10 ng/ml). The variation is 6.2% for concentrations in the78–1048 �g/l range. Tissue samples (100–200 mg) are destructedith nitric acid and 3 ml of H2O is added. This solution is diluted
ix times and 20 �l is injected. The LOQ is <100 ng/�g tissue.
.11. Statistics
All concentrations shown is this paper are depicted asedian ± standard error. Lung and serum concentrations are deter-ined in function of time and calculated as areas under the curve
AUC). The AUC values and the median concentrations at each sin-le time point were compared between the different groups usingnpaired Student’s t-test using SPSS version 15.0. Statistical signif-
cance was accepted at p < 0.05.
. Results
All interventions were performed without technical problemsnd mortality.
Fig. 2. Gemcitabine lung concentrations during SPAP and IV infusion. *p < 0.05 com-pared to IV (peak concentrations).
3.1. Gemcitabine
SPAP with gemcitabine resulted in peak gemcitabine concen-trations in lung tissue at 10 min (SPAP gemci: p = 0.017, SPAPgemci/carbo: p = 0.008) followed by diminishing concentrationsthat remained significantly higher compared to IV gemci up to30 min (SPAP gemci: p = 0.04, SPAP gemci/carbo: p = 0.006) (Fig. 2).Furthermore, AUC of SPAP gemci (p = 0.02) and SPAP gemci/carbo(p < 0.001) was significantly higher compared to IV gemci from startof the experiment towards termination at 45 min (Table 1).
Fig. 3 shows the serum gemcitabine concentrations during IVand SPAP with gemcitabine. Significantly higher peak gemcitabine
B.P. van Putte et al. / Lung Cancer 65 (2009) 208–213 211
Fig. 3. Gemcitabine serum concentrations during SPAP and IV infusion. *p < 0.05compared to IV (peak concentrations). $p < 0.05 compared to SPAP GCB.
ss(
u
3
oSdIbS2hn
Fp
Fig. 6. Carboplatin serum concentrations during SPAP and IV infusion. *p < 0.05 com-pared to IV (peak concentrations).
Fig. 4. Gemcitabine lymph node concentrations during SPAP and IV infusion.
ignificantly higher compared to SPAP gemci (p = 0.03). However, noignificant differences in AUC were observed between SPAP and IVTable 1).
Gemcitabine lymph node concentrations as well as the AUC val-es did not significantly differ between SPAP and IV (Fig. 4).
.2. Carboplatin
Fig. 5 shows the carboplatin lung concentrations in functionf time. Peak concentrations were achieved at 10 min for bothPAP groups followed by a plateau phase for SPAP carbo whileiminishing concentrations were measured for SPAP gemci/carbo.
V carbo resulted in the same concentration pattern as IV gemci,oth showing peak concentrations at 30 min. Compared to IV carbo,
PAP carbo only resulted in significantly higher concentrations at0 min (p = 0.018) while SPAP gemci/carbo resulted in significantlyigher concentrations at 10 min (p = 0.016) while significance wasot achieved at 20 and 30 min. However, AUC of SPAP gemci/carbo
ig. 5. Carboplatin lung concentrations during SPAP and IV infusion. *p < 0.05 com-ared to IV (peak concentrations).
Fig. 7. Carboplatin lymph node concentrations during SPAP and IV infusion. *p < 0.05compared to IV (peak concentrations).
was significantly higher (p = 0.045) compared to IV carbo from startof the experiment up to 45 min (Table 1).
Serum concentrations after SPAP carbo are characterized bya peak concentration at 10 min followed by a dilution patternthat is comparable with SPAP gemci. At 10 (p < 0.02) and 20(p < 0.04) minutes, serum concentrations were significantly higherfor both SPAP groups compared to IV carbo (Fig. 6). Furthermore,AUC was significantly higher (p = 0.021) for SPAP gemci/carbo com-pared to IV carbo (Table 1).
Fig. 7 presents the lymph node concentrations at 10 and 20 min.SPAP carbo resulted in significantly higher concentrations at 10 mincompared to IV carbo (p = 0.04). However, no significant differenceswere observed in AUC values (Table 1).
No differences in median renal carboplatin concentrations wereobserved (IV carbo: 13 �g/g; SPAP carbo: 16 �g/g; SPAP gemci/carbo: 17 �g/g).
3.3. Histology
Microscopic evaluation of lung histology did not reveal sig-nificant differences between the IV groups and the SPAP groups(Figs. 8 and 9). However, in all groups slight edema, fibrous pleu-ritis and sometimes pleural thickening was observed. No signs ofvasculitis were present. The SPAP gemci group showed a tendencytowards more edema compared to the other groups.
4. Discussion
This study compared local and systemic drug exposure afterSPAP and IV administration of a currently widely used combina-
tion of gemcitabine and carboplatin for the treatment of non-smallcell lung cancer (NSCLC). SPAP with gemcitabine and carboplatinresulted in significantly improved drug exposure to the lungs com-pared to IV. Equivalent serum concentrations for both drugs and a
212 B.P. van Putte et al. / Lung Can
tc
muriierww[altitll
cta
bip
Fa
Fig. 8. Normal lung tissue from an animal of the Carbo IV group.
rend towards higher mediastinal lymph node concentrations forarboplatin were achieved compared to IV.
In our initial description of SPAP, we studied the impact of pul-onary blood flow reduction during drug infusion on the drug
ptake into the lung. We concluded that blood flow reductionesulted in inhomogeneous drug distribution implying that a phys-ologic pulmonary blood flow during drug infusion is necessaryn order to achieve optimal drug uptake and distribution. How-ver, immediately after SPAP, washout of the drug infused occurredesulting in rapidly diminishing lung concentrations [5]. Therefore,e applied blood flow occlusion after SPAP in order to delay theashout from the lung resulting in significantly higher AUC values
6]. In the current study we confirmed our previous findings withhigher dose of gemcitabine (1250 versus 1000 mg/m2) showing
ung and serum concentrations of gemcitabine after SPAP and IVhat reached peak concentrations at 10 min from start of the exper-ment followed by diminishing levels that remained significant upo 30 min compared to IV. From our previous studies, we know thatung concentrations are already diminishing at 10 min while peakevels were achieved at 2 min [5,6].
In this experiment, SPAP with carboplatin resulted in the sameoncentration graphs compared with gemcitabine either for lungissue and serum suggesting that tissue uptake occurs by diffusions we concluded in a former paper [5].
However, in order to study pharmacokinetic interactionsetween gemcitabine and carboplatin, we also measured gemc-
tabine concentrations when administered together with carbo-latin. Except for 10 min (p = 0.048), no significant differences were
ig. 9. Lung tissue from an animal of the Gemci IV group that shows alveolar edemand slight pleural thickening due to edema and slight chronic inflammation.
cer 65 (2009) 208–213
observed compared to gemcitabine delivery alone. This findingmore or less confirms the results of in vitro experiments in whichthe interaction between gemcitabine and cisplatin was investigated[11]. Theoretically, gemcitabine might interact with the uptake ofplatin and the binding with DNA. Otherwise, platin might interactwith the cellular uptake, the phosphorylation and the incorpora-tion into the DNA [12]. However, Van Moorsel et al. concluded thatthe best synergistic effect between the two drugs was present whengemcitabine was administered 4 h before platin [12]. These findingssuggest that pre-treatment with gemcitabine should improve thesynergistic interaction in our model.
Gemcitabine measurements in mediastinal lymph node tissuedid not reveal any differences between the different groups. Incontrast, SPAP with carboplatin resulted in significantly higher car-boplatin concentrations at 10 min while SPAP with carboplatin andgemcitabine tended to result in higher concentrations at 10 minboth compared to IV. Furthermore, extrapolation of these twographs might reveal even higher concentrations in lymph tissueat earlier time points before 10 min.
Mediastinal lymph nodes are fed by the bronchial circulationimplying that their drug exposure should be equivalent with therest of the body except for the treated lung. Our results suggest thatmediastinal drug delivery is partially dependent on the bronchialcirculation and partially by lymph drainage from the treated lung.These results confirm previous findings from a phase 1 study on iso-lated lung perfusion with melphalan. In this study the pulmonarycirculation of the affected lung was connected to a separate cir-cuit for high dose melphalan administration. Despite high lungconcentrations, high mediastinal lymph node concentrations werealso observed while no melphalan was detected in the systemiccirculation [9]. In the current study, carboplatin is possibly trans-ported from the interstitial lung tissue towards the intrapulmonaryand finally to the mediastinal lymph nodes. From the gemcitabinelymph node concentrations, we conclude that gemcitabine trans-port from the lung interstitium towards the mediastinal lymphnodes is apparently more time-consuming compared with carbo-platin transport.
In conclusion, SPAP with carboplatin and gemcitabine or thecombination of both showed a superior uptake profile into the lungwhile systemic exposure was equivalent compared to IV. Mediasti-nal lymph node concentrations are comparable with IV or tendedto be higher after SPAP with carboplatin. Therefore, we advocate aphase 1 trial for patients suffering from NSCLC studying SPAP car-boplatin with gemicitabine IV in a neo-adjuvant setting in patientswith stage 3 NSCLC.
References
[1] Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics 2007. CACancer J Clin 2007;57:43–66.
[2] Spiro SG, Silvestri GA. One hundred years of lung cancer. Am J Resp Crit CareMed 2005;172:523–9.
[3] Betticher DC. Adjuvant and neoadjuvant chemotherapy in NSCLC: a paradigmshift. Lung Cancer 2005;50:S9–16.
[4] Robinson LA, Ruckdeschel JC, Wagner Jr H, Stevens CW. American Col-lege of Chest Physicians. Treatment of non-small cell lung cancer-stageIIIA: ACCP evidence-based clinical practice guidelines (2nd edition). Chest2007;132:243–65.
[5] Jett JR, Schild SE, Keith RL, Kesler KA. American College of Chest Physicians.Treatment of non-small cell lung cancer, stage IIIB: ACCP evidence-based clin-ical practice guidelines (2nd edition). Chest 2007;132(September (3 Suppl)):266S–76S.
[6] Splinter TA, van Schil PE, Kramer GW, Van Meerbeeck J, Gregor A, RocmansP, Kirkpatrick A. Randomized trial of surgery versus radiotherapy in patientswith stage IIIA (N2) non small-cell lung cancer after a response to inductionchemotherapy.EORTC 08941. Clin Lung Cancer 2000;2:69–72.
[7] Maas KW, El Sharouni SY, Smit EF, Schramel FMNH. Sequencing chemotherapy,radiotherapy and surgery in combined modality treatment of stage III non-smallcell lung cancer. Curr Opin Pulm Med 2007;13:297–304.
[8] Van Putte BP, Grootenboers M, Van Boven WJ, et al. Pharmacokinetics ofgemcitabine when delivered by selective pulmonary artery perfusion for thetreatment of lung cancer. Drugs Metab Dispos 2008;36:676–81.
g Can
[
B.P. van Putte et al. / Lun
[9] Grootenboers MJ, Schramel FM, Van Boven WJ, Pasterkamp G, Folkerts G, VanPutte BP. Selective pulmonary artery perfusion with gemitabine and blood flowocclusion and dose escalation. Lung Cancer, 2008 Jul 24 [Epub ahead of print].
10] De Boeck G, Van Cauwenberghe K, Eggermont AM. Determination of melpha-lan and hydrolysis products in body fluids by GC–MS. High Res Chromatogr1997;12:697–700.
[
[
cer 65 (2009) 208–213 213
11] Peters GJ, Bergman AM, Ruiz Van Haperen VWT, et al. Interaction between cis-platin and gemcitabine in vitro and in vivo. Semin Oncol 1995;22(11 suppl):72–9.
12] Van Moorsel CJA, Veerman G, Bergman AM, Guechev A, Vermorken JB, Post-mus PE, Peters GJ. Combination chemotherapy studies with gemcitabine. SeminOncol 1997;24(7 suppl). S1-17-S7-23.
