www.thelancet.com/oncology Vol 9 December 2008 1181

Review

Part II: Liver function in oncology: towards safer chemotherapy useKathryn M Field, Michael Michael

The liver is fundamentally important in drug metabolism. In oncology, the astute clinician must not only understand the meaning and limitations of commonly ordered liver biochemical tests, but also be aware of which anticancer agents might induce liver dysfunction, and of the strategies for appropriate dosing of patients with pre-existing liver dysfunction. In part I of our Review, we highlighted both the importance and inadequacies of identifying serum biochemical liver abnormalities in oncology; we also discussed a lack of routine formal investigation of liver function. We summarised chemotherapy-related hepatotoxicity and other causes of liver toxic eff ects in patients with cancer. Here in part II, we discuss trials that have specifi cally assessed chemotherapy dosing strategies in the setting of overt biochemical liver dysfunction and we note their recommendations. Furthermore, we review other assessments of liver metabolic and excretory function, particularly in the setting of chemotherapy drug handling. We discuss the potential use of these metabolic probes in practice.

IntroductionPrimum non nocere—to fi rst, do no harm—is a fundamental law of medicine learnt by all clinicians. In oncology, chemotherapy is given to patients with the inherent premise of an anticancer eff ect; however, some of the most toxic medications in clinical practice are used, and there can be a narrow margin between benefi t and risk. Their therapeutic index is made even narrower by the presence of pre-existing organ impairment such as liver or kidney dysfunction, which can increase the risk of drug toxic eff ects. Alternatively, liver impairment could reduce the activation of prodrugs, reducing the chance of a meaningful response. Liver injury and the resulting biochemical abnormalities are common in patients with cancer, and can be due to metastases (fi gure 1), concomitant drugs, or past liver damage such as cirrhosis (fi gure 2). The choice of anticancer agent and its dose must be considered carefully to meet the pledge made to patients—to do more good than harm.

Underlying liver dysfunction and anticancer agentsMany chemotherapy agents are substrates for liver uptake, metabolism, and excretion, and dysfunction of the liver can result in drug toxic eff ects (or reduced activation of prodrugs). Is the eff ect substantial? Pre-existing liver impairment (other than in Child’s class C cirrhosis) is thought to have only modest eff ects on the elimination and toxic eff ects of drugs.1 Nevertheless, a typical patient presents with several variable causes of liver impairment, including direct hepatic involvement by their disease, eff ects of cancer-induced infl ammatory cytokines, comorbidities, and concurrent medications (both prescription and complementary).

Patients with aberrant chemotherapy drug metabolism have substantial risk of severe haematological and non-haematological toxic eff ects: caution must be exercised with specifi c chemotherapy drugs. However, substantial dose reductions might reduce therapeutic eff ectiveness.

Many phase I, phase II, and phase III clinical trials have excluded patients with abnormal liver function (ie, with abnormal serum liver biochemical tests). Therefore, less is known about the most appropriate starting doses of chemotherapy in these individuals. Anticipated toxic eff ects based on trials cannot necessarily be extrapolated to this group of patients, and commonly dose reductions are empirical and not evidence-based.

Over the past decade, several trials have assessed chemotherapy drug pharmacokinetics specifi cally in patients with liver dysfunction. The trials have aimed to develop specifi c dosing guidelines on the basis of standard biochemical tests of liver function. Patients have been entered into distinct dose cohorts defi ned by a combination of standard serum liver biochemistry results. Dose can be escalated if there is no evidence of dose-limiting toxic eff ects. Of note, the numbers in the cohorts are small,

Lancet Oncol 2008; 9: 1181–90

Division of Haematology and Medical Oncology, Peter MacCallum Cancer Centre, Melbourne, VIC, Australia (K Field MBBS Hons, M Michael MBBS Hons)

Correspondence to: Dr Michael Michael, Division of Haematology and Medical Oncology, Peter MacCallum Cancer Centre, Locked Bag 1, Melbourne, VIC 8006, AustraliaMichael.Michael@petermac.org

For Part 1: Liver function in oncology: biochemistry and beyond see Lancet Oncol 2008; 9: 1092–101

Figure 1: Liver metastases in a patient with stage IV colorectal cancerImage is colour-volume rendering of a CT scan with anterior cut plane. Note two large liver metastases (arrows).

1182 www.thelancet.com/oncology Vol 9 December 2008

Review

and these trials are only a small evidence base for use of these agents in liver dysfunction. Table 1 summarises the main drugs for which pharmacokinetic studies have been done in the setting of liver dysfunction,2–21 and shows dose recommendations from the trial researchers on the basis of available evidence. Some of the more commonly used drugs that have been assessed in this setting are discussed below.

PaclitaxelA member of the taxane family, paclitaxel is mainly metabolised by the liver through the cytochrome 2C8 and 3A4 pathways; it undergoes biliary excretion through the action of ABCB1 (ATP binding cassette subfamily B member 1).

A prospective study18 of toxic eff ects and pharmaco-kinetics in 81 patients grouped into three cohorts according to liver dysfunction showed that patients with abnormal bilirubin (ie, >25 μmol/L) or aspartate aminotransferase levels 2-times or higher than the upper limit of normal had more toxic eff ects (particularly myelo suppression) compared with other groups and warranted routine dose-reduction. Compared with historical controls, those who received a 3-h paclitaxel infusion (the most common method of administration) had protracted plasma paclitaxel concentrations that exceeded 0∙05 μmol/L—the pharmaco-kinetic measure associated with reduction in neutrophils.

DocetaxelAlso a taxane, docetaxel undergoes hydroxylation by cytochrome P450 3A4 and is mainly faecally excreted. A prospective study6 assessed the association between serum liver biochemistry, cytochrome P450 3A4 activity (via the erythromycin breath test), and docetaxel clearance in patients with cancer. Patients were divided into cohorts on the basis of baseline serum liver biochemistry, ranging from normal to severely abnormal, and received diff erent initial docetaxel doses. Drug clearance was lower in patients with moderate to severely abnormal liver tests than in those with normal to mildly abnormal tests. Five of eight patients with normal liver tests but with cytochrome P450 3A4 activity lower than the norm had reduced docetaxel clearance. The researchers concluded that liver biochemical tests alone do not explain variability in docetaxel clearance.22

GemcitabineA pyrimidine analogue, gemcitabine is metabolised intracellularly by nucleoside kinases, and mainly excreted in urine. In one study, gemcitabine was analysed in a cohort of 40 patients with liver impairment.12 Patients with elevated aspartate aminotransferase (ie, ≥2-times the upper limit of normal) did not have increased toxic eff ects with gemcitabine. However, patients with elevated bilirubin levels (ie, >27 μmol/L) had substantial liver dysfunction induced by gemcitabine and dose reductions were recommended. There were no apparent pharmacokinetic diff erences between groups.

IrinotecanA topoisomerase I inhibitor prodrug that is metabolised to the active SN38 by carboxyl esterase, irinotecan excretion is mainly biliary after inactivation from glucuronide con-jugation by hepatic uridine diphosphate glucurono-syltransferase-1A1 (UGT1A1) to generate SN38G. Irino tecan is also inactivated through oxidation of its bipiperidine

A

B

Figure 2: Healthy liver (A) and liver cirrhosis (B) (A) Note histology of healthy liver: portal tract contains bile duct, hepatic artery, and portal vein (upper right); central vein shown (lower left); and hepatocytes line sinusoids in centre of image. Cords and plates of hepatocytes are one-cell thick between sinusoids. Haematoxylin and eosin stain, magnifi cation ×200. (B) Disrupted liver architecture with fi brous tissue (deep red). Nodules of hyperplastic parenchyma encircled by mature fi brous septa. Portal tracts have become incorporated into septa, and no normal central veins remain. Van Gieson stain, magnifi cation ×100.

www.thelancet.com/oncology Vol 9 December 2008 1183

Review

sidechain by cytochrome P450 3A4,23 generating the inactive APC (7-ethyl-10-[4-N-(5-amino pentanoic-acid)-1-piperidino] carbonyloxycamptothecin) and NPC (7-ethyl-10-[4-amino-1-piperidino] carbonyloxycamptothecin).

Two phase 1 studies14,15 assessed the pharmacokinetics of irinotecan given every 3 weeks in patients with varying degrees of liver dysfunction. Raymond and colleagues15 grouped patients according to serum

bilirubin levels. Elevated bilirubin (ie, >1∙5-times upper limit of normal) was associated with irinotecan toxic eff ects (grade 4 febrile neutropenia and diarrhoea). Increased bilirubin and alkaline phosphatase were associated with decreased irinotecan clearance (and that of its metabolites SN38, SN38G, and APC).15 The authors concluded that decreased clearance was related to decreased biliary excretion.

Eff ect of liver impairment Dose modifi cation Ref

Capecitabine Increased aspartate aminotransferase or bilirubin: no relation to pharmacokinetics or toxic eff ects

No dose adjustment needed 2

Cisplatin and carboplatin No published studies Unlikely to need dose adjustments; mainly renal excretion 3

Cyclophosphamide No changes in clearance No dose adjustments needed 4

Docetaxel Increased risk of neutropenia, mucositis, and deathIncreased bilirubin, with or without raised transaminases: 12–27% decreased drug clearance

Use not recommended if bilirubin is more than upper limit of normal, or if ratio of aspartate aminotransferase to alanine aminotransferase is >1·5-times upper limit of normal and alanine phosphatase is >2·5-times upper limit of normal

5,6

Doxorubicin Myelosuppression; mucositis Bilirubin <51 μmol/L: normal doseBilirubin 34–51 μmol/L: decrease dose by 50%Bilirubin 51–85 μmol/L: decrease dose by 75%Bilirubin >85 μmol/L: withhold treatment

4777

Epirubicin Aspartate aminotransferase more sensitive marker of clearance than bilirubin

Consult dose guidelines based on levels of aspartate aminotransferase 8

Erlotinib Increased aspartate aminotransferase or bilirubin Aspartate aminotransferase ≥3-times upper limit of normal or bilirubin 17–120 μmol/L: 50% dose reduction

9

Etoposide Mild to moderate impairment: no pharmacokinetic eff ectSevere impairment: myelosuppression; mucositisDecreased albumin increases unbound drug concentration and increases haematological toxic eff ects

Unclear (increased renal clearance might compensate) 10

Fluorouracil Increased bilirubin: no relation to toxic eff ects No dose adjustment needed 11

Gemcitabine Increased aspartate aminotransferase alone: no increase in toxic eff ectsIncreased bilirubin: deterioration in liver function

Usual dose: 1000 mg/m2Increased aspartate aminotransferase: no dose change neededIncreased bilirubin: reduce dose by 20% (ie, to 800mg/m2) and increase if tolerated

12

Imatinib No notable pharmacokinetic diff erences or increased toxic eff ects Stop treatment if hepatotoxicity develops; should probably not rechallenge 13

Irinotecan Increased aspartate aminotransferase alone: no increase in toxic eff ectsIncreased bilirubin: neutropenia and diarrhoea

Increased aspartate aminotransferase: no dose change Increased bilirubin: reduce dose: 3-weekly irinotecan (usual dose 350 mg/m2 every 3 weeks) • Bilirubin <1·5-times upper limit of normal: 350 mg/m2 • Bilirubin >1·5–3-times upper limit of normal: 200 mg/m2 • Bilirubin >3-times upper limit of normal: irinotecan not recommendedWeekly irinotecan (usual dose 125 mg/m2 for 4 of 6 weeks) • Bilirubin 1·5–3-times upper limit of normal and ratio of aspartate aminotransferase to alanine aminotransferase <5-times upper limit of normal: 60 mg/m2 • Bilirubin 3·1–5-times upper limit of normal and ratio of aspartate aminotransferase to alanine aminotransferase <5-times upper limit of normal: 50 mg/m2 • Bilirubin <1·5-times upper limit of normal and ratio of aspartate aminotransferase to alanine aminotransferase 5·1–20-times upper limit of normal: 60 mg/m2 • Bilirubin 1·5–3-times upper limit of normal and ratio of aspartate aminotransferase to alanine aminotransferase 5·1–20-times upper limit of normal: 40 mg/m2

14–16

Oxaliplatin Increased bilirubin, aspartate aminotransferase, or alkaline phosphatase: no eff ect on drug clearance or neurotoxicity

No dose adjustment 17

Paclitaxel Increased aspartate aminotransferase or bilirubin increases myelosuppression

Reduce dose if increased aspartate aminotransferase or increased bilirubin 18

Sorafenib Clearance does not diff er between patient cohorts Bilirubin ≤1·5-times upper limit of normal: 400 mg twice a dayBilirubin 1·5–3-times upper limit of normal: 200 mg twice a dayBilirubin 3–10-times upper limit of normal: sorafenib not recommendedAlbumin <25 g/L: 200 mg daily

19

Topotecan No obvious eff ects No dose adjustment needed if bilirubin 29–84 μmol/L 20

Vinorelbine Increased bilirubin decreases drug clearanceVolume of liver aff ected correlates with clearance

Suggested dose:Bilirubin 2·1–3-times upper limit of normal: reduce dose by 50%Bilirubin >3-times upper limit of normal: reduce dose by 75%Diff use liver metastases: decrease dose by 50% (irrespective of bilirubin concentration)

21

Table 1: Anticancer agents specifi cally studied in setting of liver dysfunction

1184 www.thelancet.com/oncology Vol 9 December 2008

Review

Venook and colleagues14 found that those with increased aspartate aminotransferase (ie, ≥3-times upper limit of normal) but normal bilirubin did not have toxic eff ects or changes in drug clearance. However, those with increased bilirubin (ie, 17∙1–120 μmol/L) had dose-limiting toxic eff ects, with decreased irinotecan and SN38 clearance.

Schaaf and colleagues16 assessed irinotecan pharma-cokinetics in patients who were prospectively classifi ed into four groups on the basis of baseline serum bilirubin, alanine aminotransferase, and aspartate aminotransferase, and who were given low-dose weekly irinotecan ranging from 40 mg/m² to 75mg/m². Compared with controls, all groups had reduced irinotecan clearance (by 30–70%) and an increased dose-adjusted SN38 area under the curve (AUC, 33–216%). There was a strong linear correlation between irinotecan AUC or SN38 AUC and serum bilirubin, which plateaued with increasing aspartate aminotransferase. These studies14–16 and others24 show that irinotecan clearance correlates with serum bilirubin more so than with other biochemical markers of liver function.

ErlotinibAn oral epidermal growth factor receptor-1 associated tyrosine kinase inhibitor for use in non-small-cell lung cancer, erlotinib is mainly metabolised by liver cytochrome P450 3A4 and to a lesser degree by cytcochrome P450 1A2.25

A phase I study9 analysed this drug’s pharmacokinetics in patients with liver or kidney dysfunction. The study showed a longer half-life, reduced clearance, and increased likelihood of dose-limiting toxic eff ects (particularly with a bilirubin rise to ≥1∙5-times baseline [n=3] and grade 4 diarrhoea [n=1]) in the 36 patients with liver dysfunction (defi ned as aspartate aminotransferase ≥3-times upper limit of normal, with or without albumin <25 g/L [n=6], or bilirubin 17–120 μmol/L [n=30]).

This study recommended a 50% dose reduction in patients with this level of liver dysfunction. A pharmaco-kinetic analysis26 of more than 1000 patients who received erlotinib found that total bilirubin, in addition to α-1 acid glycoprotein and smoking status, were the most important factors that aff ected drug clearance.

FluorouracilThis drug is eliminated by liver and peripheral metabolism through dihydropyrimidine dehydrogenase. Fluorouracil has been assessed11 in 64 patients with solid tumours who had liver dysfunction (ie, bilirubin 25–85 μmol/L or >85 μmol/L) in a phase I setting. Patients received a 24-h infusion of fl uorouracil and leucovorin on a weekly regimen. The study recorded no relation between bilirubin and fl uorouracil clearance or toxic eff ects. This fi nding suggests that infused fl uorouracil is fairly safe to use in patients with liver dysfunction, although regular monitoring with liver tests is advised.11

However, there have been reports of substantial toxic eff ects when fl uorouracil has been given to patients with

serum bilirubin levels higher than 85 μmol/L,27 and case reports exist of this drug potentially contributing to liver dysfunction.28

SorafenibAn orally available multi-kinase inhibitor, sorafenib is eff ective against renal-cell carcinoma and hepatocellular carcinoma. It is mainly metabolised in the liver through both an oxidative pathway mediated by cytochrome P450 3A4 and also UGT1A9-mediated glucuronidation; most is faecally excreted.

A phase I study19 of sorafenib in patients with liver dysfunction who were divided into groups on the basis of bilirubin levels, albumin levels, and renal function found that although drug clearance did not diff er between groups, dose-limiting toxic eff ects occurred with increasing bilirubin levels and decreased albumin (ie, <25 g/L).

Other methods to estimate liver functionIt would be best to establish an individualised chemo-therapy dosing nomogram on the basis of patient phenotypic or genomic metabolic, excretory, and pharma-co dynamic diff erences. The current method of chemo-therapy dose determination on the basis of body surface area does not take into consideration variability between patients and has poor correlation to drug clearance.

Alternative, but feasible, methods for safe and eff ective chemotherapy dosing are needed. Table 1 shows trials and their recommended routine dose reductions in patients with overt biochemical liver dysfunction; however, they do not predict dose adjustments that might be needed because of individual variability in liver-disposition pathways that are not functionally tested by serum biochemistry. Given such limitations, several dosing strategies have been assessed, many of which have been reported in the general pharmacology literature and are used in critical care, liver transplantation, or assessment of residual liver function when planning liver resection. Data are emerging for these tests in the setting of chemotherapy administration.

Many alternative tests are in vivo or indirect probes of hepatic cytochrome P450 activity. Of note, diff erent probes for the same enzyme do not always correspond to each other, given diff erences in substrate avidity. Moreover, other factors are involved in drug metabolism or excretion, including regulation and activity of other Phase I enzyme reactions (eg, oxidation/reduction and hydrolysis) and Phase II metabolic reactions (eg, glucuron idation, sulphation, or glutathione conjugation), and biliary excretion through ABC (ATP-binding cassette) membrane pumps (eg, ABCB1, ABCC1, ABCC2, and ABCG2).29

The applicability of such probes to clinical practice has been questioned.30 However, to date they relate to pharmacokinetics for many anticancer agents, and many believe they will continue to be used in research, drug development, and eventually in routine clinical practice.31 These probes are outlined in table 2.32–55

www.thelancet.com/oncology Vol 9 December 2008 1185

Review

Breath testsThe Carbon-14 N-methyl-erythromycin breath test is an in vivo probe of liver enzyme function. It is a quantitative measure of hepatic cytochrome P450 3A4 activity, and measures phenotypic variance in enzyme activity due to genetic polymorphisms, disease factors, or concomitant drugs.

Erythromycin is a substrate for cytochrome P450 3A4, but it is also transported by ABCB1. Hepatic cytochrome P450 3A4 demethylates intravenous ¹⁴C-(N-methyl)-erythromycin to generate ¹⁴CO₂. The test is done by intravenous administration of trace ¹⁴C-erythromycin and measurement of amount of exhaled ¹⁴CO₂. It has been the most widely assessed probe in oncology.

A study33 of 134 patients with cancer analysed factors that aff ected cytochrome P450 3A activity by use of the erythromycin breath test. Enzyme activity varied by up to 14-times in patients. In a multivariate analysis, enzyme activity was not signifi cantly aff ected by patient demographics, but liver function combined with the concentration of the acute-phase reactant α-1 acid glycoprotein accounted for about 18% of overall variation in cytochrome activity (p<0∙001).33 The researchers

concluded that baseline demographic, physiological, and genetic polymorphisms have a minor eff ect on phenotypic cytochrome P450 3A activity in patients with cancer.

DocetaxelDocetaxel is one of the most commonly studied anticancer agents for these probes.56 Erythromycin breath-testing has been extensively investigated for optimisation of docetaxel dosing.

One study34 measured baseline hepatic cytochrome P450 3A4 activity by use of the erythromycin breath test in 21 patients who subsequently received docetaxel every 3 weeks. The breath-test measurement (proportion of ¹⁴CO₂ exhaled in 1 h) was the best predictor of docetaxel clearance compared with serum liver biochemistry and α-1 acid glycoprotein, accounting for 67% of variability between patients. Two patients with the lowest breath-test measurement had the lowest docetaxel clearance and the most severe toxic eff ects.

A model based on this breath-test measurement and serum albumin has been assessed for tailored docetaxel dosing in 28 patients with advanced breast cancer.57 Baker and colleagues22 found that erythromycin breath-testing

Metabolic pathway assessed Chemotherapy drugs assessed

Clinical relevance References

Breath tests

Erythromycin breath test Cytochrome P450 3A4 DocetaxelIrinotecanVinorelbine

Predicts clearance more accurately than liver biochemical testsDoes not predict clearanceDoes not predict clearance

32–36

2-¹³C-uracil breath test and 2-¹³C-fl uorouracil breath test

Dihydropyridine dehydrogenase Fluorouracil Sensitive for determination of dihydropyridine dehydrogenase activity and therefore prediction of fl uorouracil-related toxic eff ects

37,38

Drug-clearance tests

Midazolam plasma concentration Cytochrome P450 3A DocetaxelIrinotecanVinorelbine

Low clearance associated with severe neutropeniaCorrelates with irinotecan clearanceFractional survival of neutrophils correlates with clearance

36,39–42

Indocyanine green plasma disappearance rate

Liver perfusion and excretion Irinotecan Area-under-curve of irinotecan and metabolite SN38 correlates with indocyanine green plasma disappearance rate

43

Dexamethasone plasma concentration

Cytochrome P450 3A DocetaxelVinorelbine

Can predict clearance in females but not malesGood predictor of clearance

44,45

Urinary steroid metabolite excretion Cytochrome P450 3A Docetaxel Associated with drug clearance 46,47

Antipyrine clearance test Several cytochrome P450 enzymes Measure of overall liver oxidative capacity

EpirubicinDocetaxel, cisplatin

Associated with clearance and epirubicin-related neutropeniaCorrelates with neutrophil nadir

48,49

Imaging

Technetium-99-sestamibi ([99mTc]MIBI)

Biliary transporter-mediated excretion; mainly assesses ATP-binding cassette family member ABCB1, but also ABCC1 and ABCC2

VinorelbineIrinotecanImatinib

Independent predictor of clearanceDecreased clearance possibly associated with increased SN38 exposureCan identify particular ABCB1 genotypes associated with imatinib elimination, but drug clearance not signifi cantly associated with MIBI clearance

41,42,50,51

Technetium-99- acetanilidoiminodiacetic acid ([99mTc]DIDA) and technetium-99-disofenin ([99mTc]DISIDA)

Biliary transporter-mediated excretion; assesses ABCC1 and ABCC2

Irinotecan Decreased clearance possibly associated with increased SN38 exposure 50

Pharmacogenomics

Single nucleotide polymorphisms Polymorphisms of drug-metabolising enzymes

Multiple—eg, fl uorouracil, paclitaxel, vinorelbine, and docetaxel

Ideal for personalised medicine, but expensive; needs ongoing validation studies

35,52–55

Table 2: Alternative liver-function tests in chemotherapy setting

1186 www.thelancet.com/oncology Vol 9 December 2008

Review

was a more accurate predictor of docetaxel clearance than was serum biochemical testing of liver function. The association between the breath-testing and docetaxel clearance was confi rmed with a weekly docetaxel regimen.58 Erythromycin breath-testing could eventually lead to more selective use of docetaxel in practice. However, this test is not associated with vinorelbine clearance.35

FluorouracilToxic eff ects with this drug might be related to defi ciency or reduced activity of dihydropyridine dehydrogenase, which mediates catabolism of up to 80–90% of the drug to 5-fl uorodihydrouracil and after several other steps to ammonia and CO₂. Up to 5% of the general population have complete or partial defi ciency of dihydropyridine dehydrogenase,59 and its defi ciency in patients with cancer is associated with severely reduced fl uorouracil clearance and increased toxic eff ects (some fatal).

However, not all patients who present with severe fl uorouracil-induced toxic eff ects are dihydropyridine-dehydrogenase defi cient.54 The 2-¹³C-uracil breath test is a phenotypic screen for dihydropyridine-dehydrogenase defi ciency in patients who receive fl uorouracil and has 96% specifi city and 100% sensitivity.37 Furthermore, in one study,38 a 2-¹³C-fl uorouracil breath test for assessment of dihydropyridine-dehydrogenase activity identifi ed patients with severe and moderate toxic eff ects.

Drug-clearance tests Midazolam clearanceAn alternative method of assessment of hepatic cytochrome P450 3A activity, midazolam is almost entirely metabolised by this enzyme family and is not an ABCB1 substrate. It has been analysed in patients with cancer receiving docetaxel. A study39 of 56 patients with advanced cancer with normal liver function found that high midazolam concentration or low clearance (indicating low cytochrome activity) was associated with severe neutropenia and febrile neutropenia.

Another study36 of 30 patients compared the midazolam-clearance test with the erythromycin breath test for assessment of irinotecan clearance. Cytochrome P450 3A4 activity varied by seven-times between patients on erythro-mycin breath-testing, but was not associated with irinotecan clearance. Midazolam clearance varied four-times, but was strongly associated with irinotecan clearance. This parallel could arise because erythromycin is rapidly metabolised by cytochrome P450 3A4, and ABCB1 might play a part in its disposition compared with irinotecan.40

Plasma and urinary steroid clearanceAs an in-vivo probe of cytochrome P450 3A metabolism, specifi cally in patients with cancer, this method has been studied as a predictor of vinorelbine clearance in 20 patients.45 Measurement of dexamethasone plasma clearance and alkaline phosphatase in a covariate model signifi cantly decreased variability between individuals in

vinorelbine clearance, by contrast with dosing based on body surface area. Dexamethasone as a probe for cytochrome P450 3A4 predicted docetaxel clearance in females, but not males.44

Steroid metabolite excretion in urine has been used as a surrogate of cytochrome P450 3A4 activity. 24-h urinary 6-β-hydroxycortisol level after cortisol admin-istration is signifi cantly associated with docetaxel clearance.46 Of 50 patients who were randomly assigned a docetaxel dose based on body surface area or an individualised dose based on clearance estimated from the probe, the latter was able to decrease the pharmacokinetic variability of docetaxel.47

Indocyanine green plasma disappearance rateIndocyanine green is a water-soluble dye given intravenously, which is almost completely eliminated unchanged into bile without enterohepatic circulation. Elimination is dependent on hepatic blood fl ow, cellular uptake, and excretion.60 As a marker of liver perfusion and function, it is a good predictor of survival in critically ill patients.60 This dye is thought to be more sensitive than serum bilirubin or enzyme tests for assessment of liver dysfunction and prediction of prognosis,61 and has been used to assess liver functional reserve before resection.62

The probe has been assessed in the setting of irinotecan use, where AUC signifi cantly correlates with indocyanine green plasma disappearance rate; AUC of SN38, its active metabolite, also correlates with disappearance rate.43

However, to our knowledge there have been no published investigations of the effi cacy of this test for prediction of toxic eff ects from, or pharmacokinetics of, other anticancer agents.

Antipyrine clearance test Antipyrine is an analgesic that is metabolised by several cytochrome P450 enzymes (1A2, 2B6, 2C, and 3A4), and is widely used as a measure of overall liver oxidative capacity.63 Antipyrine has proved useful for assessment of liver meta-bolic function in various liver diseases—before and after transplantation and for assessment of drug interactions.

The test is simple and relatively inexpensive. In patients with cancer, Gurney and colleagues48 found that the antipyrine-clearance test correlated with epirubicin AUC and clearance, and with drug-induced neutropenia. Activated partial thromboplastin time, international normalised ratio, and serum bilirubin correlated with some epirubicin pharmacokinetics, but aminotransferase levels did not. Importantly, body surface area was not associated with pharmacokinetic parameters or degree of neutropenia. A relation between antipyrine clearance and neutrophil nadir has also been shown in patients who received docetaxel and cisplatin in combination.49

Liver functional imagingLiver functional imaging aims to provide a non-invasive method of assessing drug uptake and excretion by the liver.

www.thelancet.com/oncology Vol 9 December 2008 1187

Review

This imaging uses radiotracers that have the same liver cellular transport carriers as those for a particular cytotoxin.

Biliary transporter-mediated excretion has an important role in drug clearance by the liver; several enzymes are involved including ABCB1, ABCC1, ABCC2, and ABCG2. Their substrates include vinca alkaloids, anthracyclines, taxanes, and irinotecan.51 Variable activity of these enzymes could contribute to variation between patients in drug clearance from the liver and systemic circulation.

Functional imaging by use of technetium-99-sestamibi ([⁹⁹mTc]MIBI), an ABCB1 substrate (and to a lesser extent ABCC1 and ABCC2) has been assessed in patients with cancer.51,52 Comparisons of liver [⁹⁹mTc]MIBI elimination rate with ABCB1 genotype in 66 patients found 12-times variation in the elimination constant for [⁹⁹mTc]MIBI, which did not correlate with test results of serum biochemical liver function. There was a signifi cant association between some common single nucleotide polymorphisms in ABCB1 and elimination constant.51 This nuclear imaging could be a pretreatment indicator of drug clearance.

[⁹⁹mTc]MIBI liver imaging has been assessed with vinorelbine treatment in a study of 41 patients.42 The study used the midazolam-clearance test (to assess cytochrome P450 3A phenotype), [⁹⁹mTc]MIBI (to assess ABCB1 phenotype), and genotype testing for both enzymes to predict vinorelbine clearance and frequency of myelosuppression. [⁹⁹mTc]MIBI clearance (a product of liver volume and MIBI elimination rate constant) was an independent predictor of vinorelbine clearance, which signifi cantly correlated with fractional survival of neutrophils. Midazolam clearance did not correlate with vinorelbine pharmacokinetics or toxic eff ects. An alternative method of vinorelbine dosing could be a fi xed dose guided by these pre-treatment investigations of liver function.42

The oral tyrosine kinase inhibitor imatinib is mainly metabolised by liver cytochrome P450 3A enzymes and eliminated unchanged in bile, presumably via ABCB1 and interaction with ABCG2.64 Several researchers have assessed the pharmacogenomics of enzymes in imatinib metabolism, but genetic testing can be expensive. Therefore, alternative probes for phenotypic diff erences in enzyme activity are under investigation. 22 patients treated with imatinib were assessed for pharmacokinetic data by use of erythromycin breath-testing, midazolam clearance (to assess cytochrome P450 3A), [⁹⁹mTc]MIBI liver clearance, and genotyping for single nucleotide polymorphisms in CYP3A and ABCB1.41 Erythromycin breath-testing and midazolam clearance were associated with imatinib clearance, but [⁹⁹mTc]MIBI liver clearance was not. However, a reduction in [⁹⁹mTc]MIBI clearance was seen in patients with some ABCB1 genotypes, suggesting that this test might benefi t some individuals.

Two types of liver nuclear imaging have been assessed in relation to irinotecan pharmacokinetics: [⁹⁹mTc]DIDA (acet anilidoimino diacetic acid) or

[⁹⁹mTc]DISIDA (disofenin)—both of which are substrates for ABCC1 and ABCC2—and [⁹⁹mTc]MIBI. These substrates are excreted by the same biliary transporters involved in excretion of irinotecan and its active metabolite SN38. An exploratory study50 in 21 patients with colorectal cancer who received irinotecan used [⁹⁹mTc]MIBI, iminodiacetic acid analogue-imaging, and relevant genotyping to identify correlates with drug pharmacokinetics. For both nuclear probes, the proportion retained at 1 h correlated linearly with SN38 AUC, suggesting that reduced tracer clearance could be associated with increased SN38 exposure which could be a predictor of irinotecan toxic eff ects. Increasing grade of neutropenia was associated with reduced iminodiacetic acid hepatic-extraction fraction (a measure of drug uptake by the liver), but not signifi cantly so. Increased liver uptake and rapid clearance of MIBI was associated with severity of diarrhoea.50 However, this study analysed few patients and further investigation of the effi cacy of functional imaging is warranted.

Scoring systems for severe liver diseaseGastroenterologists and hepatologists commonly assign a score to patients with chronic liver disease, on the basis of validated systems, to estimate outlook in terms of disease burden and ability to withstand the stress of major surgery. These systems generally use a combination of biochemical and clinical assessments, but are used less often in oncology. Nevertheless, awareness of alternative methods of assessing liver functional impairment can be useful. If remaining liver capacity might necessitate transplantation, then chemotherapy should probably not be given. In patients with cancer other than hepatocellular carcinoma, transplantation is rare because malignant disease is almost always a contraindication.

Child-Turcotte-Pugh scoreThis score is a combined measure based on extent of clinical ascites, encephalopathy, and laboratory indices including albumin, international normalised ratio, and bilirubin. It was originally developed to assess the ability of patients with cirrhosis from alcohol use (fi gure 2) and oesophageal varices to withstand portacaval shunt surgery to reduce the risk of bleeding.65,66 However, except for hepatocellular carcinoma, the Child-Turcotte-Pugh score is not generally used in oncology to determine chemotherapy dosing. A phase I study9 assessed erlotinib in liver or kidney dysfunction and found no relation between score and erlotinib clearance, although increased aspartate aminotransferase or bilirubin predicted drug toxic eff ects.

Model for end-stage liver disease (MELD) This scoring system was developed in 2000,67 and has been used since 2002 to prioritise patients for liver transplantation68 by use of a mathematical equation that incorporates bilirubin, international normalised ratio, and serum creatinine. In oncology, the model for end-stage

1188 www.thelancet.com/oncology Vol 9 December 2008

Review

liver disease has mainly been used to prioritise patients with hepatocellular carcinoma for transplantation; in the USA it has led to an increase in the proportion of transplants due to hepatocellular carcinoma.69 Survival after transarterial chemo embolisation for hepatocellular carcinoma is associated more closely with Child-Turcotte-Pugh score than with the score from the model for end-stage liver disease.70

PharmacogenomicsPharmacogenomics has been studied with much interest with regard to chemotherapy drugs and has expanded in recent years because of increased knowledge from studies such as the Human Genome Project. Tests can detect single nucleotide polymorphisms and defi ciencies in enzymes that are commonly involved with liver meta-bolism (eg, cytochrome P450 and UGT1A1) and excretion (eg, ABCB1). Many studies have shown associations between particular genotypes and drug clearance or metabolism. Genetic factors that are thought to contribute to drug toxic eff ects include dihydropyridine-dehydrogenase defi ciency resulting in toxic eff ects from fl uorouracil, and UGT1A1 defi ciency resulting in irinotecan toxic eff ects.

Genetic heterogeneity means that although quantitative enzyme defi ciency might not be present, polymorphisms can lead to variable clearance in individuals as identifi ed by genetic or phenotypic tests. Drug eff ectiveness and toxic eff ects might be predicted by genotyping of pharmaco-dynamic pathways.71,72 However, it would be simplistic to assume that all variation in drug handling or eff ect can be described by the characterisation of one metabolic or excretory pathway, and the following need consideration: other pathways; diff erences in host and tumoral drug sensitivity; the paraneoplastic eff ects of the tumour on host metabolic phenotype; and variation in populations or between ethnic groups.

Not all pharmacogenomic studies have shown signifi cant associations with drug pharmacokinetics. Fluorouracil dosing on the basis of genotyping its main metabolic enzyme (dihydropyridine dehydrogenase) is one example. Activity of this enzyme and the presence of a common splice-site mutation in patients with severe fl uorouracil-induced grade 3–4 toxic eff ects54 showed only 60% had decreased enzyme activity in peripheral-blood mononuclear cells, 42% of which were heterozygous for the variant and one patient homozygous. The splice-site mutation was also present in 4% of patients with normal dihydropyridine-dehydrogenase activity.54 Therefore, defi ciency of this enzyme is not solely due to inactivating mutations.52,54,73 Moreover, toxic eff ects from fl uorouracil occur in patients with normal dihydropyridine-dehydrogenase activity, suggesting that other genes of pyrimidine catabolism might have a role.

Variability between individuals in paclitaxel pharmaco-kinetics was not explained by alleles for CYP2C8, CYP3A4, CYP3A5, and ABCB1 from DNA analyses of 97 patients.74 A study of 25 patients that assessed vinorelbine

pharmacokinetics found no relation between cytochrome P450 3A5 activity or erythromycin breath-testing and vinorelbine clearance or toxic eff ects.35

Therefore, current pharmacogenomic analyses might aid only the dosing for some anticancer agents and possibly selected patients. For example, alleles in ABCB1 are associated with docetaxel-induced neuropathy, grade of neutropenia, and overall survival in patients with prostate cancer.55

The ideal use of such tests would be to reduce the frequency and severity of adverse drug reactions. Genetic variation is thought to cause more than 50% of these reactions.75 Although costs might prevent genetic testing, money could be saved from treatment of adverse eff ects, and so too could lives, by further development of alternative methods of assessing liver function and predicting appropriate dose on an individual basis.

ConclusionThe past decade has seen increased recognition of the limitations of current biochemical tests of liver function to predict hepatotoxicity and other systemic toxic eff ects from chemotherapy. Furthermore, understanding and character-isation of genetic polymorphisms has been enhanced; these can be analysed genotypically and phenotypically to accurately characterise and predict eff ectiveness and toxic eff ects of anticancer agents.

Given these encouraging fi ndings, why is there not a stronger push to use these newer, more eff ective tests of liver function and drug metabolism more rapidly and commonly in oncology? First, more research in larger populations is needed to show safe and eff ective chemotherapy dosing on the basis of adjustments for liver synthetic and enzyme function with the new tests. Population models must be validated, which incorporate all factors that are known to aff ect frequency of toxic eff ects (including genotyping and phenotyping of pharma-codynamic pathways, extent and type of past treatment, and other individual characteristics). Second, randomised phase III trials are needed to show that dosing by genotype–phenotype nomograms reduces the risk of drug toxic eff ects while maintaining eff ectiveness, and is cost-eff ective compared with empirical dose adjustments based on bedside serum liver biochemistry. Commonly, recruitment of patients to these trials can be diffi cult, although appropriate patients are commonly seen in general oncology practice.

Until then, clinicians will be reluctant to change practice. Chemotherapy dosing based on body surface-area has been the mainstay of cancer treatment for several decades. Although serum liver biochemistry is commonly regarded as a marker or predictor of possible metabolic aberration, and although appropriate trials for safer use of chemotherapy in the setting of biochemical abnormalities continue to be done, current biochemical tests are not an adequate assessment of liver metabolic function.

www.thelancet.com/oncology Vol 9 December 2008 1189

Review

In the future, alternative assessments will likely be used to individualise chemotherapy dosing for a particular patient’s metabolic capacity. Data from phase III clinical trials are mainly based on dosing according to body surface area, and it is understandable that clinicians and regulatory agencies might be reluctant to move away from these strategies. However, these studies commonly exclude patients with liver impairment, those who are obese, or elderly people—ie, individuals who need the tailored dosing that is desired.

Confl icts of interestThe authors declared no confl icts of interest.

AcknowledgmentsWe thank Jeremy Frank (Western Hospital, VIC, Australia) for his

assistance in the preparation and editing of fi gure 1; and Chris Dow

(Western Hospital) for his assistance in the preparation of fi gure 2.

References1 Schenker S, Martin RR, Hoyumpa AM. Antecedent liver disease

and drug toxicity. J Hepatol 1999; 31: 1098–105.

2 Twelves C, Glynne-Jones R, Cassidy J, et al. Eff ect of hepatic dysfunction due to liver metastases on the pharmacokinetics of capecitabine and its metabolites. Clin Cancer Res 1999; 5: 1696–702.

3 Eklund JW, Trifi lio S, Mulcahy MF. Chemotherapy dosing in the setting of liver dysfunction. Oncology (Williston Park) 2005; 19: 1057–63.

4 Donelli MG, Zucchetti M, Munzone E, D’Incalci M, Crosignani A. Pharmacokinetics of anticancer agents in patients with impaired liver function. Eur J Cancer 1998; 34: 33–46.

5 Clarke SJ, Rivory LP. Clinical pharmacokinetics of docetaxel. Clin Pharmacokinet 1999; 36: 99–114.

6 Hooker AC, Ten Tije AJ, Carducci MA, et al. Population pharmacokinetic model for docetaxel in patients with varying degrees of liver function: incorporating cytochrome P4503A activity measurements. Clin Pharmacol Ther 2008; 84: 111–18.

7 Benjamin RS, Wiernik PH, Bachur NR. Adriamycin chemotherapy—effi cacy, safety, and pharmacologic basis of an intermittent single high-dosage schedule. Cancer 1974; 33: 19–27.

8 Dobbs NA, Twelves CJ, Gregory W, Cruickshanka C, Richards MA, Rubens RD. Epirubicin in patients with liver dysfunction: development and evaluation of a novel dose modifi cation scheme. Eur J Cancer 2003; 39: 580–86.

9 Miller AA, Murry DJ, Owzar K, et al. Phase I and pharmacokinetic study of erlotinib for solid tumors in patients with hepatic or renal dysfunction: CALGB 60101. J Clin Oncol 2007; 25: 3055–60.

10 Joel SP, Shah R, Clark PI, Slevin ML. Predicting etoposide toxicity: relationship to organ function and protein binding. J Clin Oncol 1996; 14: 257–67.

11 Fleming GF, Schilsky RL, Schumm LP, et al. Phase I and pharmacokinetic study of 24-hour infusion 5-fl uorouracil and leucovorin in patients with organ dysfunction. Ann Oncol 2003; 14: 1142–47.

12 Venook AP, Egorin MJ, Rosner GL, et al. Phase I and pharmacokinetic trial of gemcitabine in patients with hepatic or renal dysfunction: Cancer and Leukemia Group B 9565. J Clin Oncol 2000; 18: 2780–87.

13 Eckel F, von Delius S, Mayr M, et al. Pharmacokinetic and clinical phase II trial of imatinib in patients with impaired liver function and advanced hepatocellular carcinoma. Oncology 2005; 69: 363–71.

14 Venook AP, Enders Klein C, Fleming G, et al. A phase I and pharmacokinetic study of irinotecan in patients with hepatic or renal dysfunction or with prior pelvic radiation: CALGB 9863. Ann Oncol 2003; 14: 1783–90.

15 Raymond E, Boige V, Faivre S, et al. Dosage adjustment and pharmacokinetic profi le of irinotecan in cancer patients with hepatic dysfunction. J Clin Oncol 2002; 20: 4303–12.

16 Schaaf LJ, Hammond LA, Tipping SJ, et al. Phase 1 and pharmacokinetic study of intravenous irinotecan in refractory solid tumor patients with hepatic dysfunction. Clin Cancer Res 2006; 12: 3782–91.

17 Doroshow JH, Synold TW, Gandara D, et al. Pharmacology of oxaliplatin in solid tumor patients with hepatic dysfunction: a preliminary report of the National Cancer Institute Organ Dysfunction Working Group. Semin Oncol 2003; 30 (suppl 15): 14–19.

18 Venook AP, Egorin MJ, Rosner GL, et al. Phase I and pharmacokinetic trial of paclitaxel in patients with hepatic dysfunction: Cancer and Leukemia Group B 9264. J Clin Oncol 1998; 16: 1811–19.

19 Miller AA, Murray DJ, Owzar K et al. Pharmacokinetic (PK) and phase I study of sorafenib (S) for solid tumors and hematologic malignancies in patients with hepatic or renal dysfunction (HD or RD): CALGB 60301. Proc Am Soc Clin Oncol 2007; 25: abstr 3538.

20 O’Reilly S, Rowinsky E, Slichenmyer W, et al. Phase I and pharmacologic studies of topotecan in patients with impaired hepatic function. J Natl Cancer Inst 1996; 88: 817–24.

21 Robieux I, Sorio R, Borsatti E, et al. Pharmacokinetics of vinorelbine in patients with liver metastases. Clin Pharmacol Ther 1996; 59: 32–40.

22 Baker S, Ten Tije A, Carducci M, et al. Evaluation of CYP3A activity as a predictive covariate for docetaxel clearance. Proc Am Soc Clin Oncol 2004; 22: abstr 2006.

23 Mathijssen RH, van Alphen RJ, Verweij J, et al. Clinical pharmacokinetics and metabolism of irinotecan (CPT-11). Clin Cancer Res 2001; 7: 2182–94.

24 van Groeningen CJ, Van der Vijgh WJ, Baars JJ, Stieltjes H, Huibregtse K, Pinedo HM. Altered pharmacokinetics and metabolism of CPT-11 in liver dysfunction: a need for guidelines. Clin Cancer Res 2000; 6: 1342–46.

25 Rakhit A, Pantze MP, Fettner S, et al. The eff ects of CYP3A4 inhibition on erlotinib pharmacokinetics: computer-based simulation (SimCYP) predicts in vivo metabolic inhibition. Eur J Clin Pharmacol 2007; 8: 31–41.

26 Lu JF, Eppler SM, Wolf J, et al. Clinical pharmacokinetics of erlotinib in patients with solid tumors and exposure-safety relationship in patients with non-small cell lung cancer. Clin Pharmacol Ther 2006; 80: 136–45.

27 Ansfi eld FJ, Schroeder JM, Curreri AR. Five years clinical experience with 5-fl uorouracil. JAMA 1962; 181: 295–99.

28 Matsumaru K IK, Wakui N, Masao S. A case of liver dysfunction without elevation of serum transaminase induced by 5-fl uorouracil (5-FU) administration for colon cancer with liver metastasis. Acta Hepatologica Japonica 2006; 47: 398–404.

29 Kurnik D, Wood AJ, Wilkinson GR. The erythromycin breath test refl ects P-glycoprotein function independently of cytochrome P450 3A activity. Clin Pharmacol Ther 2006; 80: 228–34.

30 Benet LZ. There are no useful CYP3A probes that quantitatively predict the in vivo kinetics of other CYP3A substrates and no expectation that one will be found. Mol Interv 2005; 5: 79–83.

31 Kharasch ED, Thummel KE, Watkins PB. CYP3A probes can quantitatively predict the in vivo kinetics of other CYP3A substrates and can accurately assess CYP3A induction and inhibition. Mol Interv 2005; 5: 151–53.

32 Rivory LP, Slaviero K, Seale JP, et al. Optimizing the erythromycin breath test for use in cancer patients. Clin Cancer Res 2000; 6: 3480–85.

Search strategy and selection criteria

Data for this Review were obtained from a PubMed search and additionally from the American Society of Clinical Oncology abstracts and vitual meeting online source. No lower limit to time period was applied to the search; the upper limit was to June 2008. The search terms were: “liver and chemotherapy”; “hepatotoxicity”; “chemotherapy and hepatotoxicity”; and “liver functional imaging”. The search was limited to articles published in English. Abstracts or reports from meetings were included only if they were directly relevant to the discussion and published data were unavailable.

1190 www.thelancet.com/oncology Vol 9 December 2008

Review

33 Baker SD, van Schaik RH, Rivory LP, et al. Factors aff ecting cytochrome P-450 3A activity in cancer patients. Clin Cancer Res 2004; 10: 8341–50.

34 Hirth J, Watkins PB, Strawderman M, Schott A, Bruno R, Baker LH. The eff ect of an individual’s cytochrome CYP3A4 activity on docetaxel clearance. Clin Cancer Res 2000; 6: 1255–58.

35 Schott AF, Rae JM, Griffi th KA, Hayes DF, Sterns V, Baker LH. Combination vinorelbine and capecitabine for metastatic breast cancer using a non-body surface area dosing scheme. Cancer Chemother Pharmacol 2006; 58: 129–35.

36 Mathijssen RH, de Jong FA, van Schaik RH, et al. Prediction of irinotecan pharmacokinetics by use of cytochrome P450 3A4 phenotyping probes. J Natl Cancer Inst 2004; 96: 1585–92.

37 Mattison LK, Ezzeldin H, Carpenter M, Modak A, Johnson MR, Diasio RB. Rapid identifi cation of dihydropyrimidine dehydrogenase defi ciency by using a novel 2-13C-uracil breath test. Clin Cancer Res 2004; 10: 2652–58.

38 Fourie J, Wood T, Posey J, Modak A, Diasio B. The 2-13C-5-fl uorouracil breath test (FUBT) as a novel, rapid method for assessment of dihydropyrimidine dehydrogenase (DPD) activity in cancer patients: initial characterization and comparison to the 2-13C-uracil breath test (UraBT). Proc Am Soc Clin Oncol 2007; 25: abstr 2551.

39 Alexandre J, Rey E, Girre V, et al. Relationship between cytochrome 3A activity, infl ammatory status and the risk of docetaxel-induced febrile neutropenia: a prospective study. Ann Oncol 2007; 18: 168–72.

40 Kinirons MT, O’Shea D, Kim RB, et al. Failure of erythromycin breath test to correlate with midazolam clearance as a probe of cytochrome P4503A. Clin Pharmacol Ther 1999; 66: 224–31.

41 Gurney H, Wong M, Balleine RL, et al. Imatinib disposition and ABCB1 (MDR1, P-glycoprotein) genotype. Clin Pharmacol Ther 2007; 82: 33–40.

42 Wong M, Balleine RL, Blair EY, et al. Predictors of vinorelbine pharmacokinetics and pharmacodynamics in patients with cancer. J Clin Oncol 2006; 24: 2448–55.

43 Miya T, Goya T, Fujii H, et al. Factors aff ecting the pharmacokinetics of CPT-11: the body mass index, age and sex are independent predictors of pharmacokinetic parameters of CPT-11. Invest New Drugs 2001; 19: 61–67.

44 Puisset F, Chatelut E, Sparreboom A, et al. Dexamethasone as a probe for CYP3A4 metabolism: evidence of gender eff ect. Cancer Chemother Pharmacol 2007; 60: 305–08.

45 Puisset F, Dalenc F, Chatelut E, et al. Dexamethasone as a probe for vinorelbine clearance. Br J Clin Pharmacol 2005; 60: 45–53.

46 Yamamoto N, Tamura T, Kamiya Y, Sekine I, Kunitoh H, Saijo N. Correlation between docetaxel clearance and estimated cytochrome P450 activity by urinary metabolite of exogenous cortisol. J Clin Oncol 2000; 18: 2301–08.

47 Yamamoto N, Tamura T, Murakami H, et al. Randomized pharmacokinetic and pharmacodynamic study of docetaxel: dosing based on body-surface area compared with individualized dosing based on cytochrome P450 activity estimated using a urinary metabolite of exogenous cortisol. J Clin Oncol 2005; 23: 1061–69.

48 Gurney HP, Ackland S, Gebski V, Farrell G. Factors aff ecting epirubicin pharmacokinetics and toxicity: evidence against using body-surface area for dose calculation. J Clin Oncol 1998; 16: 2299–304.

49 Nishio M, Matsuda M, Ohyanagi F, et al. Antipyrine test predicts pharmacodynamics in docetaxel and cisplatin combination chemotherapy. Lung Cancer 2005; 49: 245–51.

50 Michael M, Thompson M, Hicks RJ, et al. Relationship of hepatic functional imaging to irinotecan pharmacokinetics and genetic parameters of drug elimination. J Clin Oncol 2006; 24: 4228–35.

51 Wong M, Evans S, Rivory LP, et al. Hepatic technetium Tc 99m-labeled sestamibi elimination rate and ABCB1 (MDR1) genotype as indicators of ABCB1 (P-glycoprotein) activity in patients with cancer. Clin Pharmacol Ther 2005; 77: 33–42.

52 Ezzeldin H, Diasio R. Dihydropyrimidine dehydrogenase defi ciency, a pharmacogenetic syndrome associated with potentially life-threatening toxicity following 5-fl uorouracil administration. Clin Colorectal Cancer 2004; 4: 181–89.

53 Ezzeldin HH, Diasio RB. Genetic testing in cancer therapeutics. Clin Cancer Res 2006; 12: 4137–41.

54 Van Kuilenburg AB, Meinsma R, Zoetekouw L, Van Gennip AH. High prevalence of the IVS14 + 1G>A mutation in the dihydropyrimidine dehydrogenase gene of patients with severe 5-fl uorouracil-associated toxicity. Pharmacogenetics 2002; 12: 555–58.

55 Sissung TM, Baum CE, Deeken J, et al. ABCB1 genetic variation infl uences the toxicity and clinical outcome of patients with androgen-independent prostate cancer treated with docetaxel. Clin Cancer Res 2008; 14: 4543–49.

56 Baker SD, Sparreboom A, Verweij J. Clinical pharmacokinetics of docetaxel: recent developments. Clin Pharmacokinet 2006; 45: 235–52.

57 Schott A, Taylor J, Baker L. Individualized chemotherapy dosing based on metabolic phenotype. Proc Am Soc Clin Oncol 2001; 20: (abstr 306).

58 Slaviero KA, Clarke SJ, McLachlan AJ, Blair EY, Rivory LP. Population pharmacokinetics of weekly docetaxel in patients with advanced cancer. Br J Clin Pharmacol 2004; 57: 44–53.

59 Lu Z, Zhang R, Carpenter JT, Diasio RB. Decreased dihydropyrimidine dehydrogenase activity in a population of patients with breast cancer: implication for 5-fl uorouracil-based chemotherapy. Clin Cancer Res 1998; 4: 325–29.

60 Sakka SG, Reinhart K, Meier-Hellmann A. Prognostic value of the indocyanine green plasma disappearance rate in critically ill patients. Chest 2002; 122: 1715–20.

61 Sakka SG. Assessing liver function. Curr Opin Crit Care 2007; 13: 207–14.

62 Lee SG, Hwang S. How I do it: assessment of hepatic functional reserve for indication of hepatic resection. J Hepatobiliary Pancreat Surg 2005; 12: 38–43.

63 Engel G, Hofmann U, Heidemann H, Cosme J, Eichelbaum M. Antipyrine as a probe for human oxidative drug metabolism: identifi cation of the cytochrome P450 enzymes catalyzing 4-hydroxyantipyrine, 3-hydroxymethylantipyrine, and norantipyrine formation. Clin Pharmacol Ther 1996; 59: 613–23.

64 Shukla S, Sauna ZE, Ambudkar SV. Evidence for the interaction of imatinib at the transport-substrate site(s) of the multidrug-resistance-linked ABC drug transporters ABCB1 (P-glycoprotein) and ABCG2. Leukemia 2008; 22: 445–47.

65 Child CG, Turcotte JG. Surgery and portal hypertension. In: Child CG, ed. Liver and potal hypertension. Philadelphia: WB Saunders, 1964: 50–64.

66 Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973; 60: 646–69.

67 Malinchoc M, Kamath PS, Gordon FD, Peine CJ, Rank J, ter Borg PC. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. Hepatology 2000; 31: 864–71.

68 Wiesner R, Edwards E, Freeman R, et al. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology 2003; 124: 91–96.

69 Ioannou GN, Perkins JD, Carithers RL Jr. Liver transplantation for hepatocellular carcinoma: impact of the MELD allocation system and predictors of survival. Gastroenterology 2008; 134: 1342–51.

70 Brown DB, Fundakowski CE, Lisker-Melman M, et al. Comparison of MELD and Child-Pugh scores to predict survival after chemoembolization for hepatocellular carcinoma. J Vasc Interv Radiol 2004; 15: 1209–18.

71 Deeken JF, Figg WD, Bates SE, Sparreboom A. Toward individualized treatment: prediction of anticancer drug disposition and toxicity with pharmacogenetics. Anticancer Drugs 2007; 18: 111–26.

72 Smith NF, Figg WD, Sparreboom A. Pharmacogenetics of irinotecan metabolism and transport: an update. Toxicol In Vitro 2006; 20: 163–75.

73 Ezzeldin HH, Lee AM, Mattison LK, Diasio RB. Methylation of the DPYD promoter: an alternative mechanism for dihydropyrimidine dehydrogenase defi ciency in cancer patients. Clin Cancer Res 2005; 11: 8699–705.

74 Henningsson A, Marsh S, Loos WJ, et al. Association of CYP2C8, CYP3A4, CYP3A5, and ABCB1 polymorphisms with the pharmacokinetics of paclitaxel. Clin Cancer Res 2005; 11: 8097–104.

75 Phillips KA, Veenstra DL, Oren E, Lee JK, Sadee W. Potential role of pharmacogenomics in reducing adverse drug reactions: a systematic review. JAMA 2001; 286: 2270–79.