Sex and Common Germline Variants Affect the Toxicity Profile and Pharmacokinetics of Alectinib: A Nationwide Cohort Study in Patients With ALK-Positive NSCLC

Niels Heersche, MD,a,b Daan A. C. Lanser, MD,a,b,c
M. Benthe Muntinghe-Wagenaar, MD,d Ma Ida Mohmaed Ali, PharmD,e
Ezgi B. Ulas, BSc,f Tessa M. A. Trooster, BSc,b Evert de Jonge, BSc,b
Esther Oomen-de Hoop, PhD, Marthe S. Paats, MD, PhD, Idris Bahce, MD, PhD,f
Sander Croes, PhD, Lizza E. L. Hendriks, MD, PhD,h
Anthonie J. van der Wekken, MD, PhD,d Anne-Marie C. Dingemans, MD, PhD,
Alwin D. R. Huitema, PhD,e,ij Ron H. N. van Schaik, PhD,b
Ron H. J. Mathijssen, MD, PhD,a\* G. D. Marijn Veerman, MD, PhDa,c
Department ofMedical Oncology,rasmusMCCancerIstitutEasmusniversityMeical ente,Rotterdmh
Netherlands
DeparmentflnicalhmistrrauivesityMdilntetrdmThthl
Departmentfonaryediinersmusancestitutrasmusiversitdicalenteottrdh
Netherlands
epartment f PumonaryMedicineUniversityof Groningen,UniversityMedicalCenteGroningenGroningenTh
Netherlands
DeparmentfharmacyandharmacologyTheNetherlandCancernstituteAmsterdamTheNetherland
DeparentfunaryMdicimstrdmiveritdicalntmstrdamthl
DepartmentfClinicalPharmacyoxicolgyMastrichtUniversityMedicalCenterARIMchoolforCardvasula
Disease,Maastricht，TheNetherlands
eparmnfPnrydiinastrihtiversitdilentchoolflyndeprdt
Maastricht，TheNetherlands
DeparmentfmacolgyrincessMaximCentfrPaditricncologyechthethln
DeparmentfClinicalharmacyniversityMedicalentetrechttrechtniversitytrechthethrlnd

Received25June2024;revised300ctober2024;accepted 25November2024
Availableonline-29November2024

ABSTRACT

Introduction:Alectinib,a small-molecule kinase inhibitor, is used as first-line treatment for ALK-positive \left(A L K+\right) NSCLC. Albeit generally well-tolerated, a considerable subset of patients requires dose adjustments due to drugrelated toxicity.Single-nucleotidepolymorphisms in genes relatedtothemetabolism of alectinibmayupfrontidentify patients at risk for toxicity.

Methods: In this multicenter observational cohort study in patients with advanced A L K+ NSCLC receiving alectinib treatment, we investigated the association between toxicity, pharmacokinetics,and key genetic variants in ABCB1, CYP3A4,PPARQ,POR,and CYP3A5.Data on demographics,adverse events,and alectinib trough levels were collected from five hospitals.

Results: Among 215 patients, 47% experienced severe toxicity.Womenexperiencedmore severe toxicity(female

\*Corresponding author.

Drs.MathijssenandVeerman sharesenior authorship.

Previouspresentation:This workwaspresented at the ESMo2024 Annual Congress in Barcelona, Spain on September 13-17, 2024.

Address for correspondence: Ron H.J. Mathijssen, MD, PhD, DepartmentofMedical Oncology,ErasmusMCCancerInstitute,ErasmusUniversityMedicalCenter,Dr.Molewaterplein 40,3015GD,Rotterdam， The Netherlands.E-mail:a.mathijssen@erasmusmc.nl

Cite this article as:HeerscheN,LanserDAC,Muntinghe-Wagenaar MB, etal.Sexandcommongermlinevariantsaffectthetoxicityprofileand pharmacokineticsofAlectinib:a nationwidecohort studyinpatients withALK-positiveNSCLC.JThoracOncol2025;20:475-486.

⊚ 2024InternationalAssociationfortheStudyofLungCancer. PublishedbyElsevierInc.Thisisanopenaccessarticleunderthe CCBYlicense(http://creativecommons.org/licenses/by/4.0/).

ISSN:1556-0864

https://doi.org/10.1016/j.jtho.2024.11.025 versus male: 56% versus 34% p=0.001 )and had +35% higher alectinib trough levels (p<0.001) .Homozygous carriers of the PPAR- α 209G{>}A variant exhibited a higher incidence of grade greater than or equal to3toxicity (38%) compared with patients who carried at least one wild-type allele (11%) (p=0.004) . This remained significant after Bonferroni correction.Patients whoexperiencedsevere toxicity had +18.5% C 95% confidence interval: 2.9%* 36.6% p=0.019. )higher trough levels.

Conclusions: Female patients encounter more severe toxicity due to higher alectinib exposure,which warrants further exploration.PPAR- α 209G{>}A significantly increased relevant alectinib-induced toxicity,most likely due to an increasein alectinibexposure.Pretreatment testingfor genetic variantswithasubsequentdosereductioncouldprovidea viable approach to reduce alectinib-related toxicity.

⊚ 2024 International Association for the Study of Lung Cancer. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons. org/licenses/by/4.0/).

Keywords: Alectinib; NSCLC; Single-nucleotide polymorphisms; Toxicity;Pharmacokinetics

Introduction

Alectinib, a small-molecule kinase inhibitor (SMKI), has been approved as a first-line treatment for metastatic NSCLC harboring ALK driver aberrations and is usedfavorablyover otherfirst-line ALK SMKI options (such as ceritinib, brigatinib, and lorlatinib).1 Its advent marked a pivotal development in managing advanced A L K+ NSCLC, evidenced by a median progression-free survival (PFS) of 35 months and a remarkable 5-year overall survival(os)rate of 63%.^{2,3} In addition, alectinibis beneficial as an adjuvant treatmentfor resected A L K+ NSCLC in terms of disease-free survival.4

An exposure-response relationship was found in patients with A L K+ NSCLC treated with alectinib，with larger tumor size reduction? and extended PFS among patients exhibiting a median trough concentration (C_{{trough,s s})} above 435~ng/mL, compared with those falling below this threshold.6 Currently, a prospective randomized controlled trial (NCT05525338) is in progress toinvestigate thepotential clinical benefit of pharmacokinetic-guideddosingthroughtherapeutic drug monitoring (TDM) for alectinib.7

On average, alectinib has a favorable tolerability profile,with predominant toxicities encompassinglow-grade cardiovascular,cutaneous, and gastrointestinal adverse events.38.9 Nonetheless, these toxicities frequently require dose reductions or temporary treatment interruptionin 20% to 52% of patients on alectinib.2,9-12

Currently, there is limited evidence for an exposure-toxicity relationshipfor alectinib.In a small studyof 53 patients,thosewhoexperiencedsevere toxicityexhibitedonaverage 35% higher alectinib \mathbf{C_{{trough},s s}} levels.13 Moreover, alectinib-related adverse events are best resolved with dose reductions,89,14 inferring that a relationship with exposure is present,especiallybecausealectinibhasdose-linear plasma concentrations.14

Alectinib is primarily metabolized into the active metabolite M4.Both compounds are for 40% to 50% metabolized by cytochrome P450 3A4 (CYP3A4),14-16 meaning thatchangesinCYP3A4functionalitywill most likely affect systemic exposure toboth.Because CYP3A4 has up to a 100-fold interpatient variability in enzyme activity,17 in part attributable to genetic variety, geneticfactorsinfluencingCYP3A4might alter alectinib pharmacokinetics,and subsequently survival and severe toxicity.

Several single-nucleotide polymorphisms (SNPs) have been associated with CYP3A4 activity (Fig. 1). First, the C Y P3A4^{*}22 variant (rs35599367; _{\mathsf{c}.15389\mathsf{C}>\mathsf{T}} resultsinreducedCYP3A4proteinlevels and a decreasein enzyme activity of 35% 17-19 Likewise, the 209G{>}A (rs4253728) SNP in PPAR- α decreases CYP3A4 activity with 21% 18 The PPAR- α gene encodes the nuclear receptor peroxisome proliferator-activatedreceptor alpha and plays a major role in regulation of CYP3A4 transcription and expression.18 Conversely, the ^{*}28 variant [rs1057868; c.1508C>T) inPORcorrelateswitha1.6- fold gain of function in CYP3A4 activity.2° In addition, CYP3A5 is an important iso-enzyme to CYP3A4,21 meaning it could contribute to alectinibmetabolism as well.CYP3A5,however,is not expressed in most White patients due to ahigh minor allelefrequency of theloss of function ^{*}3 (rs776746; \mathbf{c.6986A{>}G} allele.2 Finally, ATP-binding cassette B1 (ABCB1) is an important ubiquitous cellular efflux transporter for which M4 is a substrate (Fig. 1B).14 Consequently, M4 exposure might be affected by the 3435C>T (rs1045642) SNP in ABCB1, which reduces ABCB1 functionality.23

In light of these considerations, this study aims to assess theinfluenceofSNPsinABCB1,CYP3A4,CYP3A5, PPAR- α, and POR on alectinib-induced severe toxicity and pharmacokinetics.

Material and Methods

Study Design

This is a nationwidemulticenter observational cohort study performed in four academic hospitals (Erasmus MC Cancer Institute Rotterdam, University Medical Center Groningen, Maastricht University Medical Center, and Amsterdam University Medical Center) and one tertiarycancer center (Netherlands Cancer Institute, Amsterdam)in the Netherlands.

Figure1.Metabolismof alectinib andpresumedinfuenceof SNPsaccording toliteratureAlectinibisprimarilymetabolizd byCYP3A4intoitsmajrmetaboliteM(1A).oldsimilaranti-neplasticpotentialasalectinbandisasubtrateforh ABCB1effuxtransporter(B).CYP3A5mightmetabolizealectinibtoalesserextendaswellTheblueboxesinAindicat therespectiveinfluence of fourSNPsin the genes POR,CYP3A4,PPAR- α and CYP3A5.A blunt arrow indicates a reduction in enzymatic activity,whilst a sharp arrow in combination with a plus (\sp{\bullet}+\sp{\bullet}) sign indicates an increase in enzymatic activity.

12 months. Previous systemic treatment was allowed, including chemotherapy,immunotherapy,and other ALK kinase inhibitors.All investigators were blinded for genotypingresultsuntil alldatawerecollected.

StudyPopulationandDataCollection

Adult patients whowere treated with alectinibfor advanced A L K+ NSCLCuntilJanuary2023wereeligible for inclusion. Patients were excluded if no germline material for SNP genotyping was available. Data cutoff wasJanuary2024 to ensure a minimalfollow-uptime of

Basic demographic data,previous treatment lines, alectinib dose,co-medication,dose modifications and treatment interruptions,reason for treatment discontinuation,andadverseeventswerecollectedfromelectronichealthrecords.Thereasonfor dosemodification or treatmentinterruptionwas noted as well.If multiple reasons were given, all were registered.Other toxicities were only collected if they encompassedgrade3 or higher adverse events according to theCommon TerminologyCriteriaforAdverse Eventsversion5.0 (CTCAE).24 Part of the clinical data was collected retrospectively (Supplementary Fig.1).

Materials used for genotyping and/or clinical data werepreviously collected in one of thefollowing studies or initiatives:CodeGeno(MEC 02-1002),START-TKI (MEC 16-643; https://www.clinicaltrials.gov/study/ NCT05221372),Cancer Center Amsterdam -Liquid Biopsy Center,Maastricht 2019-1018-A-10,and the OncoLifeS Databiobank Groningen (https://umcgresearch.org/ w/oncolifes).Pharmacokinetic samplesin theNetherlands Cancer Institute were collected routinely as standard of care.An overview is presented inSupplementaryFigure 1.

For pharmacokinetic analysis， only steady-state plasma concentrations(i.e.,after at least 1 week on alectinib)were eligible.Based on the time of last alectinib administration and time of sampling,plasma concentrationswereextrapolatedto12hours afteralectinib intake to represent steady-state trough concentrations (Ctrough,ss) using a previously published method.25 Becausealectinibhasatimeuntilmaximalconcentration of approximately 5 hours,14 only samples drawn at least5hours after thelastintake were included.To aid comparison,mean trough concentrations were dose corrected to the starting dosage of alectinib ( 600~mg twice daily), based on the dose linearity of alectinib.14

DNA was extracted from plasma or whole blood using Maxwell kits (Promega, Madison, Wl; AS1840,AS1520, and AS4500). Genotyping was performed for ABCB1 (rs1045642),CYP3A4(rs35599367),PPAR- α (rs4253728), POR(rs1057868),and CYP3A5(rs776746) with TaqMan polymerase chain reaction (PCR) using TaqMan mix (Thermo Fisher Scientific, Waltham, MA; ref: 4401890) in combination with the specific SNP assays (Thermo Fisher Scientific）using Quantstudio software (Thermo Fisher Scientific; version 1.5.1). If an inadequate signal was obtained, additional analysis using digital droplet PCR using a \ensuremath{Qx}2\ensuremath{00} Droplet reader (Bio-Rad Laboratories, Hercules, CA) and Quantsoft software(Thermo Fisher Scientific;version 1.7.4.0917) was performed.

EthicalApproval

Primaryethicalapprovalof the studyprotocolwas provided by the ethics committee of the Erasmus University Medical Center (MEC 2022-158).In addition, local ethics committees approved the study protocol in all participating sites (The Netherlands Cancer Institute—IRBdm23-071; AmsterdamUniversityMedical Center—UVB23-0144;University Medical Center Groningen—OLS048-202211091; Maastricht University Medical Center—2019-1080-A-11).

Study EndPoints

Mainstudyendpointswere relationshipsbetween the five mentioned SNPs and the incidence of severe toxicity,incidence of CTCAE grade3 or higher adverse events, and alectinib pharmacokinetics (mean extrapolated \mathbf{C_{{trough,{ss}}}}, dose corrected to 600~{mg} twice daily). Severe toxicity was defined as incidence of grade3 or higher adverse events according to CTCAE or any toxicity resulting in dose reduction,(temporary)cessation of treatment, or hospitalization.Alectinib trough levels were also studied in relationship with the development of severe toxicity.

Statistical Analysis

Median follow-up time was calculated using the reverse Kaplan-Meier method as described by Schemper and Smith.26 Differences in the prevalence of toxicity among the genotypes and baseline characteristics were tested using a chi-square test. Significant predictors of toxicityweresubsequentlyenteredinamultivariable logistic model using backward selection.Individual mean \mathbf{C_{{trough,{ss}}}} levels, dose corrected to alectinib 600~{mg} twice daily,were log transformed for analysis,as we assumed \mathbf{C_{{trough,{ss}}}} levels to follow a log-normal distribution,27 and were subsequently compared using an unpaired t test. Next, the exponents were taken from the resulting differences and accompanying confidence intervals (Cis) to obtainthegeometricmeanratioanditsCI.Thesewere then converted to relative differences(RD)in percentages bysubtracting onefrom theratio and multiplyingit by 100%

Genotypes were tested inboth dominant models (i.e., wild-typeversusheterozygous andhomozygous variant carriers) and recessive models (ie.,wild-type and heterozygous variant carriersversushomozygousvariant carriers), unless the minor allele frequency was either less than 10% or more than 90% .Combined testing of genotypes was allowed once significant associations were established.

Statistically significant associations with any of the five SNPs were validated using internal validation by bootstrapping 1000 samples.28 For the SNPs,a Bonferroni correction was applied to correct for multiple testing. Because we made eight comparisons using the five SNPs, p less than 0.00625 was considered significant. For all other analyses, p less than 0.05 was considered significant. All statistical tests were performed using SPSS for Windows (IBM,Armonk,NY; version 28.0.1.0).

Results

A total of 215patientswithmetastatic A L K+ NSCLC were included in this study.Median follow-up time was

44 months (interquartile range [IQR]: 39-49 months). An overview of baseline characteristics isprovided in Table 1. Most patients were White (86%) and received alectinib as first-line ALK-targeted treatment (73%) Genotyping was successful in 91% to 99% of the patients, depending on the SNP tested. All genotypes were inHardy-Weinberg equilibrium.Supplementary Table 1 presents the prevalence and corresponding minor allele frequencies (MAFs) of the respective genotypes. Genotypeswereevenlydistributed acrosssex aswell as over all otherbaseline characteristics.

Toxicity

Of 215 patients, 100 patients (47%) experienced severe toxicity,consequentlynecessitatingdosereduction of alectinib in 92 patients ( 43% of all patients) (Table 2). The most common reasons for dose reduction were hepatotoxicity (ie.,elevated liver enzymes or increases in bilirubin level), renal toxicity (i.e., decreased glomerular filtration rate), and myalgia. Moreover, 12 patients 6% of all patients)prematurely discontinued treatment due to alectinib-induced toxicity that remained unresolved after dose adjustment, most often due to fatigue and SMKI-induced pneumonitis. Notably, 30 patients 14% of all patients)experiencedCTCAE grade3 or higher toxicity.Most prevalentgrade3 orhigher toxicitieswerehepatotoxicity( 4% of all patients), renal impairment( 2% of all patients),and skin rash ( 2% of all patients).Female patients had a higher incidence of severetoxicity (56%) compared withmalepatients (34%) (p=0.001) .None of the other baseline characteristics wereassociatedwiththeincidenceofalectinib-related adverse events.

Interestingly, a higher incidence of CTCAE grade 3 or higher toxicity was found for patients who were homozygous carriers of the PPAR- α 209G{>}A variant.Although wild-type and heterozygous patients exhibited an incidence of CTCAE grade 3 or higher toxicity of 12% among homozygous variant carriers, the incidence was significantly higher,being 38% (GG and GA versus AA; p=0.004) .This remained significant after both bootstrapping and Bonferroni correction.Among patients with at least one risk allele, a higher incidence of grade 3 orhigher toxicitywasfound as well(GGversus GA and AA: 9% versus 20% p=0.024. 0.For the broader severe toxicity,a similar pattern was identified,with an incidence of 44% among wild-type and heterozygous patients, compared with 75% among patients carrying two risk alleles (GG and GA versus AA; p=0.018{~} ).Although these two findings remained significant after bootstrapping 1o00 samples,they did not retain significance afterBonferronicorrection.NoneoftheotherSNPswere significantly associatedwith severetoxicity.The prevalence of toxicity per genotype is found in Table 3.

Inmultivariableanalysis,bothsex andPPAR- α 209G{>}A genotype remained significantly associated with severe toxicity.The odds ratio forfemale sex was 2.56 CI=1.44{-}4.55 p=0.001\Omega 0.For homozygous carriers ofthePPAR- *α 209G{>}A SNP,the odds ratio was 4.06 CI=1.23{-}13.41 p=0.021\AA 1

Pharmacokinetics

Assessablealectinibplasmaconcentrationswere available for 133 patients, exhibiting a dose-corrected geometric mean alectinib \mathbf{C_{{trough},s s}} of 616~{ng/mL} (coefficient of variation [CV]=44% 0.Patients who experiencedseveretoxicityhadsignificantlyhigher alectinib \mathbf{C_{{trough,{ss}}}} levels compared with patients who had not experienced severe toxicity \scriptstyle(673\ \ng/mL [CV=38%] versus 568~ng/mL [CV=47%] ， RD=18.5%, CI=2.9%- 36.6% p\ =\ 0.019. 0.In addition, female patients had higher dose-corrected alectinib \mathbf{C_{{trough},s s}} levels compared with male patients (699~ng/mL [CV=38%] versus 517 {ng/mL} [CV=46%] p<0.001\AA (Supplementary Fig. 2).

Homozygous variant carriers of PPAR- α 209G{>}A had ahigherdose-correctedgeometricmeanalectinib \mathbf{C_{trough,s s}} level (\mathsf{G G}+\mathsf{G A} versus AA: 601~{ng/mL} versus 781~ng/mL CV=45% and 33%, respectively; RD= 29.9% ： CI=1.1%{-66.8%}, p=0.041] .A similar differencewasfoundin carriers of the C Y P3A4^{*}22 allele, who had a 30.4% [CI=4.0%-63.6%] higher geometric mean alectinib \mathbf{C_{{trough,{ss}}}} level {{(^{*}}\boldsymbol{1}\mathord{\left/{\vphantom{{(^{*}}\boldsymbol{1})^{*}\boldsymbol{1}}}\right.\kern-\nulldelimiterspace}{^{*}\boldsymbol{1}}} versus ^{*}1/{}^{*}22 and ^{*}22/ ^{*}22 ： 603~ng/mL versus 786~ng/mL respectively; p= 0.022). These results remained statistically significant after internal validation using bootstrapping, although not after Bonferroni correction.None of the other SNPs were associated with pharmacokinetics (Table 4).Post hoc analyses combining both PPAR- α 209G{>}A and C Y P3A4^{*}22 genotypes (i.e., wild-types versus PPAR- α AA and/or C Y P3A4^{*}22~^{*}1/^{*}22 or ^{*}22/{^{*}22} )revealed a sustained and likewise differencein dose-correctedgeometric mean alectinib \mathbf{C_{{trough,{ss}}}} level of 29.6% \operatorname{CI}{}= 7.4%-56.3%, p=0.007 ;after bootstrapping) (Fig. 2). In addition, a combined haplotype analysis of PPAR- α 209G{>}A and C Y P3A4^{*}22 using a risk score revealed additive effects of at-risk alleles for toxicity and pharmacokinetics (Supplementary Tables 2 and 3).

Discussion

This large multicenter study revealed that patients treated with alectinib who harbor a common germline genetic polymorphisminPPAR- α experience significantly more CTCAE grade 3 or higher toxicity,with a clinically relevant absolute higher incidence of +26% Thisfinding remained significant after bothinternal validation and Bonferroni correction.The higher prevalence of toxicity may be explained by the finding that these patients also exhibit an alectinib \mathbf{C_{trough,s s}} level that is 30% higher compared with patientswho arewild type for PPAR- α .Especially because, in our study, patients whoexperiencedseveretoxicityhad 18.5% higher alectinib troughlevels on average( \mathbf{673}~ng/mL versus 568~ng/mL . Interestingly, in patients harboring at least one C Y P3A4^{*}22 allele, a 30% higher alectinib \mathbf{C_{trough,s s}} levelwasalsofound.The association of bothPPAR- α 209G{>}A and C Y P3A4^{*}22 with alectinib exposure did not retain significance after Bonferroni correction,likely due toa relativelylimitedsamplesize ofvariantcarriers and due totheconservative nature of the Bonferroni correction for multiple testing.?9 Nonetheless,the +30% higher trough concentrations that were identified among variant carriers aredefinitelyconsidered clinicallyrelevant and could account for the higher toxicity rates found within this group. Notably, 18% of all studied patients had riskgenotypesfor eitherPPAR- α 209G{>}A and/or C Y P3A4^{*}22 .Furthermore,this is the first study to find and address the considerabledifferencebetween the sexes in the incidence of severe toxicity and exposure toalectinib.Female patients have more( 56% versus 34% ）severe toxicity,probably due to a 35% higher alectinib exposure compared with male patients.

Thefindings of this study are especially relevant as 43% of patients treated in this study received a dose reduction,emphasizing that the current starting dosefor alectinib—twice daily 600~~mg could be disproportionate for many patients.This percentage isinline with those reported by numerous other trials, which described dose reductions occurring in 20% to 52% of all patients.29-12 This is in stark contrast with the comparativelymuchlower number of patients whohad tocease their treatment with alectinib due to unresolvable toxicity (6%) .Although most alectinib-related adverseeventsencompasslow-gradeandreversible toxicity, these can still pose a substantial burden on the quality of life of patients. In particular because patients endure a prolonged response on treatment.

Figure.Distributionofdsecorretedalectinbsteadystatetroughlevelsstratifiedprgentypeistributionofindivi mean alectinib trough levels, dose corrected to 600~\mathsf{m g} alectinib twice daily \scriptstyle(17=133) .The height on the \mathsf{y} -axis corresponds with alectinib trough levels.Left groupindicates wild-type patientsfor both PPAR- α 209G>A and C Y P3A4^{*}22 ，whereasthe right group represents patients having at least one of the at-risk genotypes (i.e.wild-types versus PPAR- α AA and/orCYP3A4 ^{\ast}1/{}^{\ast}22 or ^{*}22/{^{*}22} .Abbreviations:Ctrough:individual mean alectinib steady state troughconcentration,dose corrected to 600~\mathsf{m g} twice daily, expressed inng/mL; ng/mL: nanogramper militer; n: number Symbols: Black dots:nosevere toxicity experienced;blacktriangles:did experience severe toxicity;\*\*: \mathsf{p}<0.01 ;red dotted line represent the provisional therapeutic threshold for alectinib (i.e. 435~{ng/mL)} ; black lines represent the mean concentrations per group.

Together, these findings underline the necessity for a personalizeddosingstrategyforpatientsinitiatingalectinib treatment, consequently decreasing the burden of unnecessaryseveretoxicityandsubsequentdose alterations.Most severe alectinib-induced toxicity occurs within the first months after treatment initiation and is usually quickly resolved after dose reduction.8.9,13 Preemptivemeasures aimed atidentifying individualspredisposedtoexcessive drug exposure couldtherefore offer a proactive approach to optimizing treatment outcomes.3 Hence, pretreatment genotyping for PPAR α 209G{>}A and C Y P3A4^{*}22 could foster personalized dosing strategies tailored toindividual geneticprofiles.

Genotype-guided dosing has already been implemented in clinical practice,most notably for the chemotherapeutic agents 5-FU (and its pro-drug capecitabine) and irinotecan.31-35 Evidence has demonstrated that genotype-informed dosing for these drugs mitigates toxicity and results in less interpatient variationinpharmacokinetic profileswhile decreasinghealth care expenditure.31-34 Consequently, multiple guidelines andregulatory authorities strongly support theuse of pre-emptive_ genotyping for treatment with these agents.32,33,35

In the case of alectinib,using CYP3A4 andPPAR- α genotyping,patients who are prone to alectinib overdosing can potentially be identified pre-emptively.36 Alectinib is administered in capsules of 150~mg; hence, normalization ofpharmacokinetic exposure canbe achieved in patients expressing at-risk genotypes by reducing the initial dose with just one capsule (i.e., a dose reduction of 25% 450~mg twice daily instead of 600~{mg} twice daily). In our cohort, a theoretical 25% dosereduction in thepatients atriskwould cause three patients (13%) to exhibit troughlevels below the 435 {ng/mL} efficacy threshold,6 compared with one patient (4%) when starting at a regular dose.This indicates that measuring steady-state troughlevels afterimplementation of genotype-guided dosing of alectinib would still be required.

Interestingly,our study alsoidentified increased rates of toxicity,and higher alectinib levels, in female patients compared with male patients. This was not reportedin previous pharmacokinetic studies of alectinib.6,10,16 Because PPAR α and CYP3A4 genotypes were evenlydistributedamongfemale andmalepatients in our study,these geneticfactors cannot explain the differences between the sexes.A possible hypothesis for thisfinding could however be theincreased distribution volume for lipophilic drugs in women.37 Pre-emptive dose reduction in female patients could potentially also reduce toxicity rates.Nonetheless,when applying a theoretical 25% dose reduction to all female patients, a considerable number of patients drops below the efficacy threshold,resulting in 31% of all female patients having subtherapeutic alectiniblevels.Thiscompares with only 9% when not applying a dose reduction. Consequently,wewould notrecommend standard dose reductions in all female patients,as we deem the rate of underexposed patients too high. In comparison to the atrisk genotypes, the trough concentrations among female patients exhibited agreater interpatient variation (Supplementary Fig. 2), in turn rendering it less suitable for a fixed dose reduction,as illustrated by the hypothetical situation discussed previously.Nevertheless, these findings signify the importance of being extra attentive of toxicity when treating female patients with alectinib. Conversely,physicians should be aware of the increasedrisk of undertreatment inmalepatients, especiallywhen dosereductions areperformed.

The results of our study are in line with previous researchon theinterplaybetweenPPAR- α and CYP3A4. PPAR α encodes the nuclear receptor peroxisome proliferator-activated receptor alpha and isconsidered a direct determinant of CYP3A4 gene expression.18 Carriers of the 209G{>}A SNP have a decrease in CYP3A4 functionality of 21% per variant copy.18 Moreover, associationsbetweenPPAR- α 209G{>}A andtreatment outcomes on other CYP3A4 metabolized drugs, such as tacrolimus, are reported.38

This study was conducted in a mostly White population. Hence,SNPs were selected which are prevalent among the White population. However, for other populations,different SNPs arelikely toberelevant due to variance in allele frequencies.For instance, the CYP3A5 ^{*}6 and ^{*}7 variants, which are primarily present in patients of Sub-Sahara African and Afro-American ancestry, arelikely tobeimpactfulinpatients of thoserespective ethnicities.39 In addition, inclusion of these and other rare variants,including C Y P3A4^{*}20. could have helped reduceunexplainedvariabilityinpharmacokinetic exposure and toxicity in the current analysis. However, we opted not to analyze these SNPs as tolimit the extent of multiple testing. Regarding extrapolation of our results,PPAR- *α 209G{>}A and C Y P3A4^{*}22 are most prevalent in Europe,indicating that these specificvariants might have less impact at the population level in other ethnicities.40,41

A limitation of our analysis is the partly retrospective collection of the toxicity data.Althoughmost treatmentrelated events(e.g.,dose modifications andsevere toxicity)were reliablydescribed within the electronic health records, low-grade toxicities were less accurately documented. Hence, data regarding low-grade toxicity were not specifically collected，although these data wouldberelevantfor overallquality of lifeinpatientson alectinib,as those adverse events are more frequent than grade ^{3+} toxicities. Moreover, this might also be a reason why this study was unabletoidentifya relationshipwith toxicityfor the C Y P3A4^{*}22 variant.Interestingly,the impact of C Y P3A4^{*}22 on toxicity has already been established for other SMKIs and many other drugs, which are primarily metabolized by CYP3A4.42 Hence, prospective conformation of therelationshipsbetween both C Y P3A4^{*}22 andPPAR- α 209G{>}A and alectinib exposure andtoxicityisrequiredbeforepre-emptive genotyping for these twoSNPs canbeimplemented in daily practice.

Furthermore,the current analysis categorizes toxicity on the basis of its severity, irrespective of the specific type of toxicity.Therefore,our hypothesis presumes that all toxicity occurs due to pharmacokinetic overexposure toalectinibas aresultof ageneticallyreducedmetabolism. Nevertheless, when evaluating other SMKIs, this does not necessarily seem to be the case.43 For instance, exposure-toxicityrelationships forlorlatinibandceritinibwere established distinctly withhypercholesterolemia and hyperglycemia, respectively.44,45 This indicates that a similar principle is likely applicable to alectinib as well.Consequently,the significant association that we identified between pharmacokinetic exposure and all-type toxicity suggests that toxicities that do express anexposure-dependentmechanismcouldbe associated even more strongly with systemic alectinib concentrations.This stronger association could be masked by the inclusion of non-exposure-dependent toxicities in the current analysis.

The follow-up time in this study was too short to analyze PFS and OS in a clinically meaningful way. Therefore,potentialassociations between the studied SNPsand survival outcomeswere not studied at this moment and are currently still awaited.Analysis after maturation ofthe currentcohortcouldprovideinsights intowhetherpatientsharboringcommongeneticvariants in PPAR- α andCYP3A4havedifferentsurvival outcomes.

In conclusion,this studyidentified increasedrates of toxicity and higher alectinib trough levels in female patients.In addition,a clinically 26% absoluteincrease in the prevalence of CTCAE grade3 or higher alectinibinduced toxicity was identified in patients homozygous for thePPAR- α 209G{>}A variant.These patients had a clinically relevant 30% higher exposure to alectinib.A similar 30% higher exposure was found in patients carrying at least one C Y P3A4^{*}22 allele variant. These resultsencourage theimplementationof pre-emptive genotyping for these genes and warrant a 25% dose reduction, if one of the at-risk genotypes is identified in a patient.This couldoffer a practicalsolutiontothe frequently required dose reductions in patients treated with alectinibfor oncogene-addicted NSCLC.

Ezgi Ulas: Data curation, Investigation, Project administration, Writing - review & editing.

Tessa Trooster: Data Curation, Formal analysis, Investigation, Writing - review & editing.

Evert de Jonge: Investigation, Writing - review & editing.

Esther Oomen-de Hoop:Formal analysis,Investigation, Methodology, Writing - review & editing.

Marthe Paats: Investigation, Writing - review & editing.

Idris Bahce: Investigation, Writing - review & editing.

Sander Croes: Data curation, Investigation, Project administration, Writing - review & editing.

Lizza Hendriks: Investigation, Writing - review & editing.

Anthonie van der Wekken: Investigation, Writing - review & editing.

Anne-Marie Dingemans: Conceptualization, Investigation, Methodology,Writing -review & editing.

Alwin Huitema: Investigation, Writing - review & editing.

Ron van Schaik:Conceptualization,Investigation, Methodology, Supervision, Writing - review & editing.

Ron Mathijssen:Conceptualization,Formal analysis, Funding Acquisition, Investigation,Methodology,Supervision, Writing - original draft.

Marijn Veerman: Conceptualization,Formal analysis,Investigation,Methodology,Project administration, Supervision, Writing - original draft.

Data-SharingStatement

All data sets including deidentified participant data, which are generated during and/or analyzed during thisstudy,are availablefromthe corresponding author (a.mathijssen@erasmusmc.nl)onreasonablerequest until5years after date of publication.Datarecipients are requiredtoenteraformaldata-sharingagreementthat describestheconditionsfor release and requirementsfor data transfer, storage, archiving, publication, and intellectual property with each participating center.

CRediTAuthorship Contribution Statement

Niels Heersche: Data curation, Formal analysis, Investigation, Project administration, Writing - original draft.

Daan Lanser: Formal analysis, Writing - original draft.

BentheMuntinghe-Wagenaar:Datacuration, Investigation,Project administration,Writing-review & editing.

Ma Ida Mohmaed Ali:Data curation, Investigation, Project administration, Writing - review & editing.

Disclosure

Dr.Muntinghe-Wagenaar reports receiving travel grants from Bristol-Myers Squibb. Dr. Bahce reports receiving research grants from AstraZeneca,BristolMyers Squibb,and Boehringer-Ingelheim and funds for advisory boards (all paid to institution)from AstraZeneca,Bristol-Myers Squibb，Boehringer-Ingelheim, Takeda, Pfizer, Merck Sharp & Dohme, and Regeneron. Dr. Hendriks reports receiving research funding (all paid to the institute) from Roche, GlaxoSmithKline,Novartis, Blueprint, Mirati, AbbVie, Merck Sharp & Dohme, Amgen,

Genentech,AstraZeneca,Boehringer-Ingelheim,Takeda, Merck,Pfizer,Novartis,Janssen,Anheart Bayer,Lilly, and Gilead and was an invited speaker for AstraZeneca, Bayer, Lilly, Merck Sharp & Dohme,high5oncology, Takeda,Janssen, GlaxoSmithKline, Sanofi, and Pfizer (all paid to institution), and Medtalks, Benecke, VJOncology, and Medimix (paid to self).Dr.van der Wekken reports receiving consultingfees from AstraZeneca,BoehringerIngelheim,Janssen,Lilly,Merck KGaA, Darmstadt, Germany, Novartis,Roche,Pfizer, and Takeda; acted as invited speakerforPfizerandAstraZeneca;andreports receiving researchgrants by AstraZeneca,Boehringer Ingelheim,Roche,Pfizer,and Takeda.Dr.Dingemans reports receiving grants (all paid to the institution) from Amgen.Dr. Mathijssen reports receiving unrestricted grants for investigator-initiated trials (all paid to the institution)from Astellas,Bayer,Boehringer-Ingelheim, Cristal Therapeutics,Deuter Oncology,Echo Pharmaceuticals, Nordic Pharma, Novartis, Pamgene, Pfizer, Roche, Sanofi, and Servier. The remaining authors declare no conflict of interest.

Acknowledgments

This work was supported by the Erasmus MC Personalized Medicine Fund. Figure 1A and B and the Graphical Abstract were created using Biorender (https:// biorender.com/).

Supplementary Data

Note: To access the supplementary material accompanying this article,visit the online version of theJournal of Thoracic Oncology at www.jto.org and at https://doi. org/10.1016/j.jtho.2024.11.025.

References

1.Hendriks LE,Kerr KM,Menis J, et al. Non-oncogeneaddicted metastatic non-small-celllung cancer:ESMO ClinicalPractice Guideline for diagnosis,treatment and follow-up.AnnOncol.2023;34:339-357.
2.MokT,CamidgeDR,GadgeelSM,etal.Updatedoverall survival andfinalprogression-freesurvivaldataforpatientswithtreatment-naiveadvancedALK-positivenonsmall-celllungcancerintheALEXstudy.AnnOncol.
2020;31:1056-1064.
3.Peters S,Camidge DR,Shaw AT,et al.Alectinibversus crizotinib in Untreated ALK-Positive Non-Small-Cell Lung Cancer.N EnglJ Med.2017;377:829-838.
4.Wu YL，Dziadziuszko R, Ahn JS，et al.Alectinib in resectedALK-positivenon-small-cell lungcancer.N Engl J Med.2024;390:1265-1276.
5.Center forDrugEvaluation and Research.Clinical pharmacologyandbiopharmaceuticsreview(s).https://www. accessdata.fda.gov/drugsatfda_docs/nda/2015/208434 Orig1s000ClinPharmR.pdf.Accessed September 9,2024.
6.GroenlandSL，GeelDR，JanssenJM，etal.Exposureresponseanalysesofanaplasticlymphomakinase cancer patients.Clin Pharmacol Ther. 2021;109:394-
402.
7.Meertens M,Muntinghe-Wagenaar MB, Sikkema BJ,et al. Therapeutic drug monitoring guided dosing versus standard dosing of alectinib in advanced ALK positive nonsmall cell lung cancer patients: study protocol for an international, multicenter phase IV randomized controlled trial (ADAPT ALEC). Front Oncol. 2023;13:
1136221.
8. Camidge DR, Dziadziuszko R, Peters S, et al. Updated efficacy and safety data and impact of the EML4-ALK fusion variant on the efficacy of alectinib in untreated ALK-positive advanced non-small cell lung cancer in the global Phase III ALEX study. J Thorac Oncol.
2019;14:1233-1243.
9.Dziadziuszko R, Peters S, Ruf T, et al. Clinical experience andmanagementofadverseeventsinpatientswith advanced ALK-positive non-small-cell lung cancer receiving alectinib. ESMO Open. 2022;7:100612.
10. Lin L, Barkman HJ, Smit EF, et al. Clinical relevance of high plasma trough levels of the kinase inhibitors crizotinib,alectinib,osimertinib,dabrafenib,and trametinib in NSCLC patients. Ther Drug Monit. 2024;46:73-79.
11. de Leeuw SP, Pruis MA, Sikema BJ, et al. Analysis of serious weight gain in patients using alectinib for ALKpositive lung cancer. J Thorac Oncol. 2023;18:1017-1030.
12. Bernabe-Caro R, Garrido P, Garcia-Campelo R, et al. Alectinib afterfailure to crizotinibinpatients withALKpositive non-small cell lung cancer:results from the Spanish early access program. Oncotarget. 2022;13:812-
827.
13. Pruis MA, Veerman GDM, Hassing HC, et al. Cardiac toxicity of alectinib in patients with ALK+lung cancer: outcomes of cardio-oncology follow-up.Jacc Cardiooncol.2023;5:102-113.
14. European Medicines Agency. Summary of product characteristics Alectinib. https://www.ema.europa.eu/en/ documents/product-information/alecensa-epar-productinformation_en.pdf.Accessed September 9,2024.
15. Nakagawa T, Fowler S，Takanashi K， et al.In vitro metabolism of alectinib,a novel potent ALK inhibitor, in human: contribution of CYP3A enzymes. Xenobiotica.
2018;48:546-554.
16. Morcos PN, Yu L, Bogman K, et al. Absorption, distribution, metabolism and excretion (ADME) of the ALK inhibitor alectinib: results from an absolute bioavailability and mass balance study in healthy subjects. Xenobiotica.
2017;47:217-229.
17. Okubo M, Murayama N, ShimizuM Shimada Guengerich FP, Yamazaki H.CYP3A4 intron 6 C>T polymorphism (CYP3A4\*22) is associated with reduced CYP3A4 proteinlevel and function in human liver microsomes.J Toxicol Sci.2013;38:349-354.
18. Klein K, Thomas M, Winter S, et al. PPARA: a novel genetic determinant of CYP3A4 in vitro and in vivo. Clin Pharmacol Ther.2012;91:1044-1052.
19. Zanger UM, Schwab M. Cytochrome P450 enzymes in drug metabolism: regulation of gene expression, enzyme activities，and impact of genetic variation.Pharmacol Ther. 2013:138:103-141.
20. Oneda B, Crettol S，Jaquenoud Sirot E，Bochud M, Ansermot N, Eap CB. The P450 oxidoreductase genotype is associated with CYP3A activity in vivo as measured by the midazolam phenotyping test.Pharmacogenet Genomics. 2009;19:877-883.
21. Lolodi O, Wang YM, Wright WC, Chen T. Differential regulation of CYP3A4 and CYP3A5 and its implication in drug discovery. Curr Drug Metab. 2017;18:1095-1105.
22.Rodriguez-Antona C， Savieo JL, Lauschke VM, et al. PharmVar GeneFocus: CYP3A5.Clin Pharmacol Ther. 2022;112:1159-1171.
23. Mooij MG, Nies AT, Knibbe CA, et al. Development of human membrane transporters: drug disposition and pharmacogenetics. Clin Pharmacokinet. 2016;55:507-524.
24. U.s. Department of Health and Human Services. Common Terminology Criteria for Adverse Events (CTCAE). Version 5.0.https://ctep.cancer.gov/protocoldevelopment/ electronic_applications/docs/ctcae_v5_quick_reference_ 5x7.pdf. Accessed September 9, 2024.
25. van Eerden RAG, “Oomen-de Hoop E, Noordam A, Mathijssen RHJ, Koolen SLW.Feasibility of extrapolating randomly taken plasma samples to trough levels for therapeutic drug monitoring purposes of small molecule kinase inhibitors. Pharmaceuticals (Basel). 2021;14:119.
26. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials.1996;17:343-346.
27.Lacey LF, Keene ON, Pritchard JF, Bye A. Common noncompartmental pharmacokinetic variables: are they normally or log-normally distributed?J Biopharm Stat. 1997;7:171-178.
28. Steyerberg EW, Harrell FE Jr, Borsboom GJ, Eijkemans MJ,Vergouwe Y, Habbema JD. Internal validation of predictive models: efficiency of some proceduresforlogisticregressionanalysis.JClinEpidemiol. 2001;54:774-781.
29.Armstrong RA. When to use the Bonferroni correction. Ophthalmic Physiol Opt.2014;34:502-508.
30.Heersche N, Veerman GDM, de With M, et al. Clinical implicationsofgermlinevariationsfortreatment outcome and drug resistance for small molecule kinase inhibitorsinpatientswithnon-smallcelllungcancer. Drug Resist Updat. 2022;62:100832.
31. Henricks LM, Lunenburg CATC, de Man FM, et al. DPYD genotype-guided dose individualisation of fluoropyrimidine therapy in patients with cancer: a prospective safety analysis. Lancet Oncol. 2018;19:1459-1467.
32. Lunenburg CATC, van der Wouden CH, Nijenhuis M, et al. Dutch Pharmacogenetics Working Group (DPWG) guideline for the gene-drug interaction of DPYD and fluoropyrimidines.Eur J Hum Genet.2020;28:508-517.
33. Hulshof EC, Deenen MJ, Nijenhuis M, et al. Dutch pharmacogenetics working group(DPwG)guideline for the gene-drug interaction between UGT1A1 and irinotecan. EurJHumGenet.2023;31:982-987.
34. Hulshof EC, de With M， de Man FM， et al. UGT1A1 genotype-guided dosing of irinotecan: a prospective safetyandcostanalysisinpoormetaboliserpatients.Eur J Cancer. 2022;162:148-157.
35. de With M, Sadlon A, Cecchin E,et al. Implementation of dihydropyrimidine dehydrogenase deficiency testing in Europe. ESMO Open.2023;8:101197.
36. Swen JJ, van der Wouden CH, Manson LE, et al. A 12- gene pharmacogenetic panel to prevent adverse drug reactions: an open-label,multicentre，controlled, cluster-randomised crossover implementation study. Lancet. 2023;401:347-356.
37. Whitley H, Lindsey W. Sex-based differences in drug activity. Am Fam Physician.2009;80:1254-1258.
38. Lunde I, Bremer S, Midtvedt K, et al. The influence of CYP3A,PPARA, and POR genetic variants on the pharmacokineticsoftacrolimusandcyclosporineinrenal transplantrecipients.EurJClinPharmacol.2014;70: 685-693.
39. Pratt VM, Cavallari LH, Fulmer ML, et al. CYP3A4 and CYP3A5 genotyping recommendations: a joint consensus recommendation of the association for molecular pathology， Clinical Pharmacogenetics Implementation Consortium，College of American Pathologists，Dutch Pharmacogenetics Working Group of the Royal Dutch PharmacistsAssociation,EuropeanSocietyforPharmacogenomics and Personalized Therapy, and Pharmacogenomics Knowledgebase. J Mol Diagn. 2023;25:619- 629.
40.PharmGKB.rs4253728: variant frequencies.https://www. pharmgkb.org/variant/PA166156105. Accessed October 9, 2024.
41. Zhou Y, Ingelman-Sundberg M, Lauschke VM. Worldwide distribution of cytochrome P450 alleles: a meta-analysis of population-scale sequencing projects. Clin Pharmacol Ther. 2017;102:688-700.
42. Mulder TAM, van Eerden RAG, de With M, et al. CYP3A4\*22 genotyping in clinical practice: ready for implementation? Front Genet. 2021;12:711943.
43.van der Kleij MBA, Guchelaar NAD, Mathijssen RHJ,et al. Therapeuticdrugmonitoringofkinaseinhibitorsin oncology. Clin Pharmacokinet. 2023;62:1333-1364.
44.European Medicines Agency. Assessment report: Lorviqua. https://www.ema.europa.eu/en/documents/assessmentreport/lorviqua-epar-public-assessment-report_en.pdf. Accessed 0ctober 9, 2024.
45. Center for Drug Evaluation and Research. Clinical pharmacology and biopharmaceutics review(s). https:/ /www. accessdata.fda.gov/drugsatfda_docs/nda/2014/205755 orig1s000clinpharmr.pdf.Accessed October 9, 2024.