The authors assessed the impact of germline polymorphisms on clinical outcome in patients with advanced nonsmall cell lung cancer (NSCLC) who received platinum-gemcitabine (PG) chemotherapy.
The authors assessed the impact of germline polymorphisms on clinical outcome in patients with advanced nonsmall cell lung cancer (NSCLC) who received platinum-gemcitabine (PG) chemotherapy.
In total, 137 patients with stage IIIB/IV NSCLC were included who received first-line PG chemotherapy (74% of patients received cisplatin, and 26% received carboplatin). Twenty-three germline polymorphisms that were identified in peripheral blood samples were analyzed for progression-free survival (PFS), treatment response, overall survival (OS), and toxicity.
The median PFS was 5.8 months, the median OS was 10.2 months, and 44 patients (32%) had a partial treatment response. Carriers of the excision repair cross-complementation group 1 (ERCC1) mutant thymine (T) allele had a lower treatment response rate (29% vs 52%; P = .02), shorter PFS (adjusted hazard ratio [HR], 1.60; P = .04), and shorter OS (adjusted HR, 1.54; P = .05) compared with carriers of the wild-type cytosine/cytosine (CC) genotype. The xeroderma pigmentosum group A member 10 (XPD10) mutant adenine (A) allele (adjusted HR, 0.64; P = .04) and the x-ray cross-complementing group 1 (XRCC1) mutant guanine (G) allele (adjusted HR, 0.51; P = .02) also were independent predictors of OS. Carriers of the mutant adenosine triphosphate-dependent DNA helicase Q1 (RECQ1) C allele or the mutant cytidine deaminase (CDA) C allele were more likely to experience severe leukocytopenia (26% vs 10% [P = .03] and 28% vs 11% [P = .02], respectively) compared with wild-type genotype carriers. Patients who carried the homozygous mutant glutathione S-transferase π 1(GSTP1) GG genotype were at considerable risk for severe platinum-associated polyneuropathy (18% vs 3% in wild-type vs heterozygous mutant patients, respectively; P = .01).
To the authors' knowledge, this is the first prospective study to date in patients with advanced NSCLC describing predictive germline polymorphisms not only for the clinical activity of PG chemotherapy (ERCC1, XPD10) but also for its toxicity (GSTP1, RECQ1, CDA). Nonplatinum-containing chemotherapy in carriers of the ERCC1 T allele or the XPD10 G allele should be studied prospectively. Cancer 2012;. © 2011 American Cancer Society.
Nonsmall cell lung cancer (NSCLC) is the leading cause of cancer-related deaths worldwide. Cisplatin-gemcitabine (CG) is a frequently used chemotherapy regimen in patients with advanced NSCLC based on a favorable efficacy and tolerability profile.1, 2 Overall mortality is reduced by 10% with CG compared with other platinum-containing regimens, resulting in an absolute survival benefit at 1 year of 3.9%.1 DNA repair pathways, such as nucleotide excision repair (NER) and base excision repair (BER), are critical for the activity of platinum-based chemotherapy.3 Excision repair cross-complementation group 1 (ERCC1), xeroderma pigmentosum group A (XPA), and xeroderma pigmentosum group D (XPD) are key enzymes of the NER pathway. The cytosine/cytosine (C/C) genotype in codon 118 of ERCC1 was considered a favorable marker of survival in some studies, 4-6 but not in others.7-10 The guanine-to-adenine substitution at codon 312 (G312A) of XPD has not been identified to date as a predictor of clinical outcome, 4, 6 except in 1 study that was presented only as an abstract.11 There are conflicting data on the effect of the G allele in codon 399 of x-ray cross-complementing group 1 (XRCC1) on clinical outcome in patients with advanced NSCLC.11-14 The thymine (T) allele in codon 241 of XRCC3 was identified as a favorable marker of survival in 1 clinical study.7 For gemcitabine, different drug pathway-associated gene polymorphisms have been described, 15, 16 but their clinical relevance remains largely unknown. The objective of the current study was to quantify the impact of selected drug pathway-associated gene polymorphisms on chemotherapy-associated toxicity and clinical outcome in patients with advanced NSCLC receiving standard first-line platinum-gemcitabine chemotherapy.
This multicenter prospective study was initiated in 2005. Eligible patients had histologically or cytologically documented, advanced (stage IIIB or stage IV) or recurrent NSCLC and were considered for first-line chemotherapy with gemcitabine and cisplatin. Patients with contraindications for cisplatin, such as renal function impairment, polyneuropathy, or tinnitus, received carboplatin. Other inclusion criteria included age ≥18 years; an Eastern Cooperative Oncology Group (ECOG) performance status ≤2; measurable or evaluable disease according to Response Evaluation Criteria in Solid Tumors (RECIST); and adequate hematologic, hepatic, and renal function. Main exclusion criteria included radiotherapy within the last 3 weeks before study entry, patients unsuitable for adequate follow-up, and evidence of uncontrolled systemic disease. The protocol was approved by local independent ethics committees, and the study was conducted in accordance with the Declaration of Helsinki, the Guidelines for Good Clinical Practice, and the International Conference on Harmonization Tripartite Guideline. All patients provided written informed consent.
Gemcitabine was administered intravenously at a dose of 1250 mg/m2 over 30 minutes on days 1 and 8 followed by either cisplatin at 75 mg/m2 as a 4-hour intravenous infusion or carboplatin over 30 minutes on day 1. Carboplatin was given at an area under the concentration-time curve of 5 according to the modified Calvert formula, and the glomerular filtration rate was calculated according to the Cockroft-Gault formula. Cycles were repeated every 3 weeks for a maximum of 6 cycles unless there was evidence of disease progression or unacceptable toxicity. The primary study objective was to correlate known functional polymorphisms of DNA-repair genes and drug-metabolizing genes with progression-free survival (PFS) in patients with advanced NSCLC who were receiving first-line platinum-gemcitabine chemotherapy.
Tumor status was determined every second treatment cycle during study treatment by computed tomography scans and every 2 months thereafter until disease progression. Tumor response was assessed according to RECIST, in which a complete response is defined as the disappearance of all target lesions, a partial response is defined as a decrease ≥30% in the sum of the greatest dimension of target lesions from baseline, progressive disease is defined as an increase ≥20% in the sum of the greatest dimension of target lesions from baseline, and stable disease is defined as neither sufficient shrinkage to qualify for a partial response nor sufficient increase to qualify for progressive disease.17 Adverse events were assessed at weekly intervals, were graded using the National Cancer Institute Common Toxicity Criteria for Adverse Events version 3.0, and were coded according to the Medical Dictionary for Regulatory Activities (International Federation of Pharmaceutical Manufacturers and Associations, Geneva, Switzerland). Hematologic measurements were performed at weekly intervals. Patients were followed until death.
The following germline mutations were analyzed in peripheral blood from all patients: XRCC1 A399G; XRCC3 C241T; XPD10 G312A; XPD23 A751C; ERCC1 C118T; adenosine triphosphate-dependent DNA helicase Q1 (RECQ1) A159C; DNA repair and recombination protein RAD54-like (RAD54L) C157T; deoxycitidine kinase (DCK) promoter variants −C360G, −T289A, −G243T, −C201T, −G135C, and −G125T; cyclin-dependent kinase G261A, C300T, and C364T; cytidine deaminase (CDA) A79C; solute carrier family 28 (sodium-coupled nucleoside transporter) member 1 (SLC28A1) G1543A and T1576C; SLC28A2 A283C; ribonucleotide reductase M1 (RRM1) promoter variant −C37A; glutathione S-transferase μ 1 (GSTM1) deletion; and glutathione S-transferase π 1 (GSTP1) A313G. DNA amplification was performed in a PTC-200 Thermocycler (MJ Research, Waltham, Mass). DNA sequencing was performed using the BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands) on an ABI Prism 3100 DNA analyzer (Applied Biosystems). For sequence alignment, the SEQSCAPE bioinformatics software package (version 2.1; Applied Biosystems) was used. Hardy-Weinberg equilibrium was evaluated for each genotype. The investigators performing genetic analyses (V.D.D. and P.H.M.S.) were blinded to patient characteristics and clinical outcomes.
The primary study endpoint was PFS, which was calculated from the start of chemotherapy to the first documented disease progression (defined according to RECIST17) or death, whichever occurred first. Patients without documented progression or death were censored at the last tumor assessment. Secondary endpoints included objective tumor response, overall survival (OS), and chemotherapy-associated adverse events. The plan for the study was to have 75% power to detect a 25% improvement in PFS at 9 months with a type-I error of 5% (2-sided test) when the expected survival at 9 months was 20% in patients with the nonfavorable genotype and ≥30% in the patients with the favorable genotype. Demographic and clinical information was compared across genotypes using Fisher exact tests (for binary variables) and Wilcoxon-type tests for trend (for ordinal variables with ≥3 levels). The association between PFS or OS and the genotypes was estimated using the method of Kaplan and Meier and was assessed using log-rank tests initially with 2 degrees of freedom, and subsequently with 1 degree of freedom after grouping heterozygotes with either wild type or homozygotes. Median follow-up was calculated using OS duration with deaths censored. Cox proportional hazards models for PFS were used to determine the impact of the genotypes on PFS and OS, adjusting for patient age, sex, tumor stage, performance status, and smoking status. All tests of significance were 2-sided. Given the explorative nature of this study, no multitest adjustments were made. We considered associations significant at the .05 level. All analyses were performed using Stata 11.0 software (Stata Corp., College Station, Tex).
In total, 146 patients with advanced NSCLC from 4 centers were consented between November 2005 and November 2008. Patients were recruited in the Netherlands Cancer Institute (n = 133), the Free University Medical Center Amsterdam (n = 4), the Kennemer Hospital Harlem (n = 3), and Alkmaar Medical Center (n = 6). We excluded 6 patients with inadequate source data, 2 patients with no evaluable tumor, and 1 patient with stage IIB NSCLC, for a final study population of 137 patients (Table 1). The first-cycle platinum compound was cisplatin in 101 patients (74%) and carboplatin in 36 patients (26%). Cisplatin was changed to carboplatin in 15 patients for renal impairment (8 patients), peripheral neuropathy (6 patients), or cisplatin-related hearing impairment (1 patient). The median follow-up was 3 years and 4 months, the median PFS was 5.8 months, and the median OS was 10.2 months. Treatment response was a partial response in 44 patients (32%), stable disease in 79 patients (58%), progressive disease in 11 patients (8%), and not evaluable in 3 patients (2%). Tumor response was measurable in 105 patients (77%).
|Clinical Characteristic||No. of Patients (%)||Median, mo||HR [95% CI]|
|Men||77 (56)||9.7||1.15 [0.79-1.67]|
|1||78 (57)||9.4||1.84 [1.23-2.75]|
|2||7 (5)||5.2||4.29 [1.80-10.22]|
|IV||95 (69)||9.7||1.11 [0.72-1.72]|
|Squamous cell carcinoma||14 (10)||6.8||1.31 [0.66-2.62]|
|Large cell carcinoma||68 (50)||10.8||0.82 [0.54-1.25]|
|Undifferentiated or NOS||11 (8)||9.5||0.89 [0.39-2.02]|
|Current smoker||70 (51)||10.2||1.34 [0.78-2.30]|
|Former smoker||44 (32)||7.3||1.46 [0.81-2.62]|
|Creatinine clearance, mL/min|
|Serum bilirubin, μmol/L|
All polymorphisms were in Hardy-Weinberg equilibrium. The DCK promoter polymorphisms −C360G and −C201T were fully linked. No associations were identified between genotypes and patient age, sex, performance status, or disease stage.
Carriers of the mutant ERCC1 T allele had a lower PFS compared with carriers of the wild-type ERCC1 CC genotype (hazard ratio [HR], 1.63; 95% confidence interval [CI], 1.04-2.55; P = .03) (Table 2). Carriers of the ERCC1 TT genotype had the lowest PFS (6.1 months), carriers of the CT genotype a medium PFS (7.1 months), and carriers of the CC genotype the highest PFS (9.7 months) (Fig. 1). The 9-month PFS rate was 16% for carriers of the ERCC1 TT genotype, 24% for the carriers of the CT genotype, and 48% for carriers of the CC genotype. Carriers of the mutant XPD10 A allele had a higher PFS compared with carriers of the wild-type GG genotype (HR, 0.70; 95% CI, 0.49-1.00; P = .05). The 9-month PFS rate was 16% for carriers of the XPD10 GG genotype, 30% for carriers of the GA genotype, and 33% for carriers of the AA genotype. Significantly lower PFS was observed among patients who had stage IV disease compared with patients who had stage IIIB disease (HR, 1.56; 95% CI, 1.07-2.28; P = .02), ever-smokers compared with never-smokers (HR, 2.48; 95% CI, 1.53-4.03; P < .01), and patients who had an ECOG performance status of 1 or 2 compared with patients who had a performance status of 0 (HR, 1.62; 95% CI, 1.13-2.31; P < .01). ERCC1 remained a strong predictor of PFS in the adjusted model (Table 3).
|Treatment Response||Progression-Free Survival||Overall Survival|
|Genotype||Total No. (%)||No. of PRs (%)||Pa||Median [95% CI], mo||HR [95% CI]||Pb||Median [95% CI], mo||HR [95% CI]||Pb|
|XRCC1 codon 399|
|AA genotype||17 (13)||5 (31)||4.8 [1.4-7.3]||Ref||.68c||6 [2.3-9.3]||Ref||.12c|
|AG genotype||63 (48)||18 (32)||.68||6.3 [5.3-8.1]||0.91 [0.53-1.58]||.74||10.4 [9.4-13.7]||0.62 [0.34-1.11]||.10|
|GG genotype||51 (39)||17 (35)||5.2 [3.5-7.6]||1.03 [0.59-1.82]||.91||10.8 [7.3-15.9]||0.56 [0.30-1.01]||.05|
|XRCC3 codon 241|
|CC genotype||49 (36)||9 (20)||4.7 [3.2-7.4]||Ref||.57c||10.4 [6.5-14.9]||Ref||.13c|
|CT genotype||64 (47)||27 (46)||.24||6.8 [5.6-8.1]||0.82 [0.56-1.21]||.32||10.3 [9.1-13.7]||1.12 [0.74-1.71]||.57|
|TT genotype||23 (17)||6 (27)||4.8 [3.0-5.5]||1.21 [0.73-2.02]||.46||9.3 [4.0-11.7]||1.54 [0.87-2.72]||.14|
|XPD10 codon 312|
|GG genotype||53 (39)||12 (24)||4.9 [4.2-6.0]||Ref||.14c||8.2 [5.2-10.4]||Ref||.02c|
|GA genotype||57 (41)||20 (37)||.08||7.0 [5.2-8.4]||0.67 [0.46-0.99]||.04||10.9 [9.3-14.9]||0.59 [0.39-0.91]||.01|
|AA genotype||27 (20)||10 (43)||5.6 [3.4-9.5]||0.77 [0.48-1.23]||.28||13.7 [7.8-21.2]||0.56 [0.33-0.95]||.03|
|XPD23 codon 751|
|AA genotype||51 (37)||16 (33)||5.8 [4.4-6.3]||Ref||.41c||8.7 [6.4-9.7]||Ref||.14c|
|AC genotype||71 (52)||22 (34)||1.00||5.8 [4.8-7.4]||0.84 [0.58-1.21]||.34||10.4 [8.3-15.7]||0.8 [0.54-1.20]||.29|
|CC genotype||15 (11)||4 (31)||5.6 [2.9-8.5]||0.85 [0.47-1.52]||.58||13.2 [9.3-17.9]||0.66 [0.35-1.24]||.20|
|ERCC1 codon 118|
|CC genotype||25 (18)||12 (52)||9.0 [5.2-12.2]||Ref||.01c||14.9 [7.8-21.6]||Ref||.05c|
|CT genotype||60 (44)||18 (33)||.02||5.8 [4.5-8.1]||1.47 [0.91-2.36]||.12||10.2 [7.3-12.4]||1.17 [0.68-1.99]||.55|
|TT genotype||51 (38)||12 (24)||4.9 [3.5-6.3]||1.86 [1.13-3.07]||.01||9.5 [6.6-10.9]||1.68 [0.98-2.90]||.06|
|RECQ1 codon 159|
|AA genotype||49 (36)||14 (30)||5.5 [3.7-6.3]||Ref||.24c||9.2 [6.6-11.5]||Ref||.67c|
|AC genotype||70 (52)||23 (35)||.45||6.3 [4.8-7.2]||0.86 [0.59-1.24]||.42||10.9 [9.3-14.4]||0.94 [0.62-1.42]||.78|
|CC genotype||16 (12)||5 (42)||8.2 [3.4-10.9]||0.75 [0.42-1.35]||.34||8.7 [3.8-14.9]||1.2 [0.62-2.32]||.57|
|RAD54L codon 157|
|CC genotype||102 (75)||30 (32)||6.0 [5.0-7.2]||Ref||.68c||10.2 [9.1-11.7]||Ref||.98c|
|CT genotype||31 (23)||11 (38)||.94||4.9 [3.4-6.7]||1.17 [0.77-1.78]||.45||10.3 [5.2-17.1]||0.99 [0.64-1.54]||.98|
|TT genotype||3 (2)||0 (0)||—||0.86 [0.27-2.73]||.79||—||1.03 [0.32-3.29]||.95|
|CC/CC genotype||135 (98.5)||42 (33)||.45||5.8 [4.9-6.8]||Ref||10.3 [9.2-11.7]||Ref|
|CG/CT genotype||2 (1.5)||0 (0)||—||11.1 [2.54-48.]||<.01||—||13.3 [2.82-61.1]||<.01|
|DCK codon 300|
|CC genotype||120 (87)||37 (33)||5.9 [4.8-7.0]||Ref||.85c||10.4 [9.3-12.4]||Ref||.24c|
|CT genotype||16 (12)||5 (33)||.76||5.3 [3.0-9.0]||0.99 [0.59-1.67]||.97||6.8 [4.0-22.2]||1.25 [0.68-2.30]||.45|
|TT genotype||1 (1)||0 (0)||—||1.81 [0.25-13.11]||.56||—||4.03 [0.54-29.9]||.17|
|CDA codon 79|
|AA genotype||63 (46)||18 (32)||5.3 [3.9-6.8]||Ref||.81c||9.5 [6.0-11.7]||Ref||.58c|
|AC genotype||59 (44)||18 (32)||.80||5.8 [4.8-7.0]||1.02 [0.71-1.46]||.93||10.3 [8.2-14.5]||0.83 [0.56-1.23]||.36|
|CC genotype||14 (10)||5 (36)||7.1 [4.2-9.3]||1.09 [0.60-1.97]||.78||10.8 [5.4-17.1]||0.96 [0.52-1.78]||.9|
|SLC27A1 codon 1543|
|GG genotype||36 (27)||8 (25)||5.5 [3.0-9.0]||Ref||.53c||9.8 [5.8-13.7]||Ref||.57c|
|GA genotype||74 (55)||24 (35)||.35||5.8 [4.8-7.3]||0.94 [0.63-1.41]||.77||9.7 [7.8-11.5]||1.11 [0.72-1.71]||.63|
|AA genotype||25 (18)||9 (36)||5.2 [3.5-7.0]||1.15 [0.68-1.93]||.61||11.7 [6.9-17.1]||0.84 [0.49-1.45]||.54|
|SLC28A2 codon 283|
|AA genotype||60 (45)||16 (29)||6.0 [4.9-7.2]||Ref||.71c||10.2 [6.9-11.7]||Ref||.86c|
|AC genotype||63 (47)||24 (49)||.79||5.8 [4.2-7.4]||0.98 [0.69-1.40]||.92||11.0 [8.2-14.4]||0.95 [0.64-1.41]||.81|
|CC genotype||11 (8)||2 (18)||4.5 [2.0-8.4]||1.25 [0.66-2.40]||.49||8.3 [2.0-17.9]||1.03 [0.52-2.06]||.91|
|CC genotype||66 (50)||21 (33)||5.3 [4.4-6.8]||Ref||.93c||9.7 [6.9-11.0]||Ref||.74c|
|CA genotype||56 (43)||18 (35)||.54||6.0 [4.2-8.2]||0.97 [0.67-1.39]||.86||9.8 [6.6-14.4]||0.98 [0.66-1.47]||.95|
|AA genotype||9 (7)||1 (13)||5.6 [1.2-10.6]||1.00 [0.49-2.02]||.99||12.6 [1.2-23.3]||0.87 [0.43-1.79]||.72|
|Wild type||80 (58)||20 (27)||.08||6.3 [4.9-7.6]||Ref||10.2 [7.3-11.5]||Ref|
|Gene deletion||57 (42)||22 (41)||5.6 [4.5-6.8]||0.97 [0.69-1.38]||.88||10.2 [8.2-15.7]||1.13 [0.77-1.64]||.52|
|GSTP1 codon 313|
|AA genotype||55 (42)||20 (38)||7.0 [4.5-8.2]||Ref||.97c||12.4 [6.6-15.9]||Ref||.15c|
|AG genotype||60 (45)||18 (32)||.18||5.3 [4.2-6.3]||1.38 [0.92-1.97]||.12||9.8 [8.2-11.0]||1.34 [0.89-2.02]||.15|
|GG genotype||17 (13)||3 (20)||6.0 [4.2-9.3]||0.85 [0.48-1.49]||.57||9.1 [1.6-16.2]||1.32 [0.72-2.42]||.37|
|Covariate||No. of Patients (%)||HR [95% CI]||Log-Rank P|
|ERCC1 codon 118|
|CC genotype||25 (18)||Ref|
|CT or TT genotype||112 (82)||1.62 [1.01-2.61]||.05|
|XPD10 codon 312|
|GG genotype||53 (39)||Ref|
|GA or AA genotype||84 (61)||0.92 [0.63-1.34]||.66|
|IV||95 (69)||0.81 [0.55-1.18]||.27|
|1 or 2||82 (62)||1.52 [1.05-2.19]||.03|
|Ever-smoker||114 (83)||2.28 [1.40-3.71]||<.01|
Carriers of the wild-type ERCC1CC genotype had a significantly higher response rate compared with carriers of the T allele (52% vs 29%; P = .02) (Table 2). Although carriers of the wild-type XPD10 GG genotype had a lower response rate compared with carriers of the A allele, the difference was not significant (24% vs 38%; P = .09). A higher quantitative tumor response was associated significantly with both the ERCC1 C allele and the XPD10 A allele (Fig. 2).
Carriers of the mutant RECQ1 C allele were more likely to experience severe leukocytopenia (26% vs 10%; P = .03) or to have dose reductions because of myelosuppression (48% vs 24%; P = .01) compared with carriers of the wild-type RECQ1 AA genotype (Table 4). Similarly, carriers of the mutant CDA C allele were more likely to experience severe leukocytopenia (28% vs 11%; P = .02) or to have chemotherapy dose reductions because of myelosuppression (53% vs 24%; P < .01) compared with carriers of the wild-type CDA AA genotype. Patients carrying the homozygous mutant GSTP1 GG genotype were more likely to experience severe polyneuropathy compared with patients carrying the GSTP1 AA genotype or AG genotype (18% vs 3%; P = .01).
|Covariate||Total No. of Patients (%)||No. With Severe LCP (%)||Pa||No. of Dose Modifications Because of Myelosuppression (%)||Pa||No. With Severe PNP (%)||Pa|
|RECQ1 codon 159|
|AA genotype||49 (36)||65 (10)||.03||12 (24)||.01||2 (4)||.51|
|AC or CC genotype||88 (64)||23 (26)||42 (48)||5 (6)|
|CDA codon 79|
|AA genotype||63 (46)||7 (11)||.02||15 (24)||<.1||4 (6)||.70|
|AC or CC genotype||74 (54)||21 (28)||39 (53)||3 (4)|
|GSTP1 codon 313|
|AA or AG genotype||115 (84)||23 (20)||.39||47 (41)||.47||3 (3)||.01|
|GG genotype||22 (16)||5 (23)||7 (32)||4 (18)|
|Women||60 (44)||6 (10)||.01||18 (30)||.01||2 (3)||.46|
|Men||77 (56)||22 (29)||36 (47)||5 (7)|
|<65||95 (69)||14 (15)||.02||32 (34)||.05||4 (4)||.67|
|65||42 (31)||14 (33)||22 (52)||3 (7)|
At the time of data analysis, 17 patients were still alive. Carriers of the mutant ERCC1 T allele had a shorter OS compared with carriers of the wild-type CC genotype (9.5 months vs 14.9 months; HR, 1.50; 95% CI, 1.02-2.21; P = .04). Carriers of the mutant XPD10 A allele had a higher OS compared with carriers of the wild-type XPD10 GG genotype (11.0 months vs 7.2 months; HR, 0.58; 95% CI, 0.39-0.85; P < .01). Patients with an ECOG performance status of 1 or 2 had a shorter OS compared with patients who had an ECOG performance status of 0 (HR, 1.93; 95% CI, 1.30-2.86; P < .01). In the adjusted model, the ERCC1 mutant T allele (HR, 1.54; 95% CI, 1.00-2.39; P = .05), the XPD10 mutant A allele (HR, 0.64; 95% CI, 0.42-0.98; P = .04), and the XRCC1 mutant G allele (HR, 0.51; 95% CI, 0.29-0.91; P = .02) independently predicted OS.
Clinical studies on predictive molecular markers in patients with advanced NSCLC who receive platinum-based chemotherapy have focused mainly on tumor gene or protein expression.18, 19 Although some strong associations have been described for tumor expression of ERCC120-22 and RRM123-25 and clinical outcome, to date, these associations have not been implemented into clinical decision making. One reason for this may be difficulty handling tumor material and the need for specialized laboratories to analyze gene expression. A second reason may be that often only small tumor samples are available in patients with advanced NSCLC. In these cases, available tissue may not be adequate for tumor gene expression profiling. Conversely, germline gene polymorphisms are easy to measure, constant over time, and represent an ideal tool for developing markers in patients with advanced NSCLC. In the current study, the mutant ERCC1 T allele was associated with a lower response to platinum-gemcitabine chemotherapy and with lower PFS and OS rates. In addition, the mutant XPD10 A allele was associated with better response to treatment and with higher PFS and OS rates, but the difference was significant only for OS. Accordingly, stratifying patients for ERCC1 and eventually XPD10 may be of clinical interest when tumor shrinkage is important, such as in the neoadjuvant setting. For toxicity, the mutant C alleles of both RECQ1 and CDA were associated with more frequent leukocytopenia and the need for dose reductions; and, to our knowledge, this has not been described previously. In a pharmacokinetic subanalysis of 37 patients who were included in the same study, we demonstrated that the CDA mutant C allele (CDA*2) results in greater exposure to gemcitabine, which subsequently may result in more severe leucozytopenia.26 In addition, the GSTP1 GG genotype was associated with severe polyneuropathy. Similar results have been described for GSTP1 in patients with colorectal cancer receiving oxaliplatin-based chemotherapy, 27 but not in others.28, 29 In the current study, 18% of carriers of the GSTP1 GG genotype experienced severe platinum-associated peripheral polyneuropathy. If this is confirmed, then these patients may be candidates for receiving carboplatin instead of cisplatin to avoid severe polyneuropathy.
Eight major clinical trials have assessed ERCC1 mutations in patients with lung cancer. Ryu and colleagues studied 109 patients with advanced NSCLC who received cisplatin-based chemotherapy and observed that carriers of the ERCC1 CC genotype had an improved OS compared with carriers of the CT or TT genotypes (486 days vs 281 days).6 Similarly, Isla and colleagues studied 62 patients with advanced NSCLC receiving docetaxel-cisplatin chemotherapy, and found carriers of the ERCC1 CC genotype to have an improved OS.4 Kalikaki and colleagues assessed 119 patients with NSCLC who received platinum-based chemotherapy.5 In that study, although carriers of the ERCC1 C allele were more likely to respond to chemotherapy, this did not translate into an improved OS.5 No correlation between the ERCC1 genotype and clinical outcome was observed in 4 other studies.7-10 Some studies specifically examined the impact of ERCC1 mutations on treatment response in patients with advanced NSCLC who received platinum-based chemotherapy.4-6, 30-32 In the study by Li and colleagues, the T allele of ERCC1 was associated with improved treatment response in 115 patients with advanced NSCLC who received platinum-based chemotherapy, but the difference was relatively small (50% response rate in carriers of the CT genotype vs 43% in carriers of the CC genotype).32 In the study by Kalikaki and colleagues, the ERCC1 C allele was associated with improved response (34.7% vs 5.5% in carriers of the homozygous mutant TT genotype).5 No association was observed between the ERCC1 codon 118 polymorphism and treatment response in other studies.4, 6 Recently, Horgan and colleagues pooled data from 90 trials and identified no significant association between ERCC1 codon 118 mutations and OS in Caucasian patients with lung cancer who received platinum chemotherapy; however, the T allele was a significant predictor of poor outcome in Asian patients who received cisplatin-based chemotherapy.33 Potentially, differences or imbalances in second-line or later treatment may have influenced the correlation between ERCC1 genotypes and OS in the various studies. For example, carriers of the ERCC1 CC genotype tended to have a better disease-free survival compared with carriers of the CT or TT genotypes in the study by Takenaka and colleagues, but this did not translate into a better OS.8 In the studies by Zhou et al10 and de las Penas et al, 7 no information was available on the correlation between ERCC1 genotypes, treatment response, and PFS.
In our study, carriers of the XRCC1 G allele had a lower OS, similar to what was described by Giachino and colleagues in 203 patients with NSCLC and 45 patients with small cell lung cancer, although the difference was of borderline statistical significance.12 This was confirmed further in a study by Gurubhagavatula and colleagues, who observed that the XRCC1 G allele was associated with a worse OS in 103 patients with advanced NSCLC who received platinum-based chemotherapy.11 In addition, those authors reported that a higher number of variant alleles within XPD codon 312 and XRCC1 codon 399 were associated with a lower OS.11 Two Asian studies did not identify any association between XRCC1 polymorphisms and clinical outcome in patients with advanced NSCLC who received platinum-based chemotherapy.13, 14 Overall, the data are consistent on the association of the XRCC1 G allele with a worse clinical outcome in patients with advanced NSCLC who receive platinum-based chemotherapy. Therefore, patients carrying the XRCC1 mutant G allele may be good candidates for receiving nonplatinum-containing chemotherapy. Two clinical studies assessed the predictive value of the RRM1 promoter polymorphism in patients with advanced NSCLC. The study by Kim and colleagues included 97 patients with advanced NSCLC who received platinum-gemcitabine chemotherapy.31 The response rate to platinum-gemcitabine chemotherapy was significantly higher in carriers of the heterozygous mutant RRM1 AC genotype compared with carriers of the RRM1 wild-type allele (65.5% vs 42.6%). Similar to our study, no such relation was observed in a second study in which presence of the RRM1 −C37A mutant allele was determined in 62 patients with advanced NSCLC who received gemcitabine-based chemotherapy.30
The strengths of our current study are its prospective nature, the assessment of candidate genes with a strong pathophysiologic association to the metabolism or the mode of action of the anticancer drugs, and the variety of clinical endpoints assessed. Limitations of the study include the fact that tumor mutation status was not assessed, and no genome-wide analytical approach was chosen, potentially resulting in missing other yet unknown genetic predictors for platinum-based chemotherapy in this group of patients. Furthermore, it remains unclear whether the association of gene polymorphisms with clinical outcome is predictive for patients who receive platinum-based chemotherapy or prognostic for patients with advanced NSCLC, because no patient group without chemotherapy treatment was available for comparison.
In conclusion, to our knowledge, this is the first prospective study to date in patients with advanced NSCLC describing predictive germline polymorphisms not only for platinum-gemcitabine clinical activity (ERCC1, XPD10) but also for toxicity (GSTP1, RECQ1, CDA). Nonplatinum-containing chemotherapy in carriers of the ERCC1 T allele or the XPD10 G allele should be studied prospectively.
M. Joerger is supported by a fellowship grant funded by the European Society of Medical Oncology, by a Novartis-International Union Against Cancer Translational Cancer Research Fellowship funded by Novartis AG, and by a research grant from the Swiss National Science Foundation (PBBSB-102331).
M. Joerger received a Translational Cancer Research Fellowship funded by Novartis AG.