The genotype frequencies of the different polymorphisms are shown in Table 3 . Regarding the ERCC1 (Asn118Asn) polymorphism, although the T/T genotype has been characterized as the "variant" by convention, in our study the T/T genotype was found in higher fre-quencies and was used as a reference in all analyses. Among the 8 SNPs studied, the ERCC1 (Asn118Asn) was in Hardy-Weinberg disequilib-rium. Moreover, among the 5 SNPs, namely ERCC1 (Asn118Asn), ERCC1 (C8092A), XPD (Asp312Asn), XPD (Lys751Gln) and XRCC1 (Arg399Gln), located on chromosome 19q13.3, the SNPs of ERCC1 showed the strongest linkage disequilibrium (D' = 0.814; r2 = 0.208; P < .00001). The A allele of ERCC1 (C8092A) was associated with the C allele of ERCC1 (Asn118Asn). There were no significant differences between the presence of the various polymorphisms and age, gender, disease stage, or histologic type (P > .05, by Fisher's exact test [data not shown]). Two (1.7%), 5 (4.2%), 14 (11.8%), 23 (19.3%), 24 (20.2%), 27 (22.7%), 21 (17.6%) and 3 (2.5%) patients harbored 0, 1, 2, 3, 4, 5, 6 or 7 polymorphic variants, respectively, of the genes analyzed. Two patients harbored only common variants, whereas 117 patients had at least one polymorphic variant.
Association Between Polymorphisms and Response to Platinum-based Chemotherapy
Thirty-seven patients experienced a partial response (PR) whereas 25 experienced stable disease (SD) and 57 experienced progressive disease (PD). As shown in Table 3, the ERCC1 Asn118Asn polymorphism was significantly associated with response to platinum-based chemotherapy (P = .012). Patients bearing the favorable ERCC1 118 genotypes (C/C, C/T) were more likely to respond to platinum-based chemotherapy compared with those with the unfavorable genotype (T/T). Multi-variate analysis adjusted for gender, tumor histology and disease stage revealed that the ERCC1 Asn118Asn polymorphism (TT vs. CT/CC; OR, 0.10; 95% CI, 0.013–0.828; P =.033) was independently as-sociated with an objective response to platinum-based chemotherapy. Conversely, there was no significant association between the response rate and the other (ERCC1 8092, XPD 312, XPD 751, XRCC1 399, GSTP1, GSTM1 and GSTT1) genotypes (Table 3).
Association Between Polymorphisms and Overall Survival
The median OS of patients treated with platinum-based chemotherapy as first-or second-line treatment was 9.9 and 12.6 months, respectively (P = .173). Because there was no difference in the patient characteristics between first-and second-line treatment groups, all patients were pooled and analyzed together. There was no significant association between the various clinicopathological parameters, such as gender, tumor stage, and histology, and patients' overall survival (data not shown).
Patients' genotype analysis demonstrated that those who were ho-mozygous for the ERCC1 C8092 allele had a worse OS compared with those carrying at least one A allele (C/A or A/A) (Log-rank test, P = .009; Table 4 ; Figure 1A). Even when considering the C/A and A/A genotypes separately, with median survival times of 13 and 15.9 months, respec-tively, the log-rank test comparing all three groups was still significant (P = .031). In addition, there was a statistically significant difference in OS associated with the XRCC1 Arg399Gln polymorphism (log-rank test, P = .001; Table 4 ; Figure 1B). The median survival times of codon 399 G/G, G/A, and A/A genotypes were 14.8, 11.3, and 7.1 months, respectively. Furthermore, an improved OS was observed in patients with the GSTT1 present genotype (n = 107) compared with patients with the GSTT1 null genotype (n = 6; Breslow test, P = .046). The median survival times of GSTT1 present and null genotype patients were 11.3 and 4.3 months, respectively ( Table 4 ). No other association could be identified between the remaining polymorphisms and patients' overall survival.
Kaplan Meier Estimates of Survival of Patients with Non Small Cell Lung Cancer
(A) Genotypes of ERCC1 C8092A (CC vs. CA/AA); (B) genotypes of XRCC1 N399Q (AA vs. GA vs. GG); and (C) presence of increasing numbers of polymorphic alleles of ERCC1 (C8092A, N118N).
Multivariate Cox regression analysis adjusted for tumor histology, stage of disease, and polymorphisms of DNA repair genes revealed a significant effect of ERCC1 C8092A (CC vs. CA/AA; HR, 1.7; 95% CI, 1.8-2.8; P = .02) and XRCC1 Arg399Gln (AA vs. GA/GG; HR 3.1; 95% CI, 1.4-6.8; P = .005) on patients' overall survival.
We also examined the association between the number of ERCC1, XPD, and GST polymorphisms and OS. The presence of ERCC1 polymorphic variants was associated with longer OS; median survival times were 7.46, 10.23, and 14.13 months for patients with common alleles, one polymorphic variant, and two polymorphic variants, respectively (log-rank test, P = .011; Χ2 for trend, P = .003; Figure 1C). In a Cox regression model adjusted for the clinical parameters and the number of SNPs in GSTs, XPD, and ERCC1, the presence of two ERCC1 polymorphic variants was an independent prognostic factor associated with longer survival (HR, 2.5; 95% CI, 1.26-4.96; P = .009).
Association Between Polymorphisms and Toxicity
The GSTT1 genotype was associated with toxicity because 76% (n = 13) of patients carrying the GSTT1 present genotype developed grade 3 or higher neutropenia compared with 24% (n = 4) of those carrying the null genotype (P = .006). No other significant association was found between the other genotypes and hematological or nonhematological toxicity.
Clin Lung Cancer. 2009;10(2):118-123. © 2009 other
Cite this: DNA Repair Gene Polymorphisms Predict Favorable Clinical Outcome in Advanced Non–Small-Cell Lung Cancer - Medscape - Mar 01, 2009.