OL Oncology Letters 1792-1074 1792-1082 D.A. Spandidos 10.3892/ol.2022.13649 OL-25-02-13649 Articles Side effects of tyrosine kinase inhibitors therapy in patients with non-small cell lung cancer and associations with EGFR polymorphisms: A systematic review and meta-analysis ObradovicJasmina 1 TodosijevicJovana 2 JurisicVladimir 3 Department of Sciences, Institute for Information Technologies, University of Kragujevac, 34000 Kragujevac, Republic of Serbia Institute of Biology and Ecology, Faculty of Sciences, University of Kragujevac, 34000 Kragujevac, Republic of Serbia Faculty of Medical Sciences, University of Kragujevac, 34000 Kragujevac, Republic of Serbia Correspondence to: Professor Vladimir Jurisic, Faculty of Medical Sciences, University of Kragujevac, 69 Svetozara Markovića, 34000 Kragujevac, Republic of Serbia, E-mail: jurisicvladimir@gmail.com 02 2023 23 12 2022 25 2 62 13092022 05122022 Copyright © 2022, Spandidos Publications 2022

Rash and diarrhea are common side effects of tyrosine kinase inhibitor (TKI) therapy administered to patients with non-small cell lung cancer (NSCLC). The polymorphisms of the epidermal growth factor receptor (EGFR) gene may be a potential predictor of these side effects. The aim of the present meta-analysis was to examine the association of EGFR polymorphisms and TKI-associated toxicities. Electronic databases (PubMed, Scopus and ISI Web of Science) were searched for relevant studies. According to the inclusion and exclusion criteria, a search of the databases identified 4,918 results, among which 6 clinical trials were obtained with 1,318 patients with NSCLC. A total of 9 EGFR single nucleotide polymorphisms (SNPs) associated with TKI toxicity were identified including, rs11568315, rs712829, rs712830, rs2227983, rs2075102, rs2293347, rs11977388, rs4947492 and rs884225. The data associated with skin toxicity from rs11568315, rs712829 and rs712830 were analyzed in the present meta-analysis. Data from rs11568315 were also analyzed in relation to diarrhea. Among all the examined SNPs, statistically significant results were obtained under the dominant genetic model for CA repeats in rs11568315 (SS vs. SL+LL) with skin toxicity. The long CA repeat (SL+LL) carriers were more likely to experience skin toxicity associated with TKIs (P=0.005). By contrast, there was no significant result for diarrhea (P=0.661) under dominant genetic model for CA repeats.

 EGFR variant polymorphism non-small cell lung cancer tyrosine kinase inhibitor therapy toxicity Ministry of Education, Science and Technological Development, Serbia 451-03-68/2022-14/200378 The present study was supported by the Ministry of Education, Science and Technological Development, Serbia (no. OI 175056 under agreement number 451-03-68/2022-14/200378). Introduction

As the most prevalent form of lung cancer, non-small cell lung cancer (NSCLC) is reported to be one of the deadliest types of cancer in the world, with 2,206,771 newly diagnosed cases and 1,796,144 new deaths recorded in 2020 (1,2). In patients with advanced NSCLC, platinum-based chemotherapy is the first line treatment, but it is usually cytotoxic and has a short progression-free survival (PFS) time of 3–5 months and an overall survival (OS) time of ≤10 months (3). Targeted therapy has been developed to prevent epidermal growth factor receptor (EGFR) activation (3–6). EGFR is a transmembrane protein and a potent transducer of altered signals in tumor cells. There are two ways of blocking the EGFR: Either by blocking the ligand from binding to the receptor extracellular domain with anti-EGFR monoclonal antibodies (cetuximab) or by reversibly binding the small molecule tyrosine kinase inhibitors (TKIs) to the receptor intracellular tyrosine kinase domain (4,5). Thus, the introduction of the first TKI generation drugs, gefitinib and erlotinib, resulted in markedly higher treatment response rates (73.7% for TKI compared with 30.7% for chemotherapy), and the median PFS time increased to 10–13 months for patients with NSCLC (6–8).

For accurate therapeutic decisions to be made in the management of patients with NSCLC, it is essential to find molecular markers that can identify patients who will respond most effectively to treatment. Promising molecular identifiers include mutations in the EGFR gene. In patients carrying exon 19 deletions and point mutations in exon 21, a significant clinical benefit following treatment with TKIs was observed (9,10). However, acquired resistance in connection with EGFR T790 mutations limited the efficacy of the EGFR-TKI (11,12). The role of polymorphisms [including single nucleotide polymorphisms (SNPs) and short tandem CA repeats] of the EGFR gene as another potential molecular target that improved clinical outcomes is well-established (13–16). Recently, a meta-analysis elucidated that among rs712829 (−216G>T), rs11568315 (CA repeat), rs2293347 (D994D) and rs4947492, −216G>T and variable CA repeat polymorphisms significantly affected OS and PFS time in gefitinib- or erlotinib-treated patients with NSCLC (17).

EGFR-TKI therapy is associated with side effects, primarily in the form of skin or gastrointestinal toxicities (e.g., skin rash or diarrhea). Although skin toxicities are not lethal or dose-limited, they frequently occur with EGFR-TKIs and affect patient quality of life (18). Among usual skin toxicities, such as xerosis, pruritus, paronychia, mucositis and increased growth of eyelashes or facial hair, skin rash is the most prevalent (19–22). Notably, patients with NSCLC that develop skin rashes are better responders to EGFR-TKI therapy and have a longer median overall survival time (18–23). EGFR SNPs have been examined in association with survival in NSCLC (17,18); they may provide insight into therapy outcomes, particularly the potential side effects associated with TKIs (23–28). Literature analysis discovered notable inconsistency in previously published reports. While some studies found associations with EGFR genotypes and TKI toxicity (23–26), others did not (27). Additionally, previous meta-analyses investigated EGFR mutations, but not EGFR polymorphisms and therapy side effects in patients with NSCLC (29,30), or toxicity in relation to radiotherapy (31). With regards to these discrepancies and the role of EGFR SNPs as potential determinants of treatment outcome, the aim of the present meta-analysis was to determine whether the molecular mechanisms involving EGFR SNPs were associated with EGFR-TKI therapy side effects.

 Materials and methods <sec> <title>Search strategy and study selection

The present study was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (PRISMA) (32). The systematic search for the relevant studies was performed using electronic databases, PubMed, Scopus and ISI Web of Science. Searches were performed considering EGFR polymorphisms and side effects of TKI therapy in patients with NSCLC. The search had the following retrieval strategy for the PubMed database: [(‘receptor, epidermal growth factor’ (MeSH Terms) OR EGFR (All Fields)) AND (gene(tiab) OR ‘polymorphism, genetic’(MeSH Terms)) AND (‘carcinoma, non-small-cell lung’ (MeSH Terms) OR NSCLC (All Fields)) AND (‘drug therapy’ (Subheading) OR treatment (All Fields) OR ‘erlotinib hydrochloride’ (MeSH Terms) OR ‘gefitinib’ (MeSH Terms) OR TKI OR ‘TK inhibitors’ OR ‘tyrosine kinase inhibitors’ OR ‘Tyrosine-kinase inhibitor’) AND response (All Fields)) OR Prognosis (MeSH)) OR toxic (MeSH)) OR toxicity (MeSH)) OR side effect (MeSH)) AND (humans (MeSH))]. The Scopus and ISI Web of Science databases were also searched with necessary modifications to the PubMed search query. The full search string is available from the corresponding author upon request. Finally, additional studies were searched for in the bibliographies of the selected eligible studies or reviews.

 Selection criteria

All studies fulfilling the following inclusion criteria were eligible: i) Studies published from January 1, 2009, to February 13, 2019; ii) studies published in English; iii) studies involving human subjects; iv) patients >18 years old with histopathologically confirmed NSCLC who received EGFR-TKI therapy and v) clinical trials or observational studies that investigated associations between EGFR polymorphisms and any side effects of TKI therapy. In the systematic review, studies were excluded based on the following criteria: i) Meta-analyses, editorials, letters, commentaries, systematic or narrative reviews; ii) not in the English language; iii) duplicate publications or studies involving animal or cell experimental models; iv) studies investigating EGFR polymorphisms and TKI adverse effects but not reporting their associations; v) single study reports of EGFR polymorphisms associated with TKI toxicities (skin toxicity or diarrhea), or other side effects (such as hepatotoxicity) due to being unable to make comparisons due to the lack of data from other studies and vi) randomized control trial (RCT) studies that did not report genotype numbers data, even though the odds ratio (OR) was reported.

 Data extraction

Extracted studies from the electronic databases were first merged and duplicates were removed. A total of 2 authors (JO and JT) independently performed a manual search of titles and abstracts of potentially eligible studies according to the inclusion and exclusion criteria. Any discrepancies were resolved by discussion or by consulting the third author (VJ). Finally, the following data were extracted from the full texts based on the prior determined datasheet: The first author, year of publication, country, study type, study period, number of patients, median age, sex and ethnicity of patients, percentage of smokers, clinical stage, histology, median follow-up (in months), TKI treatment dosage, additional therapy, toxicity assessment, adverse effects of treatment, available EGFR genotype, variant location, SNP database identifier and number of patients/genotype.

 Quality assessment

The Newcastle-Ottawa Quality Assessment Scale (NOS) (33) for cohort studies and the Jadad Scale for RCTs (34) were used to assess the methodological quality of the studies included. For the NOS scale, the overall maximum quality score was 9 points; for the Jadad Scale, the score was 5 points. The reviewers (JO and JT) independently evaluated the quality of the studies with discrepancies resolved by consensus.

 Statistical analysis

When ≥2 studies had available EGFR polymorphic genotypes associated with TKI therapy side effects, meta-analysis was conducted. To examine heterogeneity between the eligible studies, Cochran's Q statistics and I2 statistics were applied. I2 was interpreted as follows: 0, no heterogeneity; 25, low heterogeneity; 50, moderate heterogeneity and 75%, high heterogeneity (35). The random effect model was used when there was significant heterogeneity between studies (P<0.05; I2>50%), otherwise, the fixed effect model was applied (36). Galbraith's plot was used to identify potential sources of heterogeneity (37). If heterogeneity was present, subgroup analyses of OR were conducted according to the available EGFR SNPs. The dominant genetic model (wild-type homozygote vs. heterozygote + mutant homozygote) of all three EGFR SNPs (rs11568315, rs712829 and rs712830) was used to calculate OR. The available adverse effects for the analysis were skin toxicity (skin rash) and gastrointestinal toxicity (diarrhea). For comparison, the adverse effects were combined and used as any grade vs. the absence of adverse effects. Sensitivity analysis was also performed to determine whether the results would be affected by excluding the study with the smallest sample size. The publication bias of the enrolled studies was tested with Begg's and Egger's tests, as well as funnel plots (38,39). P<0.05 was considered to indicate a statistically significant difference. STATA software package v.15 (StataCorp LP) was used for all statistical analysis.

 Results <sec> <title>Study selection

The initial search of databases identified 4,918 results (PubMed, 881; ISI Web of Science, 395; Scopus, 3,642; Fig. 1). An additional study was included after reading the bibliographies of the full-text articles. After merging into the single datasheet and removing duplicates, 4,036 studies remained. Of these, 3,980 were excluded and 56 full-text articles were used to assess eligibility. Of these 56 articles, 50 were excluded due to not fulfilling the inclusion criteria. Finally, 6 clinical trials were included in the systematic review which contained 1,318 patients with NSCLC. A total of 4 studies were included in the meta-analysis.

 Characteristics of the studies

The six studies from the search included four cohort studies (23,25,26,28) containing 316 patients and 2 RCTs (24,27) containing 1,002 patients. The studies were published from 2009–2017, with sample sizes ranging from 52–760 patients. A total of two studies were from Asia (Taiwan and China), two from Europe (Germany and Italy), one from Canada and one RCT was from a consortium of counties (Canada, Italy, South Korea and Brazil). The number of male patients in the studies was 33–67%. The percentage of smokers was 12–76%, while the median age was 56–68 years. Most of the patients had adenocarcinoma histology and were in clinical stages IV, IIIB and IIIA (23–28). Only one study reported a median follow-up of 12 months (23). The EGFR-TKI therapy type for patients with NSCLC in all examined reports was gefitinib (250 mg/day) and erlotinib (150 mg/day), except in one study where cetuximab (250 or 500 mg/m2) or panitumumab (6 mg/kg) was prescribed (25). Additionally, four studies reported patients that had been previously treated with cisplatin (24–27). Adverse effects were skin (rash) and gastrointestinal toxicity (diarrhea and hepatotoxicity). Toxicity assessment was conducted using the National Cancer Institute's Common Terminology Criteria for Adverse Events (40). The quality of the studies was rated acceptable using the NOS and the Jadad scale (33,34). Adequacy of follow-up was the lowest rated aspect. The characteristics and quality assessment of the included studies are presented in Table I.

A total of nine EGFR SNPs (rs11568315, rs712829, rs712830, rs2227983, rs2075102, rs2293347, rs11977388, rs4947492 and rs884225) relative to TKI toxicity was identified in the literature search, which were provided by seven studies (23–28,41). Of these, four studies reported exact numbers of patients/genotype of EGFR SNPs (rs11568315, rs712829 and rs712830) associated with TKI-caused toxicity and were included in quantitative synthesis (23,25,26,28). Genotypes for all EGFR SNPs for the meta-analysis were merged according to the dominant genetic model. The data for the EGFR SNPs genotype and skin toxicity, diarrhea or hepatotoxicity caused by EGFR TKI therapy are presented in Table II. Certain studies or sets of data were excluded from further analyses. The reasons are outlined below.

Due to lack of data from other studies for comparison, one study was excluded from further analysis, although the study did identify rs884225, a 3′-untranslated region variant c.*774T>C associated with EGFR TKI toxicity (41). Similarly, data for the hepatotoxicity, as well as for four EGFR SNPs were not included (rs2075102, rs2293347, rs11977388 and rs4947492) (27). An RCT study reported pre-calculated ORs for three examined EGFR SNPs, but without precise numbers of patients per genotype (24), and was therefore included in the qualitative, but not the quantitative, analysis. Consequently, another RCT study was excluded from quantitative analysis (27) to avoid comparison between observational and RCT studies, and prevent potential heterogeneity. If zeros present in patient genotype numbers for rs2227983 and rs11568315 interfered with computation or if there were insufficient data for analysis, studies were excluded from quantitative synthesis (23–25,28). In the literature search, three other studies explored EGFR TKI toxicity, as well as EGFR SNPs, but failed to find any associations between them (42–44).

 Side effects of EGFR-TKIs

There was notable inconsistency in the scientific reports describing the association between EGFR SNPs and TKI toxicity in patients with NSCLC. While some articles reported evidence of association with skin toxicity (23–26) or severe diarrhea (26), one article found no association with skin or gastrointestinal toxicities (27).

In patients treated with gefitinib, there was a significant association between SS genotype in CA repeat polymorphism and early G2/3 skin rash (P=0.031), meaning these patients were more likely to develop early G2/3 rash (23). Despite this, the EGFR polymorphisms −216G>T and R521K were not associated with early G2/3 rash (P=0.104 and P=0.720, respectively) (23). Another study on patients treated with erlotinib found a similar result for three EGFR polymorphisms and skin rash: −191C>A, −216G>T and CA repeats (P=1.00, P=0.13 and P=0.34, respectively) (24). Only the EGFR −216/-191GC haplotype was associated with the appearance of skin rash (P=0.029) (25). Nevertheless, the absence of association with skin rash was evidenced for the single EGFR SNPs −191C>A (P=0.62), −216G>T (P=0.147) and CA repeats (P=0.36) (25). Diarrhea was a less frequent toxicity and no significant association between any of the EGFR SNPs or haplotypes with diarrhea was observed (25). However, in another study, severe diarrhea occurred in patients with NSCLC treated with gefitinib, most frequently in carriers of −191C>A, −191A>A (P<0.0001) and −216G>G genotypes (P<0.01) (26). There was no significant association between EGFR CA repeat polymorphisms and skin or gastrointestinal toxicity, nor any association between EGFR polymorphism and skin toxicity (26).

 Toxicity

The most common adverse effects associated with TKIs in treating advanced NSCLC were skin toxicity (78.36%) and diarrhea (20.75%; Table II). One study reported hepatotoxicity (0.89%) (27), but the study was excluded since there were no data from other studies for comparison. Among the studies available for the meta-analysis, gefitinib (250 mg/day) or erlotinib (150 mg/day) were predominant. For data available for genotypes relative to skin toxicity, the OR and 95% confidence interval (CI) were calculated and their effect was summarized in the quantitative synthesis (Fig. 2). This involved three EGFR SNPs (rs11568315, rs712829 and rs712830) obtained from three studies for skin toxicity (23,25,26). Of these, two examined rs11568315 and diarrhea (26,28). The pooled OR for skin toxicity and rs11568315, rs712829 and rs712830 was 1.17 (95% CI, 0.63–2.18; P=0.616) with moderate heterogeneity (I2=57.4%; P=0.022; Fig. 2).

To test heterogeneity, random effect model and subgroup analyses were performed. Subgroup analysis for skin toxicity showed that the OR for rs11568315 was 2.72 (95% CI, 1.34–5.49; P=0.005) without heterogeneity (I2=0.0%; P=0.533). A statistically significant result for skin toxicity (z=2.785 and P=0.005) were obtained under the dominant genetic model for rs11568315 (SS vs. SL + LL). OR for rs712829 was 0.81 (95% CI, 0.28–2.36; P=0.700) with moderate heterogeneity (I2=65.1%; P=0.057) and OR for rs712830 was 0.62 (95% CI, 0.29–1.35; P=0.229) with no heterogeneity (I2=0.0%, P=0.625; Fig. 3). Data for diarrhea was only available for rs11568315 (data not shown). It was tested in two studies using the fixed effects model (OR, 1.21; 95% CI, 0.52–2.82), with no evidence of heterogeneity (I2=0.0%; P=0.422) and without statistically significant association (P=0.661) (26,28).

 Publication bias and sensitivity analysis

The results of the sensitivity analysis regarding toxicity were relatively stable. The overall effective size was not affected by exclusion of each of the studies, even by a study with a smaller sample size (OR, 6.87; 95% CI, 1.17–2.28; Fig. 4A) (23). The funnel plot for EGFR SNPs and TKI skin toxicity in patients with NSCLC was roughly symmetric (Fig. 4B). Begg's funnel plot and Egger's regression test (P=0.545) were used to test the publication bias, but no significantly different results were obtained (Fig. 4C and D). Similarly, the funnel plot revealed no potential bias of rs11568315 (CA repeat) and TKI-caused diarrhea (data not shown). Galbraith's plot identified no source of heterogeneity relative to skin toxicity (data not shown).

 Discussion

The present systematic review involved the analysis of two RCTs and four cohort studies to test the association of EGFR polymorphisms with the potential toxicity of TKI therapy regimens in patients with NSCLC. A total of 1,123 patients per genotype were observed with any TKI-associated toxicity. A total of four studies provided data for the meta-analysis (23,25,26,28), while six were involved in quality analysis (23–28). In the literature search, nine EGFR SNPs relative to TKI toxicity were identified: rs11568315, rs712829, rs712830, rs2227983, rs2075102, rs2293347, rs11977388, rs4947492 and rs884225 (23–28,41). Of these, enough data was available for three (rs11568315, rs712829 and rs712830) to be included in the meta-analysis.

Our recent meta-analysis showed that CA repeat polymorphism and −216G>T significantly affected survival in patients with NSCLC treated with TKI (17). In light of the inconsistency of previous reports (23–31), the present meta-analysis was performed to extend our previous findings and to analyze the effect of the EGFR polymorphisms and TKIs on NSCLC.

A number of studies in the present review founded an association between some EGFR polymorphisms and TKI-related skin toxicity (23,25,26) or diarrhea (26,27). Contradictory results were also detected in previous studies published before 2009 (45–47). The most common TKI adverse effects in the present meta-analysis were skin toxicity and diarrhea, which were 78.36% and 20.75%, respectively (concerning any grade of toxicity vs. no toxicity). They were separately analyzed in the meta-analysis. The pooled OR for three EGFR SNPs (rs11568315, rs712829 and rs712830) was 1.17 (95% CI, 0.63–2.18) without a statistically significant overall effect on skin toxicity (P=0.616). In further analysis, a moderate overall heterogeneity (I2=57.4%; P=0.022) was observed. To explore the heterogeneity further, a subgroup analysis was performed and the random effect model was applied. The subgroup analysis involved three EGFR SNPs (rs11568315, rs712829 and rs712830) concerning skin toxicity (23,25,26). The source of heterogeneity (I2=65.1%; P=0.057) was likely due to the −216G>T (rs712829) polymorphism (26). The CIs were overlapping the line of no effect for all three studies, suggesting the result was not statistically significant. A total of two studies favored the GG genotype for −216G>T (rs712829) and skin toxicity (23,25), which contrasts the GT+TT genotype favored by Giovannetti et al (26). Most importantly, there was no heterogeneity for the other two SNPs examined (rs11568315, I2=0.0%; P=0.533; rs712830, I2=0.0%; P=0.625).

Chemotherapy is the first-line treatment for patients with NSCLC, but notable improvements in the response rate have been observed following the application of the TKIs gefitinib and erlotinib (6–8,48). However, resistance, as well as adverse effects, is common in this therapy regimen. Typical side effects of the drugs used in NSCLC treatment (for both monoclonal antibodies and small molecule TKIs) are skin rash and diarrhea (49,50). Since the EGFR is commonly affected by somatic mutations in altered neoplastic cells and the EGFR gene is highly polymorphic, the potential cause of those toxic manifestations of drugs may be EGFR genetic variability (11,13–16). SNPs or microsatellite tandem repeats are typically found in the EGFR promoter region and intron 1. These notably affect EGFR gene expression and may mediate response to TKI therapy. A CA single sequence repeat polymorphism (rs11568315) is located in EGFR intron 1 and it usually comprises 14–21 variable short tandem repeats. The shorter allele is associated with increased EGFR expression and carriers of this polymorphism are better responders to TKI therapy and have prolonged overall survival time (13,14,47,51–54).

Among side effects of TKI therapy, typical skin rash manifestations were in the form of papules and pustules on the scalp, face, neck and upper trunk. To the best of our knowledge, the mechanism of skin rash development has not yet been elucidated. One hypothesis is that there is a genetic susceptibility for rash development, where altered EGFR expression alters the TKI response (11,13–16,28). Another is that poor vascularization of the tumor tissue and drug concentrations at a level that does not inhibit tumor growth may cause a skin rash by over-saturation of EGFR (18,55). There is evidence of a significant association between skin rash and an improved outcome in patients with NSCLC (18,56). Skin rash has been reported to be a predictor of tumor response (25) and EGFR CA repeat is a valuable predictor of early G2/3 rash (23). Previous studies have reported that lower number CA repeat carriers develop skin toxicity when treated with gefitinib (13,23,57), while other studies did not (24–28,45,47,53). In another study where patients with NSCLC were treated with erlotinib, SL allele length was associated with a higher risk of diarrhea (46). In the present meta-analysis, the pooled OR values for CA repeats (rs11568315) and skin toxicity were 2.72 (95% CI, 1.34–5.49). A significant association with skin toxicity was evident under the dominant genetic model. Namely, heterozygote and long alleles (SL + LL) or prevalently long CA repeat carriers were more likely to develop TKI-related skin toxicity (P=0.005). However, it is probable that short CA carriers would be less likely develop skin rash. There was no association between CA repeats and diarrhea (P=0.661).

The other well-examined SNPs, −191C>A (rs712830) and −216G>T (rs712829) polymorphisms, are located in the EGFR promoter region and are associated with enhanced EGFR mRNA expression (14,53,58). A previous meta-analysis revealed that any genotype with T allele for −216G>T showed an association with higher response and disease control rates and longer PFS and OS times than GG homozygote carriers (59). Another meta-analysis elucidated that the −216G>T polymorphism significantly affected OS and PFS times in patients with NSCLC treated with gefitinib or erlotinib (17). Both of the aforementioned polymorphisms are reported to be in linkage disequilibrium (D'=1.0) (25). Considering their association with toxicities, a study reported the haplotypes showing association with the appearance of skin rash (25). Other studies reported that the T allele of −216G>T was significantly associated with high-risk of TKI-induced skin rash (24) or diarrhea (14). An association between −216G>T and −191C>A with grade >1 diarrhea has also been reported (26). Contrary to these findings, the present meta-analysis observed no significant association for EGFR SNPs −216G>T and −191C>A with skin toxicity (P=0.700 and P=0.229, respectively), which agreed with the findings from previous studies (23,24,27,45).

The advantage of the present meta-analysis over previous meta-analyses is the examination of commonly used TKIs (such as gefitinib and erlotinib) and their toxicity, while other studies involved a single therapeutic agent (60,61). Other analyses investigated EGFR mutations, not EGFR polymorphisms (29,30) or their association with toxicity (59), or they only investigated toxicity in relation to radiotherapy (31). The present meta-analysis has certain limitations. Firstly, some studies included in the analysis had small sample sizes so consistent conclusions could not be obtained, as with the RCTs that have larger sample sizes. A total of two RCTs were excluded from the meta-analysis (24,27), since one study alone did not provide enough data to be tested. In particular, exact numbers of patients with NSCLC with each EGFR SNP genotype were not reported and the study presented only pre-calculated data for OR (24). The aforementioned RCTs obtained low NOS scores in the quality analysis, whereas the other studies included in the present meta-analysis had relatively good scores. Also, the results of the present meta-analysis were not adjusted for other factors (i.e., demographic factors), although the majority of the studies did not report ethnicity for the examined subjects. Potential bias in the results may be due to the absence of a consensus in the literature of an exact number of CA repeats when reporting short vs. long alleles. The linkage disequilibrium between examined SNPs was not taken into account and selection bias may be present.

In conclusion, the results of the present meta-analysis revealed that out of nine EGFR SNPs related to TKI side effects, rs11568315, rs712829 and rs712830 were associated with skin toxicity. NSCLC carriers of long CA repeats (rs11568315, SL + LL) were more likely to develop TKI-associated skin toxicity than short CA repeats (rs11568315, SS). To establish clear inter-individual benefits of TKI therapy, future RCTs that include a broader genetic panel are required to determine genetic susceptibility to TKI-induced toxicity in patients with NSCLC.

 Acknowledgements

Not applicable.

 Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

 Authors' contributions

Conceptualization and supervision of the study were conducted by VJ. The selection of papers, formal analysis, investigation and writing were conducted by JO. The acquisition of data was conducted by JT. All authors have read and approved the final manuscript. JO, JT and VJ confirm the authenticity of all the raw data.

 Ethics approval and consent to participate

Not applicable.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References SiegelRLMillerKDFuchsHEJemalACancer statisticsCA Cancer J Clin71733202110.3322/caac.2165433433946 SungHFerlayJSiegelRLLaversanneMSoerjomataramIJemalABrayFGlobal cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countriesCA Cancer J Clin71209249202110.3322/caac.2166033538338 YangZYLiuLMaoCWuXYHuangYFHuXFTangJLChemotherapy with cetuximab versus chemotherapy alone for chemotherapy-naive advanced non-small cell lung cancerCochrane Database Syst Rev17CD009948201425400254 HirschFRHerbstRSOlsenCChanskyKCrowleyJKellyKFranklinWABunnPAJrVarella-GarciaMGandaraDRIncreased EGFR gene copy number detected by fluorescent in situ hybridization predicts outcome in non-small-cell lung cancer patients treated with cetuximab and chemotherapyJ Clin Oncol2633513357200810.1200/JCO.2007.14.011118612151 PirkerRPereiraJRvon PawelJKrzakowskiMRamlauRParkKde MarinisFEberhardtWEEPaz-AresLStörkelSEGFR expression as a predictor of survival for first-line chemotherapy plus cetuximab in patients with advanced non-small-cell lung cancer: Analysis of data from the phase 3 FLEX studyLancet Oncol133342201210.1016/S1470-2045(11)70318-722056021 MaemondoMInoueAKobayashiKSugawaraSOizumiSIsobeHGemmaAHaradaMYoshizawaHKinoshitaIGefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFRN Engl J Med36223802388201010.1056/NEJMoa090953020573926 ZhouCWuYLChenGFengJLiuXQWangCZhangSWangJZhouSRenSErlotinib versus chemotherapy as first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer (OPTIMAL, CTONG-0802): A multicentre, open-label, randomised, phase 3 studyLancet Oncol12735742201110.1016/S1470-2045(11)70184-X21783417 LeeCKDaviesLWuYLMitsudomiTInoueARosellRZhouCNakagawaKThongprasertSFukuokaMGefitinib or erlotinib vs chemotherapy for EGFR mutation-positive lung cancer: Individual patient data meta-analysis of overall survivalJ Natl Cancer Inst109djw279201710.1093/jnci/djw279 LynchTJBellDWSordellaRGurubhagavatulaSOkimotoRABranniganBWHarrisPLHaserlatSMSupkoJGHaluskaFGActivating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinibN Engl J Med35021292139200410.1056/NEJMoa04093815118073 PaezJGJännePALeeJCTracySGreulichHGabrielSHermanPKayeFJLindemanNBoggonTJEGFR mutations in lung cancer: Correlation with clinical response to gefitinib therapyScience30414971500200410.1126/science.109931415118125 YunCHMengwasserKETomsAVWooMSGreulichHWongKKMeyersonMEckMJThe T790M mutation in EGFR kinase causes drug resistance by increasing the affinity for ATPProc Natl Acad Sci U S A10520702075200810.1073/pnas.070966210518227510 GoldbergSBOxnardGRDigumarthySMuzikanskyAJackmanDMLennesITSequistLVChemotherapy with Erlotinib or chemotherapy alone in advanced non-small cell lung cancer with acquired resistance to EGFR tyrosine kinase inhibitorsOncologist1812141220201310.1634/theoncologist.2013-016824072220 AmadorMLOppenheimerDPereaSMaitraACusatisGIacobuzio-DonahueCBakerSDAshfaqRTakimotoCForastiereAHidalgoMAn epidermal growth factor receptor intron 1 polymorphism mediates response to epidermal growth factor receptor inhibitorsCancer Res6491399143200410.1158/0008-5472.CAN-04-103615604284 LiuGGurubhagavatulaSZhouWWangZYeapBYAsomaningKSuLHeistRLynchTJChristianiDCEpidermal growth factor receptor polymorphisms and clinical outcomes in non-small-cell lung cancer patients treated with gefitinibPharmacogenomics J8129138200810.1038/sj.tpj.650044417375033 JungMChoBCLeeCHParkHSKangYSKimSKChangJKimDJRhaSYKimJHLeeJHEGFR polymorphism as a predictor of clinical outcome in advanced lung cancer patients treated with EGFR-TKIYonsei Med J5311281135201210.3349/ymj.2012.53.6.112823074112 Winther-LarsenAEbertEBMeldgaardPSorensenBSEGFR gene polymorphism predicts improved outcome in patients with EGFR mutation-positive non-small cell lung cancer treated with erlotinibClin Lung Cancer20161166.e161201910.1016/j.cllc.2019.02.01130898568 JurisicVVukovicVObradovicJGulyaevaLFKushlinskiiNEDjordjevićNEGFR polymorphism and survival of NSCLC patients treated with TKIs: A systematic review and meta-analysisJ Oncol20201973241202010.1155/2020/197324132256580 Peréz-SolerRSaltzLCutaneous adverse effects with HER1/EGFR-targeted agents: Is there a silver lining?J Clin Oncol2352355246200510.1200/JCO.2005.00.691616051966 SegaertSVan CutsemEClinical signs, pathophysiology and management of skin toxicity during therapy with epidermal growth factor receptor inhibitorsAnn Oncol1614251433200510.1093/annonc/mdi27916012181 AgeroALDuszaSWBenvenuto-AndradeCBusamKJMyskowskiPHalpernACDermatologic side effects associated with the epidermal growth factor receptor inhibitorsJ Am Acad Dermatol55657670200610.1016/j.jaad.2005.10.01017010747 HuJCSadeghiPPinter-BrownLCYasharSChiuMWCutaneous side effects of epidermal growth factor receptor inhibitors: Clinical presentation, pathogenesis, and managementJ Am Acad Dermatol56317326200710.1016/j.jaad.2006.09.00517141360 BianchiniDJayanthAChuaYJCunninghamDEpidermal growth factor receptor inhibitor-related skin toxicity: Mechanisms, treatment, and its potential role as a predictive markerClin Colorectal Cancer73343200810.3816/CCC.2008.n.00518279575 HuangCLYangCHYehKHHuFCChenKYShihJYLinZZYuCJChengALYangPCEGFR intron 1 dinucleotide repeat polymorphism is associated with the occurrence of skin rash with gefitinib treatmentLung Cancer64346351200910.1016/j.lungcan.2008.09.00918995924 LiuGChengDDingKMaitreALLiuNPatelDChenZSeymourLShepherdFATsaoMSPharmacogenetic analysis of BR.21, a placebo-controlled randomized phase III clinical trial of erlotinib in advanced non-small cell lung cancerJ Thorac Oncol7316322201210.1097/JTO.0b013e31824166c122237259 ParmarSSchumannCRüdigerSBoeckSHeinemannVKächeleVSeeringerAPaulTSeufferleinTStinglJCPharmacogenetic predictors for EGFR-inhibitor-associated skin toxicityPharmacogenomics J13181188201310.1038/tpj.2011.5122158333 GiovannettiEZucaliPAPetersGJCortesiFD'InceccoASmitEFFalconeABurgersJASantoroADanesiRAssociation of polymorphisms in AKT1 and EGFR with clinical outcome and toxicity in non-small cell lung cancer patients treated with gefitinibMol Cancer Ther9581593201010.1158/1535-7163.MCT-09-066520159991 KimLSaiegMDi MaioMGalloCButtsCCiardielloFFeldRChengDGebbiaVBurgioMABiomarker analysis of the phase 3 TORCH trial for first line erlotinib versus chemotherapy in advanced non-small cell lung cancer patientsOncotarget85752857536201710.18632/oncotarget.1572528915692 MaYXinSHuangMYangYZhuCZhaoHZhangYChenLZhaoYLiJDeterminants of Gefitinib toxicity in advanced non-small cell lung cancer (NSCLC): A pharmacogenomic study of metabolic enzymes and transportersPharmacogenomics J17325330201710.1038/tpj.2016.3127089937 PetrelliFBorgonovoKCabidduMLonatiVBarniSRelationship between skin rash and outcome in non-small-cell lung cancer patients treated with anti-EGFR tyrosine kinase inhibitors: A literature-based meta-analysis of 24 trialsLung Cancer78815201210.1016/j.lungcan.2012.06.00922795701 WoHHeJZhaoYYuHChenFYiHThe efficacy and toxicity of gefitinib in treating non-small cell lung cancer: A meta-analysis of 19 randomized clinical trialsJ Cancer914551465201810.7150/jca.2335629721056 WangXXuYTangWLiuLEfficacy and safety of radiotherapy plus EGFR-TKIs in NSCLC patients with brain metastases: A meta-analysis of published dataTransl Oncol1111191127201810.1016/j.tranon.2018.07.00330032006 MoherDLiberatiATetzlaffJAltmanDGPreferred reporting items for systematic reviews and meta-analyses: The PRISMA statementBMJ339b2535200910.1136/bmj.b253519622551 Ottawa Hospital Research InstituteNOS ManualAvailable from:http://www.ohri.ca/programs/clinical_epidemiology/nos_manual.pdfMarch82021 JadadARMooreRACarrollDJenkinsonCReynoldsDJGavaghanDJMcQuayHJAssessing the quality of reports of randomized clinical trials: Is blinding necessary?Control Clin Trials17112199610.1016/0197-2456(95)00134-48721797 HigginsJPTThompsonSGDeeksJJAltmanDGMeasuring inconsistency in meta-analysesBMJ327557560200310.1136/bmj.327.7414.55712958120 DerSimonianRLairdNMeta-analysis in clinical trials revisitedContemp Clin Trials45139145201510.1016/j.cct.2015.09.00226343745 GalbraithRFA note on graphical presentation of estimated odds ratios from several clinical trialsStat Med7889894198810.1002/sim.47800708073413368 BeggCBMazumdarMOperating characteristics of a rank correlation test for publication biasBiometrics5010881101199410.2307/25334467786990 EggerMSmithGDSchneiderMMinderCBias in meta-analysis detected by a simple, graphical testBMJ315629634199710.1136/bmj.315.7109.6299310563 National Cancer InstitueCommon Terminology Criteria for Adverse Events (CTCAE) Version 4Available athttps://stacks.stanford.edu/file/druid:nw036fx4646/CTCAE_4.03_2010-06-14_QuickReference_5×7.pdfFebruary122021 RuanYJiangJGuoLLiYHuangHShenLLuanMLiMDuHMaCGenetic association of curative and adverse reactions to tyrosine kinase inhibitors in chinese advanced non-small cell lung cancer patientsSci Rep623368201610.1038/srep2336826988277 NieQYangXNAnSJZhangXCYangJJZhongWJLiaoRQChenZHSuJXieZWuYLCYP1A1*2A polymorphism as a predictor of clinical outcome in advanced lung cancer patients treated with EGFR-TKI and its combined effects with EGFR intron 1 (CA)n polymorphismEur J Cancer4719621970201110.1016/j.ejca.2011.04.01821616658 TiseoMRossiGCapellettiMSartoriGSpiritelliEMarchioniABozzettiCDe PalmaGLagrastaCCampaniniNPredictors of gefitinib outcomes in advanced non-small cell lung cancer (NSCLC): Study of a comprehensive panel of molecular markersLung Cancer67355360201010.1016/j.lungcan.2009.04.02119473722 ChilingirovaNHammoudehZBalabanskiLIvanovSVazharovaRNikolovaDKurtevaGTonchevaDChilingirovPTruSight cancer sequencing panel reveals pharmacogenetic variants associated with sensitivity to chemotherapy in lung cancerMemo-Magazine of European Medical Oncology93038201610.1007/s12254-015-0244-2 CusatisGGregorcVLiJSpreaficoAIngersollRGVerweijJLudoviniVVillaEHidalgoMSparreboomABakerSDPharmacogenetics of ABCG2 and adverse reactions to gefitinibJ Natl Cancer Inst9817391742200610.1093/jnci/djj46917148776 RudinCMLiuWDesaiAKarrisonTJiangXJanischLDasSRamirezJPoonkuzhaliBSchuetzEPharmacogenomic and pharmacokinetic determinants of erlotinib toxicityJ Clin Oncol2611191127200810.1200/JCO.2007.13.112818309947 TiseoMCapellettiMDe PalmaGFranciosiVCavazzoniAMozzoniPAlfieriRRGoldoniMGalettiMBortesiBEpidermal growth factor receptor intron-1 polymorphism predicts gefitinib outcome in advanced non-small cell lung cancerJ Thorac Oncol311041111200810.1097/JTO.0b013e3181861d6718827605 CappuzzoFLigorioCJännePAToschiLRossiETrisoliniRPaioliDHolmesAJMagriniEFinocchiaroGProspective study of gefitinib in epidermal growth factor receptor fluorescence in situ hybridization-positive/phospho-Akt-positive or never smoker patients with advanced non-small-cell lung cancer: The ONCOBELL trialJ Clin Oncol2522482255200710.1200/JCO.2006.09.430017538169 O'ByrneKJBondarenkoIBarriosCEschbachCMartensUKortsikYHCelikIStrohCPirkerRMolecular and clinical predictors of outcome for cetuximab in non-small cell lung cancer (NSCLC): Data from the FLEX studyJ Clin Oncol2780078007200910.1200/jco.2009.27.15_suppl.8007 SuXLacoutureMEJiaYWuSRisk of high-grade skin rash in cancer patients treated with cetuximab-an antibody against epidermal growth factor receptor: Systemic review and meta-analysisOncology77124133200910.1159/00022975219622903 DubeySStephensonPLevyDEMillerJAKellerSMSchillerJHJohnsonDHKolesarJMEastern Cooperative Oncology GroupEGFR dinucleotide repeat polymorphism as a prognostic indicator in non-small cell lung cancerJ Thorac Oncol1406412200610.1097/01243894-200606000-0000517409891 GebhardtFBürgerHBrandtBModulation of EGFR gene transcription by secondary structures, a polymorphic repetitive sequence and mutations-a link between genetics and epigeneticsHistol Histopathol15929936200010963135 HanSWJeonYKLeeKHKeamBHwangPGOhDYLeeSHKimDWImSAChungDHIntron 1 CA dinucleotide repeat polymorphism and mutations of epidermal growth factor receptor and gefitinib responsiveness in non-small-cell lung cancerPharmacogenet Genomics17313319200710.1097/FPC.0b013e328011abc017429313 IchiharaSToyookaSFujiwaraYHottaKShigematsuHTokumoMSohJAsanoHIchimuraKAoeKThe impact of epidermal growth factor receptor gene status on gefitinib-treated Japanese patients with non-small-cell lung cancerInt J Cancer12012391247200710.1002/ijc.2251317192902 DienstmannRBrañaIRodonJTaberneroJToxicity as a biomarker of efficacy of molecular targeted therapies: Focus on EGFR and VEGF inhibiting anticancer drugsOncologist1617291740201110.1634/theoncologist.2011-016322135123 WackerBNagraniTWeinbergJWittKClarkGCagnoniPJCorrelation between development of rash and efficacy in patients treated with the epidermal growth factor receptor tyrosine kinase inhibitor erlotinib in two large phase III studiesClin Cancer Res1339133921200710.1158/1078-0432.CCR-06-261017606725 PereaSOppenheimerDAmadorMCusatiGBakerSTakimotoCMaitraAIocobuzio-DonahueCHidalgoMGenotypic bases of EGFR inhibitors pharmacological actionsJ Clin Oncol2230053005200410.1200/jco.2004.22.90140.3005 LiuWInnocentiFWuMHDesaiAADolanEMCookEHJrRatainMJA functional common polymorphism in a Sp1 recognition site of the epidermal growth factor receptor gene promoterCancer Res654653200510.1158/0008-5472.46.65.115665278 ZhangHXTangYWangLWeiSXLiuQXLiFYuanXLEGFR-216 G/T polymorphism as a predictor of clinical outcomes in advanced non-small cell lung cancer patients treated with EGFR-TKIs: A meta-analysisInt J Clin Exper Med910273102802016 BiaoxueRHuaLWenlongGShuanyingYEfficacy and safety of icotinib in treating non-small cell lung cancer: A systematic evaluation and meta-analysis based on 15 studiesOncotarget78690286913201610.18632/oncotarget.1350927893423 YiLFanJQianRLuoPEfficacy and safety of osimertinib in treating EGFR-mutated advanced NSCLC: A meta-analysisInt J Cancer145284294201910.1002/ijc.3209730613959

Flowchart of the study selection process. WOS, Web of Science; PRISMA, Preferred reporting items for systematic reviews and meta-analyses.

Forest plot of pooled odds ratio and 95% CI of three epidermal growth factor receptor single nucleotide polymorphisms relative to tyrosine kinase inhibitor-caused skin toxicity. *rs11568315 (CA repeat), **rs712829 (−216G>T), ***rs712830 (−191 C/A). Weights are from the random effects model. CI, confidence interval; DL, Der Simonian and Laird method.

Forest plot of subgroup analysis of three EGFR SNPs relative to tyrosine kinase inhibitor-caused skin toxicity. *rs11568315 (CA repeat), **rs712829 (−216G>T), ***rs712830 (−191 C/A). Weights and subgroup heterogeneity test are from the random effects model. CI, confidence interval; DL, Der Simonian and Laird method; EGFR, epidermal growth factor receptor; SNP, single nucleotide polymorphism.

Funnel plots of potential bias of skin toxicity relative to EGFR single nucleotide polymorphisms and sensitivity analysis. (A) Individual studies were omitted but did not significantly modify the overall effect. (B) Funnel plot appeared to be roughly symmetrical. (C) Begg's funnel plot suggested no publication bias. (D) Egger's test (P=0.545) indicated that publication biases did not exist. *rs11568315 (CA repeat), **rs712829 (−216G>T), ***rs712830 (−191 C/A). CI, confidence interval; logor, log of odds ratio; theta, the effect estimate; s.e., standard error; SND, standard normal deviate.

Characteristics of the included studies.

First author, year Country Study type Study period Patients, n Median age, years (range) Males, % Ethnicity, % Smokers, % Clinical stage (%)a Histology (%) Median follow-up, months TKI treatment (dose) Additional therapy Toxicity assessment Adverse effect Overall quality score (Refs.)
Huang et al, 2009 Taiwan Cohort May 2005-September 1, 2006 52 66 (39–86) 33 NR NR IIIB (14), IV (87) Adenoc-arcinoma (96), Other (4) 12 Gefitinib (250 mg/day) NR NCI CTCA E v3.0c Skin NOSh: 7 (23)
Giovannetti et al, 2010 Italy Cohort NR 96 64 (NR) 57 NR 69 IIIB (9), V (91) Adenoc-arcinoma (56), Other (44) NR Gefitinib (250 mg/day) 85% of patients were previously treated with chemot-herapyb NCI CTC manual versiond Skin, diarrhea NOS: 7 (26)
Liu et al, 2012 Canada RCT Phase III NR 242 60 (NR) 64 East Asian (6), Other (94) 76 IIIB (NR), IV (NR) Adenoc-arcinoma (54), Other (46) NR Erlotinib (NR) 93% patients were previously treated with chemot-herapy NCI v2.0.e Skin Jadadi: 4 (24)
Parmar et al, 2013 Germany Cohort September 2008-November 2010 109 68 (43–86) 67 NR 12; 49.5 former IIIA (2); IIIB (11), IV (83), Unknown (5) NR NR Gefitinib (250 mg/day), erlotinib (150 mg/day), cetuximab (250 mg/m2, with weekly or biweekly or 31% patients were previously treated with chemot-herapy panitumumab (6 mg/kg) biweekly NCI CTC v3.0f Skin NOS: 6 (25)
Kim et al, 2017 Canada, Italy, South Korea and Brazil RCT Phase III NR 760 62 (56–66) 66 Caucasian (96), East Asian (3), Other (1) 26; 53 former IIIB (11), IV (89) Adenoc-arcinoma (56), Other (44) NR Erlotinib (NR) 50% of patients were treated with erlotinib, followed by cisplatin and gemcita bine at progression; 50% were treated with cisplatin and gemcita bine, followed by erlotinib at progression NR Skin, diarrhea Jadad: 3 (27)
Ma et al, 2017 China Cohort 2011-2014 59 56 (31–77) 49 NR 61 IIIB or IV (64) Adenoc-arcinoma (90), Other (10) NR Gefitinib (250 mg/day) NR CTCA E v4.0g Skin, diarrhea, hepat otoxicity NOS: 7 (28)

Data is provided to 2 significant places.

(23–28),

additional 127 chemotherapy-treated/gefitinib-non-treated patients with NSCLC were used as a comparison,

(23),

(26),

(24),

(25),

(28),

(33),

(34). CTC, Common Toxicity Criteria; CTCAE, Common Terminology Criteria for Adverse Events; NCI, National Cancer Institute; NOS, Newcastle-Ottawa Quality Assessment Scale; NR, not reported; RCT, randomized clinical trial; TKI, tyrosine kinase inhibitor.

EGFR genotypes and adverse effects of tyrosine kinase inhibitor therapy.

A, Patients with skin toxicity (n=880, 78.36%)
dbSNP-ID Variant type, location, and/or consequence First author, year Genotyping platform used Genotype Total number of patients, n Patients with skin toxicity, n ORa 95% CIa za P-valuea (Refs.)
rs11568315 Intron variant, Huang et al, 2009 PCR and SS 7 5 Ref. 2.13 0.032 (23)
g.55020560_5502 direct SL 26 8 6.87 1.1734- 70 6
0561AC[n] sequencing LL 19 4 5 40.2797
Parmar et al, 2013 Real-Time SS 22 19 Ref. 1.23 0.217 (25)
PCR and SL 58 41 2.27 0.6156- 30 7
16-capillary electrophoresis LL 29 23 6 8.4150
or KASPar
Giovannetti et al, 2010 TaqMan SS 27 18 Ref. 1.71 0.087 (26)
PCR SL-LL 60 28 2.28 0.8863- 00 2
57 5.8947
Ma et al, 2017 Sequenom SS 6 6 NA NA NA NA (28)
Massarray SL 20 14
system LL 16 10
Kim et al, 2017 Sanger SS 38 28 Ref. 0.15 0.874 (27)
sequencing and SL-LL 96 72 0.93 0.3961- 80 7
Taqman PCR 2.1994
Liu et al, 2012 PCR and direct sequencing LL-S- NR NR 0.60 0.2–1.9 NA NA (24)
rs712829 5′ UTR variant, Huang et al, 2009 PCR and direct GG 45 14 Ref. 0.61 0.540 (23)
g.5031G>T, - sequencing GT 5 1 0.60 0.1186- 20 6
216G>T TT 2 2 22 3.0567
Parmar et al, 2013 Real-Time GG 49 33 Ref. 1.92 0.054 (25)
PCR and GT 48 39 0.41 0.1670- 00 9
16-capillary electrophoresis TT 12 11 25 1.0188
or KASPar
Giovannetti et al, 2010 TaqMan PCR GG 30 19 Ref. 1.41 0.158 (26)
GT-TT 57 27 1.91 0.7752- 00 7
92 4.7513
Kim et al, 2017 Sanger sequencing and GG 40 32 Ref. 0.91 0.358 (27)
Taqman PCR GT-TT 83 60 1.53 0.62–3.82 90 2
Liu et al, 2012 PCR and direct sequencing GT-GG NR NR 2.00 0.5–8.3 NA NA (24)
rs712830 5′ UTR variant, Parmar et al, 2013 Real-Time CC 79 59 Ref. 0.58 0.561 (25)
g.5056A>C, - PCR and CA 30 24 0.73 0.2637- 00 7
191C>A 16-capillary electrophoresis or KASPar AA 0 0 75 2.0624
Giovannetti et al, 2010 TaqMan CC 72 36 Ref. 1.16 0.245 (26)
PCR CA-AA 15 10 0.5 0.1554- 20 1
1.6089
Kim et al, 2017 Sanger sequencing and CC 100 72 Ref. 1.44 0.147 (27)
Taqman PCR CA- AA 23 20 0.38 0.1062- 80 7
57 1.4007
Liu et al, 2012 PCR and direct sequencing CA-CC NR NR 1.00 0.1–6.8 NA NA (24)
rs2227983b Missense variant, Parmar et al, 2013 Real-Time PCR and GG 57 49 Ref. 2.45 0.014 (25)
(1562G>A, 16-capillary electrophoresis GA 47 29 3.24 1.2657- 10 2
R497K) or KASPar AA 5 5 26 8.3073
Giovannetti et al, 2010 TaqMan GG-GA 75 38 0.58 0.1585- 0.79 0.425 (26)
PCR 69 2.1734 80 0
AA 11 7 Ref.
Huang et al, PCR and GG 12 4 Ref. 0.16 0.867 (23)
2009 direct GA 28 8 1.12 0.2832- 70 1
sequencing AA 12 5 5 4.4695
B, Patients with diarrhea (n=233, 20.75%)
Patients
Variant type, Total with
location, and/or number of diarrhea
dbSNP-ID consequence First author, year Genotyping platform used Genotype patients, n toxicity, n ORa 95% CIa za P-valuea (Refs.)
rs11568315 Intron variant, Giovannetti et al, 2010 TaqMan assay SS 26 13 Ref. 0.76 0.446 (26)
g.55020560_550 SS-LL 49 20 1.45 0.5569- 10 6
20561AC[n] 3.7751
Ma et al, 2017 Sequenom SS 6 2 Ref. 0.50 0.612 (28)
Massarray SL 20 9 0.62 0.1012- 60 8
system LL 16 7 5 3.8585
Kim et al, 2017 Sanger sequencing and SS 38 15 Ref. 0.12 0.902 (27)
Taqman PCR SL-LL 96 39 0.95 0.4424- 20 5
2.0535
rs712829 5′ UTR variant, Kim et al, 2017 Sanger sequencing and GG 40 18 Ref. 0.94 0.346 (27)
g.5031G>T, - Taqman PCR GT-TT 83 30 1.44 0.6711- 10 6
216G>T 55 3.1131
rs712830 5′ UTR variant, Kim et al, 2017 Sanger sequencing and CC 100 37 Ref. 0.95 0.339 (27)
g.5056A>C, - Taqman PCR CA-AA 23 11 0.64 0.2570- 50 4
191C>A 1.597
rs2227983b Missense variant, Giovannetti et al, 2010 TaqMan PCR GG-GA 74 26 2.76 0.7163- 1.47 0.139 (26)
(1562G>A, 92 10.7057 60 8
R497K) AA 10 6 Ref.
C, Patients with hepatotoxicity (n=10, 0.89%)
Patients
Variant type, Total with
location, and/or Genotyping number of hepatoto
dbSNP-ID consequence Author, year platform used Genotype patients, n xicity, n ORa 95% CIa za P-valuea (Refs.)
rs1156 Intron variant, Ma et al, 2017 Sequenom Massarray SS 6 1 Ref. 0.40 0.682 (28)
8315 g.55020560_550 system SL 21 6 0.62 0.0640- 90 7
20561AC[n] LL 16 3 22 6.0507

Pre-calculated values according to available data for EGFR SNP genotypes,

rs11543848 was merged with rs2227983. CI, confidence interval; dbSNP-ID, single nucleotide polymorphism database identifier); rs11568315 SS, rs712829 GG, rs712830 CC, rs2227983 GG: wild-type homozygotes; rs11568315 SL, rs712829 GT, rs712830 CA, rs2227983 GA: heterozygotes; rs11568315 LL, rs712829 TT, rs712830 AA, rs2227983 AA: mutant homozygotes; NA, not applicable; Ref., reference value; NR, not reported; OR, odds ratio; EGFR, epidermal growth factor receptor; UTR, untranslated region.