Predictive value of EGFR amplification and EGFRvIII mutation in EGFR-targeted therapy for recurrent glioblastoma: a systematic review Authors: Dylan Evans, Joanne Voisey, Jennifer H Gunter, Fiona Rae Published: 2026 ABSTRACT AbstractIntroductionGlioblastoma (GBM) is the most aggressive and biologically heterogeneous tumor of the central nervous system, associated with dismal prognosis and frequent recurrence. Amplification of the epidermal growth factor receptor (EGFR) and EGFRvIII mutation are common alterations, yet their prognostic significance remains unclear. This systematic review evaluated whether clinical evidence supports a predictive association between EGFR amplification or EGFRvIII mutation and response to EGFR-targeted therapy in adults with recurrent GBM.MethodsPubMed, Embase and the Cochrane Library were searched (2010–2025) for studies of adults with recurrent GBM treated with EGFR-targeted therapies, with outcomes stratified by EGFR amplification and/or EGFRvIII. Screening, data extraction and risk-of-bias assessment were performed independently.ResultsTwelve studies (565 patients) met inclusion criteria. EGFR amplification was assessed in 11 studies and EGFRvIII in six using FISH, qPCR, RT-PCR, or NGS. Median overall survival ranged from 5.7 to 10.3 months and progression-free survival from 1.7 to 6.0 months, with no consistent survival benefit observed.ConclusionEGFR amplification and EGFRvIII mutation have not demonstrated reliable predictive value for EGFR-targeted therapy in recurrent GBM. The available evidence is largely derived from heterogeneous, single-arm studies, limiting robust assessment of biomarker specific treatment response and therefore these alterations are not recommended to be used to guide off-trial treatment decisions in recurrent GBM.Protocol Registration: www.crd.york.ac.uk/prospero identifier is CRD420251071466.Keywords: Recurrent glioblastoma, EGFR-targeted therapy, EGFR amplification, overall survival, progression-free survival Introduction Glioblastoma (GBM) is the most common and aggressive primary malignant brain tumor in adults, accounting for over half of central nervous system-related tumor deaths. The 2021 World Health Organization (WHO) Classification of the Central Nervous System defines GBM as “a diffuse, astrocytic glioma that is IDH-wildtype and H3-wildtype and has one or more of the following histological or genetic features: microvascular proliferation, necrosis, telomerase reverse transcriptase (TERT) promoter mutation, epidermal growth factor receptor (EGFR) gene amplification, +7/−10 chromosome copy-number changes” [1]. Unfortunately, the median overall survival is 12–15 months, with only 10% of patients achieving long-term survival greater than 2 years, despite aggressive treatment regimens including surgical resection, radiotherapy and chemotherapy [2,3]. GBM recurrence is virtually inevitable, occurring rapidly after initial therapy, with approximately 75–90% of patients ultimately experiencing disease relapses [4,5]. Recurrent glioblastoma (rGBM) is associated with an unfortunate prognosis, with median survival after progression ranging from 6 to 9 months and no established standard of care to guide second-line treatment decisions [4,6]. Given these challenges, there has been growing interest in developing molecularly targeted therapies and identifying predictive biomarkers such as EGFR amplification and EGFRvIII, which have been implicated in tumor progression and resistance to conventional therapies [7,8]. EGFR is a transmembrane receptor tyrosine kinase that plays a critical role in the pathogenesis of GBM by activating multiple downstream signaling pathways [8]. EGFR signaling includes activation of the PI3K/AKT, RAS/MAPK and STAT3 pathways, which regulate cell proliferation, survival, and progression [9]. EGFR amplification is defined as an increase in the number of copies of the EGFR gene and is one of the most frequent genetic alterations in GBM, occurring in approximately 35–45% of cases. It is often accompanied by overexpression of the receptor protein, which has been reported in up to 60% of glioblastomas [8,9]. However, definitions of amplification have varied widely across studies, with thresholds only more recently beginning to standardize, such as requiring amplification in over 50% of tumor cell nuclei by FISH or more than 5 gene copies per cell by NGS [10]. Another common feature of EGFR dysregulation is the occurrence of constitutively active variants, most notably EGFRvIII, which results from an in-frame deletion of exons 2 through 7. This deletion removes amino acids 6 to 273, including much of the ligand-binding extracellular domain and produces a receptor with weaker kinase activity compared to ligand-activated wild-type [11]. EGFRvIII has been identified in approximately one-third of EGFR-amplified GBMs, where it almost always occurs as a co-amplification event with wild-type EGFR, suggesting that the presence of both forms may confer a growth advantage [12]. Other molecular markers, most notably O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation have well established prognostic significance, including after progression [13,14]. Hypermethylation of the MGMT promoter reduces DNA repair capacity, increasing tumor sensitivity to alkylating agents such as temozolomide (TMZ). Given its prognostic value, MGMT status is frequently used to stratify clinical trial cohorts and can influence treatment selection, thus making it an important confounder when interpreting the predictive value of EGFR alterations [15]. Consequently, these alterations are associated with aggressive tumor behavior and therapeutic resistance. Despite their prevalence, EGFR amplification and EGFRvIII have not been incorporated into any clinical guidelines, and their use has remained confined to research and trial settings [16]. In part, this lack of standardization and clinical adoption helps explain the inconsistent outcomes reported with EGFR-targeted therapies, highlighting the need to better define their predictive value [8]. Multiple therapeutic strategies have been evaluated to target EGFR alterations in GBM, including small-molecule tyrosine kinase inhibitors (TKIs), monoclonal antibodies, vaccines and antibody drug conjugates (ADCs) [17]. Early phase and preclinical studies enforce a strong rationale for these treatment approaches, given the high frequency of EGFR amplification and the oncogenic potential of EGFRvIII [17,18]. However, clinical trials have consistently demonstrated limited and inconsistent efficacy. First-generation TKIs such as gefitinib and erlotinib reversibly inhibit the tyrosine kinase domain of EGFR but have demonstrated limited clinical benefit, with most trials reporting no significant improvement in survival outcomes [19]. Second-generation TKI’s including afatinib and dacomitinib were designed to block multiple ErbB family receptors but failed to achieve a notable response in survival outcomes, though EGFRvIII positive patients experienced a greater response to the treatments with improved survival outcomes compared to EGFRvIII negative patients [19]. Third and fourth-generation TKI’s such as osimertinib, CM93, and ERAS-801 have shown potential in preclinical models but clinical data, particularly in the recurrent GBM setting remain limited [18]. Moreover, monoclonal antibodies and ADCs like cetuximab and depatuxizumab mafodotin (depatux-m or ABT-414) have not demonstrated consistent efficacy in clinical studies, with most monotherapy trials showing no significant improvements in survival outcomes compared to standard therapy [19]. However, there is promising evidence that combination therapy of depatux-m with TMZ does noticeably improve survival outcomes over TMZ alone [20]. Mechanisms, including intra-tumoral heterogeneity, variable EGFRvIII mutation, and limited drug delivery across the blood-brain barrier (BBB), have contributed to the inconsistent treatment outcomes [19]. While these observations depict the considerable effort that has gone into developing EGFR-targeted therapies, most reviews and clinical trials have focused on newly diagnosed GBM or have combined populations of primary or recurrent disease, thus making it difficult to determine the predictive value of EGFR amplification and EGFRvIII mutation in rGBM. Consequently, the lack of focused analysis underscores the need for a systematic review of the current evidence to determine whether EGFR alterations can reliably guide treatment decisions in rGBM. This systematic review aims to evaluate whether the available clinical evidence supports a predictive association between the presence of EGFR amplification, EGFRvIII mutation, or their co-occurrence and objective response rate (ORR), progression-free survival (PFS), and overall survival (OS) in adult patients with rGBM treated with EGFR-targeted therapies. Accordingly, the analysis is restricted to studies reporting outcomes stratified by EGFR amplification and/or EGFRvIII mutation. By analyzing biomarker-stratified outcomes across studies, this review seeks to clarify the relevance of EGFR alterations for guiding treatment decisions exclusively in the rGBM setting rather than from mixed or newly diagnosed populations. Synthesis across therapeutic classes and trial designs reveals consistent evidentiary limitations, particularly the absence of appropriate comparator arms and substantial heterogeneity in biomarker assessment, which limits the ability to establish predictive utility. Ultimately, this analysis aims to delineate the current limits of the available evidence for EGFR-guided therapy in rGBM and highlights key considerations for the design of future biomarker-driven clinical trials. Methods Search strategy A comprehensive and systematic search of the literature was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines of 2020. Electronic databases, including PubMed, Embase, and the Cochrane Library, were leveraged to retrieve relevant studies published between January 2010 and 31 May 2025. To enhance the precision and coverage of the search, vocabulary terms like Medical Subject Headings (MeSH) in PubMed and Cochrane, and Emtree in Embase were combined with free-text keywords. The search terms were designed to capture studies related to recurrent glioblastoma, EGFR amplification and EGFRvIII mutation, and EGFR-targeted therapies like tyrosine kinase inhibitors, monoclonal antibodies, antibody-drug conjugates, and vaccines. Specifically, the search strategies integrated disease specific terms (e.g., “glioblastoma,” “recurrent,” “progressive”) along with biomarker and intervention terms (“EGFR,” “EGFRvIII,” “erlotinib,” “gefitinib,” “cetuximab,” “ABT-414,” among others) using Boolean operators “AND” and “OR” to refine the scope of the search. Additionally, filters were applied to restrict results to studies involving adult human participants, published in English, and designed as interventional or observational clinical studies. Full search strategies, keywords, indexing terms, and applied filters for each database were provided in the Supplementary Material (Table S1). The final database searches were rechecked on 7 June 2025 to confirm no additional studies had been missed. The protocol for this systematic review was registered with PROSPERO (CRD420251071466) on 27 August 2025 and is publicly available at https://www.crd.york.ac.uk/PROSPERO/view/CRD420251071466. Eligibility criteria Inclusion Criteria Studies including adult patients (>18 years) with histologically confirmed recurrent glioblastoma, classified as IDH-wildtype where reported. Studies including adult patients (>18 years) with histologically confirmed recurrent glioblastoma, classified as IDH-wildtype where reported. Studies reporting the assessment of EGFR amplification and/or EGFRvIII expression using validated molecular techniques. Studies reporting the assessment of EGFR amplification and/or EGFRvIII expression using validated molecular techniques. Studies evaluating EGFR targeted therapies (including combination treatments). Studies evaluating EGFR targeted therapies (including combination treatments). Peer-reviewed original research articles, including phase I-III clinical trials, prospective or retrospective cohort studies, or observational studies reporting outcomes stratified by biomarker status. Peer-reviewed original research articles, including phase I-III clinical trials, prospective or retrospective cohort studies, or observational studies reporting outcomes stratified by biomarker status. Studies reporting at least one outcome measure, such as objective response rate, progression-free survival, overall survival, time to progression, or time to treatment failure. Studies reporting at least one outcome measure, such as objective response rate, progression-free survival, overall survival, time to progression, or time to treatment failure. Full text available in English. Full text available in English. Exclusion Criteria Studies exclusively investigating newly diagnosed glioblastoma without separate analysis of recurrent glioblastoma. Studies exclusively investigating newly diagnosed glioblastoma without separate analysis of recurrent glioblastoma. Studies including pediatric patients (<18 years) or non-glioblastoma populations. Studies including pediatric patients (<18 years) or non-glioblastoma populations. Studies limited to IDH-mutant glioblastoma. Studies limited to IDH-mutant glioblastoma. Non-EGFR targeted treatments (e.g., chemotherapy or radiotherapy alone). Non-EGFR targeted treatments (e.g., chemotherapy or radiotherapy alone). Case reports or case series with fewer than 10 patients. Case reports or case series with fewer than 10 patients. Preclinical, animal, or in vitro studies. Preclinical, animal, or in vitro studies. Non- original research articles. Non- original research articles. Articles not published in English or which the full text is not available. Articles not published in English or which the full text is not available. Study selection All search results were imported into Covidence systematic review software for screening and selection of studies [21]. The titles and abstracts were independently screened by two reviewers (DE, FR) to identify any eligible studies. Full-text articles were then retrieved and evaluated by both reviewers using the predefined inclusion and exclusion criteria to confirm eligibility. Any discrepancies in the determined eligibility of a study were resolved through a discussion involving an independent third reviewer (JV). One full-text article was obtained via interlibrary loan from another institution to appropriately examine eligibility. No automation tools, external machine classifiers, or crowdsourcing methods were used in the screening process. The full study selection process, which includes the number of articles screened, excluded, and included, is documented in the PRISMA flow diagram (Figure 1). PRISMA 2020 flowchart of screening and selection process. Data extraction Data were extracted by one primary reviewer (DE) for all included studies, with a second reviewer independently verifying each extraction. The second reviewer role was divided among three team members (JV, JG, FR), with each reviewer assigned a specific group of studies to verify independently. A data extraction template was designed and tailored via Covidence to consistently record relevant data points, including EGFR biomarker status (including detection method and tumor tissue source where reported), details of EGFR-targeted interventions, and clinical outcomes like ORR, PFS, and OS. Any discrepancies that arose were resolved through discussion or the adjudication of a third reviewer if necessary. When multiple publications of the same study were identified, decision rules were applied to include the most up-to-date article. Risk of bias assessment The risk of bias was evaluated independently for each included article by two reviewers using the Cochrane Risk of Bias 2 (RoB2) tool for randomized controlled trials and The Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) for non-randomized studies [22,23]. The RoB2 tool assessed bias across domains including the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and the selection of reported result, with each domain rated as “low risk,” “some concerns,” and “high risk,” and an overall judgment based on the highest level or elevated when lesser domains were judged to have a compounding effect. On the other hand, ROBINS-I domains included risk of bias of confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of reported results. Each domain was assessed a risk rating of “low,” “moderate,” “serious,” or “critical risk of bias,” with the overall risk of bias judgment derived according to the ROBINS-I guidelines. Any disagreements were resolved by discussion, or when needed, a third reviewer was used to resolve the dispute. The results were presented using a traffic light plot created with the “robvis” package in R software (Supplementary File Figure S1 & S2) [24]. Ultimately, a formal GRADE assessment of overall certainty was not performed due to the small number of eligible studies and substantial heterogeneity in study design, biomarker assessment methods, and outcome reporting, all of which limited the feasibility of applying the GRADE framework through a standardized approach. Statistical analysis A structured narrative synthesis was conducted to summarize and compare the findings across all included studies. For the purposes of analysis, EGFR targeted therapies were defined as interventions whose primary mechanisms of action involved direct EGFR inhibition or EGFR directed binding, as described in the original publications. The extracted data were grouped by intervention class (i.e., TKIs or ADCs) and stratified by biomarker expression (EGFR amplification, EGFRvIII mutation, or co-occurrence). However quantitative meta-analysis was not feasible due to substantial heterogeneity in study design, biomarker assessment methods, and endpoint definitions across trials, which precluded valid statistical pooling of results. Descriptive statistics reported in the original publications, including medians, ranges, percentages of events, and confidence intervals where available, were extracted and presented to outline the variability in clinical outcomes across studies and subgroups. Summary tables and figures were constructed to display study characteristics, intervention details, and reported outcomes in a clear and consistent format. Results Search results A comprehensive database search yielded 238 records published between January 2010 and May 2025. The importing of records into Covidence resulted in the removal of 67 duplicates, thereby leaving 171 unique studies proceeding to the title and abstract screening, of which 94 records were excluded for not meeting the eligibility criteria. The remaining 77 articles then underwent full-text evaluation, resulting in the exclusion of 65 studies due to ineligible patient populations, non-EGFR, targeted interventions, abstract-only publications, absence of relevant outcome data, unsuitable study design, duplicate publication, or focus on EGFR expression rather than amplification. Ultimately, 12 studies met the inclusion criteria and were included in the systematic review; see Figure 1 for a breakdown of the study selection procedure. Study characteristics A total of 565 patients treated with EGFR targeted therapy are included across 12 studies published between 2010 and 2025. The majority were prospective interventional trials evaluating EGFR targeted treatments as monotherapy or in combination with TMZ or bevacizumab, including three phase I dose escalation and expansion cohorts [25–27], randomized and non-randomized phase II trials [25,28–32], and multi-arm or multi-cohort studies. Two additional studies were prospective cohort studies, one employing a precision medicine approach [33] and one retrospective cohort analysis [34]. These studies were conducted across eight countries, including the United States, Canada, Colombia, Spain, Italy, South Korea, and Australia, all of which enrolled adults with histologically confirmed rGBM. All studies examined EGFR amplification as the primary molecular alteration of interest. Three studies [25,29,31] incorporated co-occurring EGFRvIII mutation as a primary eligibility or stratification factor to evaluate its potential impact on treatment outcomes, while three others [27,30,32] assessed EGFRvIII status in exploratory analyses. The remaining studies primarily focused on EGFR amplification without stratifying outcomes by EGFRvIII mutation status. The median age across studies ranged from 52 to 62 years, with an approximate combined median age of 57 years. Among the 10 studies reporting sex distribution, a total of 243 male and 149 female patients were enrolled, corresponding to a female-male ratio of approximately 1.6:1. The number of participants per experimental arm ranged from 8 to 88, with treatments typically administered as second-line therapy following progression after surgical resection and standard chemotherapy. EGFR targeted interventions included TKIs (erlotinib, afatinib, dacomitinib, osimertinib), ADCs (depatuxizumab mafodotin), monoclonal antibodies (GC1118), and CAR T cell therapies, delivered either as monotherapy or in combination with TMZ or bevacizumab. Further study characteristics such as number of prior therapy lines, prior anti-VEGF exposure, MGMT/IDH status, and EGFR detection methods for each study are summarized in Table S2 and S3 in the Supplementary File, where available. Quality assessment Risk of bias was evaluated using the RoB2 tool for randomized controlled trials and the ROBINS-I for non-randomized studies. Of the two randomized trials, one was judged to have an overall low risk of bias, and the other had concerns due to inconsistent reporting of randomization and allocation concealment, though both demonstrated low risk for deviations from intended interventions, missing data, and outcome measurement. The remaining non-randomized studies were generally rated to have a moderate to serious risk of bias, primarily arising from confounding, participant selection, and incomplete follow-up. Measurement and reporting domains were typically low risk, indicating adequate methodological protocols once treatment commenced. Altogether, these domain level patterns suggest that the low certainty of evidence is largely attributed to design-related limitations and inter-study heterogeneity, which restrict the reliability of the observed treatment effects. Consequently, conclusions regarding the predictive value of EGFR amplification and EGFRvIII mutation should be interpreted cautiously and validated in larger standardized biomarker stratified clinical trials. Treatment regimens Across the included 12 studies, EGFR targeted therapies were administered as either a monotherapy or in combination with TMZ or bevacizumab. Among the investigated therapies, TKIs included erlotinib, administered as a monotherapy [28] or combined with bevacizumab [33]; afatinib, administered alone or with TMZ [25]; dacomitinib, which was administered as a monotherapy in patients who were either anti-VEGF naïve or had previously received bevacizumab [29,31]; and osimertinib combined with bevacizumab [34]. Erlotinib was dosed daily at mean amounts ranging from 168.7 mg to 337.4 mg, depending on enzyme-inducing antiepileptic drug use (EIADS). Afatinib was administered at 40 mg daily, with combination regimens incorporating TMZ at 75 mg/m2 on days 1–21 of 28-day cycles. Additionally, dacomitinib was administered at 45 mg daily in continuous treatment cycles, with median treatment durations of approximately 2 months. Osimertinib was given at 80 mg/day in combination with bevacizumab, 15 mg/kg intravenously every 3 weeks. The ADC, depatux-m was studied as a monotherapy and in combination with TMZ across several phase I and II trials [26,27,30,35]. Dosing ranged from 1.0 mg/kg to 1.5 mg/kg intravenously every 2 weeks, with some studies including dose escalation cohorts and extended treatment durations exceeding 9 months. The monoclonal antibody GC1118, targeting EGFR, was administered weekly as a monotherapy for up to six cycles [32]. Separately, one study evaluated an autologous bivalent CAR T cell therapy targeting EGFR and IL13Rα2, delivered as a single intracebroventricular infusion following disease recurrence [36]. Treatments were mostly administered as a second line or later therapy, typically following standard chemoradiotherapy. Reported grade 3 and above adverse events (AEs) included toxicities, diarrhea, fatigue, hematologic abnormalities, and ocular events such as keratitis and corneal epitheliopathy. One treatment-related death due to a pulmonary embolism occurred in the erlotinib monotherapy study [28]. Detailed dosing schedules, administration frequencies, and adverse event profiles are presented in Table S3 (Supplementary File). Survival outcomes Tyrosine kinase inhibitors (TKIs) Across studies evaluating TKIs, median OS ranged from 6.2 to 10.3 months. In the phase II trial of erlotinib monotherapy, median OS was 8.3 months (95% CI, 4.1–10.7) [28]. In the dacomitinib studies, the median OS varied by cohort, with 10.3 months reported in anti-VEGF naïve patients and 6.2 months in patients who were previously treated with bevacizumab [31]. The combination of osimertinib and bevacizumab in a retrospective cohort resulted in a median OS of 9.0 months (95% CI, 3.9–14.0) [34]. In a precision selected cohort treated with erlotinib plus bevacizumab, median OS was 8.0 months, although follow-up was limited [33]. Median PFS among TKI studies was generally short, with 2.1 months (95% CI, 1.1–3.4) reported with dacomitinib monotherapy [31], and up to 5.1 months (95% CI, 2.8–7.3) with osimertinib plus bevacizumab [34]. In the erlotinib plus bevacizumab cohort, median PFS was 6.0 months. Reported 6-month progression-free survival (PFS6) ranged widely from 0% in the post-bevacizumab dacomitinib cohort [31] to 64% with erlotinib plus bevacizumab [33]. The ORR varied significantly between studies, ranging from 3% to 57%, with the highest observed in the bevacizumab plus erlotinib cohort [33]. Antibody-drug conjugates (ADCs) In studies evaluating the EGFR-targeted ADC depatux-m, monotherapy was associated with a median OS of 9.3 months (95% CI, 6.6–11.7) [26] and 7.9 months (95% CI, 6.1–8.7) [27]. The addition of TMZ to depatux-m showed median OS ranging from 7.4 months (95% CI, 6.5–9.6) to 9.6 months (95% CI, 7.4–11.8) across both trials [30,35]. The median PFS for depatux-m regimens ranged from 1.7 months (95% CI, 1.4–3.3) with monotherapy to 2.7 months (2.0–3.8) when combined with TMZ [26,30]. Monoclonal antibodies The monoclonal anti-EGFR antibody GC1118, administered as a monotherapy was associated with a median OS of 5.7 months [32]. Limited PFS and response data were reported for this agent, reflecting the early-phase evaluation. CAR T cell therapy One study evaluated an intracerebroventricular bivalent CAR T-cell therapy targeting EGFR (epitope 806) and IL-13Rα2 in patients with EGFR-amplified rGBM. In this phase 1 trial, reported efficacy outcomes included a median progression-free survival of 1.9 months (90% CI, 1.1–3.4 months), an ORR of 8%, while median overall survival had not been reached at the time of analysis [36]. Outcomes stratified by biomarker status In studies that reported outcomes by biomarker status, no consistent survival benefit was observed for either EGFR amplification or EGFRvIII mutation. For instance, in the afatinib trial by Reardon et al. median PFS ranged between 0.99–2.73 months in the EGFR amplified cohorts, and 3.35–3.65 months in the EGFRvIII positive cohorts. Furthermore, the results from Sepúlveda-Sánchez et al. revealed that patients with both EGFR amplification and EGFRvIII mutation had a median OS of 6.7 months, compared with 7.8 months in those who only had EGFR amplification, with a similar PFS between the two groups (2.6 and 2.7 months). In the two studies reporting stratified outcomes, survival differences between EGFR amplification and EGFRvIII cohorts were small, variable, and did not demonstrate a consistent benefit for either biomarker. Full survival data and response rates for each study are provided in Table 1. Table 1.Survival outcomes of individual studies using EGFR targeted interventions.AuthorTherapyNBiomarker(s)Median OS (months, 95% CI)Median PFS (months, 95% CI)PFS6 (%) (95% CI)PFS12 (%) (95% CI)ORR (%)NotesYung (2010)[28]Erlotinib23EGFR amplified8.3 (4.1–10.7)NR21.7 (7.9–39.9)NR8.71 CR, 1 PR, no difference by EIAEDReardon (2015) [25]Afatinib (EGFRvIII)8EGFRvIIINR3.35NRNRNR Afatinib + TMZ (EGFRvIII)13EGFRvIIINR3.65NRNRNR Afatinib (EGFR Amp)14EGFR amplifiedNR0.99NRNRNR Afatinib + TMZ (EGFR Amp)9EGFR amplifiedNR2.73NRNRNR Van den Bent (2017) [26]Depatuxizumab Mafodotin (Depatux-M)34EGFR amplified9.3 (6.6–11.7)1.7 (1.4–3.3)28.8 (18.5–39.9)NR6.81 CR, 3 PR29EGFR amplified + EGFRvIIINR1.6 (1.4–3.3)17.2(6.3–32.7)NRNR Sepúlveda-Sánchez (2017) [29]Dacomitinib (EGFR Amp)30EGFR amplified7.8 (5.6–10.1)2.7 (2.3–3.2)13.3 (0–26.5)NR6.61 CR, 1 PRDacomitinib (EGFR Amp + EGFRvIII)19EGFR amplified + EGFRvIII6.7 (4.3–9.1)2.6 (1.8–3.4)5.9 (0–18)NR5.31 PR; cohort closed after futility assessmentVan den Bent (2019) [30]Depatux-M + TMZ88EGFR amplified (subset with EGFRvIII)9.6 (7.4–11.8)2.7 (2–3.8)NRNRNR Depatux-M86EGFR amplified (subset with EGFRvIII)7.9 (6.1–8.7)1.9 (1.9–2.2)NRNRNR Lassman (2019) [27]Depatux-M + TMZ (Dose Escalation & Expansion)60EGFR amplified + EGFRvIII7.4 (6.5–9.6)2.1 (1.1–3.4)25.2 (14.9–36.9)NR13.81 CR, 7 PRChi (2020) [31]Dacomitinib (Arm B, Anti-VEGF Naïve)30EGFR amplified (subset with EGFRvIII)10.32.317NR31 CRDacomitinib (Arm C, Post-Bevacizumab)16EGFR amplified6.21.80NRNR Padovan (2020)[35]Depatux-M + TMZ36EGFR amplified8.04 (5.3–10.7)2.1 (1.7–2.4)38NRNR Cardona (2021)[34]Osimertinib + Bevacizumab15EGFR amplified + EGFRvIII9.0 (3.9–14.0)5.1 (2.8–7.3)46.7NR13.31 CR, 1 PRChoi (2023)[32]GC111818EGFR amplified + EGFRvIII5.71.75.6 (0.3–25.8)NR5.61 CRD’Alessandris (2023)[33]Bevacizumab + Erlotinib14EGFRvIII8.06.0642957 Bagley (2025)[36]CART-EGFR-IL13Rα2 (bivalent intraventricular CAR T therapy)18EGFR amplifiedNot Reached (5.8–NE)1.9 (1.1–3.4)NRNR8Radiographic regression observed in 62%Survival metrics from each study based on the interventions.NR, Not Reported; CR, Complete Response; PR, Partial Response; OS, overall survival; PFS, progression free survival; PFS6, progression free survival at 6 months; PFS12, progression free survival at 12 months; CI, confidence interval; NR, not reported; CR, complete response; PR, partial response; EGFR, epidermal growth factor receptor; Amp, amplification; TMZ, temozolomide; VEGF, vascular endothelial growth factor; CAR T, chimeric antigen receptor T-cell; EIAED, enzyme-inducing antiepileptic drug. Survival outcomes of individual studies using EGFR targeted interventions. Survival metrics from each study based on the interventions. NR, Not Reported; CR, Complete Response; PR, Partial Response; OS, overall survival; PFS, progression free survival; PFS6, progression free survival at 6 months; PFS12, progression free survival at 12 months; CI, confidence interval; NR, not reported; CR, complete response; PR, partial response; EGFR, epidermal growth factor receptor; Amp, amplification; TMZ, temozolomide; VEGF, vascular endothelial growth factor; CAR T, chimeric antigen receptor T-cell; EIAED, enzyme-inducing antiepileptic drug. Discussion This systematic review synthesized evidence from 12 studies evaluating the predictive value of EGFR amplification and EGFRvIII expression for response to EGFR-targeted therapies in rGBM. Across the included studies, the efficacy of these interventions remained limited. Median OS ranged from 5.7 months with GC1118 monotherapy up to 10.3 months with dacomitinib in anti-VEGF naïve patients, while median PFS was most frequently between 1.7 and 3.0 months across studies of erlotinib, afatinib, dacomitinib, and depatux-m, with a few patients achieving durable disease control. Although some regimens, such as erlotinib combined with bevacizumab, demonstrated ORR as high as 57%, the median ORR across all reported studies was only 6.8%. These effects were typically transient without sustained PFS or consistent improvements in OS, thereby underscoring the challenges of achieving substantial clinical benefits in rGBM. Notably, improvements in ORR and PFS6 were inconsistently associated with OS benefit and should therefore be interpreted as surrogate, non-decisive endpoints in this setting. While most studies demonstrated only modest and short-lived benefits, a few reported higher ORR in carefully selected patient populations. Notably, D’Alessandris et al. evaluated bevacizumab combined with erlotinib in a small cohort of 14 patients selected for VEGF overexpression, EGFRvIII expression, and normal PTEN status, reporting a PFS6 of 64% and an ORR of 57%, with these outcomes being substantially higher than what most other studies reported. Although this may reflect the small cohort size, as the median OS was only 8 months with limited follow-up data, suggesting that even in this molecularly enriched cohort, a durable benefit was not established. By comparison, TMZ, historically one of the most widely used therapies for recurrent GBM, has demonstrated more consistent efficacy across phase II trials, with an accumulative PFS6 of 27.8%, ORR 14%, and a 6-month survival of 65.0% [37]. Similarly, the phase III EORTC 26101 trial demonstrated that lomustine combined with bevacizumab achieved a median PFS of 4.2 months and a median OS of 9.1 months compared to 8.6 months with lomustine alone [38]. These benchmarks generally exceed those achieved by EGFR targeted monotherapy across the included studies, suggesting that the observed benefit may not be sustained, even in a biologically selected population. These findings align with prior reviews investigating EGFR-targeted therapies and EGFR alterations in GBM, including studies that enrolled both mixed and recurrent cohorts. A Cochrane review by Lee et al. found that anti-EGFR therapies, including TKIs such as afatinib and erlotinib, and ADCs like depatuxizumab mafodotin (ABT-414), did not improve median OS, with median OS in recurrent or mixed cohorts ranging from 7 to 10 months [39]. Similarly, Zhu et al. conducted a large-scale meta-analysis involving over 4000 patients across 32 studies, found that EGFR amplification was significantly associated with poorer survival. They also reported that EGFRvIII expression was associated with worse prognosis in patients with rGBM compared to EGFRvIII-negative patients, identifying it as a potential prognostic marker in recurrent disease [40]. However, mechanistic and review literature suggests that the lack of clinical benefit from EGFR targeted therapies may be driven by multiple overlapping resistance pathways. A study by McKinney et al. demonstrated that EGFR inhibition can lead to the enrichment of an ALDH1A1 high tumor cell subpopulation with enhanced evasive capacity, increased resistance to lipid peroxidation, and protection from oxidative stress, enabling these cells to persist during treatment and continue driving tumor progression [41]. In parallel, a review by Pan and Magge highlighted additional barriers to efficacy, including compensatory activation of alternative receptor tyrosine kinases such as MET and IGF1R, and engagement of escape pathways like PI3K/Akt and MAPK [19]. They also noted that suboptimal drug delivery across the BBB, particularly into infiltrative tumor margins where the barrier remains largely intact, further restricts the clinical effectiveness of EGFR targeted therapies [19]. Collectively, these factors underscore the multifaceted nature of resistance mechanisms in rGBM and contribute to the reasoning behind why EGFR amplification and EGFRvIII mutation have yet to translate into reliable predictive biomarkers for treatment selection. This systematic review has several notable aspects that enhance its methodological strength and clinical relevance. It is one of the few reviews to focus specifically on EGFR amplified and/or EGFRvIII positive rGBM, a subgroup often underrepresented or combined with newly diagnosed populations in previous reviews, and for which no standard therapy currently exists [42]. Additionally, by restricting included studies to those of EGFR amplification and/or EGFRvIII expression using validated molecular techniques and reporting outcomes by molecular subgroup, the review offers a narrow and targeted analysis of the predictive value of a distinct population. A comprehensive and specific search strategy, guided by a PROSPERO-registered protocol that aligned with PRISMA guidelines, ensured a broad and unbiased gathering of the literature. Prioritizing clinically meaningful endpoints such as OS, PFS, and ORR offers practical insight into the therapeutic relevance of EGFR alterations across various treatment classes. Despite its focused scope and strong methodology, this review is limited by several recurring issues in the primary literature. Many studies included were early phase, single-arm, or non-randomized, with a median sample size of 18.5 patients, limiting statistical power and external validity. Biomarker assessment methods varied substantially, with EGFR amplification and EGFRvIII expression measured using different techniques (FISH, IHC, qPCR, RT-PCR, GSNAP), thus lacking standardized or validated methods, which limited comparability across studies. Using next-generation sequencing of key GBM genes may provide an assessment of combinations of multiple genetic alterations and identify mutations that guide treatment. Although ORR, PFS6, and PFS12 are standard outcome measures in oncology, they were inconsistently reported across studies, with some trials omitting them altogether. While MGMT methylation status was reported in most studies, IDH-wildtype status was often missing, and few studies explored whether these biomarkers influenced treatment outcomes. Consequently, this limited the ability to isolate the predictive role of EGFR alterations independent of other molecular factors. Furthermore, incomplete reporting of clinical variables including age, sex, and baseline performance status (KPS), further limited the ability to account for important confounding variables. Collectively, these limitations reduce confidence in the interpretation and comparability of findings across studies, highlighting the challenge of drawing firm conclusions about the predictive value of EGFR alterations in rGBM. This uncertainty reinforces the need for further research to clarify the predictive value of EGFR amplification and EGFRvIII expression in rGBM, where effective treatment options remain scarce. Future studies should prioritize biomarker-stratified clinical trials that evaluate EGFR-targeted therapies specifically in the recurrent setting. Specifically employing standardized molecular techniques like FISH for EGFR amplification and RT-PCR for EGFRvIII detection, alongside reproducible and clinically validated thresholds, thus enabling reliable comparison across studies. Moreover, prospective trial designs that incorporate predefined subgroup analysis and adjust for established prognostic variables such as MGMT methylation and IDH status will be essential to determine the true predictive value of EGFR alterations. While targeted monotherapies have only demonstrated modest clinical benefit, combination approaches such as ADCs paired with TMZ warrant further investigation. Although CAR-T therapy represents a novel and encouraging treatment modality, the current available evidence is derived from small, early phase cohorts, with immature survival data, thus results should be interpreted cautiously. Clinically, establishing whether EGFR amplification or EGFRvIII expression predicts response to targeted therapies would enable a more individualized treatment selection, guide drug development strategies, and improve therapeutic precision in this challenging recurrent disease. Conclusion Current evidence does not support EGFR amplification or EGFRvIII mutation as reliable predictive biomarkers for EGFR-targeted therapy in recurrent glioblastoma, and these alterations are not recommended to be used to guide off-trial treatment decisions based on current evidence. While some studies report improved outcomes in patients harboring these alterations, the evidence is inconsistent and limited by heterogeneity in study design, diagnostic methods, and outcome reporting. Well-designed biomarker stratified controlled trials using standardized molecular techniques are required to determine the clinical utility of EGFR alterations in guiding treatment selection for rGBM. Limitations This systematic review has several methodological limitations. Although a comprehensive search was conducted across PubMed, Embase, and Cochrane Library using both defined and free-text terms, the restriction to English language studies may have introduced a language and publication bias. Grey literature, including conference abstracts and unpublished data, was excluded, which may have led to the omission of relevant findings. Importantly, most studies included were single-arm or non-randomized and lacked the appropriate control groups, which fundamentally limits the ability to establish predictive biomarker effects. A further limitation is that EGFRvIII expression was frequently assessed using pretreatment tumor tissue, which may not reflect molecular status at recurrence. Given the evidence that EGFRvIII expression can be lost over time, which may contribute to the variability between biomarker status and treatment response. From a methodological perspective, dual independent screening and data extraction were implemented using Covidence. However, the overall certainty of the evidence was not formally assessed using a validated framework like GRADE, thereby limiting the ability to evaluate the strength and confidence of the conclusions. Additionally, the substantial heterogeneity across included studies limited the ability to perform statistical analyses. As a result, the findings were summarized narratively, which can be more susceptible to interpretation bias despite standardized comparisons. Ultimately, these limitations should be acknowledged when interpreting the evidence and conclusions of this review. Supplementary Material File 6 Supplementary Tables for Supplementary File.docx Funding Statement This paper was not funded. Authors contributions Dylan Evans—Conceptualization, data curation, formal analysis, writing—original draft. Joanne Voisey—Conceptualization, supervision, validation, writing—review & editing. Jennifer Gunter—Conceptualization, supervision, validation, writing—review & editing. Fiona Rae—Conceptualization, supervision, validation, writing—review & editing. All authors have read and approved the manuscript for submission. Disclosure statement The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties. No writing assistance was utilized in the production of this manuscript. Reviewer disclosures Peer reviewers on this manuscript have no relevant financial or other relationships to disclose. Data availability statement All data extracted and analyzed in this review are included in the article and the supplementary material file. References Papers of special note have been highlighted as either of interest (•) or of considerable interest (••) to readers An informative article outlining the molecular mechanisms of resistance that likely contribute to the limited efficacy of EGFR targeted therapies in glioblastoma. References Provides some of the most informative comparative clinical evidence currently available for EGFR targeted therapy in EGFR-amplified recurrent glioblastoma. References References Associated Data Supplementary Materials File 6 Supplementary Tables for Supplementary File.docx Data Availability Statement All data extracted and analyzed in this review are included in the article and the supplementary material file.