Real-World Outcomes of Anti-EGFR Rechallenge Guided by Liquid Biopsy in Metastatic Colorectal Cancer

Colorectal cancer is the third most frequently diagnosed cancer and the second leading cause of cancer-related mortality worldwide [1]. Despite significant advances in treatment over the past decade, up to 60% of patients diagnosed with localized colorectal cancer develop metastatic disease, with a 5-year survival rate below 20% [2]. Clinical and pathological factors such as sidedness (right versus left colon) and RAS or BRAF mutational status, dictate distinct prognosis and treatment strategies [3, 4].

The epidermal growth factor receptor (EGFR) is a transmembrane receptor with tyrosine kinase activity that mediates cell proliferation and growth. EGFR interacts with various downstream substrates, including the MAPK pathway, which involves the proto-oncogenes RAS and RAF – two of the most actionable targets in current cancer treatments [2, 5]. Up to 45% and 10% of metastatic colorectal cancers (mCRCs) harbour RAS or BRAF activating mutations, respectively, which confer intrinsic resistance to EGFR blockade [2].

Somatic RAS/BRAF mutational status, primary tumour location, and mismatch repair protein deficiency or proficiency (MMRp) are key determinants in selecting first-line treatment for mCRC [2, 3, 6]. For MMRp RAS/BRAF wild-type (wt) tumours, there is an overall survival (OS) benefit when treated in the first line with anti-EGFR antibodies (cetuximab or panitumumab) combined with a chemotherapy doublet. Data from the PARADIGM trial also demonstrated that left-sided tumours derive greater benefit from first-line EGFR blockade [3, 6‒8].

During treatment, the genomic landscape of mCRC evolves, and resistance to anti-EGFR therapies may emerge, leading to disease progression [9, 10]. KRAS-/NRAS-activating mutations and point mutations in the antibody-binding site of the EGFR extracellular domain can develop [9]. In second-line treatment, without EGFR blockade, EGFR sensitivity may be restored. One hypothesis suggests that this occurs due to the clearance of RAS and EGFR mutant clones, which can be monitored via circulating tumour DNA (ctDNA) analyses from serum liquid biopsies [11‒13]. The restoration of EGFR sensitivity enables the possibility of retreatment with anti-EGFR therapies, leading to improved outcomes in patients who meet the criteria for anti-EGFR rechallenge [9, 14].

Anti-EGFR rechallenge involves retreating mCRC patients with anti-EGFR antibodies if they previously achieved a response or SD during prior anti-EGFR therapy. This strategy requires an EGFR-free interval between two anti-EGFR-containing lines of treatment, ideally lasting 6 months [9, 14]. The single-arm CRICKET trial demonstrated increased response rates with cetuximab/irinotecan rechallenge in third-line mCRC treatment. However, evidence remains insufficient to establish this as a high-level treatment recommendation [3, 14]. A retrospective analysis of liquid biopsies collected prior to rechallenge in this trial revealed that only ctDNA RAS/BRAF wt patients responded, with superior progression-free survival (PFS) compared to ctDNA RAS-mutant patients [14].

Online supplementary Table S1 (for all online suppl. material, see https://doi.org/10.1159/000544044) summarizes the results of several phase II trials investigating the efficacy of anti-EGFR rechallenge in ctDNA RAS/BRAF wt mCRC patients. In this study, we present a real-world cohort treated with ctDNA-guided anti-EGFR rechallenge to further evaluate the clinical benefit and tolerability of this strategy in mCRC management.

This unicentric, retrospective cohort study included all patients aged 18 years or older with somatic RAS/BRAF wt mCRC treated with the combination of irinotecan and cetuximab between January 2018 and September 2023. Patients were eligible for inclusion if they met the following criteria for rechallenge: (1) completion of at least two prior lines of treatment for advanced disease, with at least one line involving an anti-EGFR antibody; (2) evidence of a previous partial response (PR), complete response (CR), or stable disease (SD) to anti-EGFR treatment, according to the Response Evaluation Criteria in Solid Tumours (RECIST) [15]; (3) anti-EGFR therapy in the line of treatment immediately preceding rechallenge. Additionally, liquid biopsy confirming KRAS, NRAS, and BRAF wt status was required prior to anti-EGFR rechallenge or within 1 month of the first cycle. Patients lacking records of prior systemic treatment (in early or advanced stages) were excluded.

Demographic and baseline patient characteristics, along with oncological treatment details, were extracted from electronic health records. These included age, sex, Eastern Cooperative Oncology Group performance status (ECOG PS), tumour staging per the AJCC 8th edition [16], and tumour sidedness. Data on toxicity were obtained from clinical reports and graded using CTCAE version 5 [17].

The primary survival outcomes were OS, defined as the time from initiation of rechallenge therapy to death from any cause, and PFS, defined as the time from anti-EGFR rechallenge initiation to either disease progression per RECIST 1.1 or death from any cause [15]. Efficacy was further evaluated through: (1) disease control rate (DCR), defined as the proportion of patients achieving CR, PR, or SD at the first radiological assessment; (2) overall response rate (ORR); defined as the proportion of patients achieving CR or PR at the first radiological assessment. The timing of response evaluations was determined at the clinician’s discretion during follow-up, with retrospective review of response criteria. Safety and tolerability outcomes were assessed by calculating the frequency and severity (grade) of toxicities.

All liquid biopsies were collected at IPOC using Streck Cell-Free DNA blood collection tubes^® (Streck, La Vista, NE, USA). The following hotspot mutations of KRAS, NRAS, and BRAF were assessed with real-time polymerase chain reaction using cobas™ KRAS Mutation Test V2 (Roche Diagnostics) and BRAF/NRAS mutation test (Roche Diagnostics): KRAS codons G12, G13, A59, Q61, K117, A146; NRAS codons G12, G13, A18; A59, Q61, K117, A146; BRAF codons V600, V61, G466, G469.

Continuous data were presented as medians with ranges (minimum and maximum values), while categorical data were summarized as absolute and relative frequencies. Survival outcomes were estimated using the Kaplan-Meier method, with medians and 95% confidence intervals (CIs) reported. Subgroup analyses were conducted for predefined efficacy predictors, including stage at diagnosis (stage IV vs non-IV), ECOG PS (0 vs. 1), age (<65 vs. ≥65), tumour sidedness (right vs. left), primary tumour site (colon vs. rectum), number of metastatic sites (1 organ vs. >1 organ), and the development of cutaneous toxicity (grade ≥2 vs. grade 1 or absent). Hazard ratios (HRs) and 95% CIs were calculated using Cox proportional hazards regression, with statistical significance set at p < 0.05. Analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA).

Between January 2018 and September 2023, 16 patients received anti-EGFR rechallenge guided by liquid biopsy with irinotecan and cetuximab at IPOC (Fig. 1). These patients were followed until December 2023, with a median follow-up of 12.8 months (range: 2.6–46.5 months).

The majority of patients were male (81%), with a median age of 63 years (range: 51–84 years), of whom 44% were aged 65 or older. Sixty-three percent had an ECOG PS of 0, and the same proportion (63%) had stage IV disease at diagnosis. Most tumours were left-sided (69%) and originated in the colon (75%). At baseline, liver metastases were present in 69% of patients, lung metastases in 63%, and 63% had metastatic disease in at least two organs. Table 1 summarizes the baseline characteristics of the study population.

The median time from diagnosis of metastatic disease to rechallenge was 28.3 months (range: 16.7–100.0 months). All patients had previously received irinotecan-based chemotherapy, specifically first-line treatment with fluoropyrimidine, irinotecan, and an anti-EGFR antibody. Rechallenge was predominantly administered in the third-line setting (88%), with a median EGFR-free interval (defined as the time between the last cycle of the first anti-EGFR-containing regimen and the initiation of rechallenge) of 9.8 months (range: 4.1–22.1 months). The characteristics of patients’ prior treatments can be found in the online supplementary Table 2.

Five patients achieved PR, and four experienced SD, resulting in a DCR of 56%. Seven patients progressed at the first response evaluation. The overall response rate (ORR) was 31%, exceeding the ORR observed in the second line of treatment (19%), which primarily involved fluoropyrimidine, oxaliplatin, and, in some cases, bevacizumab. Online supplementary Figure S1 illustrates the response rates for rechallenge compared with previous lines of treatment.

Figure 2 shows the main survival outcomes. The median PFS was 5.3 months (95% CI: 3.3–7.4 months). Among patients receiving anti-EGFR rechallenge as third-line therapy (n = 14), the median PFS was 5.9 months (95% CI: 2.0–9.8 months). The median OS was 12.7 months (95% CI: 8.3–17.1 months) and 12.9 months (95% CI: 11.8–13.9 months) for those receiving rechallenge as a third-line treatment. Two patients received rechallenge in the fourth-line setting after treatment with trifluridine/tipiracil plus bevacizumab. One of these patients achieved PR with a PFS of 4.8 months, while the other experienced disease progression.

Only 1 patient had no reported toxicity. Among the other 15 patients, it was possible to identify 62 adverse events, specified in Table 2. Most adverse events were gastrointestinal (31%), followed by skin and eye toxicity (23%) and haematological toxicity (23%). Eleven percent of all toxicities were grade 3; however, none resulted in treatment discontinuation. Acneiform rash was reported in 75% of patients.

In a subgroup analysis (Fig. 3), patients who developed grade 2 or 3 acneiform rash (n = 9) demonstrated superior PFS (median: 8.8 months, 95% CI: 4.9–12.6 months) compared with those with grade 1 or no rash (median: 4.8 months, 95% CI: 2.5–7.0 months), with a HR of 0.30 (95% CI: 0.1–1.0; p = 0.05). No other studied variables were found to predict PFS, as shown in Table 3. Due to the small sample size of patients receiving fourth-line rechallenge (n = 2), the effect of this variable on survival outcomes could not be assessed.

Our retrospective Portuguese cohort demonstrates a uniform sample of tissue RAS/BRAF wt mismatch repair proficient (MMRp) mCRC patients treated with irinotecan plus cetuximab rechallenge guided by liquid biopsy. Among the 16 included patients, 31% achieved partial response (PR) in the third line or beyond mCRC setting. With irinotecan/cetuximab rechallenge, the median PFS was 5.3 months, and the median OS was 12.7 months, without limiting toxicities.

The randomized phase III SUNLIGHT trial established the efficacy and tolerability of trifluridine/tipiracil (FTD/TPI) plus bevacizumab in refractory mCRC, irrespective of RAS/BRAF status. The combination significantly improved OS over FTD/TPI alone, with a median OS of 10.8 months vs. 7.5 months, respectively [18]. These findings, supported by real-world data [19, 20], provided level IA evidence with a score of 4 on the ESMO Magnitude of Clinical Benefit Scale, for third line and beyond treatment of mCRC [21, 22].

Although FTD/TPI + bevacizumab improves survival and achieves a DCR of 64% according to a meta-analysis [23], only 6% of patients achieve a response, and the absolute increase in median PFS is 3.2 months [18]. In terms of toxicity, 72% of patients experienced grade ≥3 adverse events, with 43% developing grade ≥3 neutropenia [18]. Anti-EGFR rechallenge, as reported in phase II trials (online suppl. Table 1), has shown higher DCRs of up to 80%, particularly when combined with FTD/TPI or irinotecan in European and Asian populations [24, 25]. Although no head-to-head trials have compared FTD/TPI with anti-EGFR or anti-VEGF antibodies, adding anti-EGFR to FTD/TPI appears to improve response rates without significantly affecting survival outcomes, as suggested by the VELO trial and a recently published retrospective cohort [25, 26].

The CITRIC trial provided the first randomized evidence for anti-EGFR rechallenge guided by liquid biopsy in third-line mCRC. Patients receiving irinotecan/cetuximab demonstrated a 13% ORR and 74% DCR, with a non-statistically significant improvement in median PFS (4.4 months vs. 2.2 months) compared to physician’s choice of treatment [27]. However, only 56% of control group patients in the CITRIC trial received FTD/TPI ± bevacizumab, and their outcomes fell below those reported in the SUNLIGHT trial [18, 27]. In contrast, our cohort demonstrated superior ORR (31%) and PFS (5.3 months). Notably, our cohort included fewer stage IV patients at diagnosis and fewer sites of metastatic disease, compared with CITRIC trial, which may partially explain these differences.

Despite CITRIC trial results, the best third-line option for liquid biopsy RAS/BRAF wt mCRC remains unknown, while the stronger level of evidence favours the FTD/TPI ± bevacizumab combination. Ongoing trials, such as the second part of the phase III FIRE-4 study (AIO KRK-0114; NCT02934529), are expected to clarify the role of irinotecan/cetuximab rechallenge in the third-line setting for RAS/BRAF wt mCRC. By prospectively selecting patients based on ctDNA mutational status, FIRE-4 also aims to define the optimal timing and methodology of liquid biopsy to monitor and address anti-EGFR resistance. Further prospective studies are also investigating the ideal drug combinations and timing for rechallenge [9].

While phase III evidence for rechallenge is lacking, it remains a valid third-line option for selected patients, particularly those requiring tumour shrinkage or unable to tolerate more toxic regimens [22]. Similar to the findings of the CAVE trial, our study suggests a potential role for skin toxicity, such as acneiform rash, as a biomarker for response to rechallenge [28]. However, this biomarker requires further validation due to conflicting data [29, 30].

Our study has several limitations, primarily related to its retrospective and unicentric nature. The small sample size limits the generalizability of the findings, and the timing of response evaluations was performed at the clinician’s discretion, introducing variability that may have affected the consistency of the results. Additionally, the retrospective collection of adverse events could have led to underreporting or incomplete documentation of toxicities. Another significant limitation is the restricted liquid biopsy analysis, which focused only on RAS and BRAF hotspot mutations, excluding broader molecular profiling.

In the era of precision oncology, the limited molecular characterization of this study represents a critical limitation. Comprehensive molecular profiling, using broader next-generation sequencing panels, is increasingly recommended to guide the treatment of mCRC. This approach extends beyond the evaluation of RAS/BRAF status to identify additional genetic alterations, such as ERBB2/MET amplifications and NTRK/ROS1/ALK/RET rearrangements, which are known to adversely impact the efficacy of anti-EGFR therapies [31, 32]. Importantly, comprehensive profiling allows for the process of “negative hyperselection,” which involves further selecting patients based on the absence of additional genetic alterations associated with primary or secondary resistance to anti-EGFR treatment. By applying this refined strategy, molecular characterization may not only identify patients likely to benefit from rechallenge but also help to exclude those at risk of treatment failure due to underlying resistance mechanisms [32, 33].

Our findings highlight the potential of irinotecan/cetuximab rechallenge guided by liquid biopsy as a viable third-line treatment for selected patients with RAS/BRAF wt mCRC. These real-world data highlight the utility of anti-EGFR rechallenge in the continuum of care for mCRC, demonstrating outcomes comparable to phase II trials, with adequate DCRs and manageable toxicity. However, further high-quality randomized trials are essential to validate these findings, determine the optimal treatment sequencing, and establish robust biomarkers such as skin toxicity and broader genomic alterations to refine patient selection and improve outcomes.

The study (TI 26/2024) was reviewed and approved by the Ethics Committee of the Portuguese Institute of Oncology of Coimbra on 28 October 2024. The Ethics Committee granted an exemption from requiring written informed consent from the patients, given the characteristics of the sample.

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

Ricardo Roque and Nuno Bonito conceptualized and designed the study. Ricardo Roque, Rita Coelho, Isabel Viera Fernandes, Gonçalo Cunha, and Marta Gonçalves preformed dataset management and data collection. Ricardo Roque, Rita Antunes Santos, Luís Guilherme Santos, and Judy Paulo performed data analysis and data interpretation. Ricardo Roque, Judy Paulo, and Nuno Bonito wrote the manuscript. All authors reviewed and approved the final version.

The anonymized data supporting the findings of this study are available from the corresponding author upon reasonable request.