Landscape of acquired resistance alterations in gastrointestinal malignancies after genomically targeted therapy

Introduction

Advances in understanding of cancer biology and increased accessibility to comprehensive genomic profiling have enabled the development of novel therapeutics targeting specific genomic alterations driving cancer growth and progression. This era of precision oncology has led to prolonged survival and improvements in quality of life for patients with many cancer types.

Several important pathways driving cancer via genetic alterations have been identified including the receptor tyrosine kinase (RTK)/RAS/mitogen-activated protein kinase (MAPK) pathway, phosphatidylinositol 3-kinase (PI3K)/Akt signaling, and the Janus kinase (JAK)-signal transducer and activator of transcription (STAT) pathway (1,2). Most of the approved targeted therapies in gastrointestinal (GI) cancers inhibit abnormal oncogenic activation of the RTK/RAS/MAP-kinase pathway. This canonical signaling pathway involves extracellular signaling molecules [such as epithelial growth factor (EGF) and fibroblast growth factor (FGF)] activating receptor tyrosine kinases (RTKs) at the cell surface, which results in sequential activation of the downstream MAPK cascade, including therapeutic targets RAS, RAF, MEK, and ERK, and regulates cell development, proliferation, survival and death (3). Defects in this pathway leading to constitutive activation have been shown to be major oncogenic drivers in many malignancies, including most GI cancers.

Therapeutics that inhibit this abnormal activation have demonstrated clinical benefit. Within GI cancers, targeted therapies with demonstrated clinical activity include anti-EGFR monoclonal antibodies (mAbs), anti-HER2 (ERBB2) agents, small molecule selective inhibitors of FGFR and IDH1 as well as combination therapies targeting BRAF V600E and KRAS G12C mutations. In nearly all cancers, resistance to targeted therapies inevitably develops despite initial clinical efficacy. This acquired resistance is defined as therapeutic resistance after an initial period of objective response or stable disease (4). One mechanism for acquired resistance includes ‘on-target’ mutations, which are secondary alterations within the drug target itself that confer resistance to targeted therapy, most commonly by affecting drug binding, while preserving oncogenic driver function. For example, in FGFR2-altered cholangiocarcinoma, FGFR2 kinase domain mutations—particularly molecular brake mutations (N550) that destabilize the inactive kinase conformation and gatekeeper mutations (V565) that create steric clashes blocking drug access—represent the dominant on-target resistance mechanism acquired after initial clinical benefit to FGFR inhibitors (5). In colorectal cancer (CRC) treated with anti-EGFR antibodies, EGFR extracellular domain mutations (e.g., S464L, G465R) emerge during cetuximab treatment and are located within the antibody-binding region, where they prevent cetuximab binding while preserving receptor function and downstream signaling (6). Another major mechanism of acquired resistance involves ‘bypass’ alterations in which tumor cells evade target inhibition by activating alternative signaling pathways that restore downstream proliferative and survival signaling independent of the inhibited oncogenic driver. In contrast to on-target resistance mutations—which predominantly impair drug binding while preserving target function—bypass mechanisms typically occur either downstream of the inhibited receptor or through parallel receptor tyrosine kinases, effectively reactivating shared signaling nodes such as the MAPK and PI3K pathways. Classic examples include emergence of KRAS mutations or HER2 amplification, mediating resistance to anti-EGFR therapies in colorectal cancer (7,8). Notably, response rates to targeted therapies have been relatively modest in GI cancers compared to other malignancies such as non-small cell lung cancer and melanoma, due to high rates of both upfront and acquired resistance.

Liquid biopsy, and specifically analysis of circulating tumor DNA (ctDNA) from peripheral blood, represents an efficient approach to assess genomic alterations in patients with cancer. This method has the advantage of being readily accessible without the need for an invasive biopsy, and since ctDNA is shed by tumors throughout the body, it may better characterize genomic tumor heterogeneity compared to tissue biopsy from a single site (9). ctDNA has therefore become an increasingly utilized method for assessing acquired genomically mediated resistance and has demonstrated the ability to identify multiple tumor subclones harbouring distinct resistance mechanisms. While this approach reflects the diversity of alterations across metastatic sites, it also highlights that not all detected alterations may be uniformly represented across all lesions, which may have implications for therapeutic sensitivity.

Understanding patterns of acquired resistance is essential for improving therapeutic options for patients receiving targeted therapy, whether anticipating resistance or adapting therapy upon its emergence. Therefore, we explored the landscape of acquired genomic resistance alterations in ctDNA across patients with GI malignancies treated with targeted therapies to identify common patterns that may inform future therapeutic approaches. We present this article in accordance with the STROBE reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1025/rc).

Methods

Patient identification and sample collection

Patients with locally advanced or metastatic GI malignancies treated at the Massachusetts General Hospital Cancer Center, Boston, Massachusetts, USA who received genomically targeted therapy [Food and Drug Administration (FDA)-approved or investigational as part of an institutional review board-approved protocol] between December 1^st, 2011, and December 5th, 2025, were evaluated for eligibility. Patients who received targeted therapy directed at genomic alterations for at least 60 continuous days were identified. Those who completed comprehensive genomic profiling with a cell-free DNA based commercial assay (Guardant360, Guardant Health) after clinical or radiologic progression of disease (at the discretion of the treating physician) on targeted therapy were included. Guardant 360 is a validated next-generation sequencing–based platform that interrogates a broad panel of cancer-related genes and detects single nucleotide variants, insertions/deletions, copy number alterations, and selected gene fusions with high analytic sensitivity. Baseline comprehensive molecular profiling was available for all included patients and was performed using either tissue- or liquid biopsy–based assays. Both baseline and post-progression genomic data were required for inclusion in the analysis.

Plasma (and tumor specimens) were collected through a systematic liquid biopsy program within the Center for Gastrointestinal Cancers. Patients were prospectively enrolled at the Massachusetts General Hospital Cancer Center. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Dana Farber/Harvard Cancer Center institutional review board (#14-046, 13-416, 18-380, 02-240). All included patients provided written informed consent for this prospective analysis; meanwhile, informed consent was waived for the retrospective review.

Characterization of genomic alterations

Genomic alterations observed in ctDNA at progression, including single nucleotide variants (SNVs), small insertions and deletions (InDels), amplifications, and loss of heterozygosity (LOH), were analyzed and characterized as putative resistance-associated alterations based on review of published literature and public databases of human genetic variants, including ClinVar and The Cancer Genome Atlas (TCGA). Benign variants and variants of uncertain significance were excluded. Pathogenic or likely pathogenic alterations were categorized based on established conceptual frameworks in the targeted therapy resistance literature (9). Alterations were classified as ‘on-target’ resistance mechanisms when they involved secondary mutations or alterations within the drug target itself, most commonly predicted to reduce therapeutic binding or target inhibition while preserving target function. Alterations occurring in parallel or downstream signaling pathways that could reactivate pathway signaling independent of the inhibited target were classified as ‘bypass’ resistance mechanisms. Of note, cis versus trans mutations were not possible to distinguish for mutations occurring in targeted genes using available data; these were assumed mechanistically based on review of literature, with trans mutations regarded as ‘bypass’.

Statistical analysis

A CoMut Plot of putative resistance associated alterations was made with the Van Allen Lab’s CoMut Python Library (Figure 1) (10). Information such as primary disease type and general treatment category was also included. Additional analyses of the noted resistance mutations, such as breakdowns of specific mutation type and frequency, were performed in Excel (Microsoft Inc., version 16).

Figure 1 Known putative resistance alterations identified in progression ctDNA testing. Genomic alterations known to be involved in adaptive resistance to targeted therapies were identified in testing upon or soon after progression on first targeted therapy. These were then categorized based on mutation type (SNV, InDel, amplification) and location (bypass, on-target). Each colored cell represents at least one mutation of that type in the patient of that column. The total number of identified resistance mutations is then tallied in the bar graph. BTC, biliary tract cancer; CRC, colorectal cancer; ctDNA, circulating tumor DNA; GEC, gastroesophageal cancer; InDel, insertion and deletion; SNV, single-nucleotide variant.

Results

One hundred and six patients met the inclusion criteria and were included in the analysis. These included biliary tract cancer (BTC, n=45), colorectal cancer (CRC; n=42), gastroesophageal cancer (GEC; n=14) and other tumor types (n=5; 1 carcinoma of unknown primary, 1 duodenal adenocarcinoma, 2 pancreatic ductal adenocarcinomas, 1 signet ring cell adenocarcinoma of the appendix). Therapeutic targets were BRAF V600E (n=14), EGFR (n=23), FGFR (n=45), HER2 (n=9), KRAS (n=3) and other miscellaneous targets or combinations of targets such as KRAS and EGFR (n=4), IDH1 (n=2), MET (n=2), MET and EGFR (n=1), ERK (n=1), HER3 (n=1), NTRK-CSFR (n=1). Patterns of cancer type and the associated resistance mutations detected are summarized in Figure 1 and Figure S1. As expected, CRC most frequently received targeted therapy against BRAF V600E and EGFR, while BTC received therapy most frequently against FGFR2, and GEC received therapy most frequently targeting HER2. Consistent with known biology of GI cancers, putative resistance mutations were primarily detected either in RTKs or the downstream MAPK pathway. Overall, mutations were heterogenous in frequency, type, and specific mutated genes within and across cancer types and targets, although distinct patterns emerged.

Notably, at least one putative resistance associated alteration was detected in 56/106 patients (53%), with 21/106 (20%) showing 1 alteration and 35/106 (33%) with 2 or more alterations (Figure 2A). 12 patients had 5 or more resistance mutations identified, with 2 patients having more than 10 resistance mutations each (Figure 1, Figure 2A). These data demonstrate that ctDNA frequently identifies genomic mechanisms of resistance and often multiple distinct subclonal resistance events across GI cancers. Still, 50/106 patients (47%) had no known genomic pathogenic resistance alteration detected, suggesting alternative mechanisms of resistance not captured by the ctDNA panel used.

Figure 2 Frequency of overall resistance alterations. (A) All patients were categorized into either having no noted resistance mutations, one, or multiple. Nearly half (47%, 50/106) of all patients did not develop a noted resistance alteration. Patients developed multiple noted resistance alterations at slightly more frequency than developing a single alteration, 33% (35/106) vs. 20% (21/106) respectively. (B) The patterns of developing zero, one, or multiple resistance alterations was evaluated for each primary cancer type. (C) The patterns of developing zero, one, or multiple resistance alterations was also evaluated for each targeted therapy target. BTC, biliary tract cancer; CRC, colorectal cancer; GEC, gastroesophageal cancer; n, number of patients in each cohort; pts, patients.

Among subtypes of GI cancers, both CRC and BTC had a substantial number of resistance alterations detected (Figure 2B). In CRC, 55% (23/42) of patients had at least one resistance-associated alteration detected, and most of these patients harbored multiple alterations. Similarly, in BTC, 55% (25/45) had at least one resistance-associated alteration, with the majority also demonstrating more than one alteration. In these cancers, for whom acquired genomic alterations is a critical mechanism of resistance, subclonal heterogeneity appears common. In contrast, alterations were less commonly detected in GEC (43%) or other cancers (40%). Examining acquired alterations by target (Figure 2C) showed a similar pattern. A majority of BRAF (57%), EGFR (48%), and FGFR2 (62%) targeted cancers had at least one resistance mutation detected, with many demonstrating more than 1. In contrast, HER2 targeted cancers were less likely (22%) to have a resistance alteration detected on ctDNA, all of which were a single alteration. Other miscellaneous alterations had too few patients to meaningfully characterize patterns. ctDNA therefore had varying, but still meaningful utility to detect acquired resistance across cancer types and targets.

Next, we characterized the distribution of genomic variant type, including SNVs, amplifications, and InDels, across cancer type and target (Figure 3). Of the 164 total resistance alterations detected, 82% (134/164) were SNVs, 17% (28/164) were amplifications, and 1% (2/164) InDels (Figure 3). By cancer type, a majority 66% of resistance alterations in CRC were SNVs, while 31% were amplifications. Similarly, 92% of resistance alterations in BTC and 89% in GEC cancers were SNVs while only 8% and 11% respectively were amplifications.

Figure 3 Characterization by type of resistant genomic variation: amplification, SNV, and InDel noted upon progression. Every noted resistance mutation was categorized as either an amplification, SNV, or InDel. The percentage of each category was evaluated for the whole cohort, primary cancer type and by targeted therapy target. BTC, biliary tract cancer; CRC, colorectal cancer; GEC, gastroesophageal cancer; InDel, insertion and deletion; n, number of resistance alterations identified in progression ctDNA testing; SNV, single-nucleotide variant.

Among resistance alterations observed in BRAF and EGFR targeted cancers, 75% (12/16) and 64% (21/33) respectively were acquired SNVs, while 25% (4/16) and 30% (10/33) respectively were acquired resistance amplifications (Figure 3). 5% (2/40) of noted resistance mutations in EGFR targeted cancers were InDels, detected in only a single CRC patient. This patient developed both an on-target insertion in EGFR (Q432_H433insK) and a bypass deletion in MAP2K1 (E102_I103del). In contrast, 93% (89/96) of resistance alterations in FGFR targeted cancers were SNVs, while both (100%) resistance alterations detected in HER2-targeted cancers were amplifications. Interpretation of the latter result is limited significantly by a small sample size. Thus, overall, while SNVs were predominant across cancer types and targets except HER2, amplifications were observed more frequently in patients with CRC and GEC or BRAF and EGFR targeted therapy than in patients with BTC or patients receiving FGFR2 targeted therapy. These results offer potential insight into the mechanisms that lead to the emergence of acquired resistance among these GI cancers.

Resistance alterations may be broadly divided into bypass and on-target alterations, with significant implications for strategies that may overcome or suppress development of acquired resistance. We therefore characterized each mutation as bypass or on-target as described in the methods to examine differences across cancer types and targets. Overall, approximately half of mutations (86/164, 52%) were characterized as on-target while the remaining half (78/164, 47%) were bypass (Figure 4A). Interestingly however, in CRC the majority, 69% (40/58), were bypass alterations, while in BTC 61% (51/83) were on-target alterations; 89% (16/18) alterations in GEC were on-target, though this cohort was relatively small. When categorized by target type, a similar pattern was observed, unsurprisingly given the distribution of targets in respective malignancies. 88% (14/16) of resistance alterations in BRAF targeted treatment and 70% (23/33) of resistance alterations in EGFR targeted treatment were bypass alterations (predominantly in CRC) while 66% (63/96) of resistance alterations in FGFR targeted treatments were on-target alterations (predominantly in BTC).

Figure 4 Frequency of bypass and on target resistance alterations. (A) Every noted resistance alteration was categorized as either bypass or on-target. The percentage of each category was evaluated for the full cohort, primary cancer type and treatment target. In the CRC cohort 69% (40/58) were bypass alterations while in the biliary cohort 61% (51/83) were on target alterations (n = total number of resistance mutations identified in progression ctDNA testing). (B) Of the 56 patients with at least one resistance mutation (n), 23 (41%) were CRC, 25 (44%) were BTC, 6 (11%) were GEA, and 2 (3%) were a different type of cancer. 34% (19/56) only developed on target resistance alterations, 41% (23/56) only developed bypass alterations, and 25% (14/56) developed both. (C) Every noted resistance alteration was categorized as RAS or not. A majority of the bypass resistance alterations noted were either KRAS or NRAS mutations, with KRAS being more common than NRAS (n = total number of resistance alterations across full cohort). BTC, biliary tract cancer; CRC, colorectal cancer; ctDNA, circulating tumor DNA; GEA; GEC, gastroesophageal cancer.

Importantly, multiple classes of resistance mechanisms were observed in some patients. Of the 56 patients with at least 1 resistance alteration noted, 25% (14/56) developed both on-target and bypass alterations while 34% (19/56) only developed on-target resistance alterations and 41% (23/56) only developed bypass alterations (Figure 4B). Patients with CRC more frequently exclusively developed bypass alterations rather than only on-target alterations or both, while BTC patients more frequently exclusively developed on-target alterations. When considering the heterogeneity in frequency and type of mutation, these patterns suggest distinct biologic mechanisms of acquired resistance. The heterogeneity of bypass resistance was especially prominent in CRC, highlighting the complex challenge of overcoming resistance to targeted therapies in this population.

Given that various components of the MAPK pathway, especially KRAS, are described both as important oncogenic drivers and major mechanisms of resistance, we explored the frequency of the canonical RAS gene family members as resistance alterations. The primary method of bypass resistance was an emerging KRAS mutation in 42% (33/78). Mutations and amplifications upstream as well as downstream of RAS were seen as the main bypass alteration in 40% (31/78) followed by NRAS alterations in 18% (14/78). (Figure 4C). In contrast, only 8/86 (9%) on-target alterations were in the KRAS gene, driven in part by the small number of KRAS targeted patients in the cohort, while the remaining 78/86 (91%) on-target alterations were observed in non-RAS genes. No on-target alterations were detected in NRAS or HRAS. While alterations were seen in NRAS as a bypass mechanism, HRAS did not emerge as a mechanism of bypass resistance in this dataset. Given the recent development of novel classes of (K)RAS inhibitors currently in clinical testing, the prevalence of RAS mutations as a mechanism of acquired resistance supports the potential to incorporate RAS inhibitors into future therapeutic strategies to overcome resistance in GI malignancies.

Discussion

In this study, we examined the patterns of acquired genomic resistance alterations detected in ctDNA in patients with GI malignancies receiving genomically targeted therapy. Our data demonstrate the utility of ctDNA testing to detect and characterize acquired resistance to targeted therapy in GI cancers, detecting at least one resistance alteration in approximately half of patients, with half of these harboring more than one mutation. The data further highlight the heterogeneity in frequency, type, and mechanism of resistance by cancer type and target. Discernible patterns emerge when analysing by specific cancer type (e.g., BTC vs. CRC) and target (FGFR vs. EGFR, BRAF etc.). In particular, the contrast is striking between BRAF or EGFR targeted therapy, primarily in CRC, and FGFR2 targeted therapy, primarily in BTC. In addition, our data confirm the critical importance of RAS alterations as the dominant mechanism of bypass resistance in GI cancer patients receiving targeted therapy.

In CRC, with mainly EGFR and BRAF targeted therapies, a majority of patients with detected resistance alterations have >1 mutation detected and bypass alterations in the MAPK pathway were notably prominent. This observation reflects the known biology of this tumor type, which demonstrates frequent heterogenous acquired resistance in the MAPK pathway in response to inhibition at various points of the pathway such as EGFR, KRAS, and BRAF. In our dataset, EGFR inhibition resulted in emergence of resistance with downstream activation of the MAPK pathway, prominently with acquired mutations in KRAS (G12V/C/A, Q61H), BRAF (V600E, D594N) and MAP2K1 (K57N/T). Mutations in KRAS and NRAS are known to be one of the most prominent ‘bypass’ alterations resulting in acquired resistance to anti-EGFR therapy, as also seen in our data (7,11). Amplifications in KRAS and MET, other genomic alterations associated with acquired resistance to anti-EGFR therapy, were also seen (12,13). In addition, although a minority compared to bypass alterations, mutations were noted in the EGFR extra cellular domain (I491T, S492R, S464L, G465V), which can emerge under selective pressure and interfere with binding of anti-EGFR agents, resulting in resistance to anti-EGFR mAb therapy (6). These on-target resistance alterations with anti-EGFR therapy are much rarer compared with that observed with anti-FGFR therapy; ‘bypass’ resistance remains the predominant mechanism of acquired resistance to EGFR therapy.

Reactivation of MAPK pathway via ‘bypass’ resistance is also the key mechanism of acquired resistance to anti-BRAF therapy. BRAF inhibitor combinations (typically with anti-EGFR and/or anti-MEK therapy) still lead to pathway reactivation via acquired mutations in KRAS (G12C/D/R, G13D etc.), NRAS (Q61R/K), MAP2K1 (K57N) as well as EGFR, KRAS and MET amplifications, as demonstrated in prior studies (14-16). Thus, despite combination therapies, typically combining BRAF, MEK and EGFR inhibition, MAPK reactivation remains a prominent mechanism of acquired resistance in BRAF V600E mutant CRC. These observations, consistent with our data, demonstrate the importance of optimal vertical inhibition of the MAPK pathway in CRC to suppress acquired resistance and highlights the challenge of adapting targeted therapy to acquired resistance in this population.

In BTC with anti-FGFR therapy targeting FGFR2 fusions or rearrangements, resistance alterations were predominantly on-target, and nearly all were SNVs. Prominently noted were ‘gatekeeper’ mutations which cause steric hindrance preventing drugs from reaching the ATP-binding pocket and ‘molecular brake’ mutations which maintain the kinase in an active conformation. Notable SNVs in our dataset were FGFR2 N549D/H/K/T and V564F/I/L. Next-generation covalently-binding molecules such as futibatinib are now approved, which may overcome some ‘gatekeeper’ mutations that confer resistance to earlier generation ATP-competitive inhibitors such as pemigatinib (5,17). In vitro studies have demonstrated the potential for novel molecules to overcome acquired on-target resistance to initial FGFR targeted therapy (5). Development of molecules with higher isoform selectivity (FGFR2 specific), such as the novel FGFR2-inhibitor lirafugratinib, aim to spare the adverse effects like hyperphosphatemia associated with FGFR1–3 inhibition allowing higher doses and more efficacious FGFR2 inhibition while suppressing on-target acquired resistance (18). In addition, mechanisms for efficacy of these FGFR inhibitors in FGFR2 fusions/rearrangements but not mutations or amplifications need to be determined to broaden therapeutic potential.

Our study also demonstrates the feasibility and increasing utility of on-treatment ctDNA evaluation to adapt future therapy. Molecular profiling with ctDNA may predict response or early resistance and may guide next therapies at progression. In patients with CRC receiving anti-EGFR therapy, ctDNA has been demonstrated to predict benefit from rechallenge with anti-EGFR therapy in those patients without detectable resistance alterations after a period off therapy. In contrast, patients with detectable resistance alterations typically do not benefit from repeat therapy (19). This paradigm is less well established in BRAF V600E or KRAS G12C mutated CRC, although response to repeat therapy has been reported (20,21). Given the frequency and heterogeneity of KRAS and NRAS mutations as mechanisms of resistance, the strategy of using novel pan-RAS inhibitors to suppress or overcome acquired resistance is promising. There are several pan-RAS inhibitor molecules currently in clinical trials and results are eagerly awaited (22).

It is also notable that a significant fraction of all patients (47%) had no apparent acquired genomic resistance detected, emphasizing that alternative, non-genomic, mechanisms of resistance remain critical. These include plasticity of tumor cells such as epithelial-mesenchymal transition (EMT), epigenetic changes which can alter gene expression without changing the underlying DNA sequence as well as protective changes in the tumor microenvironment (23). Recent preclinical studies have shown that cell states may also influence response to targeted therapy. Mesenchymal and basal-like cell states in KPC and PDX models displayed increased response to KRAS inhibition compared to the classical state (24). These non-genomic mechanisms need to be explored further and treatment strategies targeting these should be incorporated along with genomically targeted treatments.

This study has several limitations, including its retrospective design and modest sample size. Certain targeted therapy subgroups (e.g., HER2 and KRAS directed therapies) were underrepresented, limiting cohort-specific conclusions. Biliary tract cancers were overrepresented, reflecting institutional clinical trial availability and practice patterns during the study period. In addition, ctDNA testing was performed more frequently in later years as targeted therapy trials expanded, and the cohort was assembled prior to widespread clinical adoption of certain recently approved agents (e.g., KRAS G12C inhibitors in CRC). These factors may preferentially capture patients treated at tertiary centers with access to clinical trials; however, as the primary objective was molecular resistance characterization rather than prognostic assessment, we believe these biases do not materially alter the mechanistic conclusions. Given the exciting advancements in RAS inhibition and the need to identify mechanisms of resistance in this now ‘druggable’ RAS space, this area requires further exploration. Exclusive use of ctDNA in our cohort to detect resistance could also miss certain molecular mechanisms of resistance such as gene fusions, which have higher sensitivity on tissue biopsies. Lastly, we focused on genomic resistance mechanisms which can be detected via ctDNA and therefore our work lacks the ability to look at non-genomic mechanisms of resistance which form a prominent portion of these resistance mechanisms.

Conclusions

In this real-world cohort of gastrointestinal cancers treated with targeted therapy, ctDNA profiling revealed diverse and target-dependent resistance mechanisms, underscoring the biologic heterogeneity of acquired resistance. Distinct patterns emerged across tumor types, with on-target alterations predominating in some contexts and bypass pathway reactivation in others, highlighting the need for mechanism-specific therapeutic strategies. Notably, the absence of detectable genomic resistance in a substantial subset of patients emphasizes the importance of investigating non-genomic and adaptive mechanisms. Collectively, these findings support the clinical utility of liquid biopsy for resistance characterization and inform future strategies aimed at preventing or overcoming acquired resistance in GI malignancies.

Acknowledgments

We thank Sean Bannon and Chloe Simmons for assistance with data collection. A previous version of this work has been presented at the European Society for Medical Oncology Gastrointestinal Cancers Annual Congress, Munich, Germany, June 2024.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1025/rc

Data Sharing Statement: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1025/dss

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1025/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-1-1025/coif). B.C. has served on an advisory board for Guardant Health. M.S. has served in a consultant/advisory role for Bristol Myers Squibb, Astellas, Bayer, Sedgwick Claims Management, Encapsulate Bio, SOPHIE, PRIME Education and Talem Health. A.P. has served as a consultant/advisory board member for Checkmate, Eli Lilly, Pfizer, Inivata, Natera, and Biofidelity; serves as a SAB member and holds equity in C2I Genomics; and has received research funding from Puretech, PMV Pharmaceuticals, Plexxikon, Takeda, BMS, Novartis, Genentech, Daiichi Sankyo, and Mirati. S.K. has received consulting fees/served on advisory boards for Astellas, Merck, Gilead, BeiGene, I-Mab, Elevation Oncology, AstraZeneca, Daiichi Sankyo, Amgen, Eisai, Taiho Oncology, BMS, Gilead, Signet Therapeutics, Boehringer-Ingelheim, and EsoBiotec; honoraria from Merck; served on DSMB for Sanofi-Aventis; and holds equity in MBrace Therapeutics. R.C. has served as a consultant for Abbvie, Array Biopharma/Pfizer, Asana Biosciences, Astex Pharmaceuticals, Avidity Biosciences, BMS, C4 Therapeutics, Cogent Biosciences, Elicio, Erasca, FOG Pharma, Guardant Health, Ipsen, Kinnate Biopharma, Mirati Therapeutics, Navire, Nested Therapeutics, N-of-one/Qiagen, Novartis, nRichDx, Remix Therapeutics, Revolution Medicines, Roivant, Syndax, Taiho, Tango Therapeutics, Zikani Therapeutics; received research funding from Invitae, Lilly, Novartis, OnKure, Pfizer, and Relay Therapeutics; is a cofounder/equity holder/scientific advisory board member of Alterome Therapeutics and Sidewinder Therapeutics; and serves as a scientific advisory board member and equity holder in Avidity Biosciences, C4 Therapeutics, Cogent Biosciences, Erasca, Interline Therapeutics, Kinnate Biopharma, Nested Therapeutics, nRichDx, Remix Therapeutics, and Revolution Medicines. The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Dana Farber/Harvard Cancer Center institutional review board (#14-046, 13-416, 18-380, 02-240). All included patients provided written informed consent for this prospective analysis; meanwhile, informed consent was waived for the retrospective review.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat Med 2004;10:789-99. [Crossref] [PubMed]
2. Sanchez-Vega F, Mina M, Armenia J, et al. Oncogenic Signaling Pathways in The Cancer Genome Atlas. Cell 2018;173:321-337.e10. [Crossref] [PubMed]
3. Bahar ME, Kim HJ, Kim DR. Targeting the RAS/RAF/MAPK pathway for cancer therapy: from mechanism to clinical studies. Signal Transduct Target Ther 2023;8:455. [Crossref] [PubMed]
4. Chen C, Di Bartolomeo M, Corallo S, et al. Overcoming Resistance to Targeted Therapies in Gastrointestinal Cancers: Progress to Date and Progress to Come. Am Soc Clin Oncol Educ Book 2020;40:161-73. [Crossref] [PubMed]
5. Goyal L, Saha SK, Liu LY, et al. Polyclonal Secondary FGFR2 Mutations Drive Acquired Resistance to FGFR Inhibition in Patients with FGFR2 Fusion-Positive Cholangiocarcinoma. Cancer Discov 2017;7:252-63. [Crossref] [PubMed]
6. Arena S, Bellosillo B, Siravegna G, et al. Emergence of Multiple EGFR Extracellular Mutations during Cetuximab Treatment in Colorectal Cancer. Clin Cancer Res 2015;21:2157-66. [Crossref] [PubMed]
7. Misale S, Yaeger R, Hobor S, et al. Emergence of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal cancer. Nature 2012;486:532-6. [Crossref] [PubMed]
8. Martin V, Landi L, Molinari F, et al. HER2 gene copy number status may influence clinical efficacy to anti-EGFR monoclonal antibodies in metastatic colorectal cancer patients. Br J Cancer 2013;108:668-75. [Crossref] [PubMed]
9. Parikh AR, Leshchiner I, Elagina L, et al. Liquid versus tissue biopsy for detecting acquired resistance and tumor heterogeneity in gastrointestinal cancers. Nat Med 2019;25:1415-21. [Crossref] [PubMed]
10. Crowdis J, He MX, Reardon B, et al. CoMut: visualizing integrated molecular information with comutation plots. Bioinformatics 2020;36:4348-9. [Crossref] [PubMed]
11. Diaz LA Jr, Williams RT, Wu J, et al. The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature 2012;486:537-40. [Crossref] [PubMed]
12. Bardelli A, Corso S, Bertotti A, et al. Amplification of the MET receptor drives resistance to anti-EGFR therapies in colorectal cancer. Cancer Discov 2013;3:658-73. [Crossref] [PubMed]
13. Pietrantonio F, Vernieri C, Siravegna G, et al. Heterogeneity of Acquired Resistance to Anti-EGFR Monoclonal Antibodies in Patients with Metastatic Colorectal Cancer. Clin Cancer Res 2017;23:2414-22. [Crossref] [PubMed]
14. Ahronian LG, Sennott EM, Van Allen EM, et al. Clinical Acquired Resistance to RAF Inhibitor Combinations in BRAF-Mutant Colorectal Cancer through MAPK Pathway Alterations. Cancer Discov 2015;5:358-67. [Crossref] [PubMed]
15. Corcoran RB, André T, Atreya CE, et al. Combined BRAF, EGFR, and MEK Inhibition in Patients with BRAFV600E-Mutant Colorectal Cancer. Cancer Discov 2018;8:428-43. [Crossref] [PubMed]
16. Kopetz S, Murphy DA, Pu J, et al. Molecular profiling of BRAF-V600E-mutant metastatic colorectal cancer in the phase 3 BEACON CRC trial. Nat Med 2024;30:3261-71. [Crossref] [PubMed]
17. Goyal L, Shi L, Liu LY, et al. TAS-120 Overcomes Resistance to ATP-Competitive FGFR Inhibitors in Patients with FGFR2 Fusion-Positive Intrahepatic Cholangiocarcinoma. Cancer Discov 2019;9:1064-79. [Crossref] [PubMed]
18. Subbiah V, Sahai V, Maglic D, et al. RLY-4008, the First Highly Selective FGFR2 Inhibitor with Activity across FGFR2 Alterations and Resistance Mutations. Cancer Discov 2023;13:2012-31. [Crossref] [PubMed]
19. Sartore-Bianchi A, Pietrantonio F, Lonardi S, et al. Circulating tumor DNA to guide rechallenge with panitumumab in metastatic colorectal cancer: the phase 2 CHRONOS trial. Nat Med 2022;28:1612-8. [Crossref] [PubMed]
20. Ji J, Wang C, Fakih M. Rechallenge With BRAF and anti-EGFR Inhibitors in Patients With Metastatic Colorectal Cancer Harboring BRAF(V600E) Mutation Who Progressed on Cetuximab and Encorafenib With or Without Binimetinib: A Case Series. Clin Colorectal Cancer 2022;21:267-71. [Crossref] [PubMed]
21. Morris L, Pappas L, Pandey A, et al. Anti-KRAS G12C and Anti-EGFR Rechallenge in Chemotherapy-Refractory KRAS G12C-Mutated Colorectal Cancer: A Case Report. JCO Precis Oncol 2024;8:e2400547. [Crossref] [PubMed]
22. Caughey BA, Strickler JH. Targeting KRAS-Mutated Gastrointestinal Malignancies with Small-Molecule Inhibitors: A New Generation of Breakthrough Therapies. Drugs 2024;84:27-44. [Crossref] [PubMed]
23. Bell CC, Gilan O. Principles and mechanisms of non-genetic resistance in cancer. Br J Cancer 2020;122:465-72. [Crossref] [PubMed]
24. Dilly J, Hoffman MT, Abbassi L, et al. Mechanisms of Resistance to Oncogenic KRAS Inhibition in Pancreatic Cancer. Cancer Discov 2024;14:2135-61. [Crossref] [PubMed]