Impact of Tissue-Agnostic Approvals for Patients with Gastrointestinal Malignancies

Abstract

Gastrointestinal malignancies encompass a broad range of tumors with limited treatment options, particularly for advanced disease. With the development and implementation of next-generation sequencing in routine practice, molecular-targeting therapies have been increasingly incorporated into the treatment paradigm for various cancers. Several drugs have achieved tissue-agnostic regulatory approvals, which offer promising biomarker-driven therapy options for patients with advanced gastrointestinal malignancies. In this review, we focus on the clinical evidence of recent drug approvals for neurotrophic tyrosine receptor kinase (NTRK) fusion, microsatellite instability-high (MSI-H) phenotype, tumor mutation burden-high (TMB-H), BRAF V600E and rearranged during transfection (RET), in the context of gastrointestinal malignancies. We also highlight the future landscape of tissue-agnostic targets such as Her2/neu, fibroblast growth factor receptor (FGFR), and neuregulin-1 (NRG1).

The era of tissue agnostic approvals

Gastrointestinal malignancies encompass a broad range of tumors arising from organs along the digestive tract, including the pancreas and hepatobiliary system. Despite numerous advances in the treatment of advanced and metastatic gastrointestinal malignancies, the disease is often incurable, with 5-year survival rates falling well below 20% across various sites^i,ii. Conventional treatment approaches for gastrointestinal tumors usually involve cytotoxic chemotherapy, which tumors eventually acquire resistance to through various mechanisms, thus, investigation into alternative treatment approaches such as those targeted to specific molecular characteristics of the tumor has been of great interest.

The development of next-generation sequencing (NGS) technologies has facilitated the identification of driver mutations, shifting the treatment paradigm from a purely histology-based approach to one that incorporates precision-medicine. Genome driven precision oncology introduction of molecularly-based treatment has significantly changed the standard-of-care such as for the treatment of non-small cell lung cancer (NSCLC) for which targeted therapies are administered in the frontline for metastatic disease [1]. Specific molecular aberrations have a high concordance with tumor tissue of origin, while others have been identified across numerous solid tumors. In theory, treatments tailored towards these molecular characteristics would induce a response regardless of tumor histology, which has been the basis of tissue-agnostic drug development. Immune checkpoint inhibitor pembrolizumab was the first drug to receive approval by the US FDA for a tissue agnostic indication in 2017 for the treatment of microsatellite instability high (MSI-H) tumors and, more recently, dostarlimab-glxy received FDA approval for a similar indicationⁱⁱⁱ [2,3]. Pembrolizumab received an approval for another tissue-agnostic indication – high tumor mutational burden (TMB-H) [3,4]. Larotrectinib and entrectinib received tissue-agnostic approvals in 2018 and 2019 respectively for pan-cancers harboring NTRK fusions [5,6].

In this review, we focus on the prevalence, detection methods, and data regarding the efficacy tissue agnostic indications of drugs targeting NTRK fusions, MSI-H phenotype, TMB-H status, and BRAFV600E in gastrointestinal malignancies (Figure 1, Key figure). In addition, promising tissue agnostic targets and therapies currently in development, and their relevance to gastrointestinal tumors, are highlighted.

Figure 1. Key Figure. Current and future tissue agnostic therapies in GI malignancies.

Various drugs have been approved or are in development for tissue-agnostic approvals which are relevant to patients with gastrointestinal malignancies. Enterectenib and Larotrectinib have tissue agnostic approvals to target neurotrophic tyrosine receptor kinase (NTRK) fusions. Pembrolizumab and dostarlimab-glxy have tissue agnostic approvals to target deficient mismatch repair (dMMR)/microsatellite instability high (MSI-H) cancers while pembrolizumab also has a tissue agnostic indication for tumors with high tumor mutational burden (TMB). Selpercatinib and Pralsetinib have tissue agnostic indications for RET fusion positive cancers and Vemurafenib/Dabrafenib for cancers with BRAF V600E mutation. Alofanib, pemigatinib, and bemarituzumab have promising tissue agnostic activity in gastrointestinal tumors with fibroblast growth factor receptors (FGFR) fusions and Zenocotuzumab is a promising agent with activity in Neuregulin 1 (NRG1) fusion positive cancers. Human Epidermal Growth Factor Receptor 2 (HER2) targeted agents have several FDA approvals for gastrointestinal malignancies.

Neurotrophic tropomyosin receptor kinase (NTRK) fusions

NTRKs [notably NTRK1 (chromosome 1q23.1), NTRK2 (chromosome 9q21.33), and NTRK3 (chromosome 15q25.3)] are genes that encode tropomyosin receptor kinase (TRK) proteins that regulate neuronal development and function. Chromosomal rearrangements can result in somatic NTRK gene fusions that lead to the expression of chimeric TRK fusion proteins that promote oncogenesis [7]. Such TRK fusions can be identified in several adult cancers including lung, breast, melanoma, sarcoma, and gastrointestinal malignancies. The prevalence of kinase mutations in gastrointestinal malignancies varies based on tumor type and molecular characteristics of the tumor. NTRK fusions have been reported to be present in 0.2-0.3% of unselected colorectal cancers, with a significantly higher prevalence noted among RAS/BRAF-WT, MSI-H tumors [8-11]. Moreover, NTRK Fusions have been reported in a small number of cases of cholangiocarcinoma (3.6%) and pancreatic cancer (0.3%), however they are less prevalent or absent in other gastrointestinal malignancies such as gastric and esophageal carcinomas [12-15].

There is no universal standard test for the identification of TRK fusions, however several methods are available to detect rearrangements: fluorescence in situ hybridization (FISH), RT-PCR, RNA- or DNA-based NGS, and immunohistochemistry (IHC) [16]. FISH and RT-PCR have several disadvantages which limit their use for the detection of NTRK fusions in gastrointestinal tumors; FISH requires separate assays or multiplex assays for the three NTRK genes and is unable to identify the gene fusion partner, while RT-PCR can detect only known fusion partners and breakpoints. Thus, NGS, which evaluates somatic genetic alterations across hundreds of genes, may be the preferred methodology given its high sensitivity and ability to provide other important molecular information such as the presence of other oncogenic drivers and tumor mutation burden. RNA-sequencing has advantages over DNA-sequencing including de novo detection of gene fusion transcripts that have not been previously described, and increases the sensitivity of detection in low tumor purity samples, however, RNA samples extracted from formalin-fixed, paraffin-embedded (FFPE) have a propensity to be of poor-quality due to degradation [17]. IHC has been proposed as a screening test due to its wide availability, low-cost, and high sensitivity and specificity to detect NTRK fusions, with positive results warranting further sequencing as part of a two-step process [18].

NTRK inhibitors have been FDA approved for a tissue agnostic indication, after demonstration of clinical efficacy regardless of tumor organ of origin, histology, or fusion partner. Larotrectinib, a selective inhibitor or TRK A, B, and C, was approved by the FDA in November 2018 following initial results of three parallel Phase I/II trials with 55 enrolled patients which demonstrated an overall response rate (ORR) of 75%, with median duration of response and PFS not reached [19]. In an updated analysis, which included 218 patients with 21 different tumor types represented, the ORR was 75% and complete responses were achieved in 22% of patients with a median PFS of 35.4 months and median overall survival (OS) was not reached [20]. In a published pooled analysis of Larotrectinib across multiple tumor types, 14 patients with gastrointestinal malignancies were assessed (Table 1). Responses were seen among patients with colorectal cancer, cholangiocarcinoma, and pancreatic cancer (ORR 43%), including a 50% response rate among patients with colorectal cancer.

Table 1.

Summary of study characteristics and efficacy of FDA-approved histology-agnostic drugs, including specific information on gastrointestinal tumors when available

Biomarker	Drug	Study population	Overall efficacy	Efficacy in GI malignancies
NTRK fusions	Larotrectinib
	Drilon et al.[19]	Pediatric and adult patients with TRK-fusion positive tumors	ORR (all): 41/55 (75%) CR=7/55 (13%) PR=34/55 (62%) PD=5/55 (9%) Not evaluated =2/55 (4%)	Colon n=4; ORR 50% (2 PR, 1 SD) Cholangiocarcinoma n=2 (1 PR, 1 PD); ORR 50% Appendix n=1 (SD) Pancreatic n=1 (PR)
	Hong et al.[20]	Pediatric and adult patients with locally advanced/metastatic solid tumors with NTRK fusion (via NSG or FISH). (107 adults, 52 pediatric)	ORR (all)=121/153 (79%) ORR (adults)= 74/102 (73%) CR=24/153 (16%) PR=97/153 (63%) SD=19/153 (12%)	Colon n=8; ORR 50% (4 PR, 4 SD) Cholangiocarcinoma n=2 (1 PR, 1 PD) Pancreas n=2 (1 PR) Appendix n=1 (1 SD), Hepatocellular n=1
	Entrectinib
	Demetri et al.[21]	Adults with NTRK-fusion positive sold tumors	ORR=74/121 (61%) CR=19/121 (16%) PR=55/121 (46%) SD=13/121 (11%) PD=13/121 (11%)	Colon n=10; ORR 20% Pancreatic n=4; ORR 75% Cholangiocarcinoma n=1; ORR 100%
dMMR/MSI-H	Pembrolizumab
	Le et al.[28]	Adult patients with dMMR CRC and non-CRC solid tumors	ORR=43/86 (50%) CR=18/86 (21%) PR=28/86 (33%)	Colon n=40; ORR 53% (5 CR, 16 PR) Ampullary n=4; ORR 25% (1 CR, 1 SD, 1 PD) Cholangiocarcinoma n=4; ORR 25% (1 CR, 3 SD) Gastroesophageal n=5; ORR 60% (3 CR, 2 PD) Pancreas n=8; ORR 63% (2 CR, 3 PR, 1 SD) Small Intestine: 5; ORR 80% (2 CR, 2 PR, 1 PD)
	Le et al.[30]	Adult patients with metastatic dMMR CRC	ORR=41/124 (33%), 7 CR and 34 PR, 26 SD	All CRC patients
	Marabelle et al.[29]	Adult patients with advanced non-CRC dMMR solid tumors	ORR=80/233 (34%) CR=23/233 (10%) PR=57/233 (24%) SD=42/233 (18%)	Gastric n=24; ORR 46% (4 CR, 7 PR) Cholangiocarcinoma n=22; ORR 41% (2 CR, 7 PR) Pancreatic n=22; ORR 18% (1 CR, 3 PR) Small Intestine n=19; ORR 26% (3 CR, 2 PR) Anal n=1
	Dostarlimab-gxly
	Andre et al.[31]	Adults with dMMR advanced solid tumors	ORR=41/106 (39%)	Colon n=69; ORR=36% Small Intestine n=12; ORR=33% Pancreatic n=4; ORR 0% Liver n=2; ORR 50% (1 PR, 1 PD) Biliary n=1 (CR) Esophageal n=1 (PD) Gallbladder n=1(CR)
TMB-H	Pembrolizumab
	Marabelle et al.[40]	Adults with advanced solid tumors, progressed on one or more line of standard therapy with TMB >= 10 mut/Mb	ORR=30/102 (29%) CR=4/102 (4%) PR=26/102 (26%)	Anal Cancer n=14; ORR (7%)
	Cristescu et al.[41]	Adults enrolled in KEYNOTE studies across multiple advanced solid tumors with TMB >= 175 mut/exome	ORR: 136/433 (31%)	Gastric Cancer n=53; ORR 34%
BRAF V600	Dabrafenib and Trametinib
	Salama et al.[51]	BRAF V600E mutated refractory/advanced malignancies	ORR=11/29 (38%) CR=0/29 SD=11/29 PD=2/29	Cholangiocarcinoma n=4; ORR 75% Pancreatic n=1; SD
	Vemurafenib
	Subbiah et al.[46]	BRAF V600 mutated advanced/refractory malignancies	ORR=56/172 (33%) CR=5 PR=51 SD=65 PD=35	Colon n=10; ORR 0% Cholangiocarcinoma n=9; ORR 50% Pancreatic n=1; ORR 0% Esophageal n=1; ORR 0%
	Dabrafenib and Trametinib, FDA approval label dataix	BRAF V600 mutated advanced/refractory malignancies	ORR=54/131 (41%)	Biliary tract n=48; ORR 46% Small bowel adenocarcinoma n=4, ORR 50% Pancreatic adenocarcinoma n=3, ORR 0% Neuroendocrine carcinoma of colon n=2; ORR 0% Adenocarcinoma of anus n=1; ORR 0%
HER2	Trastuzumab/Pertuzumab
	Meric-Bernstam et al.[60]	Previously treated HER2-amplified metastatic colorectal cancer	ORR=18/57 (32%) CR=1 PR=17	Colorectal cancer n=57, ORR 32%
	Javle et al.[62]	Previously treated HER2-positive advanced biliary tract cancer	ORR=9/39 (23%) PR=9 PD=10	Biliary tract n=39, ORR 23%
	Fam-trastuzumab Deruxtecan
	Shitara et al.[58]	Previously treated HER2-positive advanced gastric cancer	ORR=61/119 (51%) CR=11 PR=50 SD=42 PD=14	Gastric n=119, ORR 61%
	Siena et al.[61]	Metastatic HER-2 colorectal cancer progressed on 2 prior lines of therapy	ORR=24/53 (45%) CR=1 PR=23 SD=20 PD=5	Colorectal adenocarcinoma n=53, ORR=45%
	Ohba et al.[63]	Previously treated HER-2 positive biliary tract cancer	ORR= 8/22 (36%)	Biliary tract cancer n=22, ORR=36%
RET Fusion	Selpercatinib
	Subbiah et al.[69]	Patients with advanced RET fusion positive non-NSCLC/Thyroid solid tumors	ORR=18/41 (47%)	Pancreatic n=11; ORR 55% Colon n=10; ORR=20% Small Intestine n=1; ORR=100% Cholangiocarcinoma=1; ORR=100%
	Pralsetinib
	Subbiah et al.[70]	Patients with advanced RET fusion positive non-NSCLC/Thyroid solid tumors	ORR=13/23 (57%)	Pancreatic n=4; ORR 100% Colon n=2; ORR 0% (2 SD) Cholangiocarcinoma m=3; ORR 67% Gastric n=1 (PD)

Entrectinib, an orally bioavailable tyrosine-kinase inhibitor (TKI) targeting TRKA, TRKB, TRKC, as well as c-Ros oncogene 1 (ROS1) and anaplastic lymphoma kinase (ALK) was assessed in an integrated analysis of three Phase I/II clinical trials including 121 patients (16 with gastrointestinal malignancies) and demonstrated similar efficacy; ORR was 61% including 16% with complete response with a median duration of response of 20 months [21]. As with Larotrectinib, there were responses noted among patients with colorectal cancer, pancreatic cancer, and cholangiocarcinoma (Table 1). Moreover, both NTRK inhibitors larotrectinib and entrectinib were noted to be relatively safe, with only 2% and 6.5% of patients discontinuing larotrectinib and entrectinib respectively due to treatment-related adverse events.

With the relatively high response rates and favorable toxicity profile observed with NTRK inhibition compared with standard lines of therapy for refractory and metastatic gastrointestinal tumors, NTRK inhibitors appear to be a promising option for patients with tumors harboring NTRK fusions and warrant integration into the diagnostic molecular panel for colorectal cancer, pancreatic cancers, and cholangiocarcinoma; with increased availability and adoption of NGS, NTRK fusions can be readily identified in order to select candidates for targeted therapy, and should be considered for use early in the refractory setting.

MSI-H/dMMR

Microsatellites are short tracts of DNA in which there are multiple repeats (on the order of 5-50) of 1 to 5 base pair motifs. The mismatch repair (MMR) system is one of several cellular DNA repair mechanisms that functions to maintain the count and integrity of microsatellites during cell division. Deficiencies in the MMR (dMMR) system lead to cells being unable to regulate the length of their microsatellites during cell division, a term called microsatellite instability (MSI-H). This leads to hypermutability and disease states such as cancers. There are multiple methodologies to evaluate for dMMR/MSI-H which are readily available; IHC for the MMR proteins MLH1, MSH2, MSH6, and PMS2 is sensitive and specific for identify this alteration in addition to PCR tests for the assessment of microsatellite alterations [22]. Both methodologies have been successfully implemented in clinical trials and in routine clinical practice. NGS methods to assess for the presence of microsatellite instability have been developed and have been found to have high concordance with IHC and PCR-based testing [23].

Lynch Syndrome is characterized by germline defects in mismatch repair proteins and is associated with a significantly increased lifetime risk for the development of colorectal cancer, accounting for approximately 3% of colorectal cancer diagnoses [24]. However, the majority of dMMR/MSI-H colorectal cancer arise from somatic changes in one of the MMR genes; the phenotype is present in 15-20% of early-stage colorectal cancers and 3.5-6.5% of metastatic cases [25]. dMMR/MSI-H is present in a significant proportion of other gastrointestinal malignancies including gastric adenocarcinoma (19%) and rectal adenocarcinoma (5.7%) [26]. A small percentage of esophageal cancers (1.6%) and cholangiocarcinomas (1.3%) are dMRR/MSI-H, however this alteration is rare in pancreatic adenocarcinoma (1-2%).

It has been hypothesized that the increased somatic mutation burden seen in dMMR/MSI-H tumors result in the production of immunogenic antigens, which enhance tumor response to immune checkpoint inhibition. The landmark KEYNOTE-016 study by Le et al. demonstrated a 53% response rate (21% with complete response) among patients with various dMMR/MSI-H solid tumors receiving pembrolizumab [27,28]. Moreover, the responses were durable, as PFS and OS were not reached at follow-up, and responses were seen among patients with various gastrointestinal malignancies including colorectal cancers, cholangiocarcinoma, gastroesophageal, pancreatic, and small bowel cancers (Table 1). Compelling results from this proof-of-concept trial spurred additional investigation into the efficacy of anti-PD1 therapy in dMMR solid tumors and pooled results from five clinical trials enrolling 149 patients (90 with CRC and 49 with other tumors) resulted in FDA approval for a tissue agnostic indication in May 2017 [3]. Among the included trials, KEYNOTE-158 enrolled 233 patients across 27 tumor types (including gastrointestinal malignancies such as gastric, cholangiocarcinoma, pancreatic, and small bowel cancers) and noted an ORR of 34% with pembrolizumab and median duration of response had not been reached [29]. KEYNOTE-164 assessed pembrolizumab in patients with metastatic dMMR/MSI-H CRC refractory to multiple lines of chemotherapy and noted an ORR of 33%, and a duration of response greater than or equal to 12 months was seen in 95% of these patients [30]. Another anti-PD-1 antibody dostarlimab-glxy was assessed in patients with dMMR tumors, majority of which were gastrointestinal in origin, and ORR was 39% with 91% maintaining response at 12 months [31]. With these findings, the FDA granted approval to dostarlimab-glxy for tumor agnostic indication in August 2021 [2].

The tissue agnostic approvals of pembrolizumab and dostarlimab-glxy have had a meaningful impact on the treatment landscape for gastrointestinal tumors, as these treatments are effective and dMMR/MSI-H gastrointestinal tumors are prevalent. Colorectal cancers paved the way for adoption of anti-PD-1 therapy, and it is now considered the first line treatment option for patients with dMMR/MSI-H metastatic disease^iv. Moreover, anti-PD-1 therapy is also considered a first line option for metastatic dMMR/MSI-H small bowel adenocarcinomas, and as a consideration for first line in patients with pancreatic cancer and poor performance status but is reserved as a subsequent treatment option for patients with gastric cancer, and hepatobiliary cancers^{v,vi,vii,viii}. Compared with colorectal and gastric adenocarcinomas, it is notable that the response rates among patients with dMMR/MSI-H pancreatic cancer was low in the pivotal KEYNOTE studies, which suggest that these tumors may be more resistant to immune checkpoint blockade. Nevertheless, routine testing for mismatch repair status should be implemented for most patients with advanced gastrointestinal tumors to identify candidates for anti-PD-1 therapy.

Tumor mutational burden (TMB)

TMB is a quantification of the total number of somatic mutations per megabase (mut/Mb) of the genome examined. Early data suggested that, across different tumor types, tumors with a high mutational burden (TMB-H) were more responsive to immune checkpoint inhibition (ICI) [32,33]. It has been hypothesized that TMB-H tumors generate more neoantigens and thus increase T-cell reactivity, enhancing tumor response to anti-PD-(L)1 and anti-CTLA-4 therapy. While the gold-standard for assessment of TMB is whole exome sequencing (WES) by NGS, targeted gene panels such as the FoundationOne CDx have been developed due to lower cost and complexity. Nevertheless, there is considerable variability among commercial NGS panels with respect to estimation of TMB [34]. The Friends of Cancer Research TMB Harmonization Project has attempted to address such variability by developing a tool using The Cancer Genome Atlas data to promote reproducibility and comparability across TMB assays [35].

While the mutation load of gastrointestinal malignancies are typically lower than typically seen in melanomas and lung cancers, there are a relatively high prevalence of TMB-H tumors among several gastrointestinal tumor pathologies [36,37]. Sale et al. evaluated 4,125 tumors across 14 gastrointestinal cancer sites and identified TMB-H (when defined as ≥17 mut/Mb) to be present among 14.6% of right-sided colon cancers and 10.2% of small bowel adenocarcinomas, followed by 8.3% of anal cancers and 8.3% of gastric cancers. Pancreatic adenocarcinoma and esophageal adenocarcinomas were among the least likely to harbor a TMB-H phenotype, representing 1.4% and 1.9% of evaluated samples respectively [38]. Notably, there was a high correlation between microsatellite instability and TMB among gastrointestinal malignancies, with the exception of anal cancer, in which all cases were microsatellite stable, and some gastric cancers as well; this is contrary to what has been observed in other cancer types such as melanoma and lung cancer where TMB-high tumors are prevalent, but MSI-H is not [37]. POLE/POLD mutations have also been frequently associated with high TMB [39].

A preplanned retrospective analysis of KEYNOTE-158 was performed to evaluate the role of TMB as a predictive biomarker for immunotherapy across multiple solid tumors [40]. Among 790 patients who received pembrolizumab and had sufficient tissue for TMB analysis, 102 (13%) were identified as having TMB-H. The ORR was 29% for TMB ≥ 10 mut/Mb compared to 6% for the non-TMB-H group and was maintained even when MSI-H tumors were excluded. Moreover, median duration of response was not reached. Based on higher response rates to pembrolizumab among TMB-H tumors, the FDA expanded the approval of pembrolizumab in June 2020 to include unresectable or metastatic tumors with TMB-H (≥10 mut/Mb) that have progressed following prior treatment. KEYNOTE-158 included 14 patients with TMB-H anal cancer, among which 1 response was noted; notably, this was lower than the response rate among non-TMB-H patients. Data from the pan-cancer cohort of predominantly microsatellite stable (MSS) solid tumors (KEYNOTE-158) found an ORR of 27.1% for patients with MSS tumors and TMB ≥ 10 mut/Mb. In a retrospective analysis of KEYNOTE studies across 24 tumor types, WES was performed to determine TMB on DNA extracted from pretreatment tumor samples and matched normal DNA [41]. TMB-H (≥175 mut/exome) tumors were associated with improved ORR and survival with immune checkpoint inhibition; ORR among 53 patients with gastric cancer was 34% in the TMB-H group.

Seminal trials have described the efficacy of anti-PD-(L)1 therapy for patients with MSI-H tumors, most of which are presumably TMB-H, originating from the gastrointestinal system, however, there are comparatively fewer reports describing the response of MSS, TMB-H gastrointestinal malignancies to immunotherapy. In a study of MSS cancers of 14 different histologies, treatment with an ICI produced longer median progression-free survival (PFS) (26.8 vs 4.3 mo., p=0.0173) in TMB-high (≥20 mut/Mb) vs TMB-low tumors [42]. In a study of 54 advanced gastric cancers, median TMB for MSS cases was 6.6 mut/Mb and there was one MSI-H tumor. Response to anti-PD1 antibody toripalimab revealed that patients with TMB-high (≥ 12 mut/Mb) vs. low tumors had better ORR (33.3% vs. 7.1%, P=0.017), similar PFS (2.5 vs. 1.9 months, P=0.055), but significantly improved OS (14.6 vs 4.0 months, P=0.038) [43]. MSI-H tumors represent the majority of TMB-H colorectal cancer, however nearly 3% of MSS colorectal cancers are noted to be TMB-H [44].

In the context of gastrointestinal malignancies, many patients with TMB-H tumors would qualify for anti-PD-(L)1 therapy on the basis of MSI-H status, however, there are subsets of MSS stable gastrointestinal tumors such as esophageal cancer, anal cancer, and some colorectal cancers which display a high tumor mutation burden in the context of microsatellite stability for which the expanded tumor agnostic indication for immune checkpoint inhibition provides an additional treatment option [45]. While TMB-H is a promising indication for ICI therapy, there are many questions that remain unanswered including optimal TMB cut-offs for each tumor type, and evaluation of clinical efficacy in MSS gastrointestinal tumors.

BRAF V600

Beyond melanoma, BRAF V600E alterations are seen in many gastrointestinal cancers such as colorectal cancer, biliary tract cancer, appendiceal cancer and pancreatic cancer [46-48]. Although monotherapy inhibition with BRAF inhibitors did not show any activity in colorectal cancer due to the innate/adaptive resistance mechanism through EGFR pathway, it has been shown that BRAF V600E is an actionable gene in colorectal cancer in combination with EGFR inhibition [49]. Beyond colorectal cancer activity of BRAF with and without MEK inhibitors have shown activity in multiple gastrointestinal cancers like biliary tract, and pancreatic cancers [50,51].

Recently, the US FDA granted accelerated approval for dabrafenib and trametinib for the treatment of adult and pediatric patients 6 years of age and older with unresectable or metastatic solid tumors with BRAF V600E mutation who have progressed following prior treatment and have no satisfactory alternative treatment options. Among the 131 patients enrolled in BRF117019 and NCI-MATCH several GI tumor types were enrolled including cholangiocarcinoma, small intestine adenocarcinoma, pancreatic adenocarcinoma, neuroendocrine tumors, mucinouspapillary serous adenocarcinoma of peritoneum, and adenocarcinoma of anus^ix.

Human epidermal growth factor receptor (HER2)

A member of the HER protein family, HER2 is a transmembrane growth factor receptor that is amplified or overexpressed in a significant proportion of malignancies, leading to enhanced heterodimerization of the receptor with other members of the EGFR family and the propagation of signaling pathways that promote cellular proliferation [52]. HER2 expression and amplification is typically identified by using a combination of IHC and FISH; HER2 3+ staining on IHC indicates HER2 positivity while HER2 2+ staining requires further testing to evaluate ratio of HER2 gene amplification [53]. HER2-targeted therapies have significantly impacted the treatment landscape for advanced HER-2 positive breast cancer, with indications for trastuzumab plus pertuzumab, famtrastuzumab deruxtecan, adox-trastuzumab emtansine (T-DM1), lapatinib, tucatinib, and margetuximab [54]. Nevertheless, a significant proportion of malignancies beyond breast cancer harbor HER2 overexpression/amplification including gastric (20%), biliary tract (10-20%), and colorectal cancers (5%) which appear to confer benefit from these agents [55].

Beyond breast cancer, HER2-targeted therapies were first integrated into the treatment paradigm for gastric cancer after a clear benefit was demonstrated with the addition of trastuzumab to chemotherapy in advanced gastric and gastroesophageal cancers [56]. While there have been multiple negative trials in the second line for advanced gastric cancer, the development of HER-2 directed antibody drug conjugate (ADC) famtrastuzumab deruxtecan has expanded the armamentarium for recurrent gastric cancer [57,58]. In the phase II DESTINY-Gastric01, fam-trastuzumab deruxtecan was associated with a superior ORR compared to chemotherapy (515 vs 14%) and median OS (12.5 months vs 8.4 months). HER2-directed TKI tucatinib is also currently under investigation in combination with chemotherapy [59]. Similar inroads have been made in advanced/metastatic HER2 overexpressed/amplified colorectal cancer and cholangiocarcinoma, with promising activity demonstrated by the combination of trastuzumab/pertuzumab and fam-trastuzumab deruxtecan in heavily pretreated patients [60-63]. While with tumor-site specific nuances have limited the role of some HER2-directed therapies in gastrointestinal malignancies, fam-trastuzumab deruxtecan appears to have the most potent tissue-agnostic activity thus far. As additional HER2-targeted treatment successes in breast cancer are explored in gastrointestinal malignancies, the landscape will continue to evolve.

RET fusions

RET is a proto-oncogene that encodes a transmembrane tyrosine kinase receptor which is involved in renal and central nervous system development during embryogenesis [64]. Chromosomal rearrangements that generate fusion genes containing the kinase domain of RET and various upstream gene partners lead to abnormal expression and oligomerization of chimeric kinase fusion proteins, resulting in constitutively active singling and oncogenesis. RET fusions have been implicated in the pathogenesis of several malignancies including papillary thyroid cancer, and non-small cell lung cancer among others [65,66]. High response rates to RET inhibitor selpercatinib among RET-fusion positive advanced/metastatic thyroid and NSCLC (69% and 64% respectively) prompted evaluation in other solid tumors in the LIBRETTO-001 trial [67-69]. Among the 41 enrolled patients, 23 (56%) had gastrointestinal tumors and ORR was 44% and median duration of response was 18.4 months (Table 2). The RET inhibitor pralsetinib has also demonstrated high response rates among patients with RET fusion-positive tumors other than thyroid and NSCLC, with a response rate of 53% among 19 patients and median duration of response of 19 months. Notably all 3 patients with pancreatic cancer and 2 out of 3 patients with cholangiocarcinoma experienced a response [70]. Though present in a relatively small number of gastrointestinal tumors (0-2%), [71] the efficacy in RET inhibition in RET fusion-positive cancers across multiple histologies led to the recent tissue agnostic FDA approval of RET inhibitor selpercatinib.

Table 2.

Summary of study characteristics and efficacy of emerging histology-agnostic drugs, including specific information on gastrointestinal tumors when available

Biomarker	Drug	Study population	Overall efficacy	Efficacy in GI tumors
NRG1 fusions	Zenocutuzumab
	Schram et al.[76]	Patients with advanced NRG1+ pancreatic cancer, non-small cell lung cancer (NSCLC), and other solid tumors	ORR=9/33 (27%)	Pancreatic n=10; ORR 40%
FGFR	Alofanib
	Tsimafeyeu et al.[82]	Patients with advanced/metastatic gastric cancer	ORR=1/12 (8%)	Gastric cancer n=12; ORR=8%
	Pemigatinib
	Abou Alfa et al.[78]	Patients with advanced/metastatic FGF/FGFR alternated cholangiocarcinoma	ORR=38/107 (36%)	Cholangiocarcinoma n=107; ORR=36%
	Subbiah et al.[80]	Refractory advanced malignancies with and without FGF/FGFR alterations	ORR=12/128 (9%)	Cholangiocarcinoma n= 21; ORR 24% Pancreatic Cancer n=4; ORR 25% Colon n=10; ORR 0% Esophageal n=7; ORR 0%
	Futibatinib
	Meric et al.[81]	Refractory advanced malignancies with FGF/FGFR alterations	ORR=27/197 (14%)	Cholangiocarcinoma n=64; ORR 16% Gastric n=9; ORR 22% Colorectal Cancer n=5 Esophageal n=3 Gallbladder n=3
	Bemarituzumab
	Catenacci et al.[83]	Patients with locally advanced or metastatic non-HER2 positive gastric cancer	Median OS in bema+FOLFOX arm vs placebo+FOLFOX arm: 19.2 mo vs 13.5 mo	Gastric n=77

Emerging targets

Neuregulin-1 fusions

Neuregulins (NRGs) are a family of structurally related cellular growth factors that are primarily involved in the development of the nervous and cardiovascular system. Neuregulin-1 (NRG1) fusions are postulated to promote oncogenesis via constitutive activation of HER3 [72]. NRG1 fusions have been identified in variety of solid tumors, including cholangiocarcinoma (0.5%) pancreatic (0.5%), and colorectal cancers (0.1%) [73]. The identification of this driver mutation lead to the development of therapeutic strategies such as pan-HER tyrosine kinase inhibition with afatinib, which has demonstrated some success in treating solid tumors with NRG1 fusions [74,75]. Moreover, bispecific anti-HER2/HER3 monoclonal antibody Zenocutuzumab, has been under development with high response rates among patients with a variety of advanced, treatment-refractory solid tumors [76]. Among 33 evaluable patients, 8 patients (27%) had a response; notably 4 out of 10 patients with pancreatic cancer (40%) achieved a response and disease control was seen in 9 (90%) of the patients. Treatment was welltolerated with fewer than 5% of patients experiencing grade 3 adverse events. Pending further clinical efficacy and safety data, Zenocutuzumab is a promising treatment option for patients with gastrointestinal tumors harboring NRG1 fusions.

Fibroblast growth factor receptors

The FGFRs, expressed on cell membranes are involved in cell development, embryogenesis, angiogenesis, and tissue regeneration. The family contains 4 isoforms, which can be constitutively activated via genetic aberrations, promoting oncogenesis. The most common alterations include amplification, activating mutations, and rearrangements. FGFR alterations are described in a variety of solid tumors, including gastrointestinal malignancies such as cholangiocarcinoma (7%), gastric adenocarcinoma (6.7%), pancreatic cancer (4.7%), and colorectal cancer (4.4%) [77]. Given their role in promoting tumor growth, inhibition of FGFR has been evaluated in gastrointestinal malignancies. Clinical responses to FGFR inhibition in metastatic cholangiocarcinomas harboring a FGFR2 fusion/rearrangement has led to FDA approvals of pemigatinib and infigratinib [78,79]. Beyond cholangiocarcinomas, FGFR alterations are seen in multiple GI tumor types including gastric and pancreatic cancer and early signals of activity have been reported in phase 1 studies [80,81]. Efficacy and safety of FGFR inhibition in gastric cancer is being prospectively investigated with promising early results [82,83].

Concluding remarks

While gastrointestinal malignancies span a variety of tissue environments with unique biological characteristics, they often present at advanced stages and eventually develop resistance to conventional treatments, promoting the need for more effective therapies. With the introduction of clinical NGS, potentially actionable pathogenic mutations are identified with increasing frequency. There has been significant progress, which has resulted in FDA drug approvals for tissue-agnostic indications which are of particular relevance for gastrointestinal malignancies. Still, questions remain regarding whether some tissue-agnostic approvals can be appropriately applied to patients with gastrointestinal malignancies, given the relatively small number of patients with various gastrointestinal malignancies included in published studies (see Outstanding questions). There are promising new targets present across a variety of solid tumors, including gastrointestinal tumors, for which drugs are being developed in a tissue-independent fashion, and provide hope for the near future. In recognition of this progress, it is becoming increasingly important for patients with gastrointestinal malignancies to undergo molecular profiling of their tumors to have available to them the greatest number of treatment options.

Outstanding Questions.

• What is the best strategy to select patients with gastrointestinal malignancies to assess for NTRK, RET, and NRG1 fusion, given low prevalence among some histologies?

• What mechanisms are responsible for treatment resistance to immunotherapy among MSI-H/dMMR pancreatic adenocarcinoma relative to other gastrointestinal cancers?

• What is the optimal TMB cut-off for each of the gastrointestinal tumor histologies?

• What is the relevance of TMB in MSS gastrointestinal tumors?

Highlights.

• Tissue-agnostic approvals for tumors with NTRK fusion, MSI-H/dMMR phenotype, TMB, BRAF V600E and RET fusion provide treatment options for a small but significant subset of patients with gastrointestinal tumors

• Most of the clinical trials which resulted in tissue-agnostic approvals included patients with various gastrointestinal malignancies, providing real-world evidence for efficacy in this population

• Inhibitors designed to target gene fusion driven tumors have demonstrated remarkable response rates in patients with pancreatic cancer, for which few effective therapy options exist