Fusion Challenges in Solid Tumors: Shaping the Landscape of Cancer Care in Precision Medicine

Abstract

Targeting actionable fusions has emerged as a promising approach to cancer treatment. Next-generation sequencing (NGS)–based techniques have unveiled the landscape of actionable fusions in cancer. However, these approaches remain insufficient to provide optimal treatment options for patients with cancer. This article provides a comprehensive overview of the actionability and clinical development of targeted agents aimed at driver fusions. It also highlights the challenges associated with fusion testing, including the evaluation of patients with cancer who could potentially benefit from testing and devising an effective strategy. The implementation of DNA NGS for all tumor types, combined with RNA sequencing, has the potential to maximize detection while considering cost effectiveness. Herein, we also present a fusion testing strategy aimed at improving outcomes in patients with cancer.

Fusion detection in cancer is pivotal for precise diagnosis and tailored treatment, warranting their integration into clinical decision-making.

INTRODUCTION

The landscape of cancer care is increasingly allowing oncologists to make treatment recommendations on the basis of genomic drivers of tumorigenesis, thereby improving outcomes in a subset of patients. Regulatory approvals in precision medicine, pharmacogenomic biomarkers, and companion diagnostic tests are also increasing.^1,2 A recent study on rare and advanced tumors revealed that 51% of the tumors analyzed had ≥1 potentially actionable alteration.³ Nowadays, multimarker tumor panel tests that examine DNA alterations in 200-300 genes are becoming highly used in clinical settings.^4,5 This is due to the growing conviction among clinicians regarding their clinical utility, the development of guidelines recommending their implementation, and the willingness of healthcare payers to provide coverage for the costs associated with these advanced tests.^6,7 However, worldwide accessibility and interpretation of multimarker panel tests remain challenging.^8,9 Furthermore, these panels frequently identify variants of unknown significance or germline variants, which can further complicate interpretation.¹⁰ In addition, because of the low overall frequency of fusions and their varying prevalence across different tumor types, identifying patients who are most likely to benefit from use of targeted agents poses a clinical challenge.

As the field of DNA and RNA testing advances, there is a particular challenge when it comes to testing for fusions. The approval of multitarget drugs requires improving the molecular diagnostic approach, highlighting the clinical utility of moving from testing for single mutations in specific genes to using broad, multimarker tumor panels, such as next-generation sequencing (NGS), and defining the differences among DNA- and RNA-based tests for fusions.

Since the approval of crizotinib in 2013 for patients with advanced anaplastic lymphoma kinase (ALK)–rearranged non–small cell lung cancer (NSCLC), many other targetable fusions have been identified, including c-ROS oncogene 1 (ROS1); rearranged during transfection (RET); neurotrophic tyrosine receptor kinase 1, 2, or 3 (NTRK1-3); fibroblast growth factor receptor 1, 2, or 3 (FGFR1-3); platelet derived growth factor receptor alpha (PDGFRA) and platelet derived growth factor receptor beta (PDGFRB); human epidermal growth factor receptor 2 (HER2); B-Raf proto-oncogene; serine/threonine kinase (BRAF); and CRAF gene fusions.^11-14 Targeted drugs aimed at gene fusions have garnered impressive approval rates because of encouraging outcomes in clinical trials; this trend underscores the efficacy of these therapies in addressing a heterogenous group of cancer types (Table 1). In an analysis comparing therapy outcomes in patients with solid tumors who received matched targeted therapy for a fusion versus those who had received targeted therapy for nonfusion alterations, the median response rates were significantly higher for matched patients (68% for fusions compared with 50% in nonfusion; P < .0001) and the fusion therapy-matched group had a longer progression-free survival (PFS) compared with the group of patients who received unmatched therapy for fusions (11.6 months for matched compared with 4.9 months for fusion therapy unmatched; P = .034).³⁷

TABLE 1.

Fusion Targets in FDA-Approved Clinical Use

Gene Fusion	Drugs	Included Diseases	FDA Companion Diagnostic Test	References	ORR, %	mPFS, months	mOS, months
ALK	Alectinib (ALECENSA)	NSCLC	FoundationOne CDxa Ventana ALK (D5F3) CDx assayb FoundationOne Liquid CDxa	15	82.9	34.8	NR
	Brigatinib (ALUNBRIG)	NSCLC (patients who have progressed on or are intolerant to crizotinib)	Vysis ALK Break Apart FISH Probe Kitc	16	71.0	24.0	NR
	Ceritinib (ZYKADIA)	NSCLC	FoundationOne CDxa Ventana ALK (D5F3) CDx assayb	17	67.7	16.6	51.3
	Crizotinib (XALKORI)	NSCLC	FoundationOne CDxa Ventana ALK (D5F3) CDx sssayb Vysis ALK Break Apart FISH Probe Kitc	18	74	10.9	NR
	Lorlatinib (LORBRENA)	NSCLC (patients whose disease has progressed on crizotinib and ≥1 other ALK inhibitor for metastatic disease, or alectinib as the first ALK inhibitor therapy for metastatic disease, or ceritinib)	Ventana ALK (D5F3) CDx assayb	19	76.0	NR	NR
	Crizotinib (XALKORI)	IMT		20	66.7	18	NR
FGFR2 or FGFR3	Erdafitinibd (BALVERSA)	Metastatic or surgically unresectable urothelial cancer with progression during or after previous systemic therapy including anti–PD-1/anti–PD-L1 agent and not more than 2 previous lines of therapy	Therascreen FGFR RGQ RT-PCR Kite	21	45.5	5.6	12.1
FGFR2	Pemigatinibf (PEMAZYRE)	Cholangiocarcinoma	FoundationOne CDxa	22	35.5	6.9	21.1
	Futibatinibf	Intrahepatic cholangiocarcinoma		23	42	9	21.7
	Infigratinibf	Cholangiocarcinoma	FoundationOne CDxa	24	23.1	7.3	12.2
NTRK1,g NTRK2, NTRK3h,i	Larotrectinibf (VITRAKVI)	Solid tumors (patients with metastatic solid tumors or for whom surgical resection is likely to result in severe morbidity and who have no satisfactory alternative treatments or who have progressed after treatment)	FoundationOne CDxa	12,25	57j	24.6j	48.7j
	Entrectinibf (ROZLYTREK)	Solid tumors (patients with metastatic solid tumors or for whom surgical resection is likely to result in severe morbidity and who have no satisfactory alternative treatments or who have progressed after treatment)	FoundationOne Liquid CDxa FoundationOne CDxa	26,27	61.2	13.8	33.8
RET	Selpercatinibf (RETEVMO)	Solid tumors (patients whose disease has progressed on or after previous systemic treatment or who have no satisfactory alternative treatment options)	FoundationOne CDxa	28	43.9	11.1	18.0
	Selpercatinibf (RETEVMO)	NSCLC	Oncomine Dx Target Testk	13,29	84	22	NA
	Selpercatinibf (RETEVMO)	Thyroid (adult and pediatric patients 12 years and older with advanced or metastatic tumors who require systemic therapy and who are radioactive iodine–refractory, if radioactive iodine is appropriate)	Oncomine Dx Target Testk	30	69	11.2	NA
	Pralsetinib (GAVRETO)	NSCLC	Oncomine Dx Target Testk	31,32	72	13	NR
	Pralsetinib (GAVRETO)	Thyroid		32,33	55.7	25.8	NR
ROS1	Entrectinib (ROZLYTREK)	NSCLC	FoundationOne Liquid CDxa FoundationOne CDxa	34	68	15.7	47.8
	Crizotinib (XALKORI)	NSCLC	Oncomine Dx Target Testk	35	72	19.3	51.4
	Repotrectinib (AUGTYRO)	NSCLC		36	79 (TKI naïve)	35.7 (TKI naïve)	NR

NOTE. Acquired resistance mutation is not defined in the FDA label indication. ^g,h,iThese excluded biomarkers have been determined based on research and leadership decision at MD Anderson Cancer Center.

Abbreviations: ALK, anaplastic lymphoma kinase; FDA, US Food and Drug Administration; FGFR2/3, fibroblast growth factor receptor 2 or 3; FISH, fluorescence in situ hybridization; IMT, inflammatory myofibroblastic tumor; mOS, median overall survival; mPFS, median progression-free survival; NA, not assessed; NE, not estimable; NR, not reached; NSCLC, non–small cell lung cancer; NTRK1/2/3, neurotrophic tyrosine receptor kinase 1, 2, or 3; ORR, overall response rate; RET, rearranged during transfection; ROS1, c-ROS oncogene 1.

^a Foundation Medicine, Inc, Cambridge, MA.

^b Ventana Medical Systems, Inc, Oro Valley, AZ.

^c Abbott Molecular Inc, Des Plaines, IL.

^d On January 19, 2024, FDA granted full approval to erdafitinib for locally advanced or metastatic urothelial carcinoma carrying the FGFR3 genetic alterations. Previously, FDA had granted accelerated approval for the same indication in FGFR2- or FGFR3-altered urothelial cancers.

^e QIAGEN Manchester Ltd, Hilden Germany.

^f FDA-accelerated approval.

^g Excluded biomarkers: NTRK1 G595R.

^h Excluded biomarkers: NTRK3 F617L, NTRK3 G623R, and NTRK3 G696A.

ⁱ Excluded biomarker: NTRK3G623R.

^j Updated meeting abstract data, not a full publication.

^k Life Technologies Corporation, Carlsbad, CA.

NGS enables rapid and simultaneous sequencing of multiple fragments of DNA (or RNA) of varying lengths, including entire genomes, in a considerably short time and harnesses the benefits of advanced bioinformatics technology.^38,39 Large repositories of molecular alterations, such as the Catalogue of Somatic Mutations in Cancer, The Cancer Genome Atlas (TCGA), and the International Cancer Genome Consortium, are experiencing significant expansion in data recorded for gene fusions.^40-42 A 2018 study by Gao et al, using TCGA RNA-Seq data, showed that 16.5% of cancers have ≥1 driver fusion, with 6% of cases having druggable fusions. Nevertheless, the numbers of newly conducted trials and treatment approvals have risen since the publication of these data.⁴⁰ In the clinical setting, oncologists are using the findings derived from these multimarker platforms to guide their clinical decisions, recognizing that interpretation of similar results obtained from different platforms, whether DNA- or RNA-based, can lead to an interpretation bias.^43,44

This article provides a comprehensive overview on the actionability and clinical advancement of targeted agents aimed at driver fusions. It also highlights the challenges associated with fusion testing, including the evaluation of patients with cancer who could benefit from testing and devising an effective testing strategy.

ACTIONABILITY

Fusions refer to genetic rearrangements that occur either within the same chromosome (intrachromosomal) or between different chromosomes (interchromosomal).^44,45 Fusions can be classified as balanced if they do not involve a potential loss or gain of genetic information (eg, insertions, inversions, etc) or unbalanced if they do (eg, deletions).⁴⁶ Translocations are genetic fusions where a segment of DNA breaks off from one chromosome and becomes attached to a different chromosome or a different region within the same chromosome. Translocations can also be either balanced or unbalanced.⁴⁷ From this point forward, we will use the terms “translocation” and “fusion” interchangeably, but our focus will be on translocations. Other mechanisms of gene fusion formation include transcription read-through and mRNA splicing.

Fusions or translocations involve a multistep process, beginning with the occurrence of double-strand breaks at multiple locations in the chromosomes. These breaks can arise spontaneously or because of various factors, including replication errors or exogenous stress such as ionizing radiation and chemotherapeutic agents. Cells activate DNA repair mechanisms to restore genome integrity, with homologous recombination being active during the synthesis and G2 phases and nonhomologous end joining active throughout the cell cycle. However, persistent breaks that are not promptly resolved may lead to the misjoining of ends from different chromosomes, which results in a translocation or fusion.⁴⁸

Breakpoints in fusion genes do not exhibit a specific sequence preference, but they tend to occur more frequently within the introns of large genes, such as RET or ALK (Fig 1A).⁴⁹ This observation suggests that the breakage events occur randomly within fragile sites of chromosomes that are susceptible to DNA damage, resulting in high breakpoint heterogeneity.^50,51 In addition, a gene can make multiple fusion partners, typically following the principle of chromosomal proximity. Despite the significant linear distance between some genes and their fusion partners, the recombination can be facilitated by their spatial proximity in the 3-dimensional structure of the nucleus, influenced by various factors.⁵² For example, in prostate cancer, androgen signaling can bring the Transmembrane Serine Protease 2 and Erythroblast Transformation Specific (ETS)–related gene (ERG) loci into close chromosomal proximity, favoring the Transmembrane Serine Protease 2-ERG fusion.⁵³

FIG 1.

(A) Mechanism of fusion formation. An example of an intrachromosomal, in-frame, kinase-activating fusion is the EML4-ALK fusion. The ALK gene encodes a protein with distinct domains located extracellularly, in the transmembrane region, and intracellularly. Specifically, the intracellular domain is positioned at the C-terminal and contains the tyrosine kinase domain, which plays a crucial role in cell signaling pathways. To activate the TK domain, it is imperative that two ALK proteins dimerize. This process is initiated when a growth factor binds to the ligand-binding domain located in the extracellular portion of the ALK protein. This event leads to the phosphorylation of the TK domain and subsequent activation of intracellular pathways. EML4-ALK protein expression is typically minimal or even absent in lung cells. The EML4-ALK fusion oncogene includes the promoter region of EML4. Consequently, the EML4-ALK fusion protein is constitutively expressed in individuals with the fusion. This fusion usually includes the ALK portion, which contains the intracellular TK domain at the C-terminal, and the N-terminal region derived from EML4, which contains a trimerization domain. This domain enables the fusion protein to bring together three EML4-ALK proteins, resulting in TK phosphorylation and activation of intracellular molecular pathways. The figure was created using BioRender.com. (B) Testing methodologies. ALK, anaplastic lymphoma kinase; EML4, echinoderm microtubule–associated protein-like 4; FISH, fluorescence in situ hybridization; IHC, immunohistochemistry; NGS, next-generation sequencing; RT-PCR, reverse transcriptase–polymerase chain reaction; TAPE, tandem atypical propeller domain; TD, trimerization domain; TK, tyrosine kinase; TM, transmembrane.

When translocated DNA is transcribed into RNA and subsequently translated, the resulting fusion or chimeric proteins have the potential to reconfigure cellular signaling pathways.⁵⁴ A single driver fusion event has, in many cases, the capacity to drive tumor development without any additional genomic alterations.⁴⁰ Thus, fusions often occur in patients with low mutational burden and in the absence of other driver mutations, following a pattern of mutual exclusivity.^40,55

Fusion transcripts that result in a frameshift are categorized as out-of-frame fusions, whereas those that maintain the reading frame are considered in-frame fusions. Out-of-frame fusions typically do not generate functional fusion proteins, mainly because of a premature stop codon, whereas in-frame fusions will be fully translated into amino acid sequences and may give rise to fusion proteins with potential functional consequences.^56,57

Driver fusions refer to in-frame fusions that result in a gain of protein function, whereas passenger in-frame fusions do not have a significant impact on protein functionality. On the other hand, out-of-frame fusions can result in the loss of a certain tumor suppressor gene (TSG) function. For instance, in osteosarcomas, out-of-frame translocations involving p53 protein (TP53) can generate fusion proteins that lack essential functional domains or have a disrupted structure, rendering them nonfunctional and tumorigenic.⁵⁸

Driver fusion proteins can be classified into two main classes: protein kinases, such as tyrosine kinases (eg, ALK, ROS1, RET, NTRK1-3, and FGFR1-3) or serine-threonine kinases (eg, BRAF and CRAF), and transcription factors (eg, Ewing sarcoma [EWS] and MYC). However, oncogenic protein fusion formation can involve other proteins, such as growth factors (eg, NRG1).⁴⁷

Tyrosine kinase (TK) fusion originates from genomic rearrangements placing the 3′ portion of the TK gene encoding the kinase domain behind a partner gene at the 5′ position. The resulting TK protein possesses a kinase domain at the C-terminal and a ligand-binding domain and inhibitory regions in the N-terminal regions. Typically, the partner protein occupies the N-terminal region, whereas the C-terminal possesses the original TK domain. Ligand-independent activation is often facilitated by the fusion partner, which contains potential multimerization domains that can induce constitutive activation of the TK domain (Fig 1A).⁵⁹ Other mechanisms of overactivation of kinase proteins because of fusions include increased kinase expression due to alterations in upstream regulatory elements or loss of regulatory mRNA.

CLINICAL DEVELOPMENT OF TARGETED AGENTS FOR FUSIONS

The development of small-molecule kinase inhibitors has enabled the targeting of protein kinase fusions driving various cancers, as shown in Tables 1 and 2. These inhibitors are classified into seven types, depending on their binding sites within the kinase protein (eg, ATP-binding site, extracellular domain, etc) and their different inhibitory mechanisms (allosteric, covalent, or bivalent inhibition), which determine the level of TK specificity and associated toxicities.^60,61 Table 1 depicts the US Food and Drug Administration (FDA)–approved TK inhibitor drugs targeting fusions and the associated overall response rates, PFS, and overall survival.

TABLE 2.

Targeted Therapy to Fusions in Drug Development

Gene Fusion	Actionable for	Example Panels Detecting Fusion	Drugs	Development Status of Drugs
ALK	Treatment with ALK inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Vysis ALK Break Apart FISH Probe Kit, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, STGA-Fusions 2018a, Tempus xF assay, Tempus xT assay	Alectinib, brigatinib, ceritinib, crizotinib, entrectinib, gilteritinib, lorlatinib, sunitinib	FDA-approved
			Alkotinib, APG-2449, CT-707, ensartinib, gandotinib, NVL-655, repotrectinib, SIM1803-1A, SY-3505, TGRX-326, WX-0593, XZP-3621	Clinical development
BCL2	Treatment with BCL2 inhibitors	FoundationOne CDx, FoundationOne Liquid CDx	Venetoclax	FDA-approved
			APG-2575, AZD0466, BGB-11417, L-Bcl-2 antisense oligonucleotide, LP-108, LP-118, navitoclax, pelcitoclax, S65487, ZN-d5	Clinical development
BRAF	Treatment with BRAF inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xF assay, Tempus xT assay	Dabrafenib, dasatinib, encorafenib, regorafenib, sorafenib	FDA-approved
			Belvarafenib, BGB-3245, BRAF inhibitor LUT014, JZP815, KIN-2787, lifirafenib, naporafenib, PF-07284890, PF-07799933, PLX8394, TAK-580, VS-6766, XP-102	Clinical development
CCNE1	Treatment with CDK2 inhibitors	FoundationOne Liquid CDx	NA	FDA-approved
			BLU-222, CDKI AT7519, CT7001, dinaciclib, fadraciclib, milciclib, seliciclib, voruciclib, zotiraciclib	Clinical development
EGFR	Treatment with EGFR inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xT assay	Afatinib, amivantamab, brigatinib, ceritinib, cetuximab, dacomitinib, erlotinib, gefitinib, ibrutinib, lapatinib, mobocertinib, necitumumab, neratinib, osimertinib, panitumumab, vandetanib	FDA-approved
			Amivantamab/rHuPH20, anti-EGFR antibody-drug conjugate MRG003, anti-EGFR monoclonal antibody JMT101, anti-EGFR monoclonal antibody ZZ06, anti-EGFR/CD16A bispecific antibody AFM24, anti-EGFR/CD28 bispecific antibody REGN7075, anti-EGFR/c-Met bispecific antibody EMB-01, anti-EGFR/c-Met bispecific antibody MCLA-129, anti-EGFR/HER3 bispecific antibody SI-B001, BCA101, cetuximab-IR700 conjugate RM-1929, CM93, D2C7-IT, DBPR112, DZD9008, EGFR-targeting agent ABBV-637, futuximab, FWD1509, ganetespib, HA121-28, Hemay-022, icotinib, larotinib, lazertinib, lifirafenib, modotuximab, nimotuzumab, ORIC-114, pimurutamab, poziotinib, pyrotinib, rezivertinib, sapitinib, SH-1028, simmitinib, SKLB1028, SPH5030, SYN004, TAS2940, TY-9591, WSD0922-FU	Clinical development
FGFR1	Treatment with FGFR1 inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xF + assay	Brigatinib, erdafitinib, futibatinib, infigratinib, lenvatinib, nintedanib, pazopanib, pemigatinib, ponatinib, regorafenib, sorafenib, sunitinib, tivozanib	FDA-approved
			3D185, AL8326, anlotinib, AZD4547, cediranib, CFI-400945 fumarate, CPL304110, derazantinib, dovitinib, fruquintinib, gunagratinib, HMPL-453, KIN-3248, lucitanib, MAX-40279, ON123300, rogaratinib, simmitinib, sulfatinib, tasurgratinib, tinengotinib	Clinical development
FGFR2	Treatment with FGFR2 inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xF assay, Tempus xF + assay, Tempus xT assay	Brigatinib, ceritinib, erdafitinib, futibatinib, infigratinib, lenvatinib, nintedanib, pazopanib, pemigatinib, regorafenib, sorafenib, sunitinib, vandetanib	FDA approved
			3D185, AL8326, anlotinib, AZD4547, bemarituzumab, CFI-400945 fumarate, CPL304110, derazantinib, dovitinib, gandotinib, gunagratinib, HMPL-453, KIN-3248, lucitanib, MAX-40279, RLY-4008, rogaratinib, simmitinib, tasurgratinib, tinengotinib	Clinical development
FGFR3	Treatment with FGFR3 inhibitors	Therascreen FGFR RGQ RT-PCR Kit, FoundationOne CDx, FoundationOne Liquid CDx, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xF assay, Tempus xF + assay, Tempus xT assay	Brigatinib, ceritinib, erdafitinib, futibatinib, infigratinib, lenvatinib, nintedanib, pazopanib, pemigatinib	FDA-approved
			3D185, AL8326, anlotinib, AZD4547, B-701, CPL304110, derazantinib, dovitinib, gandotinib, gunagratinib, HMPL-453, KIN-3248, lucitanib, MAX-40279, rogaratinib, simmitinib, tasurgratinib, tinengotinib, TYR-300	Clinical development
FLT3	Treatment with FLT3 inhibitors	FoundationOne Liquid CDx, STGA-Fusions 2018a	Brigatinib, cabozantinib, ceritinib, fedratinib, gilteritinib, ibrutinib, midostaurin, nintedanib, pacritinib, pexidartinib, quizartinib, sorafenib, sunitinib, vandetanib	FDA-approved
			Anti-FLT3/CD3 bispecific antibody CLN-049, crenolanib, dovitinib, E6201, Flt3 ligand/anti–CTLA-4 antibody/IL-12 engineered oncolytic vaccinia virus RIVAL-01, fostamatinib, gandotinib, HM43239, lestaurtinib, MAX-40279, MRX-2843, MRX-2843, ningetinib, ON123300, Pan-FLT3/Pan-BTK multikinase inhibitor CG-806, rebastinib, sitravatinib, SKI-G-801, SKLB1028, SKLB1028, zotiraciclib	Clinical development
MET	Treatment with MET inhibitors	FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, MDA MAPPa, STGA-Fusions 2018a	Amivantamab, cabozantinib, capmatinib, crizotinib, tepotinib, tivozanib	FDA-approved
			ABN401, AL2846, amivantamab/rHuPH20, anti–c-Met ADC ABBV-400, anti–c-Met antibody-drug conjugate BYON3521, anti–c-Met antibody-drug conjugate TR1801, anti–c-Met/MMAE ADC RC108, anti-EGFR/c-Met bispecific antibody EMB-01, anti-EGFR/c-Met bispecific antibody MCLA-129, anti-MET x MET antibody drug conjugate REGN5093-M114, APL-101, BPI-9016M, emibetuzumab, glumetinib, kanitinib, merestinib, ningetinib, savolitinib, sitravatinib, telisotuzumab vedotin, XL092	Clinical development
NOTCH1	Treatment with gamma secretase inhibitors	FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, STGA-Fusions 2018a	NA	FDA-approved
			AL101, AL102, crenigacestat, LY3039478, nirogacestat, PF-03084014	Clinical development
NOTCH3	Treatment with gamma secretase inhibitors	FoundationOne Liquid CDx	NA	FDA-approved
			AL101, AL102, crenigacestat, LY3039478, nirogacestat, PF-03084014	Clinical development
NRG1	Treatment with HER2 and HER3 inhibitors	Caris MI Profile, Caris MI Tumor Seek, STGA-Fusions 2018a, Tempus xT assay	Afatinib, dacomitinib, hyaluronidase-zzxf/pertuzumab/trastuzumab, ibrutinib, lapatinib, margetuximab, mobocertinib, neratinib, pertuzumab, trastuzumab, trastuzumab deruxtecan, trastuzumab emtansine, trastuzumab/hyaluronidase-oysk, trastuzumab-ANNS, tucatinib	FDA-approved
			ACE1702, anti-EGFR/HER3 bispecific antibody SI-B001, anti-HER2 antibody-drug conjugate ARX788, anti-HER2 antibody-drug conjugate DB-1303, anti-HER2 antibody-drug conjugate DP303c, anti–HER-2 bispecific antibody KN026, anti-HER2 bispecific antibody-drug conjugate ZW49, anti-HER2 GSPT1 degrader ORM-5029, anti-HER2 monoclonal antibody BAT1006, anti-HER2 monoclonal antibody HLX22, anti-HER3 antibody-drug conjugate YL202, autologous anti-HER2 CAR macrophages CT-0508, autologous anti-HER2 CAR T cells CCT303-406, BI 1810631, BTRC4017A, CAM-H2, disitamab vedotin, DZD1516, DZD9008, entinostat, GQ1001, HER2/neu peptide vaccine GLSI-100, HER2-positive B-cell peptide antigen P467-DT-CRM197/montanide vaccine IMU-131, ICT-107, inetetamab, larotinib, MRG002, nelipepimut-S, ORIC-114, patritumab, pirotinib, poziotinib, PRS-343, pyrotinib, sapitinib, seribantumab, TAC01-HER2, TPIV100, trastuzumab duocarmazine, trastuzumab monomethyl auristatin F, trastuzumab rezetecan, U3-1402, WOKVAC, zanidatamab, zenocutuzumab	Clinical development
NTRK1	Treatment with TRK inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xF assay, Tempus xT assay	Cabozantinib, crizotinib, entrectinib, larotrectinib, lorlatinib, regorafenib, sorafenib, sunitinib	FDA-approved
			ARRY-470, CFI-400945 fumarate, lestaurtinib, merestinib, milciclib, PBI-200, repotrectinib, SIM1803-1A, SIM1803-1A, sitravatinib, taletrectinib, VC004, VMD-928	Clinical development
NTRK2	Treatment with TRK inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, MDA MAPPa, STGA-Fusions 2018a, Tempus xF + assay, Tempus xT assay	Cabozantinib, entrectinib, larotrectinib, lorlatinib, sorafenib, sunitinib	FDA-approved
			ARRY-470, CFI-400945 fumarate, gandotinib, lestaurtinib, merestinib, PBI-200, repotrectinib, SIM1803-1A, sitravatinib, taletrectinib, VC004	Clinical development
NTRK3	Treatment with TRK inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, MDA MAPPa, STGA-Fusions 2018a, Tempus xF + assay, Tempus xT assay	Entrectinib, larotrectinib, lorlatinib	FDA-approved
			ARRY-470, lestaurtinib, merestinib, PBI-200, repotrectinib, SIM1803-1A, SIM1803-1A, taletrectinib, VC004	Clinical development
PDGFB	Treatment with imatinib	BostonGene Tumor Portrait	Imatinib	FDA-approved
			NA	Clinical development
PDGFRA	Treatment with PDGFRA inhibitors	FoundationOne CDx, FoundationOne Liquid CDx Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xT assay	Axitinib, dasatinib, ibrutinib, imatinib, lenvatinib, midostaurin, nilotinib, nintedanib, olaratumab, pazopanib, ponatinib, quizartinib, regorafenib, ripretinib, sorafenib, sunitinib, tivozanib	FDA-approved
			Anlotinib, cediranib, chiauranib, crenolanib, dovitinib, famitinib, lucitanib, masitinib, radotinib, rebastinib, sitravatinib, telatinib, vorolanib	Clinical development
PDGFRB	Treatment with PDGFRB inhibitors	PDGFRB FISH Assay FoundationOne Liquid CDx Caris MI Profile, Caris MI Tumor Seek, STGA-Fusions 2018a	Axitinib, cabozantinib, dasatinib, imatinib, lenvatinib, midostaurin, nilotinib, nintedanib, pazopanib, quizartinib, regorafenib, ripretinib, sorafenib, sunitinib, tivozanib	FDA-approved
			Anlotinib, cediranib, chiauranib, crenolanib, dovitinib, famitinib, lucitanib, masitinib, ON123300, radotinib, rebastinib, SKLB1028, telatinib, vorolanib	Clinical development
RAF1	Treatment with MEK inhibitors	FoundationOne CDx, FoundationOne Liquid CDx Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a	Binimetinib, cobimetinib, selumetinib, trametinib	FDA approved
			ABM-168, ARQ 531, E6201, FCN-159, HL-085, IMM-1-104, mirdametinib, PF-07799544, pimasertib, SHR7390, TAK-733, VS-6766	Clinical development
RET	Treatment with RET inhibitors	Oncomine Dx Target test, FoundationOne CDx, FoundationOne Liquid CDx, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xF assay, Tempus xF assay xF+, Tempus xT assay	Alectinib, brigatinib, cabozantinib, ceritinib, fedratinib, ibrutinib, lenvatinib, pazopanib, ponatinib, pralsetinib, quizartinib, regorafenib, selpercatinib, sorafenib, sunitinib, vandetanib	FDA-approved
			Anlotinib, apatinib, BOS172738, dovitinib, fruquintinib, HA121-28, HS-10365, LOXO-260, sitravatinib, SY-5007	Clinical development
ROS1	Treatment with ROS1 inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa, STGA-Fusions 2018a, Tempus xF assay, Tempus xF assay xF+, Tempus xT assay	Brigatinib, cabozantinib, ceritinib, crizotinib, entrectinib, lorlatinib	FDA-approved
			Alkotinib, APG-2449, CFI-400945 fumarate, NVL-520, NVL-520, repotrectinib, SIM1803-1A, SIM1803-1A, sitravatinib, taletrectinib, TGRX-326, XZP-3621, XZP-5955	Clinical development
RSPO2	Treatment with Wnt signaling inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, MDA MAPPa, STGA-Fusions 2018a	NA	FDA-approved
			Porcupine inhibitor WNT974, Porcupine inhibitor RXC004, Porcupine inhibitor ETC-159, Porcupine inhibitor CGX1321	Clinical development
RSPO3	Treatment with Wnt signaling inhibitors	Caris MI Profile, Caris MI Tumor Seek, STGA-Fusions 2018a, STGA-Fusions 2018a	NA	FDA approved
			Porcupine inhibitor WNT974, Porcupine inhibitor RXC004, Porcupine inhibitor ETC-159, Porcupine inhibitor CGX1321	Clinical development
SYK	Treatment with SYK inhibitors	FoundationOne Liquid CDx	Midostaurin	FDA approved
			ASN002, ASN-002, entospletinib, fostamatinib, HMPL-523, lanraplenib, PRT062070, tipapkinogen sovacivec	Clinical development
ALK	Resistance to PD1/PDL1 inhibitors	FoundationOne CDx, FoundationOne Liquid CDx, Vysis ALK Break Apart FISH Probe Kit, Guardant360 CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, STGA-Fusions 2018a, Tempus xF assay, Tempus xT assay	Atezolizumab, avelumab, cemiplimab, durvalumab, nivolumab, nivolumab/relatlimab-rmbw, pembrolizumab	FDA-approved
			AK-104, AK-105, AMG 404, anti-CD47/anti–PD-L1 bispecific antibody 6MW3211, anti-CD47/anti–PD-L1 bispecific antibody IBI322, anti–CTLA-4/anti–PD-1 monoclonal antibody combination BCD-217, anti–PD-1 antibody-interleukin-21 mutein fusion protein AMG 256, anti–PD-1 monoclonal antibody 609A, anti–PD-1 monoclonal antibody SCT-I10A, anti–PD-1 monoclonal antibody Sym021, anti–PD-1 monoclonal antibody SYN125, anti–PD-1/anti-CD3 bispecific antibody ONO-4685, anti–PD-1/anti–LAG-3 bispecific antibody RO7247669, anti-PD1/CTLA4 bispecific antibody XmAb20717, anti-PD1/ICOS bispecific monoclonal antibody XmAb23104, anti–PD-1/TIM-3 bispecific antibody RO7121661, anti–PD-1/VEGF bispecific antibody AK112, anti–PD-L1 antibody-drug conjugate SGN-PDL1V, anti–PD-L1 monoclonal antibody A167, anti–PD-L1 monoclonal antibody CBT-502, anti–PD-L1 monoclonal antibody IMC-001, anti–PD-L1 monoclonal antibody ZKAB001, anti–PD-L1/anti–4-1BB bispecific monoclonal antibody GEN1046, anti–PD-L1/CD137 bispecific antibody MCLA-145, anti–PD-L1/CTLA-4 bispecific antibody KN046, balstilimab, BCD-100, bintrafusp alfa, budigalimab, camrelizumab, CK-301, CS1003, CX072, envafolimab, ezabenlimab, FAZ053, FS118, HB0030, INCA32459, LVGN3616, LY3300054, MGD013, NM21-1480, PD1-Fc-OX40L, pidilizumab, pimivalimab, retifanlimab, rilvegostomig, SAR445877, sasanlimab, serplulimab, SHR-1316, SHR-1701, SIM0237, sintilimab, sotiburafusp alfa, spartalizumab, sugemalimab, terelizumab, tislelizumab, toripalimab, volrustomig, zimberelimab	Clinical development
ESR1	Resistance to endocrine therapies	FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, STGA-Fusions 2018a	Anastrozole, bazedoxifene, elacestrant, exemestane, fulvestrant, letrozole, tamoxifen, toremifene	FDA-approved
			AC682, afimoxifene, camizestrant, ER alpha proteolysis-targeting chimera protein degrader ARV-471, estrogen receptor degrader AC699, H3B-6545, lasofoxifene, LY3484356, SAR439859	Clinical development
NOTCH2	Resistance to gamma secretase inhibitors	FoundationOne, FoundationOne Liquid CDx, Caris MI Profile, Caris MI Tumor Seek, Liquid Biopsy Panel V1a, MDA MAPPa	NA	FDA-approved
			AL101, AL102, crenigacestat, LY3039478, nirogacestat, PF-03084014	Clinical development

Abbreviations: ALK, anaplastic lymphoma kinase; BCL2, b-cell lymphoma 2; BRAF, B-Raf proto-oncogene, serine/threonine kinase; CCNE1, cyclin E1; EGFR, epidermal growth factor receptor; ESR1, estrogen receptor 1; FDA, US Food and Drug Administration; FGFR1/2/3, fibroblast growth factor receptor 1, 2, or 3; FISH, fluorescence in situ hybridization; FLT3, Fms-related receptor tyrosine kinase 3; HER2/3, human epidermal growth receptor factor 2 or 3; MDA MAPP, MD Anderson mutation analysis precision panel; MEK, mitogen-activated protein kinase; MET, MET proto-oncogene, receptor tyrosine kinase; NA, not applicable; NOTCH1/2/3, Notch receptor 1, 2, or 3; NRG1, neuregulin 1; NTRK1-3, neurotrophic tyrosine receptor kinase 1, 2, or 3; PDGFB, platelet-derived growth factor subunit B; PDGFRA/B, platelet-derived growth factor receptor Alpha or Beta; RAF1, Raf-1 proto-oncogene, serine/threonine kinase; RET, rearranged during transfection; ROS1, c-ROS oncogene 1; SPO2/3, R-Spondin 2 or 3; STGA, solid tumor genomic analysis; SYK, spleen-associated tyrosine kinase; TRK, receptor tyrosine kinase.

^a Liquid Biopsy Panel V1, MDA MAPP, and STGA-Fusions 2018 are MD Anderson Cancer Center next-generation sequencing–based tests.

Currently, there are ongoing developments in therapeutic approaches for targeting fusions different from small-molecule kinase inhibitors. For example, bispecific antibodies, such as zenocutuzumab, have demonstrated promising outcomes in phase 1 trials for NRG1-rearranged solid tumors by inhibiting the interaction between NRG1 fusion and HER3, as well as HER3 heterodimerization with HER2.⁶² In addition, various strategies are being explored to target transcription factor fusions, such as the EWS-friend leukemia integration 1 fusion including the use of small interfering RNA (siRNA) to silence the expression of EWS-friend leukemia integration 1.⁶³

CHALLENGES IN FUSION DETECTION

Challenge 1: How to Test

Solid tumors tend to display complex chromosomal rearrangements and, in this context, potentiate the occurrence of gene fusions throughout the genome.^64,65 Often, these fusions occur in tumors lacking other clear driver alterations, supporting a functional role of these rearrangements and underscoring the importance of methods capable of detecting multiple partner genes to ensure the detection of all therapeutically actionable fusions.^66,67 The constant increase in the number of new and unique partners reinforces the idea that the full spectrum of gene fusions in cancer is still incomplete and highlights the need to continue broadly profiling fusions in solid tumors. This need is especially evident given the increase in approved drugs that specifically target these rearrangements.^9,68

Testing for gene fusions has evolved rapidly in the past few years since the advent of deep sequencing assays. The evolution of sequencing techniques has led to an exponential increase in the list of known gene rearrangements. However, not all rearrangements lead to recurrent constitutive kinase activation; therefore, it is critical to understand the functional and clinical consequences of each fusion.⁵⁷ Fusion testing takes into consideration factors such as the structure of fusions and the intrinsic challenges of fusion detection, including variability in fusion partner and breakpoints and the fact that these events usually occur in introns⁶⁹ (Fig 1B).

Early Genetic Tests (one test, one marker)

The identification of fusions in specific genes was first explored through cytogenetics (karyotyping); fluorescence in situ hybridization (FISH); reverse transcriptase–polymerase chain reaction (RT-PCR), which uses a single-gene guided approach; and, in some cases, immunohistochemistry (IHC).⁷⁰ These methods offer significant advantages over more comprehensive methodologies, including global accessibility, a reduced expertise requirement for the pathology team, and a short turnover period to obtain test results. Nevertheless, the capacity of these tests to identify the full range of genomic alterations, fusions partners, and breakpoints is restricted. Another drawback to these tests is that, in each patient case, numerous genes must be analyzed. For instance, in a patient with NSCLC, assessments for several alterations, such as epidermal growth factor receptor (EGFR), ALK, ROS1, and NTRK, are required. Conducting individual gene tests for each alteration may prove to be costly and demand more tissue.

The FISH technique commonly uses break-apart or translocation probes and requires previous knowledge of the gene and gene partner, which could lead to missing cases where these rearrangements occur with partners not previously described.⁷¹ However, accurately discerning a gene fusion can be challenging because of the occurrence of false-positive and false-negative results.⁷²

RT-PCR can be used to detect the presence of a single oncogenic fusion–transcribed RNA for which the fusion partners are known, independent of the breakpoint. This approach achieves better sensitivity rates compared with histologic techniques (IHC and FISH) and holds special interest in certain situations, such as intrachromosomal rearrangements. In addition, the turnaround time of approximately 1 week for this platform, although higher than that for IHC or FISH, is significantly lower compared with that for NGS.⁷³ Since only one partner can be tested in each experiment, the utility of RT-PCR is limited when interrogating gene fusions with multiple partners, such as in the case of ALK or NTRK fusions.⁷⁴

The success of IHC is dependent on the antibody clone used and screening for expression of genes that are not expressed in normal tissues (eg, NTRK1-3 and ALK fusions) but not in other cases (eg, FGFR1-3 fusion). Despite good pan-TRK staining, IHC assays show low sensitivity for NTRK3 fusions and low specificity in certain tumors such as sarcomas. ⁷⁵

With advancements in fusion discovery and fusion partners in diverse tumor types persistently emerging, along with the identification of novel fusion partners, these approaches have become insufficient to provide optimal treatment options for patients with cancer.

Fusion Testing on the Basis of DNA- and RNA-Based NGS Panels

Genomic- and transcriptomic-based assays allow for parallel testing of multiple fusion alterations. Massive parallel sequencing offers the advantage of high depth and exon coverage. NGS panels produce a more uniform coverage across clinically relevant genes included in the assay and can also detect targetable gene fusions owing to the inclusion of breakpoint-containing introns in some test designs.⁷⁶ Currently, the number of clinically actionable fusions in lung cancers, cholangiocarcinomas, urothelial carcinomas, sarcomas, and other solid tumors exceeds the scope of what can be reasonably tested using a single-gene testing approach.⁷⁷

NGS tests either use amplicon-based (classical multiplex PCR or anchored multiplex PCR) and hybrid capture-based enrichment methods.⁷⁸ Some of the commonly used commercial assays for massive parallel testing (DNA and RNA sequencing) include the Oncomine Precision Assay on the Ion Torrent System (Thermo Fisher Scientific, Waltham, MA), QIAseq Targeted RNAscan Custom Panel (Qiagen, Hilden, Germany), and Illumina sequencing platforms (Illumina, San Diego, CA).^79,80

The technology, biochemistry, and design of DNA-based NGS testing vary in comparison with RNA-based NGS testing.^76,81 In a study on the basis of Tempus xT assay across 2,118 fusion events, RNA sequencing outperformed DNA sequencing for fusion detection, identifying 29.1% versus 4.8% fusions, respectively. In addition, combining RNA and DNA sequencing detected 76% of actionable fusions.⁸² Both DNA- and RNA-based NGS methods have limited worldwide accessibility and require specialized training. Test performance for DNA-based NGS depends on several factors reliant on material considerations and inherent to testing coverage (Table 3). Both methods can use formalin-fixed, paraffin-embedded material. However, one caveat to RNA testing is that the RNA in the sample is more susceptible to fragmentation and degradation. Poor-quality RNA in archival tissue can lead to false-negative results.⁸³ Specifically, suboptimal RNA quality will decrease the number and quality of reads, thereby reducing the chances to detect fusion transcripts and resembling the pattern that would be observed when the result of the gene fusion is a decrease in the level of transcription.⁸⁴ A clear advantage of RNA-based NGS is the potential to detect new fusion partners and to provide evidence of functionality. In specific cases of DNA-based NGS testing, the ability to detect all breakpoints/introns, such as NTRK1-3 fusions, could be limited because of the large size of introns, which can lead to false-negative test results.⁸⁵ RNA-based NGS is thus notably important in some gene fusions (eg, NTRK2) that arise from rearrangements in very long introns that harbor repetitive sequence elements that are also present elsewhere in the genome.⁸⁶ In addition, RNA sequencing provides direct evidence that the rearrangement produces a fusion expressed at the mRNA level, overcoming the limits of DNA sequencing when rearrangements appear noncanonical at the genomic DNA level. Thus, since the technique is focused on mature mRNA, it is not affected by the size of the intron or detecting a fusion when the sequence of exons is altered.⁸¹ In addition, the RNA-based approach allows the identification of chimeric RNA that results from cis- or trans-splicing and is not detectable at the DNA level.⁸⁷

TABLE 3.

Comparison of DNA- Versus RNA-Based NGS

Methodology	DNA-Based NGS	RNA-Based NGS
Worldwide accessibility	Limited. Training and expertise needed	Limited. Training and expertise needed
Material	Could be used in FFPE specimens	Could be used in FFPE specimens (fresh tissue) although RNA is more susceptible to fragmentation and degradation, especially in older material
Turnaround time	2 weeks	2 weeks
New fusion partners	No. Limited to known fusion partners of intron of interest	Yes. Possibility to detect new fusion partners
Evidence of overexpression	No	Yes
Resistance mutations to therapy	Yes	Yes (differential expression)
Affected by intron coverage	Yes	No
Sensitivity	Highly dependent on the extent and depth of coverage	Very high (assuming good RNA quality). Sensitivity may be lessened if RNA is degraded
Specificity	Highly dependent on whether the structural variant results in transcribed fusion	Very high

Abbreviations: FFPE, formalin-fixed paraffin-embedded; NGS, next-generation sequencing.

The clinical relevance of NGS-based panels is crucial, encompassing quality control and test validation through Clinical Laboratory Improvement Amendments (CLIA), which necessitate evaluating the performance characteristics of NGS. However, currently, only a handful of NGS-based panels, such as Thermo Fisher Scientific's Oncomine Dx Target Test, FoundationOne CDx, and FoundationOne Liquid CDx, have received FDA approval as companion diagnostic tests. Currently, no whole-exome or whole-genome sequencing test has full FDA approval for clinical use.

There are persistent challenges in implementing panel tests, including patient engagement in biomarker-based clinical trials and ensuring universal access to testing in patients with cancer.^88,89 As the number of commercial vendors offering NGS testing grows, it is crucial for the clinicians to stay updated on testing modalities since only some provide hybrid DNA and RNA testing.

Fusion Testing on Liquid Biopsies

Liquid biopsy offers a less invasive alternative to tissue genotyping by using various body fluids. ctDNA offers valuable information about the genetics and heterogeneity of the tumor, allowing for real-time monitoring of treatment response, prediction of relapse, mechanisms of resistance, and personalized treatment stratification.⁹⁰

In a recent study in 53,842 patients representing 66 cancer types, FoundationOneLiquid CDx detected a pathogenic rearrangement in 7,377 (14%), whereas the detection rate was notably comparable with that of tissue biopsy for cases with a tumor fraction of ≥1%. ⁹¹

While a ctDNA assay is highly sensitive in detecting single-nucleotide variations and small insertions or deletions, its sensitivity for detecting copy number variations or gene fusions may be limited. Similar to tissue-based testing, cell-free RNA (cfRNA)–based assay appears to be a more effective test than ctDNA-based assay for identifying a broad range of gene fusions and splicing variants.⁹² The widespread existence of even trace amounts of ribonucleases poses a significant challenge in assessing cfRNA, leading to an even lower RNA yield.⁹³

ctDNA assays are valuable for detecting fusions in treatment-naive patients, particularly in situations where obtaining tumor tissue for genotyping may be delayed, invasive procedures pose risks or contraindications, or only bone biopsy is feasible.⁹⁴

Challenge 2: Who to Test

The clinical problem for fusion testing lies in the fact that driver fusions are infrequent in common tumors but common in rare tumors.⁵⁷ However, fusions have a high potential for actionability, with favorable response rates and survival benefit when matched with a specific inhibitor.^37,95

It is crucial to understand the primary barriers encountered in the field of personalized oncology and, by extension, fusion identification. One study reported that the two most important factors include the cost associated with testing and accessibility to different techniques.⁹⁶ As a result, the clinical utility and cost-effectiveness of fusion testing remain controversial. Given that a single fusion gene can have multiple fusion partners and some of these fusions may be intrachromosomal, this further complicates the selection of appropriate testing techniques.

Recognizing tumors exhibiting recurrent gene fusions is a valuable approach to identifying relevant clinical scenarios where fusion testing may become crucial (Table 4). These fusions can then be used to classify a distinct molecular category of tumors. The selection process for patients who are suitable candidates for hybrid panels involves careful consideration and assessment for potential benefit.

TABLE 4.

Clinical Scenarios Where Fusion Testing May Become Useful

Rare Tumors Known to Be Driven by Fusions	Comments
Salivary gland tumors	Certain gene fusions have clinical significance in salivary gland cancers, including the MYB-NFIB fusion for adenoid cystic carcinoma, the CRTC1-MAML2 fusion commonly found in low-/intermediate-grade mucoepidermoid carcinoma, and ETV6-NTRK3 fusion observed in mammary analog secretory carcinoma97
Secretory breast cancer	NTRK fusions are prevalent in this tumor subtype71,98
ESR1 wild-type endocrine-resistant luminal breast cancer75	RET and NTRK fusions are highly prevalent across this specific tumor subtype99
Papillary and anaplastic thyroid cancer, ie, RAS/RAF/RET-negative	This testing should be performed on the tumor sample to identify molecular alterations, including RET fusions and other fusion events100
Infantile fibrosarcomas	Over 90% of cases of infantile fibrosarcomas exhibit NTRK fusions.101 Other potential alterations that may be identified are RET and BRAF gene fusions102
Fibrolamellar hepatocellular carcinoma	DNAJB1-PRKACA fusions. Also occur in oncocytic pancreatic and biliary tract cancers103
Young adults with undifferentiated carcinoma	—
Inflammatory myofibroblastic tumor	This tumor type presents ALK fusions in 50% of the cases. For ALK-negative patients, other potentially targetable fusions, like RET or ROS1, are often identified104
Dermatofibrosarcoma protuberans	Virtually all patients with this tumor type have COL1A1-PDGFB fusion. In addition, NTRK fusions may also be revealed through the analysis105
Mesenchymal tumors (eg, soft tissue sarcomas with no actionable mutations)	—
Quadruple-negative GIST (negative for KIT, PDGFRA, BRAF mutations, and SDH gene inactivation)	Specifically, NTRK, FGFR, or BRAF fusions should be investigated as they have been associated with this tumor subtype106,107
All pontine gliomas and low-grade gliomas (in particular, pilocytic astrocytoma)	The primary focus is detecting the KIAA-BRAF fusion. However, it is important to note that other molecular alterations, such as NTRK fusions, may also emerge, and they exhibit a mutually exclusive pattern with the KIAA-BRAF fusion108
Urothelial carcinomas	Should be tested via RNA NGS or FDA-approved RT-PCR for the detection of e fusions
Not-so-rare tumors
MSI-H colorectal cancer, in scenarios where MSI-H is uncommon and/or RAS/BRAF wild type	A substantial proportion of advanced colorectal cancers that are RAS and BRAF wild-type exhibit RET rearrangement. NTRK fusions are also more commonly reported in MSI-H tumors.109 In addition, fusions must be suspected in older patients or right-sided tumors110
NSCLC in younger, nonsmoking patient populations with no actionable mutation in EGFR, BRAF V600, KRAS, METex14, or ERBB2	Parallel or sequential approaches are accepted (ALK, ROS1, RET, and NTRK).67,111 NRG1 fusions are also prevalent in patients with mucinous and non–KRAS-mutated NSCLC112
Pancreatic cancer with KRAS wild type	They frequently present various fusions, such as RET, NTRK, ALK, FGFR, MET, NRG1, and RAF1113
Cholangiocarcinoma	RNA NGS is relevant for FGFR2 fusion detection, as well as NTRK and RET fusion detection, if DNA NGS did not previously detect any other molecular alterations, including HER2 amplifications, IDH mutations, and BRAF V600 mutations113,114
Common tumors with unusual histology	—
Common tumors in an unusual clinical setting	—
Tumors that have progressed to target therapies	—

Abbreviations: ALK, anaplastic lymphoma kinase; BRAF, B-Raf proto-oncogene, serine/threonine kinase; COL1A1, Collagen, type 1, alpha 1; CRTC1, CREB-Regulated Transcription Coactivator 1; ETV6, ETS Variant Transcription Factor 6; ERBB2, Erb-B2 Receptor Tyrosine Kinase 2; EGFR, epidermal growth factor receptor; ESR1, estrogen receptor 1; FGFR, fibroblast growth factor receptor; HER2, human epidermal growth factor receptor 2; METex14, MET proto-oncogene, receptor tyrosine kinase, exon 14; MAML2, mastermind like transcriptional coactivator 2; MSI-H, microsatellite instability-high; MYB, myeloblastosis; NFIB, Nuclear Factor I B; NGS, next-generation sequencing; NRG1, neuregulin 1; NTRK, neurotrophic tyrosine receptor kinase; NSCLC, non–small cell lung cancer; PDGFRA, platelet-derived growth factor receptor Alpha; PDGFB, platelet-derived growth factor subunit B; RET, rearranged during transfection; ROS1, c-ROS oncogene 1; RT-PCR, reverse transcriptase–polymerase chain reaction.

The tumor types for which fusion testing may become relevant include rare tumors where histology has a pathognomonic fusion and advanced tumors where the targeted or standard-of-care option is no longer available and for which the molecular alterations are unknown and not included in standard testing.

Challenge 3: Testing Strategy

Initial Diagnosis

In the context of fusion analysis, simultaneous DNA and RNA parallel sequencing at the time of diagnosis would be ideal to optimize fusion detection. However, given the additional cost of RNA NGS and its impact on offering standard or experimental targeted treatments to patients, it is imperative to identify flexible testing strategies on the basis of the resource availability of each hospital setting.

When determining the most suitable testing strategy for fusions during the initial diagnosis, three critical dimensions should be considered:

1. Timing of testing: The decision of whether to perform DNA sequencing and fusion detection tests simultaneously (parallel testing) or after DNA sequencing tests yield inconclusive results (sequential testing) is crucial. Both approaches have their advantages, and selecting the appropriate timing depends on the specific clinical scenario.

2. Choice of test: The choice between using RNA sequencing techniques or other tests, such as those described above, is a significant factor in fusion detection. Each method has its strengths and limitations, and the selection should be based on factors like sensitivity, specificity, and resource availability.

3. Selection of patients to test: There are several pivotal considerations for the patient population to be tested, including the following:

   1. Testing all patients, irrespective of their fusion prevalence, to ensure comprehensive genomic profiling.

   2. Prioritizing testing for patients with high-prevalence fusions (as listed in Table 2) to streamline detection efforts.

   3. Testing patients who have FDA-approved therapies available.

   4. Considering testing for patients eligible for clinical trials, enabling access to cutting-edge treatment options (Data Supplement, Table S1).

Considering these three aspects, we propose several distinct approaches that aim to optimize fusion detection and inform tumor treatment decisions while considering the economic resources available at different centers.

Testing strategy 1. Performing DNA sequencing for all patients and RNA sequencing for certain tumor types in a parallel manner.

The implementation of this strategy has the potential to maximize fusion detection while considering cost-effectiveness. In addition, it is recommended to perform RNA NGS in carefully selected patients who may harbor fusions of interest as this could render them eligible for participation in clinical trials, provided that such trials are available. We advocate for a gradual integration of this testing strategy in settings with adequate resources while acknowledging the need for further studies to establish its cost-effectiveness.

Testing strategy 2. Performing DNA sequencing for all patients, and if results show lack of clear tumor driver, then sequentially proceeding to RNA sequencing for certain tumor types.

Implementing RNA sequencing for specific tumor types, as outlined before, when DNA sequencing yields negative results, could offer a more cost-effective approach compared with the parallel approach. However, it is essential to consider that this approach may result in the potential omission of concurrent fusions that might not have been detected in a DNA panel that already detected a molecular alteration. In addition, there could be an increase in turnaround time associated with this strategy.

Testing strategy 3. Implementing whole genome sequencing/whole exome sequencing with CLIA certification, along with fusion testing strategies, such as FISH, RT-PCR, and IHC, for FDA-approved targeted therapies.

For medium- and low-resource settings, where there is limited availability for DNA NGS testing and no access to RNA NGS, we advocate for the following:

1. DNA NGS from CLIA-approved laboratories

2. RNA molecular testing (FISH and RT-PCR) or IHC for tumors with FDA-approved target therapies for fusions.

Considering the variable sensitivity and specificity of FISH, RT-PCR, and IHC testing across different fusions and tumor types, we have developed a comprehensive guide to effectively navigate this challenge (Table 4 and Fig 2).^{100,104,115-119}

FIG 2.

Detecting gene fusions with FDA-approved therapies in the absence of RNA NGS technology. ^aNTRK1-3 fusion highly prevalent tumor subtypes (FISH or RT-PCR) for NTRK fusion detection. These include but are not limited to the following tumor types: secretory breast cancer, infantile fibrosarcoma, wild-type GI stromal tumors, salivary gland tumors, thyroid, uterine sarcoma, ESR1 wild-type endocrine resistant luminal breast cancer, microsatellite instability, and high tumor mutational burden colon cancer. In addition, detection of RET fusion in highly prevalent tumor subtypes. These include but are not limited to salivary gland tumors; wild-type pancreatic cancers, sarcomas, cholangiocarcinoma, and neuroendocrine tumors; ESR1 wild-type endocrine-resistant luminal breast cancer. The figure was created using BioRender.com. ALK, anaplastic lymphoma kinase; CLIA, Clinical Laboratory Improvement Amendments; FDA, US Food and Drug Administration; FGFR2/3, fibroblast growth factor receptor 2 or 3; FISH, fluorescence in situ hybridization; IHC, immunohistochemistry; IMT, inflammatory myofibroblastic tumor; NCCN, National Comprehensive Cancer Network; NGS, next-generation sequencing; NSCLC, non–small cell lung cancer; NTRK1-3, neurotrophic tyrosine receptor kinase 1, 2, or 3; PDGFRB, platelet-derived growth factor receptor beta RET, rearranged during transfection; ROS1, c-ROS oncogene 1; RT-PCR, reverse transcriptase–polymerase chain reaction.

MOLECULAR TESTING BEYOND PROGRESSION

Secondary resistance to TK inhibitors occurs through various molecular events, including on-target mutations in the fusion gene, off-target mechanisms independent of the fusion gene, or histopathologic transformation. One example of secondary resistance is gatekeeper mutations, which are specific on-target mutations that occur at a critical residue within the kinase domain, often at the ATP-binding pocket.¹²⁰ On the other hand, solvent front mutations are on-target mutations that occur near the active site where water molecules are present and affect the ATP binding site of kinases and alter drug sensitivity. The emergence of resistance mechanisms involving KRAS^G12C mutations in oncogenic fusions with ALK, BRAF, FGFR3, RAF1, and RET represents another example of tumor adaptability to targeted therapies.¹²¹ Understanding the presence and impact of resistance mechanisms is crucial for developing effective treatment strategies to combat drug resistance.

In summary, gene fusions hold immense diagnostic and therapeutic value in solid tumors, warranting their integration into clinical decision making, especially in cancers with recurrent fusions and during disease progression.

AUTHOR CONTRIBUTIONS

Conception and design: Jibran Ahmed, Carlos Torrado, Anca Chelariu, Jordi Rodon Ahnert

Collection and assembly of data: Jibran Ahmed, Carlos Torrado, Sun-Hee Kim, Jordi Rodon Ahnert

Data analysis and interpretation: Jibran Ahmed, Carlos Torrado, Jordi Rodon Ahnert

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).