Front. Oncol. Frontiers in Oncology Front. Oncol. 2234-943X Frontiers Media S.A. 10.3389/fonc.2026.1739598 Review ROS1-positive non-small cell lung cancer: from genomics to treatment decisions WespiserMylène * Writing – review & editing Conceptualization Writing – original draft GilleRomane Conceptualization Writing – review & editing Writing – original draft PérolMaurice Supervision Data curation Validation Writing – review & editing Conceptualization Visualization Department of Medical Oncology, Centre Léon Bérard, Lyon, France Correspondence: Mylène Wespiser, mylene.wespiser@lyon.unicancer.fr 02 02 2026 2026 16 1739598 04 11 2025 05 01 2026 15 12 2025 Copyright © 2026 Wespiser, Gille and Pérol. 2026 Wespiser, Gille and Pérol https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

ROS1 rearrangements define a distinct, targetable subset of non–small cell lung cancer (NSCLC), representing ~2% of non-squamous cases and frequently presenting with metastatic disease and CNS involvement. Multiple ROS1 tyrosine kinase inhibitors (TKIs)—from crizotinib to newer agents such as entrectinib, lorlatinib, repotrectinib, taletrectinib, and the highly selective zidesamtinib—have improved systemic and intracranial outcomes, although resistance remains inevitable and biologically diverse, involving both on-target kinase mutations and off-target mechanisms. This review synthesizes current knowledge on ROS1 biology, diagnostic strategies, therapeutic options, and resistance mechanisms. We outline ROS1 fusion architecture and signaling, highlight partner-specific features, and summarize available diagnostic modalities. In clinical practice, RNA-based next-generation sequencing (NGS), often preceded by immunohistochemistry screening, provides the most sensitive approach for fusion detection and resistance profiling. Given the expanding therapeutic landscape and increasing complexity of treatment sequencing, we adopt a pragmatic, practice-oriented framework. CNS-penetrant next-generation TKIs with activity against common resistance mutations now constitute preferred first-line therapy. Repotrectinib and taletrectinib show strong systemic and intracranial efficacy, including activity against ROS1 G2032R, whereas zidesamtinib offers high selectivity with encouraging early data. Pemetrexed-based chemotherapy remains an effective option, whereas immune checkpoint inhibitors provide limited benefit. At progression, molecular reassessment is essential to guide tailored therapy. Looking ahead, priorities include optimizing sequencing strategies, evaluating perioperative targeted approaches, and incorporating genomic monitoring to anticipate resistance. These advances are reshaping the natural history of ROS1-rearranged NSCLC and supporting a more durable, precision-driven treatment paradigm.

 clinical trial drug resistance non-small cell lung cancer ROS1 rearrangement tyrosine kinase inhibitor oncogene-addicted The author(s) declared that financial support was not received for this work and/or its publication. section-at-acceptance Thoracic Oncology Introduction

Lung cancer remains the leading cause of cancer-related mortality worldwide, accounting for more than 1.8 million deaths each year (1). Non–small cell lung cancer (NSCLC) represents approximately 85% of all lung cancer cases and encompasses a heterogeneous group of tumors with distinct molecular and histologic characteristics. Over the past two decades, the identification of oncogenic driver alterations, most notably EGFR mutations and ALK rearrangements, has profoundly transformed the therapeutic landscape of advanced NSCLC, establishing molecularly targeted therapy as a new paradigm in precision oncology. Recently discovered, ROS1 rearrangements have defined a distinct molecular subset of NSCLC since 2007 (2, 3). It occurs in approximately 2% of NSCLC, corresponding to an estimated 50,000 new cases diagnosed each year worldwide (1). ROS1 rearrangements result in the constitutive activation of the ROS1 tyrosine kinase domain, driving oncogenic signaling and tumor proliferation. Although rare co-occurrences with EGFR or KRAS mutations have been reported, ROS1 fusions are mutually exclusive to most other oncogenic drivers (4).

Over the past decade, several generations of ROS1-targeted tyrosine kinase inhibitors (TKIs), from crizotinib to newer agents such as lorlatinib, repotrectinib, taletrectinib, and zidesamtinib, have demonstrated both strong systemic and intrecranial efficacy, leading to substantial and durable responses. In this review, we aim to provide a comprehensive and up-to-date overview of the current knowledge regarding the management of ROS1-positive NSCLC as of 2025. We will discuss recent advances in our understanding of ROS1 tumor biology, the diagnostic modalities to detect ROS1 rearrangements, and how to integrate the therapeutic progress achieved through the successive generations of ROS1-targeted TKI into the treatment strategy. As treatment options continue to expand across both early-stage and metastatic disease, as well as the knowledge about resistance mechanisms, optimizing the sequencing of therapies has become increasingly relevant. This review aims to offer clinicians a broad, precise, and practically useful synthesis of the latest data while also outlining emerging directions and future challenges in ROS1-rearranged NSCLC.

 <italic>ROS1</italic> gene and fusions

The ROS1 proto-oncogene, discovered in the 1980s, resides on chromosome 6q22.1 and encodes a 2347-amino acid receptor tyrosine kinase that belongs to the insulin receptor family (5, 6). The ROS1 protein is organized as a single transmembrane pass, with a large extracellular N-terminus (exons 1–34) containing β-propeller domains and fibronectin type III repeats. It is followed by a juxtamembrane segment, a tyrosine kinase domain (exons 36–42 in the canonical isoform), and a C-terminal tail (7). Interestingly, phylogenetic analyses have demonstrated significant homology between ROS1 and ALK, sharing more than 80% sequence identity, particularly within the kinase domain (8, 9).

To date, no physiological ligand for ROS1 has been identified in humans and is still considered an orphan receptor. However, recent studies in mice have suggested potential physiologic roles, including the identification of neural epidermal growth factor–like 2 (NELL2) as a binding partner for murine ROS1 receptor (10), as well as a potential role for extracellular matrix adhesion in ROS1 activation (11). During embryologic development, wild-type ROS1 expression in mice and chicken is needed, especially for kidney collecting duct and intestine but it also persists into multiple adult tissues such as the kidney, gastrointestinal tract, cerebellum, lung, testis, thymus, and bursa (12). Experimentation showed that ROS1-knockout mice appear viable and phenotypically normal, aside from male infertility due to epididymal defects (10, 13).

The oncogenic potential of ROS1 arises from chromosomal rearrangements that fuse its 3′ tyrosine kinase domain of ROS1 to the 5′ region of partner genes. This mechanism was first uncovered in 1986 through DNA transfer from the MCF-7 human breast carcinoma cell line into NIH-3T3 fibroblasts, revealing homology to the avian v-ros oncogene (14). One year later the same team identified the FIG (also known as GOPC)–ROS1 fusion in a glioblastoma cell line, establishing the oncogenic role of ROS1 rearrangements (15). In 2007, a study of phosphotyrosine signaling profiles across 41-NSCLC using a phosphoproteomics approach identified ROS1 fusions in a separate subset, paving the way for their clinical recognition and pharmacological development (2).

In cancers, the occurrence of inter- or intra-chromosomal rearrangements may result in ROS1-dependent carcinogenesis, leading to the formation of a constitutionally active fusion protein. Mechanistically, this uncontrolled kinase activity and autophosphorylation stimulates multiple canonical pathways, such as MAPK, PI3K/AKT/mTOR, and STAT3. These events promote tumor cell proliferation, survival, and metastasis (16). In NSCLC, breakpoints typically cluster around exons 32-34, whereas the preserved kinase domain spans exons 36- 42. Intrachromosomal events (e.g., 6q22 microdeletions/inversions) are well described in glioblastoma, whereas interchromosomal translocations predominate in NSCLC (9). To date, about 30–50 ROS1 fusion partners have been identified across tumor types. In NSCLC, the most prevalent fusion partners are CD74 (~44–64%), EZR (~14–29%), SDC4 (~6–16%), SLC34A2 (~5–12%) and others rarer (TPM3, ZCCHC8, SLC6A17, GOPC, CCDC6, CTNND2, LRIG3,LIMA1, MSN,…) (4, 9, 17, 18). Other ROS1 fusion partners have also been described in non-pulmonary malignancies, highlighting the broader oncogenic role of ROS1 across different tumor types (4, 17). The oncogenicity of these fusions depends on the preservation of the ROS1 kinase domain, whereas the 5′ partner determines subcellular localization and downstream signaling (19). SDC4–ROS1 and SLC34A2–ROS1 fusions localize to endosomes and preferentially activate the MAPK pathway via SHP2, whereas CD74–ROS1 localizes to the endoplasmic reticulum and may preferentially activate the JAK/STAT3 pathway (Figure 1) (19, 20). These distinct molecular characteristics may be associated with different clinical profiles and drug sensitivity outcomes. For example, CD74–ROS1 was associated with higher rates of central nervous system metastases in a retrospective study of 36 ROS1-positive NSCLC tumors (21).

Schematic representation of oncogenic ROS1 fusion proteins and downstream signaling pathways in ROS1 rearranged NSCLC. ROS1 rearrangements generate constitutively active chimeric tyrosine kinases in non–small cell lung cancer (NSCLC), in which the ROS1 kinase domain (3′ end) fuses with the 5′ partner gene. The most frequent fusion partners include CD74, EZR, SDC4, SLC34A2, TPM3, and GOPC/FIG, each conferring distinct subcellular localization and signaling properties. In the absence of an identified physiological ligand in humans, these fusions drive ligand-independent activation of ROS1 and its downstream signaling cascades, notably the SHP2–RAS–RAF–MEK–ERK, PI3K–AKT–mTOR, and JAK–STAT pathways. These pathways promote tumor cell proliferation, survival, and epithelial–mesenchymal transition (EMT). Pie charts illustrate the relative prevalence of individual fusion partners across reported NSCLC cohorts.

Diagram illustrating ROS1 rearrangements in non-small cell lung cancer (NSCLC). It shows various molecular pathways activated by ROS1, indicating fusion partners like CD74, EZR, SDC4, SCL34A2, and others. Pathways include SHP2, PDK1, AKT, mTOR, JAK, STAT, RAS, Raf, MEK, and ERK, leading to cell survival and signaling processes. Different fusion types like ROS1-TPM3, ROS1-EZR, ROS1-SLC34A2, ROS1-SDC4, and ROS1-GOPC are depicted with a corresponding pie chart showing their prevalence. The illustration highlights the impact on cellular processes and transitions. Diagnosis of <italic>ROS1</italic> fusions

The detection of ROS1 rearrangements in clinical practice relies on complementary diagnostic approaches, each with specific advantages and limitations. It is mandatory at the metastatic stage in non-squamous NSCLC and for non-smokers in squamous-cell carcinoma. By analogy with ALK-positive NSCLC, ROS1 rearrangement detection should also be recommended in the early stages of NSCLC before making a decision of perioperative treatment, including immunotherapy, or consolidation with durvalumab in locally advanced disease.

Immunohistochemistry (IHC) is often used as an initial screening tool because of its broad availability, low cost, and rapid turnaround time. Commercial antibodies (e.g., D4D6 clone) can detect ROS1 protein expression in formalin-fixed paraffin-embedded (FFPE) tumor samples, but with a lack of specificity. False positives have been reported mainly due to non-specific cytoplasmic staining or background signal observed in normal lung tissues, including alveolar macrophages and type II pneumocytes, particularly in tumors with abundant mucin or necrosis (9, 22). In addition, false-negative results may occur (approximately 3%–5% depending on the series) due to inadequate tissue fixation, heterogeneous or low fusion protein expression, or inadequacy between antibody clones and platforms (16, 23, 24). Therefore, IHC alone is insufficient to establish a diagnosis and requires confirmatory genomic testing.

Fluorescence in situ hybridization (FISH) with a break-apart probe was for a long time regarded as the gold standard in clinical trials evaluating ROS1-targeted therapies. This single-gene assay uses two labeled fluorescents (red and green), each targeting either the 5’ or 3’ regions of the ROS1 gene. Standard criteria define positivity as the presence of split signals or atypical patterns, such as isolated 3′ signals, in at least 15% of evaluable tumor nuclei (3, 25). It enables the detection of rearrangements independent of the fusion partner. Therefore, the inability to identify the fusion partner can make this method less sensitive in cases of complex or cryptic rearrangements. Interpretation can also be challenged in samples with low tumor cellularity or mucinous histology. However, FISH remains a reliable and widely validated test for diagnosing ROS1-rearranged NSCLC (9).

The reverse transcriptase polymerase chain reaction (RT-PCR) is an alternative and relatively cost-effective method for detecting already known ROS1 fusion RNA transcripts by amplifying their expressed complementary DNA. It enables precise partner identification and structural resolution of the fusion event. Compared with FISH, RT-PCR showed good sensitivity (even in the case of low-level variant allele frequencies) but lower specificity (~85%), as requiring high-quality RNA from well-preserved RNA extracted from FFPE samples. Moreover, commercial RT-PCR tests are designed to detect a panel of the most common fusion partners (e.g., CD74, SDC4, SLC34A2, GOPC, EZR, TPM3, LRIG3; AmoyDx RT-PCR assay) and therefore miss novel or rare rearrangements (26).

Next-generation sequencing (NGS) has increasingly emerged as the reference method for comprehensive molecular profiling. DNA-based NGS panels are capable of identifying ROS1 breakpoints but may fail to detect fusions when intronic breakpoints occur in large or poorly covered regions. In contrast, RNA-based NGS directly demonstrates the expressed fusion transcript, allowing the precise identification of the partner gene with high sensitivity and specificity. Importantly, NGS simultaneously surveys several actionable oncogenic drivers, which is clinically relevant, particularly by providing information on co-mutations and the emergence of resistance mechanisms. This strategy is now often prioritized as a frontline diagnostic tool in centers with access to broad RNA-based NGS panels (23, 27).

Liquid biopsy using plasma NGS of circulating tumor DNA (ctDNA) enables the detection of oncogenic gene fusions, including ROS1 rearrangements. Although false negatives may occur in patients with low ctDNA shedding, typically those with limited intrathoracic disease or isolated brain metastases, plasma-based assays have demonstrated high specificity and increasing sensitivity for fusion detection in advanced NSCLC. Sensitivity is closely related to tumor DNA allelic fraction. In the LIBELULE trial, early ctDNA profiling identified actionable genomic alterations in approximately 13% of cases where tissue analysis was non-informative, confirming the clinical utility of liquid biopsy as a complementary approach to tissue testing when assessing fusion-driven tumors (28, 29). Furthermore, ctDNA can also be used for longitudinal monitoring of ROS1 rearrangements throughout the disease course and may identify emerging resistance mechanisms to guide subsequent therapeutic decisions.

Diagnostic workflows are adapted to local resources in daily practice (Figure 2). A pragmatic algorithm involves using IHC as an initial screen, followed by confirmation with either NGS (RNA-based preferred) or FISH, depending on the availability of the method. An NGS-first approach is increasingly favored in well-resourced settings, with IHC serving as a triage tool. This layered strategy balances efficiency, cost, and diagnostic precision to ensure the accurate identification of patients with ROS1-positive NSCLC who may benefit from targeted therapies. Looking ahead, artificial intelligence has substantial potential to support diagnostic workflows, particularly within molecular biology and pathology, where the analysis of FFPE slides and other histopathological data could be integrated with molecular profiles to enhance diagnostic precision and efficiency (30).

Diagnostic testing algorithm for the detection of ROS1 rearranged NSCLC. This flowchart summarizes a pragmatic diagnostic approach for identifying ROS1 rearrangements in non-small cell lung cancer (NSCLC). Immunohistochemistry (IHC) serves as an initial screening tool, with positive cases requiring confirmatory molecular testing by fluorescence in situ hybridization (FISH), reverse transcriptase polymerase chain reaction (RT-PCR), or next-generation sequencing (NGS)-preferably RNA-based assays. Liquid biopsy using specific plasma NGS may be considered when tissue is insufficient or unavailable, allowing non-invasive detection of ROS1 fusions.

Flowchart outlining the diagnostic pathway for lung cancer with suspected ROS1 rearrangement. It starts with clinical and radiological features, followed by histological sampling for adenocarcinoma, NSCLC NOS, or SCC in non-smokers. The pathway continues with IHC ROS1 testing, leading to two outcomes: negative (no rearrangement) or positive (molecular testing). Positive cases proceed with RT-PCR, FISH, DNA-based NGS, or RNA-based NGS, concluding with positive ROS1 rearrangement confirmation. A liquid biopsy is possible if samples are insufficient.

A reliable diagnostic framework is essential because ROS1-positive NSCLC with distinctive demographic and histologic features requires tailored therapeutic strategies.

 Clinical and histological presentation

The median age of patients with ROS1-positive NSCLC onset is younger (~45–50 years) than that of patients with NSCLC in general. In the reported series, a predominance of females, Asian patients, and never- or light-smokers was observed (3, 31). The predominant histological type is adenocarcinoma, with acinar and solid subtypes being more frequent, but lepidic, papillary, mucinous, signet ring cell subtypes of adenocarcinomas have also been described (32–34). Thyroid transcription factor-1 (TTF1) is expressed in >90% of cases. ROS1 rearrangements are rarely detected in other histological subtypes, such as squamous cell carcinoma, pleomorphic carcinoma, or large cell carcinoma.

These tumors are characterized by a high metastatic potential and are often diagnosed at stage IV (>80%), which may be linked to the rarity of genomic testing in the early stages. They present a propensity for lymphatic dissemination and a marked neurotropism, as reflected by the high incidence of central nervous system involvement at diagnosis (36-40%) and during evolution, regardless of the fusion partner (35, 36). Miliary patterns of lung metastases have been reported. Several authors have reported an increased thromboembolic risk in patients with ROS1-positive NSCLC (37).

 Treatment options

Treatment should be optimized on an individual basis, leveraging the expanding repertoire of ROS1-targeted TKIs alongside established chemotherapy options.

Tyrosine kinase inhibitors targeting ROS1

ROS1-targeted therapies are primarily classified into type I and type II agents based on their binding modalities within the kinase catalytic domain (Table 1). Type I inhibitors are ATP-competitive compounds that reversibly bind to the active conformation of the kinase, occupying the ATP-binding site and competing directly with ATP during phosphorylation. In contrast, type II inhibitors interact with the inactive (DFG-out) conformation, extending from the ATP pocket into an adjacent hydrophobic back pocket located behind the gatekeeper residue, thereby stabilizing the inactive state of the kinase. This distinct binding mode may confer a broader resistance-mutation coverage, as type II inhibitors can overcome specific solvent-front and β-sheet mutations (Cβ6), including ROS1 L2086F, which mediates resistance to several type I inhibitors (38).

Type I and II ROS1 inhibitors.

Inhibitor type ROS1 inhibitor Structure Other targets
Type I Crizotinib Noncyclic ALK, MET
Entrectinib Noncyclic NTRK, ALK
Brigatinib Noncyclic ALK
Ceritinib Noncyclic ALK
Lorlatinib Macrocyclic ALK
Repotrectinib Macrocyclic ALK, NTRK
Taletrectinib Noncyclic NTRK
Zidesamtinib Macrocyclic highly selective for ROS1
Type II Cabozantinib Noncyclic ALK, MET, AXL, KIT, RET, VEGFR2, FLT3
Foritinib Noncyclic ALK

Type I inhibitors bind to the active conformation of the kinase domain, whereas type II inhibitors recognize the inactive conformation. The table lists the main ROS1 tyrosine kinase inhibitors according to their binding mode, structure (noncyclic or macrocyclic), and known off-target kinases.

ROS1 inhibitors are also classified into successive generations, developed to overcome the ROS1 resistance mutations to earlier agents and to enhance central nervous system (CNS) penetration.

First-generation inhibitors

Crizotinib was the first approved ROS1-targeted TKI (FDA since 2016), representing a paradigm shift for treating ROS1-rearranged NSCLC. In the pivotal PROFILE-1001 trial, crizotinib achieved an overall response rate (ORR) of 72%, a median progression-free survival (PFS) of 19 months, and a median overall survival (OS) of 51 months in TKI-naïve patients (39, 40). These results were subsequently confirmed in multiple independent phase II studies across different geographic cohorts, including the EUCROSS European study (41), the AcSé French trial (42), the East Asia study (43), and the METROS Italian study (44), reporting consistent ORRs ranging from 65% to 80% and a median PFS between 15 and 22 months. Its limitations include poor intracranial penetration and lack of activity against resistance mutations, such as ROS1 G2032R. Common toxicities include visual disturbances, gastrointestinal symptoms, peripheral subcutaneous edema, and hepatotoxicity (see Supplementary Table 1). Crizotinib remains widely available but has been largely superseded by newer-generation TKIs with improved efficacy and intracranial activity.

Entrectinib is a first-generation multitarget TKI (ROS1, ALK, TRK) with improved brain penetration compared with crizotinib. In an integrated analysis of three phase I–II trials (STARTRK-1, STARTRK-2, ALKA-372-001), entrectinib demonstrated an ORR of 68%, a median PFS of 15.7 months, and a median OS of 47.8 months in TKI-naïve ROS1-positive NSCLC patients, with an intracranial ORR (IC-ORR) of 55% (45). However, entrectinib is ineffective against common resistance mutations such as ROS1 G2032R, D2033N, L2026M, and shows limited activity post-crizotinib. Toxicities include fatigue, dizziness, weight gain, dysgeusia, and gastrointestinal effects (see Supplementary Table 1) with rare but serious risks of congestive heart failure and QTc prolongation. Entrectinib was approved by the FDA and EMA in 2019 and 2020, respectively.

 Second-generation inhibitors

Lorlatinib is a second-generation type I ATP-competitive inhibitor with dual activity against ROS1 and ALK. It is also based on a macrocyclic chemical structure that ensures high brain penetration. In a multicenter phase I/II study, lorlatinib achieved an ORR of 62% and a median PFS of 21.1 months in TKI-naïve patients (n=69), with an IC-ORR of 64% (46). In previously treated patients (n=40), the efficacy was more modest (ORR 35%; median PFS 8.5 months; IC-ORR 50%) (46). Recently, the phase II IFCT-2003 ALBATROS study evaluated the efficacy of lorlatinib after failure of a first-line ROS1 TKI, mainly crizotinib (Duruisseaux et al. ESMO congress 2025, 1987MO). Lorlatinib demonstrated a confirmed ORR of 34% (BICR-assessed), a median PFS of 7.4 months, and a median duration of response (DOR) of 20.4 months. Among patients with measurable brain metastases, the IC-ORR reached 92% (n=12/13). Lorlatinib retains no activity against ROS1 G2032R on-target resistance mutations emerging after prior ROS1 TKI (47). Its toxicity profile includes hyperlipidemia (72% hypercholesterolemia, 66% hypertriglyceridemia), 42% neuropsychologic-related adverse effects (cognitive changes, mood alterations, speech disturbances), peripheral edema, and weight gain (44%), which may limit long-term tolerability in some patients. Although the FDA did not approve lorlatinib for ROS1-positive NSCLC, it has been included in the NCCN and ESMO guidelines since 2019.

Foritinib (SAF-189s), is a type II dual ROS1/ALK kinase inhibitor that was recently evaluated in the phase II SAF-189s-ROS1–001 study conducted in China (48). Foritinib demonstrated strong systemic activity in TKI-naïve patients (n=56), achieving an ORR of 88% (95% CI 76–95) and a median PFS of 22.1 months, with an IC-ORR of 90% in patients with measurable brain metastases (n=19/21). The ORR was 40% (95% CI 21–61) in patients previously treated with ROS1 TKIs. The safety profile was favorable, with hyperglycemia (12% grade 3) and QTc prolongation (6% grade 3) being the most common treatment-related adverse events; no neurotoxicity or sensory disturbances were reported. Foritinib is still under clinical evaluation (SAF001, NCT04237805) and has not yet obtained regulatory approval from the authorities.

Ceritinib, another second-generation type I ATP-competitive inhibitor with ALK and ROS1 activity, demonstrates superior brain penetration compared with crizotinib. In a phase II study (49), ceritinib showed an ORR of 62% and a median PFS of 9 months in mostly crizotinib-naïve patients (n=32). Gastrointestinal adverse events (nausea, diarrhea, vomiting) and hepatotoxicity dominate the toxicity profile, often requiring dose adjustments. ECG monitoring is necessary due to potential QTc prolongation. However, its clinical development in this setting has not led to regulatory approval.

Brigatinib is a second-generation type I ATP-competitive inhibitor with dual ALK/ROS1 activity and enhanced brain penetration. In the phase II BAROSSA study (50), brigatinib demonstrated an ORR of 71% and a median PFS of 12 months in mostly TKI-naïve patients, though data in the post-TKI setting remains limited. Brigatinib exhibits poor activity against the ROS1 G2032R resistance mutation. Its safety profile includes gastrointestinal symptoms, hypertension, increased creatine phosphokinase, and pulmonary adverse events (notably early-onset pneumonitis). Brigatinib is not FDA-approved for ROS1-positive NSCLC.

Cabozantinib is a type II, non-ATP competitive multikinase inhibitor targeting ROS1, ALK, MET, AXL, KIT, RET, and VEGFR. Preclinical studies suggest activity against selected ROS1 resistance mutations (e.g. ROS1 D2033N, L2086F, G2032R) (51). In addition to preclinical data, isolated case reports have reported responses to cabozantinib, alone or combined with other ROS1 TKIs, in patients harboring the ROS1 L2086F resistance mutation, with response lasting approximately 7–11 months (52–54). Nevertheless, these reports are subject to publication bias and do not provide sufficient clinical evidence to support cabozantinib in this setting. Its toxicity profile includes diarrhea, hypertension, palmar-plantar erythrodysesthesia, and cardiovascular events, consistent with other VEGFR-targeting multikinase inhibitors.

Unecritinib (TQ-B3101) is a type I ATP-competitive multikinase inhibitor developed only in China targeting ROS1, ALK, and c-MET. In phase II (NCT03972189), 111 ROS1 inhibitor-naïve patients with advanced or metastatic ROS1-positive NSCLC received 300mg twice daily. The ORR was 81%, and the median PFS was 17.2 months. Unecritinib cannot overcome ROS1 resistance mutations. Among 33 patients with baseline brain metastases, IC-ORR was 72.7%. Grade ≥ 3 treatment-related adverse events (TRAEs) occurred in 51.3% of patients, mainly elevated liver enzyme, whereas ocular and neurologic adverse events were mostly grade 1–2 (55, 56).

Repotrectinib is a macrocyclic, type I ATP-competitive inhibitor designed to target ROS1, ALK, and NTRK fusions, with high CNS penetration and activity against the ROS1 G2032R mutation. In the TRIDENT-1 trial, repotrectinib achieved a confirmed ORR (cORR) of 79% in TKI-naïve patients (n=71), an IC-ORR of 89% (n=8/9), and a median PFS of 31.1 (95%CI, 21.9–NE) months. With a median follow-up of 44.6 months, the OS rate was 56% at 4 years (57). In previously treated patients (n=56), the cORR was 41%, median PFS was 8.6 (95% CI, 5.5-14.5) months, and median OS was 25.1 (95% CI, 12.8-32.1) months (57). Repotrectinib was active against ROS1 G2032R with an ORR of 59% (n=10/17). The most common adverse events include neurological effects: dizziness (58%, 3% grade ≥3), paresthesia (30%), ataxia (20%), muscular weakness (14%), memory impairment (13%); gastrointestinal disturbance: dysgeusia (50%), constipation (26%), nausea (12%) weight increase (12%) and dyspnea (8%, <1% grade ≥3). Increased AST (aspartate transaminase) and ALT (alanine transaminase) were reported in 18% (1% grade ≥3) and anemia in 26% of cases (58). Repotrectinib has been FDA and EMA-approved since 2023 and 2024 respectively.

Taletrectinib (DS-6051b/AB-106) is a second-generation, CNS-penetrant, type I ATP-competitive inhibitor targeting ROS1 and NTRK but sparing TRKB. It has demonstrated robust activity against several resistance mutations, including ROS1 G2032R, L2026M, L1951R, and S1986F. In the pooled TRUST-I and TRUST-II analyses (59), taletrectinib achieved high efficacy in TKI-naïve patients (n=160), with a cORR of 88.8%, an IC-cORR of 76.5%, a median PFS of 45.6 months, and a median duration of response (DOR) of 44.2 months. In the post-TKI (mainly crizotinib) setting (n=113), the cORR was 55.8%, the IC-cORR 65.6% with a median PFS of 9.7 months, and a median DOR of 16.6 months. Activity against ROS1 G2032R-positive disease was confirmed with a cORR of 61.5% (n=8/13). The toxicity profile is characterized mainly by gastrointestinal adverse events (88%), transaminases elevation (AST 70.2%, ALT 66.5%), and rare neurologic events such as dizziness (16%) and dysgeusia (15%), most of which being grade 1. Treatment discontinuations occurred in 6.5% of patients. Taletrectinib was approved by the FDA in 2025.

 Third-generation inhibitors

Zidesamtinib (NVL-520) is a third-generation, macrocyclic, type I ATP-competitive inhibitor specifically designed for high selectivity against ROS1. Kinome profiling demonstrated potent inhibition of ROS1, with minimal off-target activity across 335 kinases; only ALK (2-fold weaker IC50) and five kinases (LTK, FAK, PYK2, TRKB, and FER) were inhibited within 10–50-fold of the ROS1 IC50, underscoring its high degree of selectivity (60). This selectivity reduces off-target toxicities commonly observed with earlier multitarget TKIs while preserving potent activity against ROS1-driven tumors. Its macrocyclic structure further contributes to conformational rigidity, pharmacokinetic stability, and intracranial penetration while also circumventing steric hindrance caused by solvent-front mutations such as ROS1 G2032R.

In the phase I/II ARROS-1 trial (NCT05118789), zidesamtinib demonstrated promising efficacy in TKI-naïve (61) and heavily pretreated patients (62). Among 35 TKI-naïve response-evaluable patients, the ORR reached 89%, including 9% complete responses, with a DOR of 96% at 12 months. Among six patients with measurable intracranial disease, the IC-ORR was 83%, including 67% complete responses, and no CNS progression events were observed during follow-up. In TKI-pretreated or post-chemotherapy cohorts (n=117), the ORR was 44%, with PFS of 48% at 12 months and 40% at 18 months. In patients who had received only one prior TKI, ORR reached 51%, with an estimated PFS of 68% at 12 and 18 months. Fifty-three percent of the patients had measurable intracranial metastases at enrollment. Activity against intracranial disease was notable, with an IC-ORR of 85% (n=11/13) in patients pre-treated only with crizotinib (± chemotherapy) and an IC-ORR of 48% (n=27/56) in patients with ≥2 prior ROS1 TKIs, including lorlatinib and/or repotrectinib. Importantly, zidesamtinib demonstrated potent activity against ROS1 G2032R, with an ORR of 54% (n=14/26) in patients previously treated with any ROS1 TKI, including lorlatinib and/or repotrectinib. The safety profile was favorable, with the most common TRAEs being peripheral edema (36% any, 0.7% grade 3), constipation (17% any), increased CPK (16% any), dyspnea (15%, 3% grade 3), and transaminase elevation (11%). Dose reductions due to treatment-emergent adverse events (TEAEs) occurred in 10% of patients, whereas treatment discontinuations were infrequent (2%), primarily related to pneumonia (n=3). Zidesamtinib is currently under FDA Real-Time Oncology Review for potential approval.

The various ROS1 TKI differ in dosage, administration schedule, and interactions with food and concomitant medications, as summarized in Table 2. Efficacy outcomes in naïve and pretreated patients, including intracranial efficacy and G2032R mutation resistance coverage, are outlined in Table 3.

Pharmacology and practical use of ROS1 inhibitors.

Gen Inhibitor Dose and schedule PK/Interactions Approval status
3rd Zidesamtinib 100 mg QD ± food step-down 75→50 mg (tabs 100/25 mg) CYP3A4/2C9, P-gp, BCRP, MATE1 FDA RTOR (ongoing)
2nd Taletrectinib 600 mg QD; fasting >2h step-down 400→200 mg (caps only 200 mg) CYP3A4; PPI,QTc monitoring FDA-approved 2025
Repotrectinib 160 mg QD x14d → 160 mg BID ± food; (q12h) step-down 240→160 mg (tabs only 40 mg) CYP3A4, CYP2D6 FDA-approved 2023; EMA-approved
Lorlatinib 100 mg QD ± food step-down 75→50 mg (tabs 100/25 mg) CYP3A4; lipid monitoring Guideline-listed (not FDA-approved)
Foritinib 160 mg QD; ± food step-dows 120→ 80mg CYP3A4;monitor glucose; QTc monitoring Under clinical investigation
Ceritinib 450 mg QD with high-fat meal step-down 300 mg (caps only 150 mg) CYP3A4/2C9; QTc monitoring Not approved for ROS1
Brigatinib 90 mg QD x7d → 180 mg QD ± food step-down 90 mg (tabs 180/90 mg) CYP3A4/2C8; QTc monitoring Not approved for ROS1
Cabozantinib 60 mg QD fasting >2h step-down 40→20 mg (tabs 60/40/20 mg) CYP3A4; QTc monitoring Not approved for ROS1
1st Entrectinib 600 mg QD ± food step-down 400→200 mg (caps only 200mg) CYP3A4; QTc monitoring FDA-approved 2019; EMA-approved
Crizotinib 250 mg BID ± food (q12h) step-down 200 mg BID → 200 mg QD CYP3A4 FDA-approved 2016; EMA-approved

Efficacy of ROS1 TKIs in TKI-naïve (A) and TKI-pretreated (B) ROS1-positive NSCLC.

A.
Drug Crizotinib Entrectinib Ceritinib Brigatinib Lorlatinib Foritinib Repotrectinib Taletrectinib Zidesamtinib
(reference) Phase 1 (39) Phase 1-2 (45) Phase 2 (49) Phase 2 (50) Phase 1-2 (46) Phase 2b (48) Phase 1-2 (57) Pooled phase 2 (59) Phase 1-2 (61)
n 53 168 30 28 21 56 71 160 35
ORR 72% 67.9% 67% 71.4% 62% 88% 79% 88.8% 89%
Median PFS (months) 19.3 15.7 19.3 12.0 21.1 22.1 31.1 45.6 NA
IC ORR NA 52.1%(25/48)Measurable and non-measurable 40%(2/5)Measurable and non-measurable 0(0/3)Measurable 64%(7/11)Measurable and non-measurable 90%(19/21)Measurable and non-measurable 89%(8/9)Measurable 76.5%(13/17)Measurable 83%(5/6)Measurable
B.
Drug Entrectinib Lorlatinib Lorlatinib Lorlatinib Brigatinib Foritinib Repotrectinib Taletrectinib Zidesamtinib
(reference) Phase 1-2 (45) Phase 1-2 (46) Phase 2 (ALBATROS)¤ (French RW cohort)£ Phase 2 (BAROSSA) Phase 2b (48) Phase 1-2 (TRIDENT-1) (57) Phase 2 (TRUST I-II) (59) Phase 1-2 (ARROS-1) (61)
n 18 40 50 80 19 25 56 113 117 any prior TKIs (≥ 2, with L or R, ± CT) 55 1 prior ROS1 TKI (C or E) ± CT
ORR 11.1% 35% 34%* 45% 31.6% 40% 41% 55.8% 44%* 51%*
Median DOR (months) NA 13.8 20.4 7.4 NA 7.0 17.8 16.6 78% at 12 mo 93% at 12 mo
Median PFS (months) 4.7 8.5 7.4 7.1 7.3 5.5 8.6 9.7 48% at 12 mo 68% at 12 mo
IC-ORR (n) 18.8% (3/16) 50% (12/24) 92.3% (12/13) 72% (33/46) 66.7% (4/6) 40% (6/15) 38% (5/13) 65.6% (21/32) 48% (27/56) 85% (11/13)
ROS1 G2032R coverage no no no no no no Yes ORR 59% (n=10/17) Yes ORR 61.5% (n=8/13) Yes ORR 54% (n=14/26)

It is important to note that, given the heterogeneity of the patients included in these studies, no direct comparison can be made between these different clinical trials.

*BICR-assessed (Blinded independent central review); ¤Duruisseaux et al. ESMO congress 2025. £Girard et al. ESMO congress 2022.

Figure 3 highlights the most clinically relevant toxicities of crizotinib (n=53) (40), entrectinib (n=224) (45), repotrectinib (n=426) (58), taletrectinib (n=352) (59), and zidesamtinib (n=432) (61) based on the data available to date. TRAEs occurring in ≥10% of patients are detailed in the Supplementary Table 1.

Selected TRAE (occurring in ≥10% patients, all grades) with crizotinib (39), entrectinib (45), repotrectinib (58), taletrectinib (59), and zidesamtinib (61). Light-colored bars indicate the percentage of all-grade toxicities and include dark-colored bars representing the percentage of grade ≥3 toxicities. It is important to note that, given the heterogeneity of the patients included in these studies, no direct comparison can be made between these different clinical trials.

Bar chart showing the percentage of various side effects for different drugs: crizotinib, entrectinib, repotrectinib, taletrectinib, zidesamtnib, and their respective grades. Side effects include dysgeusia, nausea, vomiting, constipation, diarrhea, peripheral oedema, weight gain, dizziness, muscular weakness, neutropenia, anemia, AST/ALT, blood creatinine, and CPK. Repotrectinib shows the highest percentage for dizziness, while crizotinib is notable for nausea and constipation. Grades three or higher are marked in darker shades. Chemotherapy

Before the advent of targeted therapies, the standard first-line treatment for advanced or metastatic ROS1-positive NSCLC consisted of platinum-based chemotherapy. Among available cytotoxic regimens, pemetrexed associated to platinum chemotherapy was the most frequently used and appeared to confer superior outcomes, achieving longer PFS (median PFS 6.8 months) and higher ORR (45-60%) than when cisplatin or carboplatin was combined with other agents (gemcitabine, docetaxel) (17, 63). Even compared with other oncogene-addicted NSCLC, patients with ROS1-fusion achieved better ORR (57.9%) and PFS (median 7.5 months) with platinum plus pemetrexed chemotherapy (64). A retrospective analysis revealed that long responses to pemetrexed (over a year) were more frequent in the ALK- and ROS1-positive population (65). Mechanistically, this sensitivity may be partly explained by the low expression level of thymidylate synthase mRNA observed in ROS1-positive tumors compared with ROS1 wild-type cancers (63).

After disease progression or in case of contraindication to pemetrexed, other cytotoxic drugs like taxane, gemcitabine, vinorelbine… may be used, following the same approach as in non–oncogene addicted NSCLC. Regarding bevacizumab, no specific data are available for ROS1-positive NSCLC. In different series, some patients received bevacizumab-based regimens, without comparative data to assess its efficacy in this molecular subgroup, and no specific safety signal was reported. The single-arm phase II GFPC 06–2018 study assessing atezolizumab with or without bevacizumab in association with platinum-pemetrexed in EGFR/ALK/ROS1-positive NSCLC failed to include ROS1-positive NSCLC in the bevacizumab cohort (66). Given the lack of specific data, the same principles applied to non–oncogene addicted NSCLC can guide the use of bevacizumab.

 Immunotherapy

Regarding immunotherapy, his use differs from that in non–oncogene-addicted NSCLC. Data on the efficacy of immune checkpoint inhibitors (anti-PD-1/PD-L1) in ROS1-positive NSCLC are limited and largely derived from small retrospective series. PD-L1 expression level appears highly variable, with some studies reporting TPS ≥50% in up to 40%–60% of cases, while others describe uniformly low expression. High expression of PD-L1 is not predictive of immunotherapy efficacy, in fact, his upregulation is oncogene-driven rather than a marker of true tumor immunogenicity. A preclinical study demonstrated in vitro that blocking the ROS1 fusion with a specific siRNA reduced the expression of both ROS1 and PD-L1 on the surface of HCC78 cells. The authors showed that ROS1 oncogene possesses distinct signaling pathways regulating PD-L1 expression, specifically through the SHP2 and c-Jun pathways (67). Tumor mutational burden is generally low (less than 10 mut/Mb), even lower than other NSCLC, including EGFR-mutant NSCLC (68–70).

Clinical activity of anti–PD-1/PD-L1 monotherapy is poor, with ORR ranging from 0% to 20% (68, 69, 71). In the ImmunoTarget study, which retrospectively evaluated the activity of PD-1/PD-L1 inhibitors across oncogene-addicted NSCLC, only 7 patients with ROS1 rearrangements were included with a single objective response observed, corresponding to an ORR of 17% (72). No specific data on PFS or OS were reported for this subgroup. More recently, a meta-analysis of 86 studies concluded that anti-PD-(L)1 as a single agent has limited efficacy in HER2/RET/ROS1-altered NSCLC (71). Overall, the available data highlight that clinical benefit from immune checkpoint inhibitors appears to be rare and inconsistent in ROS1-positive NSCLC, supporting the use of targeted therapies or chemotherapy as preferred treatment strategies in this population (73).

Similarly, very limited data are available regarding the role of chemo-immunotherapy combination in ROS1-positive NSCLC. Although not formally excluded, these patients were likely underrepresented in first-line advanced NSCLC trials. A few retrospective studies have reported chemo-immunotherapy outcomes in ROS1-positive NSCLC, with ORR ranging from 28.6% to 83%, and median PFS varying between 5.8 and 24.3 months, depending on treatment line setting (69, 74). These disparate data do not allow an accurate picture of the contribution of immunotherapy, compared with that of chemotherapy. Given the overall lack of efficacy of immune checkpoint inhibitors in ROS1-positive tumors, the use of immunotherapy in combination with chemotherapy in this setting should not be recommended, especially since the toxicity of TKIs may be increased by prior exposure to immunotherapy.

 Resistance mechanisms to ROS1 TKIs

Despite major clinical advances achieved with TKIs targeting ROS1, the development of resistance remains inevitable for most patients. Resistance mechanisms can be broadly divided into on-target (ROS1-dependent) alterations, which directly affect the kinase domain and impair inhibitor binding, and off-target (ROS1-independent) mechanisms, which reactivate downstream signaling through alternative pathways or induce phenotypic transformation (23, 54).

On-target resistance mechanisms

The most common form of resistance involves secondary point mutations within the ROS1 kinase domain that alter the drug-binding interface. Structural analyses have classified these mutations according to their location and functional impact on the kinase conformation (17, 60).

The solvent-front mutation ROS1 G2032R accounts for approximately 35-40% of all resistance events following crizotinib exposure (4, 38). This substitution replaces a small glycine residue with arginine (larger amino acid) at the entrance of the ATP-binding pocket, creating steric hindrance that prevents the TKI from fitting properly into the catalytic pocket (75). The ROS1 G2032R variant induces high-level resistance to all first- and second-generation inhibitors, including crizotinib, entrectinib, ceritinib, and lorlatinib (47, 76). Novel generation inhibitors (such as taletrectinib) or macrocyclic inhibitors (such as repotrectinib and zidesamtinib) have been rationally designed to minimize bulk within this region, thereby maintaining activity against this sterically restrictive mutation (23, 60).

The gatekeeper mutation ROS1 L2026M, occurring in approximately 5-10% of resistant cases, replaces a small amino acid (leucine) with a larger one (methionine), altering the hydrophobicity and volume of the pocket. This reduces the affinity of TKI without necessarily blocking access completely (54). The gatekeeper mutation acts as a “modified key” within the ATP site, leading to cross-resistance to several TKIs.

Although rarer, there are other clinically relevant mutations, including:

ROS1 D2033N, a hinge-binding site mutation, disrupts hydrogen bonding with inhibitors, resulting in intermediate resistance to crizotinib and lorlatinib, but is often sensitive to repotrectinib.

ROS1 S1986F/Y, activation loop mutation within the αC-helix, stabilizes the active conformation of the kinase, reducing the efficacy of ATP-competitive type I inhibitors, but certain macrocyclic or compact molecules (such as repotrectinib, talectrectinib, or zidesamtinib) can still bind effectively due to their structural flexibility.

ROS1 L2086F or L2000V alter the hydrophobic back pocket. L2086F introduces a bulky aromatic side chain (phenylalanine), which distorts the geometry of the hydrophobic pocket. As a result, ROS1 L2086F confers resistance to all type I ROS1 TKIs (including crizotinib, entrectinib, lorlatinib, repotrectinib, taletrectinib, and zidesamtinib) (18, 47). Type II inhibitors appear to be effective in this condition, based mainly on preclinical data and a few clinical data (54). Similarly, L2000V, located nearby in the C-terminal lobe, alters the local hydrophobic environment and may reduce inhibitor affinity.

Overall, these mutations confer intermediate resistance to crizotinib and entrectinib, frequent resistance to lorlatinib, and remain partially sensitive to next-generation inhibitors such as repotrectinib, taletrectinib, and zidesamtinib. Some patients acquire compound or polyclonal mutations, such as ROS1 G2032R + L2086F or L2026M + D2033N, conferring cross-resistance even to next-generation inhibitors (17).

 Off-target resistance mechanisms

When no secondary ROS1 mutation is identified, resistance is often mediated by bypass pathway activation, which re-establishes downstream signaling independent of ROS1 inhibition. Activation of the EGFR pathway has been observed in vitro and in clinical samples as an early adaptive response to ROS1 inhibition, leading to persistent activation of ERK and AKT despite effective ROS1 blockade (54, 75, 77). Similarly, MET amplification, KRAS mutations, and MAP2K1 (MEK1) activating mutations have been identified in tumor or plasma NGS analyses from patients treated with entrectinib or lorlatinib (54, 78). In some cases, PI3K–AKT–mTOR pathway reactivation has also been described, particularly in preclinical models of chronic TKI exposure (75, 79).

Beyond signaling reactivation, phenotypic transformations also contribute to TKI resistance. The most frequent is epithelial-to-mesenchymal transition (EMT), characterized by the loss of epithelial markers (E-cadherin) and upregulation of mesenchymal factors, such as vimentin and ZEB1 (80). This process promotes cellular plasticity, invasiveness, and drug tolerance. Rarely (in about 2% of cases), small-cell lung cancer (SCLC) transformation has been reported in ROS1-positive tumors following prolonged TKI exposure, mirroring patterns observed in EGFR- and ALK-driven disease (81). SCLC transformation during ROS1 TKI therapy is frequently associated with upstream molecular events, such as loss of RB1 and TP53 mutations, suggesting a lineage plasticity–driven process.

Thus, these mutational and non-mutational resistance mechanisms highlight the adaptive capacity of tumor cells under targeted therapy pressure. The spectrum of acquired ROS1 resistance mutations varies according to the ROS1 fusion partner, the type and generation of TKI used as first-line therapy, reflecting distinct structural constraints and selective pressures imposed by each inhibitor. Collectively, these findings underscore the polyclonal nature of resistance, where multiple subclones may coexist within the same tumor or between metastatic sites. This intrapatient heterogeneity emphasizes the need for repeated molecular profiling at progression to tailor subsequent lines of therapy.

 Therapeutic strategy in <italic>ROS1</italic>-positive NSCLC

The therapeutic strategy for ROS1-positive NSCLC has to integrate available targeted agents, chemotherapy options, and sequencing considerations.

Treatment recommendations should be guided by a multidisciplinary team, taking into account patient age and comorbidities. Patients should be involved in the decision-making process, with particular consideration given to their preferences and the impact of treatment-related adverse effects on quality of life.

Early and locally advanced stages

The management of the early stages of ROS1-positive NSCLC generally follows the same principles as non-oncogene-addicted NSCLC, with surgical resection and, when indicated, perioperative chemotherapy.

Recent advances have established the efficacy of perioperative chemo-immunotherapy in resectable NSCLC, demonstrating an improvement in pathological complete response [CheckMate 816 (82), KEYNOTE-671 (83), CheckMate 77T (84)] and OS benefit (CheckMate 816 (82), Keynote 671 (85)). Patients with ALK or EGFR alterations were excluded from the majority of the phase III trials, and it remains unclear whether ROS1-positive NSCLC were included in these studies. The limited efficacy of immune checkpoint inhibitors in ROS1-positive NSCLC (as discussed above) does not support the use of perioperative immunotherapy in this setting.

However, early stages of NSCLC dependent on oncogenic addiction may benefit from targeted therapies in an adjuvant setting. The ADAURA study paved the way for EGFR mutations by demonstrating that osimertinib given for 3 years after adjuvant chemotherapy, if indicated, provides a benefit in DFS (HR 0.27, 95% CI, 0.21-0.34) and OS (HR 0.49, 95% CI, 0.33-0.73), establishing a new paradigm in this situation. Similarly, after complete resection of ALK-positive NSCLC, both alectinib compared with adjuvant chemotherapy (86) and ensartinib given during 2 years after adjuvant chemotherapy compared with a placebo (Yue et al. ESMO Congress 2025, LBA66) showed a highly significant DFS benefit (HR 0.24, 95% CI, 0.13-0.45 and HR 0.20, 95% CI, 0.11-0.38, respectively).

To date, no clinical trial has established a benefit of ROS1-targeted therapies in an adjuvant setting. Several studies are ongoing. A phase III study of taletrectinib versus placebo as adjuvant therapy is ongoing (TRUST-IV, NCT07154706). A phase II trial evaluating crizotinib in the neoadjuvant setting has been completed, and results are awaited (NCT03088930). The “umbrella” phase II NAUTIKA1 trial (NCT04302025) assesses the efficacy and safety of multiple targeted therapies as perioperative (neoadjuvant and adjuvant) treatment in patients with resectable stage IB–III NSCLC, including ROS1-positive NSCLC treated with entrectinib. Preliminary results have been reported only for ALK-positive NSCLC; further outcomes are pending (87). There are a few case reports describing patients with resectable ROS1-positive NSCLC who received ROS1 TKI in the neoadjuvant setting before surgical resection (88–90). No conclusion can formally be drawn for clinical practice.

In the unresectable locally advanced stage, the treatment approach is similar to that of non–oncogene-addicted NSCLC, consisting of concurrent chemoradiotherapy. In ROS1-positive NSCLC, the preferred choice of chemotherapy would be the combination of cisplatin or carboplatin-pemetrexed, given the data available in stage IV disease. For consolidation treatment after chemoradiation, a similar rationale to that used for EGFR-mutant tumor may apply. Indeed, osimertinib demonstrated in the LAURA phase III trial a significant improvement in PFS (HR 0.16, 95% CI, 0.10-0.24) in this setting and a trend toward OS benefit, although data remain immature (91). It also significantly improved intracranial PFS (HR 0.17, 95% CI, 0.09-0.32), implying a protective effect against brain relapses (92). These findings could support the role of targeted therapies as consolidation treatment post chemoradiotherapy in oncogene-addicted NSCLC, rather than durvalumab (93), for which no specific data are available in ROS1-positive NSCLC. However, to the best of our knowledge, no clinical trial is currently ongoing to specifically address this question.

 Stage IV disease First-line treatment of stage IV ROS1-positive NSCLC

First-line treatment of advanced ROS1-positive NSCLC should ideally involve the most effective and best-tolerated TKI currently available. Similar to the situation with EGFR mutations or ALK rearrangements, the choice of first-line treatment appears crucial in altering the natural history of the disease and improving patient survival. Delaying disease progression by preventing the emergence of resistance mechanisms appears to be more effective than attempting to overcome them. The frequency of brain metastases at baseline and during the disease course warrants the use of a CNS-penetrant TKI, and alternatives to crizotinib should therefore be preferred in this setting. Interestingly, a phase III trial (NCT04603807) compares entrectinib with crizotinib with intracranial PFS as the primary endpoint. The results are not yet available to date.

Furthermore, regulatory approval status is an additional practical consideration, as not all ROS1 inhibitors are available in every country. In the ASCO Guidelines, first-line options mention crizotinib, entrectinib, or repotrectinib; if not available or tolerated, ceritinib or lorlatinib are possible alternatives (94). Updated 2025 ESMO Guidelines offer the options of crizotinib or entrectinib in the first-line setting and repotrectinib as an option (95). The NCCN Guidelines recommended first-line options include crizotinib, entrectinib, repotrectinib, and taletrectinib (96).

The use of new-generation inhibitors more potent against native ROS1 kinase, with high level of CNS penetration appears to be the best current option to improve first-line treatment and try to change the natural history of the disease. Among these, taletrectinib has demonstrated the most favorable efficacy and safety profile (see section 4.1.2); it therefore appears to be the preferred first-line option. A phase III trial is currently ongoing and aims to compare taletrectinib with crizotinib in ROS1-positive locally advanced or metastatic NSCLC previously untreated (TRUST-III, NCT06564324). The tolerance profile of repotrectinib, which has fairly similar activity, makes it more difficult to use over the long term, particularly as a first-line treatment. Zidesamtinib has also shown promising preliminary activity in phase I/II ARROS-1, although data remain immature.

 Beyond the first-line setting

Beyond the first-line setting, the treatment principles are similar to those for advanced NSCLC with oncogenic addiction. The therapeutic sequence mainly depends on the first-line agent used. Whenever possible, patients should be systematically evaluated for eligibility and offered enrollment in a clinical trial, particularly in later-line settings where the level of evidence remains limited.

Most available data concern post-crizotinib setting. Clear treatment guidelines following the first-line use of next-generation ROS1 TKIs cannot currently be defined due to the lack of long-term data.

The clinical pattern of disease progression should be assessed first. In cases of oligoprogression, definitive local therapies (e.g. stereotactic radiotherapy, surgery and thermoablation) should be considered to allow the continuation of the initial TKI beyond progression.

In cases of systemic disease progression, systematic rebiopsy or liquid biopsy is strongly advised whenever feasible. Identification of resistance mechanisms may enable treatment personalization in the presence of targetable alterations. Objective response rates observed with lorlatinib, entrectinib, repotrectinib, taletrectinib, and zidesamtinib after prior ROS1 TKI therapy, mainly crizotinib, are summarized in Table 4. These data remain limited and are derived from relatively small cohorts and should therefore be interpreted with caution. Repotrectinib is currently the preferred option for post-crizotinib treatment according to the ESMO guidelines (95). In the event of CNS progression, switching to a more CNS-penetrant TKI should be prioritized. Preferred options include zidesamtinib, taletrectinib, and repotrectinib. If these options are unavailable, lorlatinib or entrectinib may be considered.

Efficacy of subsequent ROS1 TKIs after progression on a first-line ROS1 TKI.

Drug (reference) Lorlatinib (46) Entrectinib (45) Repotrectinib (57) Taletrectinib (59) Zidesamtinib§ (61)
Post-crizotinib ORR 35% (n=14/40) ORR 11% (n=2/18) ORR 39% (n=18/46) ORR 53% (n=55/103) ORR 68% (n=19/28)
Post-entrectinib (n=1/1)   ORR 22% (n=2/9) ORR 80% (n=8/10) ORR 33% (n=9/27)
Post-repotrectinib NA NA   NA ORR 47% (n=8/17)
Post-taletrectinib NA NA NA   ORR 43% (n=3/7)

Objective response rates are shown for each treatment according to prior ROS1 TKI exposure. Data are extracted from published studies as indicated. It is important to note that, given the heterogeneity of the patients included in these studies, no direct comparison can be made between these different clinical trials.

§ORR refers to patients previously treated with a ROS1 TKI with or without chemotherapy (i.e., one or two prior lines of therapy).

The only data available for disease progression under next-generation ROS1 TKIs come from the preliminary results of the ARROS-1 trial (62). Zidesamtinib showed significant activity after next-generation ROS1 TKIs, whether or not preceded by crizotinib, with an ORR of 38% (95% CI, 26-52) after ≥2 prior ROS1 TKIs ± chemotherapy (Table 4). As zidesamtinib is still not widely available, subsequent treatment should consist of platinum-pemetrexed chemotherapy, without immunotherapy. To the best of our knowledge, no prospective studies have demonstrated the benefit of maintaining ROS1 TKI post-progression in addition to chemotherapy.

At all times during patient management, comprehensive supportive care must accompany systemic treatment to optimize patient outcomes and quality of life.

 Conclusion

ROS1-rearranged NSCLC constitutes a rare but clinically significant molecular subset, underscoring the importance of systematic detection strategies, ideally based on RNA sequencing. The treatment advances made in ROS1-positive disease mirror that of ALK-rearranged tumors, although its lower prevalence has limited the availability of large cohorts and randomized clinical trials. First-generation TKIs with mainly crizotinib and then entrectinib established the initial standard of care; however, the emergence of next-generation ROS1 inhibitors with enhanced potency, central nervous system penetration, and activity against crizotinib resistance mechanisms has markedly improved clinical outcomes. These agents now achieve median progression-free survival durations comparable to those reported with second-generation ALK inhibitors, with promising data emerging for novel compounds such as zidesamtinib, leading to expect a substantial improvement in overall survival in advanced disease settings. Despite these advances, TKIs retain inherent limitations similar with those of other oncogene-addicted NSCLCs with target-independent mechanisms of resistance and tumor heterogeneity, explaining that chemotherapy continues to play a complementary role in disease management. As with ALK, targeted therapies have reshaped the natural history of ROS1-positive NSCLC, with anticipated benefits extending into earlier disease stages. Critical knowledge gaps remain, particularly regarding resistance mechanisms to newer-generation TKIs and the optimal sequencing of available therapeutic strategies.

 Author contributions

MW: Writing – review & editing, Conceptualization, Writing – original draft. RG: Conceptualization, Writing – review & editing, Writing – original draft. MP: Supervision, Data curation, Validation, Writing – review & editing, Conceptualization, Visualization.

 Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The handling editor AA declared a past co-authorship with the author MP.

 Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

 Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc.2026.1739598/full#supplementary-material

Most frequent TRAEs (10% or more, all grades) described with crizotinib, entrectinib, repotrectinib, taletrectinib, and zidesamtinib summarizes the most frequent toxicities of crizotinib (39), entrectinib (45), repotrectinib (58), taletrectinib (59), and zidesamtinib (61) based on the data available to date. As the report and classification of AEs may vary from one trial to another, these data are not comparative.

 References Bray F Laversanne M Sung H Ferlay J Siegel RL Soerjomataram I . Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. (2024) 74:229–63. doi: 10.3322/caac.21834, PMID: 38572751 Rikova K Guo A Zeng Q Possemato A Yu J Haack H . Global survey of phosphotyrosine signaling identifies oncogenic kinases in lung cancer. Cell. (2007) 131:1190–203. doi: 10.1016/j.cell.2007.11.025, PMID: 18083107 Bergethon K Shaw AT Ignatius Ou SH Katayama R Lovly CM McDonald NT . ROS1 rearrangements define a unique molecular class of lung cancers. J Clin Oncol. (2012) 30:863–70. doi: 10.1200/JCO.2011.35.6345, PMID: 22215748 Lin JJ Shaw AT . Recent advances in targeting ROS1 in lung cancer. J Thorac Oncol. (2017) 12:1611–25. doi: 10.1016/j.jtho.2017.08.002, PMID: 28818606 Shibuya M Hanafusa H Balduzzi PC . Cellular Sequences Related to Three New onc Genes of Avian Sarcoma Virus (fps, yes, and ros) and Their Expression in Normal and Transformed Cells. J Virol. (1982) 42:143–52. doi: 10.1128/jvi.42.1.143-152.1982, PMID: 6177868 Matsushime H Shibuya M . Tissue-specific expression of rat c-ros-1 gene and partial structural similarity of its predicted products with sev protein of Drosophila melanogaster. J Virol. (1990) 64:2117–25. doi: 10.1128/jvi.64.5.2117-2125.1990, PMID: 2139140 Roskoski R . ROS1 protein-tyrosine kinase inhibitors in the treatment of ROS1 fusion protein-driven non-small cell lung cancers. Pharmacol Res. (2017) 121:202–12. doi: 10.1016/j.phrs.2017.04.022, PMID: 28465216 Robinson DR Wu YM Lin SF . The protein tyrosine kinase family of the human genome. Oncogene. (2000) 19:5548–57. doi: 10.1038/sj.onc.1203957, PMID: 11114734 Rossi G Jocollé G Conti A Tiseo M Zito Marino F Donati G . Detection of ROS1 rearrangement in non-small cell lung cancer: current and future perspectives. Lung Cancer Targets Ther. (2017) 8:45–55. doi: 10.2147/LCTT.S120172, PMID: 28740441 Kiyozumi D Noda T Yamaguchi R Tobita T Matsumura T Shimada K . NELL2-mediated lumicrine signaling through OVCH2 is required for male fertility. Science. (2020) 368:1132–5. doi: 10.1126/science.aay5134, PMID: 32499443 Acquaviva J Wong R Charest A . The multifaceted roles of the receptor tyrosine kinase ROS in development and cancer. Biochim Biophys Acta BBA - Rev Cancer. (2009) 1795:37–52. doi: 10.1016/j.bbcan.2008.07.006, PMID: 18778756 Chen J Zong CS Wang LH . Tissue and epithelial cell-specific expression of chicken proto-oncogene c-ros in several organs suggests that it may play roles in their development and mature functions. Oncogene. (1994) 9:773–80., PMID: 8108119 Sonnenberg-Riethmacher E Walter B Riethmacher D Gödecke S Birchmeier C . The c-ros tyrosine kinase receptor controls regionalization and differentiation of epithelial cells in the epididymis. Genes Dev. (1996) 10:1184–93. doi: 10.1101/gad.10.10.1184, PMID: 8675006 Birchmeier C Birnbaum D Waitches G Fasano O Wigler M . Characterization of an activated human ros gene. Mol Cell Biol. (1986) 6:3109–16. doi: 10.1128/mcb.6.9.3109-3116.1986, PMID: 3785223 Birchmeier C Sharma S Wigler M . Expression and rearrangement of the ROS1 gene in human glioblastoma cells. Proc Natl Acad Sci. (1987) 84:9270–4. doi: 10.1073/pnas.84.24.9270, PMID: 2827175 Desilets A Repetto M Yang S Drilon A . Targeting ROS1 rearrangements in non–small cell lung cancer: Current insights and future directions. Cancer. (2025) 131:e35784. doi: 10.1002/cncr.35784, PMID: 40171848 Drilon A Jenkins C Iyer S Schoenfeld A Keddy C Davare MA . ROS1-dependent cancers — biology, diagnostics and therapeutics. Nat Rev Clin Oncol. (2021) 18:35–55. doi: 10.1038/s41571-020-0408-9, PMID: 32760015 Zhou S Zhang F Xu M Zhang L Liu Z Yang Q . Novel insights into molecular patterns of ROS1 fusions in a large Chinese NSCLC cohort: a multicenter study. Mol Oncol. (2023) 17:2200–12. doi: 10.1002/1878-0261.13509, PMID: 37584407 Terrones M de Beeck KO Van Camp G Vandeweyer G . Pre-clinical modelling of ROS1+ non-small cell lung cancer. Lung Cancer. (2023) 180:107192. doi: 10.1016/j.lungcan.2023.1071925 Neel DS Allegakoen DV Olivas V Mayekar MK Hemmati G Chatterjee N . Differential subcellular localization regulates oncogenic signaling by ROS1 kinase fusion proteins. Cancer Res. (2019) 79:546–56. doi: 10.1158/0008-5472.CAN-18-1492, PMID: 30538120 Li Z Shen L Ding D Huang J Zhang J Chen Z . Efficacy of crizotinib among different types of ROS1 fusion partners in patients with ROS1-rearranged non–small cell lung cancer. J Thorac Oncol. (2018) 13:987–95. doi: 10.1016/j.jtho.2018.04.016, PMID: 29704675 Hofman V Rouquette I Long-Mira E Piton N Chamorey E Heeke S . Multicenter evaluation of a novel ROS1 immunohistochemistry assay (SP384) for detection of ROS1 rearrangements in a large cohort of lung adenocarcinoma patients. J Thorac Oncol. (2019) 14:1204–12. doi: 10.1016/j.jtho.2019.03.024, PMID: 30999109 Boulanger MC Schneider JL Lin JJ . Advances and future directions in ROS1 fusion-positive lung cancer. Oncologist. (2024) 29:943–56. doi: 10.1093/oncolo/oyae205, PMID: 39177972 Yoshida A Tsuta K Wakai S Arai Y Asamura H Shibata T . Immunohistochemical detection of ROS1 is useful for identifying ROS1 rearrangements in lung cancers. Mod Pathol. (2014) 27:711–20. doi: 10.1038/modpathol.2013.192, PMID: 24186139 Bozzetti C Nizzoli R Tiseo M Squadrilli A Lagrasta C Buti S . ALK and ROS1 rearrangements tested by fluorescence in situ hybridization in cytological smears from advanced non-small cell lung cancer patients. Diagn Cytopathol. (2015) 43:941–6. doi: 10.1002/dc.23318, PMID: 26152804 Cao B Wei P Liu Z Bi R Lu Y Zhang L . Detection of lung adenocarcinoma with ROS1 rearrangement by IHC, FISH, and RT-PCR and analysis of its clinicopathologic features. OncoTargets Ther. (2015) 9:131–8. doi: 10.2147/OTT.S94997, PMID: 26770062 Hendriks LE Kerr KM Menis J Mok TS Nestle U Passaro A . Oncogene-addicted metastatic non-small-cell lung cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann Oncol. (2023) 34:339–57. doi: 10.1016/j.annonc.2022.12.009, PMID: 36872130 Swalduz A Schiffler C Curcio H Ambasager B Le Moel G Debieuvre D . LIBELULE: A randomized phase III study to evaluate the clinical relevance of early liquid biopsy in patients with suspicious metastatic lung cancer. J Thorac Oncol. (2025) 20:437–50. doi: 10.1016/j.jtho.2024.12.011, PMID: 39694415 Dagogo-Jack I Rooney M Nagy RJ Lin JJ Chin E Ferris LA . Molecular analysis of plasma from patients with ROS1-positive NSCLC. J Thorac Oncol. (2019) 14:816–24. doi: 10.1016/j.jtho.2019.01.009, PMID: 30664990 Mayer C Ofek E Fridrich DE Molchanov Y Yacobi R Gazy I . Direct identification of ALK and ROS1 fusions in non-small cell lung cancer from hematoxylin and eosin-stained slides using deep learning algorithms. Mod Pathol. (2022) 35:1882–7. doi: 10.1038/s41379-022-01141-4, PMID: 36057739 Park S Ahn BC Lim SW Sun JM Kim HR Hong MH . Characteristics and outcome of ROS1-positive non-small cell lung cancer patients in routine clinical practice. J Thorac Oncol. (2018) 13:1373–82. doi: 10.1016/j.jtho.2018.05.026, PMID: 29883837 Dugay F Llamas-Gutierrez F Gournay M Medane S Mazet F Chiforeanu DC . Clinicopathological characteristics of ROS1- and RET-rearranged NSCLC in caucasian patients: Data from a cohort of 713 non-squamous NSCLC lacking KRAS/EGFR/HER2/BRAF/PIK3CA/ALK alterations. Oncotarget. (2017) 8:53336–51. doi: 10.18632/oncotarget.18408, PMID: 28881815 Gendarme S Bylicki O Chouaid C Guisier F . ROS-1 fusions in non-small-cell lung cancer: evidence to date. Curr Oncol Tor Ont. (2022) 29:641–58. doi: 10.3390/curroncol29020057, PMID: 35200557 Chen YF Hsieh MS Wu SG Chang YL Shih JY Liu YN . Clinical and the prognostic characteristics of lung adenocarcinoma patients with ROS1 fusion in comparison with other driver mutations in East Asian populations. J Thorac Oncol. (2014) 9:1171–9. doi: 10.1097/JTO.0000000000000232, PMID: 25157770 Woo JH Kim TJ Kim TS Han J . CT features and disease spread patterns in ROS1-rearranged lung adenocarcinomas: comparison with those of EGFR-mutant or ALK-rearranged lung adenocarcinomas. Sci Rep. (2020) 10:16251. doi: 10.1038/s41598-020-73533-y, PMID: 33005033 Patil T Smith DE Bunn PA Aisner DL Le AT Hancock M . The incidence of brain metastases in stage IV ROS1-rearranged non-small cell lung cancer and rate of central nervous system progression on crizotinib. J Thorac Oncol. (2018) 13:1717–26. doi: 10.1016/j.jtho.2018.07.001, PMID: 29981925 Zhu VW Zhao JJ Gao Y Syn NL Zhang SS Ou SHI . Thromboembolism in ALK+ and ROS1+ NSCLC patients: A systematic review and meta-analysis. Lung Cancer Amst Neth. (2021) 157:147–55. doi: 10.1016/j.lungcan.2021.05.019, PMID: 34049720 Ou SHI Hagopian GG Zhang SS Nagasaka M . Comprehensive review of ROS1 tyrosine kinase inhibitors-classified by structural designs and mutation spectrum (Solvent front mutation [G2032R] and central β-sheet 6 [Cβ6] mutation [L2086F]). J Thorac Oncol. (2024) 19:706–18. doi: 10.1016/j.jtho.2023.12.008, PMID: 38070596 Shaw AT Riely GJ Bang YJ Kim DW Camidge DR Solomon BJ . Crizotinib in ROS1-rearranged advanced non-small-cell lung cancer (NSCLC): updated results, including overall survival, from PROFILE 1001. Ann Oncol. (2019) 30:1121–6. doi: 10.1093/annonc/mdz131, PMID: 30980071 Shaw AT Ou SHI Bang YJ Camidge DR Solomon BJ Salgia R . Crizotinib in ROS1-rearranged non–small-cell lung cancer. In: The new england journal of medicine. Massachusetts Medical Society (2014). doi: 10.1056/NEJMoa1406766, PMID: 25264305 Michels S Massutí B Schildhaus HU Franklin J Sebastian M Felip E . Safety and efficacy of crizotinib in patients with advanced or metastatic ROS1-rearranged lung cancer (EUCROSS): A european phase II clinical trial. J Thorac Oncol. (2019) 14:1266–76. doi: 10.1016/j.jtho.2019.03.020, PMID: 30978502 Moro-Sibilot D Cozic N Pérol M Mazières J Otto J Souquet PJ . Crizotinib in c-MET- or ROS1-positive NSCLC: results of the AcSé phase II trial. Ann Oncol. (2019) 30:1985–91. doi: 10.1093/annonc/mdz407, PMID: 31584608 Wu YL Yang JCH Kim DW Lu S Zhou J Seto T . Phase II study of crizotinib in east asian patients with ROS1-positive advanced non–small-cell lung cancer. J Clin Oncol. (2018) 36:1405–11. doi: 10.1200/JCO.2017.75.5587, PMID: 29596029 Landi L Chiari R Tiseo M D’Incà F Dazzi C Chella A . Crizotinib in MET-deregulated or ROS1-rearranged pretreated non–small cell lung cancer (METROS): A phase II, prospective, multicenter, two-arms trial. Clin Cancer Res. (2019) 25:7312–9. doi: 10.1158/1078-0432.CCR-19-0994, PMID: 31416808 Drilon A Chiu CH Fan Y Cho BC Lu S Ahn MJ . Long-term efficacy and safety of entrectinib in ROS1 fusion–positive NSCLC. JTO Clin Res Rep. (2022) 3:100332. doi: 10.1016/j.jtocrr.2022.100332, PMID: 35663414 Shaw AT Solomon BJ Chiari R Riely GJ Besse B Soo RA . Lorlatinib in advanced ROS1-positive non-small-cell lung cancer: a multicentre, open-label, single-arm, phase 1–2 trial. Lancet Oncol. (2019) 20:1691–701. doi: 10.1016/S1470-2045(19)30655-2, PMID: 31669155 Fabbri L Di Federico A Astore M Marchiori V Rejtano A Seminerio R . From development to place in therapy of lorlatinib for the treatment of ALK and ROS1 rearranged non-small cell lung cancer (NSCLC). Diagnostics. (2023) 14:48. doi: 10.3390/diagnostics14010048, PMID: 38201357 Yang JJ Zhou J Liu SYM Li M Zhang Z Cheng Y . Foritinib in advanced ROS1-rearranged non-small-cell lung cancer in China: a multicentre, open-label, single-arm, phase 2 study. Lancet Respir Med. (2024) 12:671–80. doi: 10.1016/S2213-2600(24)00171-1, PMID: 39059398 Lim SM Kim HR Lee JS Lee KH Lee YG Min YJ . Open-label, multicenter, phase II study of ceritinib in patients with non–small-cell lung cancer harboring ROS1 rearrangement. J Clin Oncol. (2017) 35:2613–8. doi: 10.1200/JCO.2016.71.3701, PMID: 28520527 Niho S Goto Y Toyozawa R Daga H Ohashi K Takahashi T . Phase II study of brigatinib in patients with ROS1 fusion-positive non-small-cell lung cancer: the Barossa study. ESMO Open. (2024) 9:103642. doi: 10.1016/j.esmoop.2024.103642, PMID: 39018589 Katayama R Kobayashi Y Friboulet L Lockerman EL Koike S Shaw AT . Cabozantinib overcomes crizotinib resistance in ROS1 fusion-positive cancer. Clin Cancer Res. (2015) 21:166–74. doi: 10.1158/1078-0432.CCR-14-1385, PMID: 25351743 Weber U Davies KD Camidge DR . L2086F mutant ROS1-rearranged NSCLC resistant to repotrectinib responds to cabozantinib: A case report. JTO Clin Res Rep. (2024) 5:100673. doi: 10.1016/j.jtocrr.2024.100673, PMID: 39091594 Sakamoto M Patil T . Exceptional response to lorlatinib and cabozantinib in ROS1-rearranged NSCLC with acquired F2004V and L2086F resistance. NPJ Precis Oncol. (2023) 7:56. doi: 10.1038/s41698-023-00381-0, PMID: 37291202 Lin JJ Choudhury NJ Yoda S Zhu VW Johnson TW Sakhtemani R . Spectrum of mechanisms of resistance to crizotinib and lorlatinib in ROS1 fusion-positive lung cancer. Clin Cancer Res. (2021) 27:2899–909. doi: 10.1158/1078-0432.CCR-21-0032, PMID: 33685866 Lu S Pan H Wu L Yao Y He J Wang Y . MA06.04 unecritinib in patients with ROS1 positive advanced non-small cell lung cancer: updated results from a phase II trial. J Thorac Oncol. (2024) 19:S73–4. doi: 10.1016/j.jtho.2024.09.132 Lu S Pan H Wu L Yao Y He J Wang Y . Efficacy, safety and pharmacokinetics of Unecritinib (TQ-B3101) for patients with ROS1 positive advanced non-small cell lung cancer: a Phase I/II Trial. Signal Transduct Target Ther. (2023) 8:249. doi: 10.1038/s41392-023-01454-z, PMID: 37385995 Cho BC Lin J Camidge DR Nagasaka M Kim SW Solomon B . MA02.03 repotrectinib in patients with ROS1 fusion-positive (ROS1+) NSCLC: long-term follow-up from the phase 1/2 TRIDENT-1 trial. J Thorac Oncol. (2025) 20:S56–7. doi: 10.1016/j.jtho.2023.09.035 Drilon A Camidge DR Lin JJ Kim SW Solomon BJ Dziadziuszko R . Repotrectinib in ROS1 fusion-positive non-small-cell lung cancer. N Engl J Med. (2024) 390:118–31. doi: 10.1056/NEJMoa2302299, PMID: 38197815 Pérol M Li W Pennell NA Liu G Ohe Y De Braud F . Taletrectinib in ROS1+ Non–small cell lung cancer: TRUST. J Clin Oncol. (2025) 43:1920–9. doi: 10.1200/JCO-25-00275, PMID: 40179330 Drilon A Horan JC Tangpeerachaikul A Besse B Ou SHI Gadgeel SM . NVL-520 is a selective, TRK-sparing, and brain-penetrant inhibitor of ROS1 fusions and secondary resistance mutations. Cancer Discov. (2023) 13:598–615. doi: 10.1158/2159-8290.CD-22-0968, PMID: 36511802 Drilon AE Cho BC Lin JJ Solomon BJ Lin CC Langen AJd . PL02.15 pivotal ARROS-1 efficacy and safety data: zidesamtinib in TKI pre-treated patients with advanced/metastatic ROS1+ NSCLC. J Thorac Oncol. (2025) 20:S3. doi: 10.1016/j.jtho.2025.09.018 Besse B Drilon AE Cho BC Camidge DR Neal J Lin C-C . 1256MO Phase I/II ARROS-1 study of zidesamtinib (NVL-520) in ROS1 fusion-positive solid tumours. Ann Oncol. 35:S804-5. doi: 10.1016/j.annonc.2024.08.1313 Song Z Su H Zhang Y . Patients with ROS1 rearrangement-positive non-small-cell lung cancer benefit from pemetrexed-based chemotherapy. Cancer Med. (2016) 5:2688–93. doi: 10.1002/cam4.809, PMID: 27544536 Chen YF Hsieh MS Wu SG Chang YL Yu CJ Yang JCH . Efficacy of pemetrexed-based chemotherapy in patients with ROS1 fusion-positive lung adenocarcinoma compared with in patients harboring other driver mutations in east asian populations. J Thorac Oncol. (2016) 11:1140–52. doi: 10.1016/j.jtho.2016.03.022, PMID: 27094798 Liang Y Wakelee HA Neal JW . Relationship of driver oncogenes to long-term pemetrexed response in non–small-cell lung cancer. Clin Lung Cancer. (2015) 16:366–73. doi: 10.1016/j.cllc.2014.12.009, PMID: 25665893 Bylicki O Tomasini P Radj G Guisier F Monnet I Ricordel C . Atezolizumab with or without bevacizumab and platinum-pemetrexed in patients with stage IIIB/IV non-squamous non-small cell lung cancer with EGFR mutation, ALK rearrangement or ROS1 fusion progressing after targeted therapies: A multicentre phase II open-label non-randomised study GFPC 06-2018. Eur J Cancer Oxf Engl 1990. (2023) 183:38–48. doi: 10.1016/j.ejca.2023.01.014, PMID: 36801605 Cai L Duan J Qian L Wang Z Wang S Li S . ROS1 fusion mediates immunogenicity by upregulation of PD-L1 after the activation of ROS1–SHP2 signaling pathway in non-small cell lung cancer. Front Immunol. (2020) 11:527750/full. doi: 10.3389/fimmu.2020.527750/full, PMID: 33324391 Dantoing E Piton N Salaün M Thiberville L Guisier F . Anti-PD1/PD-L1 immunotherapy for non-small cell lung cancer with actionable oncogenic driver mutations. Int J Mol Sci. (2021) 22:6288. doi: 10.3390/ijms22126288, PMID: 34208111 Choudhury NJ Schneider JL Patil T Zhu VW Goldman DA Yang SR . Response to immune checkpoint inhibition as monotherapy or in combination with chemotherapy in metastatic ROS1-rearranged lung cancers. JTO Clin Res Rep. (2021) 2:100187. doi: 10.1016/j.annonc.2024.08.1313 Rangachari D VanderLaan PA Shea M Le X Huberman MS Kobayashi SS . Correlation between classic driver oncogene mutations in EGFR, ALK, or ROS1 and 22C3-PD-L1 ≥50% Expression in lung adenocarcinoma. J Thorac Oncol. (2017) 12:878–83. doi: 10.1016/j.jtho.2016.12.026, PMID: 28104537 Chen J Lu W Chen M Cai Z Zhan P Liu X . Efficacy of immunotherapy in patients with oncogene-driven non-small-cell lung cancer: a systematic review and meta-analysis. Ther Adv Med Oncol. (2024) 16:17588359231225036. doi: 10.1177/17588359231225036, PMID: 38420602 Mazieres J Drilon A Lusque A Mhanna L Cortot AB Mezquita L . Immune checkpoint inhibitors for patients with advanced lung cancer and oncogenic driver alterations: results from the IMMUNOTARGET registry. Ann Oncol. (2019) 30:1321–8. doi: 10.1093/annonc/mdz167, PMID: 31125062 Mushtaq R Cortot AB Gautschi O Mazieres J Camidge DR . PD-1/PD-L1 inhibitor activity in patients with gene-rearrangement positive non-small cell lung cancer-an IMMUNOTARGET case series. Transl Lung Cancer Res. (2022) 11:2412–7. doi: 10.21037/tlcr-22-329, PMID: 36636412 Huang Z Yan H Zeng L Xu Q Guo W Lin S . Efficacy of immune checkpoint inhibitor plus chemotherapy in patients with ROS1-rearranged advanced lung adenocarcinoma: A multicenter, retrospective cohort study. JCO Precis Oncol. (2023) 7:e2200614. doi: 10.1200/PO.22.00614, PMID: 36952645 McCoach CE Le AT Gowan K Jones K Schubert L Doak A . Resistance mechanisms to targeted therapies in ROS1+ and ALK+ Non-small cell lung cancer. Clin Cancer Res. (2018) 24:3334–47. doi: 10.1158/1078-0432.CCR-17-2452, PMID: 29636358 Awad MM Katayama R McTigue M Liu W Deng YL Brooun A . Acquired resistance to crizotinib from a mutation in CD74–ROS1. In: The new england journal of medicine. Massachusetts Medical Society (2013). doi: 10.1056/NEJMoa1215530, PMID: 23724914 Lin JJ Ritterhouse LL Ali SM Bailey M Schrock AB Gainor JF . ROS1 fusions rarely overlap with other oncogenic drivers in non–small cell lung cancer. J Thorac Oncol. (2017) 12:872–7. doi: 10.1016/j.jtho.2017.01.004, PMID: 28088512 Tyler LC Le AT Chen N Nijmeh H Bao L Wilson TR . MET gene amplification is a mechanism of resistance to entrectinib in ROS1+ NSCLC. Thorac Cancer. (2022) 13:3032–41. doi: 10.1111/1759-7714.14656, PMID: 36101520 D’Angelo A Sobhani N Chapman R Bagby S Bortoletti C Traversini M . Focus on ROS1-positive non-small cell lung cancer (NSCLC): crizotinib, resistance mechanisms and the newer generation of targeted therapies. Cancers. (2020) 12:3293. doi: 10.3390/cancers12113293, PMID: 33172113 Lee JK Lee J Kim S Kim S Youk J Park S . Clonal history and genetic predictors of transformation into small-cell carcinomas from lung adenocarcinomas. J Clin Oncol. (2017) 35:3065–74. doi: 10.1200/JCO.2016.71.9096, PMID: 28498782 Lin JJ Langenbucher A Gupta P Yoda S Fetter IJ Rooney M . Small cell transformation of ROS1 fusion-positive lung cancer resistant to ROS1 inhibition. NPJ Precis Oncol. (2020) 4:21. doi: 10.1038/s41698-020-0127-9, PMID: 32802958 Forde PM Spicer J Lu S Provencio M Mitsudomi T Awad MM . Neoadjuvant nivolumab plus chemotherapy in resectable lung cancer. N Engl J Med. (2022) 386:1973–85. doi: 10.1056/NEJMoa2202170, PMID: 35403841 Wakelee H Liberman M Kato T Tsuboi M Lee SH Gao S . Perioperative pembrolizumab for early-stage non–small-cell lung cancer. N Engl J Med. (2023) 389:491–503. doi: 10.1056/NEJMoa2302983, PMID: 37272513 Cascone T Awad MM Spicer JD He J Lu S Sepesi B . Perioperative nivolumab in resectable lung cancer. N Engl J Med. (2024) 390:1756–69. doi: 10.1056/NEJMoa2311926, PMID: 38749033 Spicer JD Garassino MC Wakelee H Liberman M Kato T Tsuboi M . Neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab compared with neoadjuvant chemotherapy alone in patients with early-stage non-small-cell lung cancer (KEYNOTE-671): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. (2024) 404:1240–52. doi: 10.1016/S0140-6736(24)01756-2, PMID: 39288781 Wu YL Dziadziuszko R Ahn JS Barlesi F Nishio M Lee DH . Alectinib in resected ALK-positive non-small-cell lung cancer. N Engl J Med. (2024) 390:1265–76. doi: 10.1056/NEJMoa2310532, PMID: 38598794 Lee JM McNamee CJ Toloza E Negrao MV Lin J Shum E . Neoadjuvant targeted therapy in resectable NSCLC: current and future perspectives. J Thorac Oncol. (2023) 18:1458–77. doi: 10.1016/j.jtho.2023.07.006, PMID: 37451404 Xiao S Shi Y Li R . Surgical treatment of IIIb (N2), ROS1(+) non-small cell lung cancer after neoadjuvant chemotherapy combined with targeted therapy: a case report. AME Case Rep. (2025) 9:83. doi: 10.21037/acr-24-151, PMID: 40761203 Chen R Wang G Zhao J Liu Y . Efficacy of neoadjuvant therapy for ROS1-positive locally-advanced lung adenocarcinoma: A case report. Am J Case Rep. (2024) 25:e945446. doi: 10.12659/AJCR.945446, PMID: 39716454 Fan M Yan W Zhuo M Yu R Qi L Yang X . Feasibility and safety of perioperative entrectinib in a patient with ROS1-positive locally advanced NSCLC: A case report. Lung Cancer Amst Neth. (2025) 202:108501. doi: 10.1016/j.lungcan.2025.108501, PMID: 40120333 Lu S Kato T Dong X Ahn MJ Quang LV Soparattanapaisarn N . Osimertinib after chemoradiotherapy in stage III EGFR-mutated NSCLC. N Engl J Med. (2024) 391:585–97. doi: 10.1056/NEJMoa2402614, PMID: 38828946 Lu S Ahn MJ Reungwetwattana T Özgüroğlu M Kato T Yang JCH . Osimertinib after definitive chemoradiotherapy in unresectable stage III epidermal growth factor receptor-mutated non-small-cell lung cancer: analyses of central nervous system efficacy and distant progression from the phase III LAURA study. Ann Oncol. (2024) 35:1116–25. doi: 10.1016/j.annonc.2024.08.2243, PMID: 39289145 Antonia SJ Villegas A Daniel D Vicente D Murakami S Hui R . Durvalumab after chemoradiotherapy in stage III non–small-cell lung cancer. N Engl J Med. (2017) 377:1919–29. doi: 10.1056/NEJMoa1709937, PMID: 28885881 Owen DH Ismaila N Ahluwalia A Feldman J Gadgeel S Mullane M . Therapy for stage IV non-small cell lung cancer with driver alterations: ASCO living guideline, version 2024.3. J Clin Oncol. (2025) 43:e2–16. doi: 10.1200/JCO-24-02785, PMID: 40014839 Hendriks LEL Cortiula F Martins-Branco D Mariamidze E Popat S Reck M . Updated treatment recommendations for systemic treatment: from the ESMO oncogene-addicted metastatic NSCLC living guideline. Ann Oncol. (2025) 36:1227–31. doi: 10.1016/j.annonc.2025.06.007, PMID: 40571160 Riely GJ Wood DE Aisner DL Loo BW Axtell AL Bauman JR . NCCN guidelines® Insights: non-small cell lung cancer, version 7.2025. J Natl Compr Cancer Netw JNCCN. (2025) 23:354–62. doi: 10.6004/jnccn.2025.0043, PMID: 40930154 GlossaryAE

adverse event

AKT

protein kinase B

ALK

anaplastic lymphoma kinase

ALT

alanine transaminase

AST

aspartate transaminase

BCRP

breast cancer resistance protein

BID

twice daily

caps

capsules

CPK

creatine phosphokinase

CYP

cytochrome P450 enzymes

DNA-NGS

DNA-based next-generation sequencing

DOR

duration of response

EMA

European Medicines Agency

EMT

epithelial–mesenchymal transition

ERK

extracellular signal-regulated kinase

ESMO

European Society for Medical Oncology

FDA RTOR

U.S. Food and Drug Administration Real-Time Oncology Review

FISH

fluorescence in situ hybridization

FLT3

Fms-like tyrosine kinase 3

Gen

generation of the tyrosine kinase inhibitor (first, second, or third)

IC-ORR

intracranial objective response rate

IHC

immunohistochemistry

JAK

Janus kinase

KIT

KIT proto-oncogene receptor tyrosine kinase

MATE1

multidrug and toxin extrusion protein 1

MEK

mitogen-activated protein kinase kinase

MET

mesenchymal–epithelial transition factor

mDOR

median duration of response

mo

months

mOS

median overall survival

mPFS

median progression-free survival

mTOR

mechanistic target of rapamycin

NA

not available

NCCN

National Comprehensive Cancer Network

NGS

next-generation sequencing

NSCLC

non–small cell lung cancer

NTRK

neurotrophic tyrosine receptor kinase

ORR

objective response rate

PFS

progression-free survival

P-gp

P-glycoprotein

PI3K

phosphatidylinositol 3-kinase

PK

pharmacokinetics

QD

once daily

RET

rearranged during transfection proto-oncogene

RNA-NGS

RNA-based next-generation sequencing

ROS1

c-ros oncogene 1 receptor tyrosine kinase

RT-PCR

reverse transcriptase polymerase chain reaction

RW

real world

SHP2

Src homology phosphatase 2

STAT

signal transducer and activator of transcription

tabs

tablets

TKI

tyrosine kinase inhibitor

Type I/II

binding mode of the inhibitor (Type I: ATP-competitive active conformation

VEGFR2

vascular endothelial growth factor receptor 2.

Edited by: Alfredo Addeo, Hôpitaux universitaires de Genève (HUG), Switzerland

Reviewed by: Chengwei Jin, Zibo Central Hospital, China

Urs Michael Weber, University of Colorado, United States