The Gap Between What Oncologists Have Been Prescribing and What the ROS1 Data Now Supports

ROS1 gene fusions occur in roughly 1% to 2% of advanced non–small cell lung cancer (NSCLC) cases, and the patients who harbor them—typically younger never-smokers with adenocarcinoma—now have access to 4 approved tyrosine kinase inhibitors (TKIs). Yet for many community oncologists, these patients arrive so infrequently that selecting among crizotinib (Xalkori), entrectinib (Rozlytrek), repotrectinib (Augtyro), and taletrectinib (Ibtrozi) requires real-time navigation of cross-trial comparisons rather than lived clinical experience. What surfaced during a recent roundtable was a practical tension: the drugs oncologists have been using most often are not the same as the drugs they now say they would choose for a new patient — and the data largely explain why.

In a virtual Case-Based Roundtable event for oncologists in the eastern United States, Bruna Pellini, MD, chief of Thoracic Medical Oncology at Baptist Health Herbert Wertheim Cancer Institute and Associate Professor at Herbert Wertheim College of Medicine at Florida International University, reviewed the efficacy and safety landscape across all approved ROS1 TKIs and moderated 2 patient cases. Pellini framed the central challenge of this disease as one of sequencing under uncertainty: with no head-to-head trials and a patient population too small for most community practices to accumulate meaningful personal experience, clinicians must rely on cross-trial extrapolation to choose among agents whose toxicity profiles are as distinct as their response curves.

The discussion centered on a 58-year-old never-smoker who presented with a persistent cough, mild dyspnea, and right-sided chest discomfort. Imaging revealed a solid 7-cm right mid-lobe mass with ipsilateral mediastinal lymphadenopathy and PET-avid uptake extending from L3 to L5 through bilateral iliac crests and ischial spine. Brain MRI was negative. Biopsy confirmed adenocarcinoma, and immunohistochemistry of the FFPE specimen showed elevated ROS1 protein with an H-score greater than 150 and a PD-L1 tumor proportion score of 95%. Molecular profiling identified a CD74-ROS1 gene fusion with no other actionable co-mutations. ECOG performance status was 1.

Before the data review, participants were asked what first-line agent their most recent ROS1-positive patient had actually received. The answers reflected a field mid-transition: entrectinib and taletrectinib were each used by a third of respondents, repotrectinib by 13%, crizotinib by 7%, and 2 participants answered "other." After the case was presented but before the trial data were reviewed, participants voted on what they would offer this patient today; 60% chose taletrectinib, 27% repotrectinib, and just 7% entrectinib. The gap between historical prescribing and current preference was the quiet story of the evening.

Pellini reviewed the foundational efficacy data in chronological order. In the PROFILE 1001 study (NCT00585195), crizotinib produced an objective response rate (ORR) of 72% (95% CI, 58%–83%) in 53 patients with ROS1-rearranged NSCLC, with a median progression-free survival (PFS) of 19.3 months (95% CI, 15.2–39.1) and a median overall survival (OS) of 51.4 months (95% CI, 29.3–not reached [NR]).¹ Those numbers remained the reference point for the next several years. The integrated analysis of the ALKA-372-001 and STARTRK-1/2 trials supporting entrectinib approval showed an ORR of 68.2% (95% CI, 57.2–77.9) in the first-line population of 85 patients, with a median PFS of 17.5 months (95% CI, 12.0–30.4) overall, falling below 12 months in the subset with baseline central nervous system (CNS) metastases. Median OS was 52.3 months (95% CI, 44.6–not evaluable [NE]), comparable to crizotinib.²

The efficacy signal shifted meaningfully with repotrectinib. In the phase 2 TRIDENT-1 trial (NCT03093116), TKI-naive patients treated in the expansion cohort achieved a confirmed ORR of 79% and a median PFS of 31.1 months (95% CI, 21.9-NE), nearly doubling the PFS seen with the earlier agents. Among patients with measurable brain metastases at baseline, the intracranial confirmed ORR was 89%. Pellini noted that when this data first appeared in 2023, the jump in PFS was striking even in the context of cross-trial comparison.³

The most recent addition to the landscape is taletrectinib, studied in the TRUST-I (NCT04395677) and TRUST-II (NCT04919811) trials.⁴ The pooled TKI-naive efficacy population of 157 patients showed a confirmed ORR of 89.8% (95% CI, 84.0%-94.1%), with an intracranial confirmed ORR of 76.5% in patients with measurable brain metastases. In the larger and more mature TRUST-I cohort, median PFS reached 49.6 months (95% CI, 34.5-NR) and median duration of response was 49.7 months (95% CI, 41.3-NR). Median OS in TKI-naive patients was not reached. Pellini described the near-overlap of median PFS and median duration of response as unusual and worth watching as follow-up continues to mature.

For participants, the conversation turned quickly from efficacy to tolerability , and here the differences among agents were more vivid than the trial tables convey. Pellini described a patient who had received entrectinib and told her he would rather take cisplatin again than remain on that drug, attributing the toxicity primarily to off-target NTRK inhibition driving CNS and peripheral neurotoxicity. Repotrectinib, which requires a 14-day once-daily lead-in before escalating to the twice-daily recommended dose, carries a similar burden.

"The paresthesia really impairs life, and the ataxia—when it happens, it is striking," Pellini said, recounting a patient in his mid 80s who required repeated hospitalizations for falls while receiving repotrectinib and ultimately could not continue any treatment. The dose reduction rates across all 4 agents are broadly similar—roughly 19% to 38% of patients require a reduction—but the type of toxicity driving those reductions is not.

Baidehi Maiti, MD, medical oncologist oncologist at Cleveland Clinic in Lakewood, Ohio, summarized what the group had coalesced around: "With taletrectinib, the cognitive impairment and the ataxia are lower than with repotrectinib, so that may be one reason to gravitate toward taletrectinib, but overall I completely agree it is the safety, efficacy, comfort level, and tolerability." CNS toxicity was identified as the most challenging adverse event to manage across ROS1 TKIs by 64% of participants. Pellini connected this back to mechanism: agents with greater off-target NTRK inhibition produce more neurotoxicity, and taletrectinib's selectivity for ROS1 over NTRK translates directly into the clinical experience. The most common treatment-emergent adverse events with taletrectinib are elevated liver enzymes in 76% of patients, diarrhea in 63%, and hemoglobin decreases in 48%—events that are largely laboratory-flagged and manageable without the quality-of-life impact of dizziness, ataxia, or memory impairment.

Adriana Alvarez, MD, medical oncologist at Cleveland Clinic in Avon, Ohio, raised a practical question the data do not directly answer: whether to reserve a newer agent for a patient who might tolerate earlier toxicity better, given that sequencing is still possible. "The only caveat is, what do we use for second line?" Pellini answered with the arithmetic: a median PFS of 49 months in the first-line taletrectinib cohort vs a median PFS of roughly 7.6 to 11.8 months in the second-line taletrectinib cohorts. Even adding a generous second-line PFS of 8 months to repotrectinib's first-line 31 months yields a theoretical total of about 39 months, still shorter than the 49-month first-line PFS achieved by taletrectinib, and that calculation assumes the patient reaches and tolerates a second line, which is not guaranteed. The group converged around the principle: use your most effective and best-tolerated drug first, because the return on sequencing diminishes faster than the curves suggest.

The high PD-L1 tumor proportion score of 95% in the case patient generated a brief but pointed exchange. Without prompting, multiple participants said it would not influence their treatment decision. Pellini agreed, characterizing the PD-L1 positivity in ROS1-driven tumors as a biological artifact of a cold tumor microenvironment rather than a signal of immunotherapy benefit. Patients with this alteration derive minimal benefit from immune checkpoint inhibitors, and no participant considered adding or substituting immunotherapy.

For a disease where most community oncologists see 2 or 3 patients per year, the question of how to access a drug urgently is not hypothetical. Pellini described two patients with significant symptom burden, including 1 with a pericardial effusion, for whom taletrectinib was delivered within 24 to 48 hours through Nuvation Bio's patient assistance program after direct outreach to the company's commercial representative. Maiti and Bachar Dergham, MD, medical oncologist at Cleveland Clinic in Independence, Ohio, noted that specialty pharmacy turnaround typically runs 1 to 2 weeks, and that copay assistance programs, though available, can add weeks of delay for patients with commercial insurance gaps. Pellini's practical message: know the company representative and escalate early.

Whether the next ROS1 TKI—zidesamtinib, currently available only on trial or through expanded access—will further shift the frontline standard remains to be seen. For now, the community oncologists in this discussion left with a clear sense that the data have moved ahead of practice patterns, and that for newly diagnosed patients, the window for using the most effective and best-tolerated agent is the first line of therapy.

Register today to join a Case-Based Roundtable near you.

DISCLOSURES: Pellini reported employment at Miami Cancer Institute; honoraria from Foundation Medicine, Merck, AstraZeneca, Regeneron, Gilead Sciences, Lilly, Bayer, Amgen, Boehringer Ingelheim, OncoHost, Jazz Pharmaceuticals; consulting or advisory roles with AstraZeneca, Regeneron, Bristol Myers Squibb/Roche, Bayer, Gilead Sciences, Foundation Medicine, OncoHost, Catalyst Pharmaceuticals, AbbVie, Boehringer Ingelheim, BlossomHill, Caris Life Sciences, Lilly, Thermo Fisher Scientific, Pfizer, Summit Therapeutics, Johnson & Johnson/Janssen; research funding from Bristol Myers Squibb, Bristol Myers Squibb Foundation (Inst), National Cancer Institute (Inst), Moffitt Cancer Center (Inst), Merck (Inst); and travel accomodations and expenses from Bristol Myers Squibb/Roche, MSD, Gilead Sciences, Regeneron.

REFERENCES

1. Shaw AT, Riely GJ, Bang YJ, et al. Crizotinib in ROS1-rearranged advanced non-small-cell lung cancer (NSCLC): updated results, including overall survival, from PROFILE 1001. Ann Oncol. 2019 Jul 1;30(7):1121-1126. doi: 10.1093/annonc/mdz131. PMID: 30980071; PMCID: PMC6637370.

2. Doebele RC, Drilon A, Paz-Ares L, et al. Entrectinib in patients with advanced or metastatic NTRK fusion-positive solid tumours: integrated analysis of three phase 1-2 trials. Lancet Oncol. 2020 Feb;21(2):271-282. doi: 10.1016/S1470-2045(19)30691-6. Epub 2019 Dec 11. Erratum in: Lancet Oncol. 2020 Feb;21(2):e70. doi: 10.1016/S1470-2045(20)30029-2. Erratum in: Lancet Oncol. 2020 Jul;21(7):e341. doi: 10.1016/S1470-2045(20)30345-4. Erratum in: Lancet Oncol. 2020 Aug;21(8):e372. doi: 10.1016/S1470-2045(20)30382-X. Erratum in: Lancet Oncol. 2021 Oct;22(10):e428. doi: 10.1016/S1470-2045(21)00538-6. PMID: 31838007; PMCID: PMC7461630.

3. Drilon A, Camidge DR, Lin JJ, et al. Repotrectinib in ROS1 Fusion-Positive Non-Small-Cell Lung Cancer. N Engl J Med. 2024 Jan 11;390(2):118-131. doi: 10.1056/NEJMoa2302299. PMID: 38197815; PMCID: PMC11702311.

4. Pérol M, Li W, Pennell NA, et al. Taletrectinib in ROS1+ Non-Small Cell Lung Cancer: TRUST. J Clin Oncol. 2025 Jun;43(16):1920-1929. doi: 10.1200/JCO-25-00275. Epub 2025 Apr 3. PMID: 40179330; PMCID: PMC12118623.