Safety and preliminary clinical activity of NVL-520, a highly selective ROS1 inhibitor, in patients with advanced ROS1 fusion-positive solid tumors

Document Type


Publication Date


Publication Title

Eur J Can


Background: Oncogenic ROS1 fusions drive various malignancies, including 1–3% of non-small cell lung cancers (NSCLC). Rationally designed ROS1 tyrosine kinase inhibitors (TKIs) that surpass the limitations of FDA/ EMA-approved (crizotinib/entrectinib) or other investigational agents are a medical need. The novel ROS1 TKI NVL-520 is highly selective and designed to avoid the neurologic toxicities associated with ROS1 TKIs that concurrently inhibit TRK (entrectinib/repotrectinib/taletrectinib). Furthermore, NVL-520 is brain-penetrant and targets a diverse array of ROS1 fusions and recalcitrant resistance mutations, including the ROS1 G2032R solvent-front mutation.

Materials and Methods: ARROS-1 (NCT05118789) is a global, tumor-agnostic, phase 1/2 trial of NVL-520. In the ongoing phase 1 dose escalation, patients are required to have a previously treated ROS1 fusion-positive solid tumor, including NSCLC treated with ≥1 prior ROS1 TKI. Primary objectives are to determine the recommended phase 2 dose (RP2D) and, if applicable, the maximum tolerated dose. Additional objectives include safety, pharmacokinetics (PK), pharmacodynamics, and preliminary activity. Response (RECIST v1.1) was investigator assessed. Data cut: June 13, 2022.

Results: Twenty patients (19 NSCLC, 1 pancreatic cancer) have received NVL-520 orally at dose levels of 25–100 mg once daily. Patients received a median of 3 (range: 1–9) prior anticancer therapies, including any ROS1 TKI (100%); investigational ROS1 TKI (85%, including lorlatinib in 55%, repotrectinib in 40%); ≥2 ROS1 TKIs (75%); any chemotherapy (80%); ≥2 lines of chemotherapy (50%). At baseline, 55% had CNS metastases and 45% had ROS1 G2032R. No dose-limiting toxicities (DLTs), dose reductions, or drug-related treatment discontinuations have been reported. All treatment-related adverse events (TRAEs) were grade 1. The only TRAE in >1 patient was nausea (n = 2). NVL-520 PK analyses demonstrated dose-dependent exposure. Among 12 efficacy-evaluable patients with ROS1 + NSCLC treated at 25–75 mg QD, 6 confirmed partial responses (PRs) were achieved. Shrinkage or resolution of intracranial metastases were observed; no patients had intracranial progression. A PR was achieved in most (n = 5/7) ROS1 G2032R-mutant cancers, including lorlatinib or repotrectinib pretreated tumors. Circulating tumor DNA analyses showed reductions of ROS1 variant allele frequency. The RP2D has not been identified and dose escalation continues.

Conclusions: NVL-520 has been well-tolerated up to 100 mg daily with favorable pharmacokinetics. Activity has been demonstrated in heavily pretreated patients (of whom 70% received ≥2 prior ROS1 TKIs plus chemotherapy), including those with brain metastases and the G2032R mutation.



First Page


Last Page