ESMO Congress 2026

Lorlatinib is a third-generation ALK/ROS1 tyrosine kinase inhibitor whose position was established stepwise from phase 1 to phase 3: the phase 1 trial completed dose escalation in advanced ALK/ROS1-rearranged NSCLC ^[4], the global multi-cohort phase 2 study confirmed overall and intracranial activity ^[3], and the pivotal phase 3 CROWN trial randomized 296 treatment-naive patients with advanced ALK-positive NSCLC to lorlatinib versus crizotinib, with 12-month progression-free survival of 78% (95% CI 70–84) versus 39% (95% CI 30–48) and a hazard ratio for disease progression or death of 0.28 ^[2]. On the resistance side, prior work showed that each ALK inhibitor carries a distinct spectrum of ALK resistance mutations, that ALK G1202R increases significantly after second-generation agents, and that models harboring ALK resistance mutations are more sensitive to lorlatinib ^[1]; reviews categorize resistance into on-target kinase-domain mutations and off-target mechanisms such as bypass activation ^[5], and an on-target case, L1198F, conferring lorlatinib resistance has been documented after later-line treatment ^[6].

At this ESMO meeting, @bensolomon1 presented the ctDNA analysis from CROWN, using the Guardant Health Infinity assay, focusing on molecular evolution under first-line lorlatinib. Key points: post-progression samples were few, but no on-target ALK resistance mutations were noted, only bypass/off-target mechanisms; patients with early progression had more baseline mutations; TP53 mutations were enriched among patients with early progression, and baseline TP53 mutation was associated with poorer prognosis; baseline detectable ctDNA and ctDNA clearance at weeks 4 and 24 were prognostic of PFS, with specific hazard ratios undisclosed. A commentator also noted that 17% of ALK-positive patients with long-term response still harbor a TP53 mutation, and proposed distinguishing whether TP53 mutations are disruptive, to understand the impact of function rather than the mutation itself on treatment outcome.

Relative to prior knowledge, this readout moves the resistance observation from later lines to the first-line setting and offers directional information: against the background that on-target mutations were common after first- and second-generation agents ^[1] and that on-target resistance such as L1198F has been reported after later-line lorlatinib ^[6], the few CROWN first-line post-progression samples showed no ALK kinase-domain mutations and only bypass mechanisms, which contrasts with the spectrum in which both on-target and off-target mechanisms coexist for third-generation inhibitors ^[5]; it also extends clinical attention from kinase-domain mutations to two prognostic dimensions, TP53 co-mutation and ctDNA clearance kinetics, while the fact that 17% of long-term responders still carry a TP53 mutation suggests its prognostic implications are not uniform and may need stratification by functional impact.

The limitations are equally clear: post-progression samples were few, so the estimate of on-target resistance mutation frequency carries limited reliability; TP53 functional stratification remains a proposal without supporting data; and PFS figures by molecular subgroup were not disclosed and await full publication.