In 1936, Jorge Luis Borges published an essay on the translators of One Thousand and One Nights: Burton, Lane, Galland, and Mardrus.1 Four men, four fundamentally different books. Galland gave eighteenth-century France a Nights polished into salon prose. Lane, working from Cairo, expurgated. Burton, vain and scholarly, footnoted his version of the collection of Middle Eastern and Indian stories with notes on Arabian sexual practice. Mardrus invented whole passages (AI, anyone?).
None of them, Borges observed, was working from the original. There was no original. The Nights had been growing for centuries through the hands of unknown copyists, and each translator added himself to the manuscript.
Four years earlier, Borges had ended an essay on the Greek translators of Homer with a sentence that has aged into a small monument: el concepto de texto definitivo no corresponde sino a la religión o al cansancio—the concept of a definitive text belongs only to religion or to fatigue.2
I thought about both essays in Chicago as I walked out of a plenary session on June 1 at the 2026 American Society of Clinical Oncology (ASCO) Annual Meeting. ASCO Immediate Past President Eric J. Small, MD, named his year The Science and Practice of Translation: Improving Cancer Outcomes Worldwide.
The meeting did what Borges said the literary world had always done. It produced successive versions of the same target. The question Borges put to readers of the Nights is the question I want to put to those of us who left Chicago carrying 10 new translations of the lung cancer text: When we retell the same target’s story, what does “closer to the original” mean when the original, the biology, is unreadable directly?
The Plenary’s Two Versions
For the first time in years, two lung trials shared the plenary slot, and they were two different kinds of translation.
LIBRETTO-432 took the ADAURA paradigm and translated it into RET. In the initial cohort, stage IB-IIIA resected RET fusion-positive non-small cell lung cancer (NSCLC), adjuvant selpercatinib produced a statistically significant improvement in event-free survival (EFS): Not reached compared with 31.8 months with placebo, with a hazard ratio of 0.172 (95% CI 0.058–0.509, p = 0.0003).
The results also showed an 83% reduction in the risk of recurrence or death at a median follow-up of 22.1 months. The 24-month EFS rate was 91.5% with selpercatinib versus 61.1% with placebo. Four recurrences or deaths occurred in the 54-patient selpercatinib arm; seventeen in the placebo arm.
The tolerability ledger was substantial: Grade 3 or higher treatment-related adverse events in 66.7% of selpercatinib patients; ALT elevation in 17.3%; AST elevation in 18.7%; hypertension in 10.7%; with 17.3% discontinuing treatment for adverse events. Goldman called it a new standard of care in early-stage RET fusion-positive lung cancer.3
This is a translation across stages. The targeted-therapy text moved from advanced disease, where it had matured, into the adjuvant setting, where the answer had started to become less ambiguous since the results of the ADAURA trial were unveiled. The open question is whether targeted-adjuvant EFS will convert to overall survival (OS).
ADAURA’s eventual OS readout closed that loop, but the negative ALCHEMIST adjuvant erlotinib trial, also presented at ASCO 2026, is a reminder that disease-free survival benefits in this space have not always translated into lives saved. With four events, LIBRETTO-432 is an early version of its own story. Expect the next translation to arrive within the next five years.
I am optimistic that the new generation of targeted agents will improve OS in other targets just as osimertinib did in the EGFR realm. Selpercatinib is the new standard in the adjuvant setting in RET (+) resectable NSCLC.
HARMONi-6 was the other plenary lung translation and the more Borgesian. Shun Lu, MD, PhD, presented OS data for ivonescimab, a first-in-class PD-1/VEGF bispecific, plus chemotherapy versus tislelizumab plus chemotherapy in 532 previously untreated patients with advanced squamous NSCLC.
Median OS exceeded 2 years for the first time in this disease, with a hazard ratio of 0.66 at the prespecified interim, against a significance boundary of 0.0049. The trial was simultaneously published in The Lancet. It was the first China-originated investigational oncology drug ever selected for the ASCO Plenary in the society’s 61-year history.4
Read carefully, HARMONi-6 is a translation problem made flesh. The comparator, tislelizumab, is itself a Chinese retranslation of the PD-1 paradigm, approved in China and the US, but not yet the US standard in lung cancer, where pembrolizumab plus chemotherapy is what most patients would typically receive.
The population was Chinese, 39% with PD-L1 tumor proportion score below 1%, 63% with centrally located tumors, and 33.8% with multi-site, liver, or brain metastases. The inclusion and exclusion criteria made it a better prognosis population than squamous cell carcinoma is in the real world.
The ivonescimab story in EGFR-mutant disease had already shown that Chinese readouts may translate imperfectly to Western populations. The Lancet publication closes the durability and peer-review gap. The generalizability question remains open.
Borges would have recognized the shape. Same compendium, different translator, different book. While this story is true, we will need to wait for confirmatory data outside China before ivonescimab becomes a new standard. Stay tuned.
Multiple Translators of the Same Target
The rest of the lung program was the same concept at scale. Around each major molecular target, two, three, or four translators had filed their versions, and the differences among them were as interesting as the agreements.
RET arrived with three. AcceleRET-Lung confirmed in a randomized phase III trial what ARROW had suggested in a single-arm study: First-line pralsetinib doubled progression-free survival (PFS) versus chemotherapy (18.7 vs. 9.0 months, p = 0.003) in advanced RET (+) NSCLC with an objective response rate (ORR) of 66% versus 42%. The sponsor terminated early in January 2025, but the infection signal—pneumonia 19% versus 6%, infection-related deaths 7.4% versus none—is the asterisk every reader should note.5
LIBRETTO-432 brought selpercatinib into the adjuvant space, with a different toxicity fingerprint, focused on hepatic and hypertensive side effects. Lunbotinib, a next-generation brain-penetrant RET inhibitor presented in a separate session, yielded ORRs of 87% in pretreated (with chemotherapy mostly) and 81% in treatment-naïve RET fusion–positive NSCLC; six complete intracranial responses were observed in each cohort.6
Same target, three translators, three distinct sentences. The reader’s task is not to declare which is the original. It is to know what each translator sees and what each is preoccupied with.
ALK arrived with three of its own translations and the cleanest demonstration of Borges’s claim. Tony S.K. Mok, MD, presented the 7-year update of CROWN. Median PFS has not yet been reached in the lorlatinib arm, with a 7-year PFS of 55% versus 3% for crizotinib.7
If a patient is progression-free at 24 months, the conditional probability of remaining progression-free at year 7 is 79%. Each prior follow-up of CROWN—3 years, 4 years, 5 years—was a version of the same trial. None was wrong. None was definitive. OS data are still pending.
ALKOVE-1 was a different translator: Neladalkib, a next-generation ALK TKI built to beat lorlatinib resistance, registered a 26% ORR after a median of three prior lines, including lorlatinib in 91% of patients.8 ARCHER was the strangest version: A prophylactic peptide vaccine targeting seven ALK resistance mutations before they had a chance to emerge, given alongside a TKI in minimal residual disease.
T-cell responses were observed in 71%. The single progression in 15 patients was not driven by an ALK mutation but by a separate pathway, an emergent KRAS G12D mutation, at once proof of the concept’s logic and a preview of its limit.9
In part 2, I will explore EGFR and KRAS-centered trials, with their own translators, then focus on the trials whose original targets moved while we were reading them, and the harder translation Dr. Small’s theme could not avoid.
References
- 1. Borges JL. Los traductores de Las 1001 Noches. In: Historia de la eternidad. Buenos Aires: Viau y Zona; 1936.
- 2. Borges JL. Las versiones homéricas. In: Discusión. Buenos Aires: Sur; 1932.
- 3. Goldman JW, et al. Adjuvant selpercatinib versus placebo in resected stage IB–IIIA RET fusion-positive NSCLC: LIBRETTO-432. J Clin Oncol. 2026;44(suppl):LBA3.
- 4. Lu S, et al. Ivonescimab plus chemotherapy versus tislelizumab plus chemotherapy in previously untreated advanced squamous NSCLC: overall survival from HARMONi-6. J Clin Oncol. 2026;44(suppl):LBA4. Simultaneously published in Lancet 2026.
- 5. Popat S, et al. Pralsetinib as first-line treatment of RET fusion-positive advanced NSCLC: phase 3 AcceleRET-Lung. J Clin Oncol. 2026;44(suppl):8504.
- 6. Zhou Q, et al. Lunbotinib (A400/EP0031), a next-generation selective RET inhibitor, in advanced RET fusion-positive NSCLC: pivotal phase II study. J Clin Oncol. 2026;44(suppl):8505.
- 7. Mok T, et al. Lorlatinib versus crizotinib as first-line treatment for advanced ALK-positive NSCLC: 7-year update from CROWN. J Clin Oncol. 2026;44(suppl):8502.
- 8. Lin JJ, et al. ALKOVE-1: efficacy and safety of neladalkib in patients with advanced ALK-positive NSCLC. J Clin Oncol. 2026;44(suppl):8503.
- 9. ARCHER investigators. ALK-Vac peptide vaccine in ALK-positive NSCLC in minimal residual disease. J Clin Oncol. 2026;44(suppl):8517.
