On Feb. 2, 2026, Nature Medicine published a paper that seemed to offer one of the simplest and cheapest ways to improve outcomes for lung cancer patients: just treat them before 3:00 PM. The study, LungTIME-C01, reported that patients with advanced non-small cell lung cancer who received immunochemotherapy in the morning or early afternoon had median overall survival of 28.0 months, compared to 16.8 months for those treated after 3:00 PM. A 58% reduction in the risk of death, achieved by nothing more than changing the time of day a drug was infused.
On June 24, 2026, Nature Medicine retracted the paper after a four-month investigation concluded that the journal could “no longer have confidence in the integrity of the results.”
What the paper claimed
The trial, registered at ClinicalTrials.gov (NCT05549037), randomized patients with advanced NSCLC to receive a PD-1 immune checkpoint inhibitor (either Merck’s Keytruda or Innovent’s Tyvyt) combined with platinum-based chemotherapy, administered either before or after 15:00 hours. The first four cycles were timed; subsequent cycles were not.
The reported numbers were striking:
- Median progression-free survival: 11.3 months (early) vs. 5.7 months (late), a hazard ratio of 0.40, meaning a 60% lower risk of progression or death.
- Median overall survival: 28.0 months (early) vs. 16.8 months (late), a hazard ratio of 0.42.
- Objective response rate: 69.5% vs. 56.2%.
The mechanistic correlate was a higher morning circulating CD8+ T-cell count and a better activated-to-exhausted CD8+ T-cell ratio in the early-treatment group, consistent with the known circadian regulation of immune function.
The paper was immediately influential. It was covered by major news outlets, cited by oncologists on social media as a practice-changing result, and held up as the first prospective randomized evidence that time-of-day matters for immunotherapy.
What went wrong
Within weeks of publication, red flags emerged. Oncologist Paolo Tarantino at Dana-Farber/Harvard and internist Anil Makam at UCSF posted critical comments on PubPeer and social media, pointing to multiple anomalies. The journal posted an editor’s note on Feb. 19, 2026, and launched an investigation.
The investigation uncovered a pattern of problems:
Protocol inconsistencies. The ClinicalTrials.gov registration showed substantial mid-trial changes to endpoints, eligibility criteria, sample size, and study design. These were inconsistently reported and modified over time. When the journal requested the original protocol, the authors submitted a translated English version dated 2022, but it contained references to studies published in 2023 and 2024. The translated protocol was anachronistic. The authors attributed this to “administrative errors” rather than producing documented prospective amendments.
Impossibly clean data. The progression-free survival curve was described by statisticians as “smoother than what is most commonly observed” in similar large Phase 3 trials, lacking the stair-step drops that come from fixed-interval imaging assessments. There was no censoring in the first year of follow-up: not a single patient dropped out. Zero adverse events led to treatment discontinuation in either arm throughout the entire trial. Comparator studies of similar size typically show approximately 4.5% discontinuation from adverse events. The rates of immune-related adverse events were similar in both arms despite the large efficacy differences, a biologically atypical finding.
Methodological concerns. Randomization was performed on the day of treatment for almost all patients, an uncommon practice. Deviations from fixed-calendar RECIST imaging schedules, due to COVID-19 delays and cycle-based timing, introduced further inconsistencies.
The editors concluded that the cumulative weight of these problems made the results unreliable. The retraction was signed by the journal editors, not by the authors.
The authors’ response
Of the 28 authors, 17 (including the lead authors Zhe Huang and Liang Zeng, and the senior authors Tony Mok, Francis Levi, Christoph Scheiermann, and Yongchang Zhang) agreed with the retraction. Eleven authors did not respond to the editors’ correspondence. As of the retraction date, no formal institutional investigation from the Hunan Cancer Hospital or Chinese authorities had been publicly announced.
What this means for chronotherapy
The concept of chronotherapy, timing drug administration to align with circadian rhythms, is biologically plausible. Circadian regulation of immune function is well-established at the molecular level. Multiple retrospective studies have suggested a signal for morning immunotherapy benefit, including the MEMOIR study in melanoma (Qian et al., Lancet Oncology, 2021) and meta-analyses pooling thousands of patients.
But the retraction eliminates the only prospective randomized trial that supported the hypothesis. As Paolo Tarantino stated: “The only prospective evidence that time-of-day matters for immunotherapy is now gone.”
The most robust counter-evidence comes from the i-TIMES study, presented at the 2026 European Lung Cancer Congress by Solange Peters of University Hospital Lausanne. This was a large-scale pooled analysis of eight international randomized controlled trials involving 3,060 lung cancer patients. Using propensity-score matching to compare early versus late immune checkpoint inhibitor administration, i-TIMES found no clinically meaningful difference in overall survival, median 17.3 months (early) vs. 16.1 months (late), failing to meet even the non-inferiority margin.
The field now finds itself in an uncomfortable position: a biologically attractive hypothesis supported by retrospective data, contradicted by the only adequately powered prospective analysis, and brought into disrepute by a retracted study whose findings were too clean to be credible.
Several prospective trials of chrono-immunotherapy remain underway. Their results will determine whether chronotherapy survives as a serious clinical question or is relegated to the category of plausible-sounding ideas that failed to survive experimental scrutiny.
The peer review question
The retraction has also revived long-standing debates about the adequacy of peer review at high-impact journals. Tarantino and others noted that the red flags, impossibly smooth survival curves, zero dropouts, anachronistic protocols, should have been caught before publication. Tarantino has called for paid professional peer reviewers rather than overburdened unpaid volunteers, arguing that the current system is ill-equipped to detect deliberate data fabrication.
Sources:
[Retraction Note] Nature Medicine (2026). DOI: 10.1038/s41591-026-04508-1
[Science AAAS] “Retraction questions claim that cancer therapy works better in morning.” June 2026. https://www.science.org/content/article/retraction-questions-claim-cancer-therapy-works-better-morning
[Fierce Pharma] “Nature retracts provocative PD-1 study tied to lung cancer survival.” June 2026. https://www.fiercepharma.com/pharma/nature-retracts-provocative-pd-1-study-tied-lung-cancer-survival-treatment-timing
[i-TIMES Study] Peters S, et al. “Administration time-of-day unlikely to be critical determinant of outcomes in lung cancer.” ELCC 2026.