This week's picks: a prostate cancer approval, 'healthy' weight, AI sleep

FDA approved talazoparib with enzalutamide for prostate cancer with DNA-repair gene mutations

On 26 October 2026 the US Food and Drug Administration approved talazoparib (Talzenna, Pfizer) together with enzalutamide for adults with metastatic prostate cancer that either has not yet been treated with newer drugs acting on the androgen pathway or still responds to them, and that carries mutations in homologous recombination repair (HRR) genes. The mutation has to be found with an FDA-authorised test; the FDA had previously approved FoundationOne CDx and FoundationOne Liquid CDx as companion diagnostics for this use.

The decision rests on TALAPRO-3, a randomised, double-blind, placebo-controlled trial in 599 patients with such mutations, in genes including BRCA1, BRCA2, ATM, CHEK2 and PALB2. Everyone received enzalutamide and continuous androgen deprivation; half also got talazoparib, half a placebo. Median time to radiographic progression or death was not reached with the combination versus 45.8 months with enzalutamide alone, a hazard ratio of 0.48 (95% CI 0.36–0.65): the risk of progression or death was roughly halved.

Overall survival data were still immature at this analysis, so it is not yet known whether the combination helps people live longer. The label warns of myelodysplastic syndrome and acute myeloid leukaemia, bone marrow suppression and harm to a fetus. The combination applies only when an FDA-authorised test finds an HRR mutation, so it is the test result that decides who can receive it.

An investigational antibody with immunotherapy halved the risk of death in head and neck cancer at an interim look, the company reports

On 28 October 2026 Genmab announced topline results of the phase 3 LiGeR-HN1 trial in people with recurrent or metastatic squamous-cell cancer of the head and neck whose tumours express PD-L1, treated first-line. Patients were randomised 1:1, in an open-label design, to petosemtamab, an investigational bispecific antibody aimed at EGFR and LGR5, plus the immunotherapy drug pembrolizumab, or to pembrolizumab alone.

At the first interim analysis, assessed by an independent data monitoring committee, the combination reduced the risk of death by 50% (hazard ratio 0.50, 95% CI 0.37–0.67) and raised the odds of the tumour shrinking about threefold (odds ratio 3.17, 95% CI 2.06–4.89), regardless of HPV status. The company reports no new safety signals. It puts median overall survival in this disease at about 10–12 months, which is the backdrop for the size of the effect.

This is a company announcement: it does not give the number of patients, the trial's median overall survival or detailed side-effect data. Genmab plans to present the full data at a future medical meeting and to approach regulators. Petosemtamab holds FDA Breakthrough Therapy designation for this use, but its safety and efficacy have not been established and it remains investigational.

Four in ten people with 'metabolically healthy' excess weight lost that status by a repeat visit

A cohort study published online in Endocrine Practice on 25 October 2026 used data from 7,395 UK Biobank participants who had been examined twice. The researchers sorted people by body mass index (normal weight versus overweight or obesity) and by metabolic health, judged both by a newer definition and by the older ATP-III criteria, and checked who had changed category by the repeat assessment.

Among people with overweight or obesity who were metabolically healthy at the start, 40.9% had become metabolically unhealthy by the newer definition and 23.7% by ATP-III. Those who made that shift had a higher risk of cardiovascular disease than people of normal weight who stayed metabolically healthy (hazard ratio 1.71, 95% CI 1.14–2.56), while those who kept their metabolic health showed no significant excess risk. Systolic blood pressure, age and BMI were the main predictors of the shift.

The study is observational: it shows an association, not that holding on to metabolic health prevents heart disease, and the abstract gives neither the length of follow-up nor the number of events. The authors read it as a case for periodic monitoring: 'healthy' excess weight is a state that can change, so blood pressure and metabolic markers are worth rechecking rather than judging once.

An open AI model staged sleep from almost any EEG setup across 28 datasets

On 30 October 2026 npj Digital Medicine published AnySleep, a deep neural network that scores sleep stages from whatever electroencephalography (EEG) or eye-movement (EOG) channels a recording happens to have, and at time steps shorter than the standard 30-second epochs. It was trained and validated on 20,589 overnight recordings, more than 200,000 hours, from 28 datasets across multiple clinics, and checked on datasets from studies not used in training.

Agreement with expert scoring, measured as average macro F1, was 0.812 on in-distribution test sets versus 0.799 for the established U-Sleep model, and 0.777 versus 0.768 on hold-out datasets: a small edge rather than a leap. At timescales of 2–8 seconds, a simple rule applied to the model's high-resolution output detected up to 52.1% of the brief awakenings, or arousals, marked by experts. Classifiers built on this finer output separated patients with obstructive sleep apnoea, narcolepsy type 1 or insomnia from healthy controls better than those built on 30-second scoring, and the authors made the model publicly available.

The analysis is retrospective. The authors write that prospective validation in routine clinical workflows is still needed and that use with wearable EEG would require more high-quality data, so for now this is a research tool for sleep laboratories and large studies.

This material is for information only and does not replace a consultation with a doctor. Nothing in it is a recommendation to start, stop or change treatment.

Sources

• U.S. Food and Drug Administration. FDA approves talazoparib with enzalutamide for metastatic androgen pathway modulation-naïve or -sensitive prostate cancer. 26 October 2026. FDA, NCT04821622

• Genmab A/S. Company announcement no. 44: Petosemtamab combination improves overall survival and objective response at first interim analysis in first-line PD-L1+ head and neck cancer. 28 October 2026. Genmab, NCT06525220

• Zeng Y, Xu XW, Hu Z, et al. Transitions in metabolically healthy obesity based on a novel definition and their association with incident cardiovascular disease: a cohort study. Endocr Pract. 2026. PMID 42927833, doi:10.1016/j.eprac.2026.10.011

• Grieger N, Raskob J, Mehrkanoon S, Bialonski S. AnySleep: a channel-agnostic deep learning system for high-resolution sleep staging in multi-center cohorts. npj Digit Med. 2026. doi:10.1038/s41746-026-03406-7