Why genomes need clinical data: new Nature study uses Hartwig real world data to test a prostate cancer biomarker

Understanding prostate cancer’s clinical heterogeneity

A new study in Nature investigates the mutational processes behind prostate cancer’s clinical heterogeneity and used realworld treatment and outcome data from the Hartwig Medical Database to ask whether those processes can improve therapy choice.

Whole genome analysis reveals eight integrated mutational footprints

The clinical journey of a patient with prostate cancer can be long and highly variable, from stable disease to rapidly lethal metastatic cancer. The reasons why have remained only partly understood.

In a study published in Nature, the Pan Prostate Cancer Group analyzed the whole genomes of 959 donors and integrated singlebase substitution, indel, copynumber and six newly defined complex structural variant signatures into eight “integrated mutational footprints” (IMFs).

Together, these explain the mutational processes at work in 85% of primary prostate cancer genomes, and four of them (present in 37% of primary tumors) were associated with more severe disease, namely a significantly shorter time to metastasis.

Why genomic insights alone cannot guide treatment decisions

That is, like many others, an important finding — but turning it into something that could change a treatment decision requires something genomics alone cannot supply. To take that next step we need to know what treatment patients received and what the outcomes were. For this, the authors used the Hartwig Medical Foundation cohort of metastatic castration resistant prostate cancer, whole genome sequencing linked to real world, retrospective treatment histories and outcomes.

Testing ARPI vs taxane using realworld evidence

The clinical question they put to that data is one that clinicians face routinely and typically must resort to a degree of trial and error. “A second line of treatment is needed — should I choose androgen receptor pathway inhibitor (ARPI) or taxane?”

Using a target trial emulation design that mimics a phase III randomized biomarker trial, the authors stratified patients by their “integrated mutational footprint” activity and compared these two regimens. Across the cohort there was no significant difference between ARPIs and taxanes, highlighting why this is a difficult decision for clinicians.

IMF6 biomarker identifies patients who benefit more from ARPI

But among patients whose tumors were biomarker positive for “IMF6”, a signature indicative of replication stress, ARPI treatment was associated with a substantially reduced risk of treatment failure — i.e. IMF6+ patients benefit more from ARPI than others.

Real world evidence as a driver for biomarker guided trials

It is important to note that these analyses are retrospective and observational, with limited subgroup sizes, warranting prospective biomarker driven validation. That is precisely how real world evidence should function: to identify, prioritize and de risk the questions worth taking into prospective studies, so that trials are built around biomarkers with a real chance of holding up and improving patient outcomes.

Hartwig’s role in enabling biomarker discovery

Hartwig makes this possible by collecting whole genome and clinical data from routine clinical procedures and with consent of patients as an integrated and standardized resource, making it freely available for approved research. Access to genomic, processed and clinical data from the Hartwig Medical Database can be requested via email. Click here to read the data request procedure.

Reference: Gruber, A. J., Olsen, A. V., Hernando, B., Cheng, K. C. L. et al. Integrated signatures define
mutational processes in prostate cancer. Nature (2026). https://doi.org/10.1038/s41586-026-10468-w

You read an article in the category Biomarker. You may also be interested in Hartwig Medical Database, Learning healthcare system, Molecular diagnostics, Personalized treatment, Re-use data, Research, Treatment, Scientific publications or Whole genome sequencing.
All news