Same platinum, different scar: why cured children's livers look genomically old

In Short. Cisplatin and carboplatin turned localized childhood liver cancer from a 20 percent five-year survival into more than 80 percent, and they also aged the DNA of remaining healthy liver to an adult-scale load of about 2,200 mutations per sample, with a liver-only signature absent from blood. Systemic platinum is rewritten by tissue metabolism and repair.

---

Hepatoblastoma is a rare liver cancer of early childhood, and for decades it killed most of the children who presented with a localized tumor. Platinum chemotherapy changed that arithmetic. Cisplatin, later joined by carboplatin, pushed five-year survival for localized disease from about 20 percent to more than 80 percent, which is why the drugs remain the backbone of care even as surgeons still take the tumor out. The combination made those operations into cures, and it created a growing population of adults who now carry whatever the same chemistry did to the liver that stayed behind.

Platinum drugs cross-link DNA until a dividing cell can no longer copy its genome cleanly, which is lethal for a tumor and incidental for every other tissue the infusion reaches. In adults, that collateral damage already leaves recognizable mutational signatures, repeating patterns of DNA letter changes that name the chemistry that made them. In children the long-term picture was thinner, even as survivorship clinics watched for late liver disease and, rarely, a second cancer. On 10 September 2026, Anna Wenger and colleagues reported in Science what platinum exposure wrote in remaining healthy liver after hepatoblastoma treatment, under DOI 10.1126/science.ady0339.

A mutational scar is tissue-readable, so a blood sample would stand in for the body only if every organ recorded the same platinum fingerprint, and survivorship monitoring could stay in the phlebotomy chair. The new measurement says the scar depends on the organ. Wenger's team surveyed blood, tumor, and non-cancerous liver after platinum, then compared those genomes with children given other drugs, children treated without prior chemotherapy, and fetal liver, using NanoSeq, a duplex method that marks both DNA strands so rare mutations in normal tissue become visible where bulk sequencing would miss them because neighboring healthy cells rarely share the same mutation.

The cohort is small and concrete. One hundred eighty-six samples of blood, liver tumor, and non-cancerous liver came from nine children after platinum. Thirty more samples came from two children with liver cancer treated without platinum. Forty-seven further samples came from children with other cancers, on platinum or off it, and from untreated children. Across that set, platinum-exposed livers carried hundreds of mutations, on average about 2,200 per normal liver sample, a burden typical of adult livers. Dual exposure to cisplatin and carboplatin wrote more mutations than cisplatin alone, and as little as one platinum cycle already lifted the load. Dose is written into the genome as a count, independently readable in residual tissue.

Blood from the same children carried fewer mutations, even though the drug had traveled the whole body. The liver also held a mutational signature missing from blood and from the other tissues surveyed. That liver-only pattern is the paper's claim about how DNA damage actually happens in a living organ: the same platinum molecule, delivered systemically, becomes a different chemical event once hepatocytes handle it. The liver is where many drugs are broken down, so a distinct hepatic fingerprint is biologically cheap to imagine and, until this assay, hard to prove in residual pediatric tissue.

Foad Rouhani, a co-senior author at the Francis Crick Institute and King's College London, put the remaining uncertainty in two biological slots. Either the drug is broken down differently in liver cells, or those cells repair platinum adducts along a different path. The team has still to sort metabolism from repair, and both explanations share a consequence. Tissue context rewrites the DNA chemistry, which means a blood-only test of mutational scars will miss the liver's version, and the liver's version is the one that has to last for decades in an organ platinum already used as a battlefield.

Adult livers accumulate mutations slowly, year by year, from ordinary metabolism and environmental chemistry. Two thousand two hundred mutations in a child's residual liver, packed into a course of platinum, compress that clock into months. The AACR abstract of the same work notes that the liver-specific signature was absent from blood in the same children, and that in-vitro work plus published cancer catalogues helped corroborate a tissue-restricted pattern. A drug that saved the organ as a patient also aged the organ as a genome, in a way blood will not report.

The mutations land in genes that matter for cancer and for liver metabolism. Gene-focused NanoSeq found a wide repertoire of variants in normal tissues, including leukaemogenic changes in blood, and Rouhani told STAT the liver samples showed cancer genes plus genes tied to long-term metabolic change. He also drew the limit the clinic needs. A mutation load is evidence of potential, and it leaves later clinical outcome open. Secondary malignancy after childhood platinum remains rare. The DNA record supplies a plausible mechanism for late liver disease and for that small second-cancer risk, while most survivors will still live without a second tumor. The map is for the residual tissue that keeps the scar, and for the clinics that will see those children as adults.

That distinction is why the companion Science perspective, by Sanjeev Vasudevan and Donald Williams Parsons at Baylor College of Medicine, asks for survivorship studies of children treated for liver cancer past the third decade of life. The practical verb the evidence earns today is monitoring, for decades, of a population platinum created by saving. Clinics already follow childhood cancer survivors. The genomic scar says the liver deserves particular, lasting attention because its mutation burden and its unique signature belong to that tissue, and they will stay invisible in a blood-only readout. Lifelong follow-up is therefore a tissue-biology instruction, earned by a countable load.

Monitoring past the third decade will still be an epidemiological job: imaging, liver enzymes, and second-cancer registries have to catch events the genome only flags as possible. What the sequence adds is a reason to keep looking at the liver specifically, including in survivors whose blood counts and circulating DNA look quiet. Designers of the next cytotoxic, meanwhile, now have a test. A candidate that still clears hepatoblastoma while writing a shallower, less liver-restricted scar would be a genuine hardware improvement, measurable in the same duplex assay on residual tissue.

The research path that follows is a design problem, and it leaves platinum on the protocol. Equally effective alternatives for hepatoblastoma remain scarce, which is why the drugs stayed after they turned a lethal localized tumor into a survivable one. Rouhani asked whether understanding the mechanism could yield next-generation cytotoxics that still kill the tumor while leaving background tissue largely untouched. Sam Behjati, co-senior author at the University of Cambridge, made the same pairing: chemotherapy is the key to curing childhood cancer, and DNA damage in normal tissues is a plausible route to late adverse effects. Protective strategies become thinkable only after the scar is mapped by organ, which is the step this paper takes.

Wenger, first author at the Wellcome Sanger Institute and the University of Gothenburg, described the aging as a clock. In a short course of treatment, children's healthy cells acquired DNA damage that would otherwise accumulate over decades and appear in middle-aged adults. Premature-aging clinics already see some of those children as grown patients, and the sequence data now sit under that clinical impression as a countable load and a liver-restricted fingerprint. Ellie Waters-Barnes, a recently qualified doctor and childhood cancer survivor who was outside the study, told the Sanger release that the hope is a future in which the drugs that cure children leave a lighter burden on the rest of a life. That hope is a research program. It starts from a fact the field can now count: hepatoblastoma survival rose because platinum was brutal enough to erase a childhood liver tumor, and the same brutality, filtered through hepatic metabolism and repair, left an adult-scale genome in tissue that still has to last a lifetime. Late effects after cure are a problem of tissue-written mutagenesis. Next drugs, and next decades of follow-up, have to start from that map.

---