Updated Favorable histology Wilms tumor risk stratification: Rationale for future Children’s Oncology Group clinical trials

What Was Done

For decades, Wilms tumor risk stratification relied heavily on what could be seen and measured at the time of surgery 0:00. Tumor weight after nephrectomy served as a proxy for disease burden — a straightforward metric that correlated with outcomes in the era when treatment protocols were being established 0:41. Histology and patient age formed the other pillars of the system, dividing patients into treatment groups based on features visible under the microscope and the clinical observation that very young and older children behaved differently 0:43. The approach was pragmatic and reproducible 0:00. It worked well enough to guide therapy intensity and achieved cure rates that made Wilms tumor a pediatric oncology success story 0:00.

What Changed

The driver was molecular biology catching up to clinical need 0:32. As genomic profiling became feasible in pediatric solid tumors, patterns emerged that cut across the old anatomic categories 0:32. Loss of heterozygosity at 11p15 and gain of chromosome 1q identified subsets of patients whose tumors behaved more aggressively than their stage or weight would predict 0:32. Lymph node involvement, long recognized as important in adult cancers, gained clearer prognostic weight in Wilms tumor as surgical sampling became more systematic 0:32. These biological markers did not replace clinical judgment — they refined it, separating patients who could be spared toxicity from those who needed more intensive treatment despite favorable-appearing disease 0:32.

The Children's Oncology Group formalized this shift in a model published in Nature in June 2025 0:09. The new stratification incorporates biological and clinical features absent from the first-generation system 0:19. Tumor nephrectomy weight, once a cornerstone, no longer appears 0:41. The framework also modified how histology and age stratify certain patients, though the specifics of those modifications are not detailed in this discussion 0:43.

Where Practice Stands Now

Risk stratification for favorable histology Wilms tumor now integrates three molecular markers — 11p15 loss of heterozygosity, 1q gain, and lymph node involvement — alongside the traditional histologic and demographic variables 0:32. The model represents a generation of accumulated trial data, not a single study 0:00. It reflects what the cooperative group learned works: which children can safely receive less chemotherapy, which need more, and which features reliably predict those differences 0:19. The system is more complex than its predecessor, requiring molecular testing that was not universally available when the old model was built 0:19. But complexity bought precision, and precision matters when the alternative is either undertreating aggressive disease or overtreating children who would do well with less 0:19.

What Remains Unsettled

This is not the final model 0:48. The discussant states plainly that more changes are expected as ongoing Children's Oncology Group trials produce results 0:48. Risk stratification for Wilms tumor has evolved many times over the years 0:00, and the 2025 update is one iteration in a continuing process 0:09. Each generation of trials tests whether the current risk groups are drawn correctly — whether patients assigned to a given intensity of treatment have outcomes that justify that assignment 0:00. When they do not, the groups are redrawn 0:00. The molecular markers now included were themselves validated in prior cohorts; the next set of markers is likely being evaluated in patients treated under the current schema 0:32. The model will change again, probably within the career span of a pediatric surgeon practicing today 0:48.

What Was Done

For years, Wilms tumor risk stratification relied on features visible to the pathologist and measurable at the time of surgery 0:00. Tumor weight at nephrectomy was a cornerstone of that system 0:41. The logic was sound: larger tumors represented more advanced disease, and weight was objective, reproducible, and available immediately after resection. Histology and age were also central to the model, used to sort patients into treatment groups 0:43. The approach worked well enough to guide therapy for thousands of children, and the Children's Oncology Group refined it iteratively over decades 0:00.

But the model had limits 0:19. Tumor weight correlates imperfectly with biology — a large tumor may be indolent, a small one aggressive 0:41. And the original stratification could not account for molecular features that were unknown or technically inaccessible when the framework was built 0:19.

What Changed

The shift came from accumulated evidence that certain genetic alterations predict outcome more reliably than size 0:19. Loss of heterozygosity at 11p15, 1q gain, and lymph node involvement are now incorporated into the updated model published by the Children's Oncology Group in Nature in June 2025 0:09 0:32. These are not new discoveries — 1q gain in particular has been studied for years — but the COG trials finally generated enough data to justify their inclusion in formal risk assignment 0:19.

Tumor nephrectomy weight, meanwhile, was removed 0:41. The decision reflects a rebalancing: when you have molecular markers that stratify more cleanly, anatomic surrogates like weight add noise rather than signal 0:19 0:41. The model also modified how histology and age are applied in certain patient subgroups, though the specifics of those modifications are not detailed here 0:43.

This is the second-generation model 0:27. The first generation relied on what could be seen and weighed; the second integrates what can be sequenced 0:19 0:27.

Where Practice Stands Now

The updated stratification is now the framework for Children's Oncology Group trials enrolling patients with favorable histology Wilms tumor 0:09. Loss of heterozygosity at 11p15, 1q gain, and lymph node status are required data points 0:32. Tumor weight is not 0:41.

This matters most at the margins — the patient whose tumor is large but molecularly favorable, or small but carries 1q gain 0:19 0:41. Those patients were misclassified under the old system 0:19. The new model should route them to the right intensity of therapy more often 0:09 0:19.

It also matters for trial design 0:09. Stratifying by molecular features allows COG to test whether patients with 1q gain need intensified therapy, or whether those without it can be spared toxicity 0:19 0:32. The old model could not ask those questions cleanly 0:19.

What Remains Unsettled

This is not the final version 0:48. The discussant notes that more changes are expected as ongoing COG trials produce additional results 0:48. That is the correct framing: risk models in pediatric oncology are living documents, updated as evidence accumulates 0:48.

Several questions remain open 0:48. First, how should the molecular markers be weighted against each other 0:32? A patient with 1q gain but no lymph node involvement is not the same as one with both, but the relative contributions are still being worked out 0:32. Second, are there other markers — copy number changes, gene expression signatures, methylation patterns — that will further refine the model 0:19? Almost certainly, but they need prospective validation before they can be incorporated 0:19. Third, how does this model perform outside the COG trial population 0:09? The Nature publication reflects North American practice; whether the same stratification applies in settings where upfront chemotherapy is standard, or where access to molecular testing is limited, is unclear 0:09.

Finally, there is the practical question of implementation 0:32. Loss of heterozygosity and 1q gain require tissue, a reference lab, and turnaround time 0:32. Not every center that treats Wilms tumor has that infrastructure in place 0:32. The model is only useful if the data it requires can be obtained reliably and quickly enough to guide treatment decisions 0:32.

The shift from weight to molecular markers is progress, but it is progress that creates new dependencies 0:19 0:41. The next iteration will likely address those dependencies — or replace them with different ones 0:48.

What Was Done

The first-generation risk stratification for favorable histology Wilms tumor relied on features visible to the pathologist and measurable at the time of nephrectomy: tumor weight, histologic subtype, patient age, and stage 0:00. Tumor nephrectomy weight served as a proxy for tumor biology — the assumption being that larger tumors carried worse prognosis independent of stage 0:41. The model was internally consistent and guided therapy assignment across decades of Children's Oncology Group trials 0:00.

It was reasonable given the tools available: histology, a scale, and the calendar 0:00. Molecular profiling was not routine, and the specific genetic events driving relapse had not been identified in sufficiently powered cohorts to justify changing treatment based on them.

What Changed

The Children's Oncology Group published updated risk stratification for patients with favorable histology Wilms tumor in Nature in June 2025 0:09. The new model incorporates biological and clinical features absent from the first generation 0:19 and removes or modifies older ones 0:27.

Three factors enter the model: loss of heterozygosity at 11P15, 1Q gain, and lymph node involvement 0:32. Loss of heterozygosity at 11P15 marks a subset of tumors with distinct molecular pathogenesis 0:32. 1Q gain, a chromosomal abnormality detectable by array or sequencing, has emerged as one of the strongest predictors of relapse in favorable histology disease 0:32. Lymph node involvement, while a staging feature, is now formally integrated into risk assignment in a way it was not previously 0:32.

Tumor nephrectomy weight no longer appears in the stratification 0:41. The weight of the tumor and the kidney together — a number that once determined whether a child received two or three drugs — has been retired 0:41. The model also modified how histology and age stratify certain patients 0:43.

The driver was accumulated evidence across successive COG studies 0:00 0:09. Certain clinical features lost their predictive power when molecular markers were included in multivariate models 0:19. Tumor weight, in particular, added little once 1Q status was known 0:41. At the same time, the feasibility of routine molecular testing improved — what required research-grade platforms previously is now achievable in many pathology labs.

Where Practice Stands

The updated model is now the framework for risk assignment in Children's Oncology Group trials 0:09. A child with favorable histology Wilms tumor will have tumor tissue analyzed for 11P15 loss of heterozygosity and 1Q gain 0:32, and those results — along with histology, age, stage, and nodal status — will determine therapy intensity 0:19 0:32. Tumor weight is recorded but does not influence treatment decisions 0:41.

This is an operational shift, not just a conceptual one. Pathology workflows must now include molecular testing with turnaround times fast enough to inform upfront therapy 0:32. Tumor banking protocols must ensure adequate tissue is available for both diagnosis and molecular studies. Informed consent discussions now include genetic testing of the tumor 0:32, and families may ask what 1Q gain means for their child — a question that did not exist in the prior era.

What Remains Unsettled

The 2025 model is explicitly provisional 0:48. One of the discussants notes that more changes are expected as ongoing COG trials produce additional results 0:48. Risk stratification for Wilms tumor has evolved many times over the years 0:00, and this iteration will not be the last.

Several questions are open. The optimal therapy for patients with 1Q gain is not yet defined — the marker identifies high-risk patients 0:32, but whether they benefit from intensified chemotherapy, different drug combinations, or radiation field modifications is under investigation. The role of other molecular markers — TP53 mutations in favorable histology tumors, for example, or microRNA signatures — remains to be determined. The model also does not yet address how to risk-stratify patients treated with preoperative chemotherapy, a standard approach in European protocols but less common in North America.

Finally, the clinical utility of the model depends on access to molecular testing 0:32. A risk stratification system that requires genomic profiling is only as equitable as the distribution of that technology. Whether the 2025 model widens or narrows outcome disparities will depend on implementation, not biology.