Matthew Jonathan Murray

Whole-genome sequencing (WGS) has been shown to improve diagnosis beyond standard-of-care genetic tests and guide precision treatment for childhood cancer. While WGS is now offered for children with suspected cancer in England, clinical workflows typically return results after several weeks whereas clinical decisions are often required within days. Here we evaluate feasibility and efficacy of a rapid WGS (rWGS) research workflow with streamlined bioinformatics to deliver clinical genomic findings (NCT07201038, Active, not recruiting). In 54 children with suspected or confirmed malignancy, rWGS reduces mean end-to-end turnaround time from 42 to 3 days, whilst increase variant detection rate. Whereby rWGS detected 147 of 155 (95%) clinically actionable variants identified in the cohort, standard of care testing detected 138 of 155 (89%) variants. In a subset of 35 patients assessed prospectively, clinicians report that in 17 (49%), rWGS either (i) facilitates faster access to targeted therapy and risk stratification, or (ii) avoids unnecessary investigations and treatment escalation. These findings indicate feasibility and suggest clinical utility of rWGS when deployed as a first line diagnostic test in pediatric oncology.

Modern cancer diagnostics is heavily reliant on molecular data. Next-generation sequencing using liquid biopsy (LB) is primarily focused on identifying predictive and prognostic biomarkers to guide systemic therapy decision making. However, additional findings including genomic variants of possible or probable germline origin are detected in a significant minority of patients (3%-18%), with significant implications for patients and their families. Pretest information provision regarding both the purpose and potential benefits of the test and additional findings on LB varies in clinical practice, with little guidance for clinicians regarding the degree of detail required regarding possible additional findings. LB requires explicit disclosure of these aspects. Implementing a standardized, context-dependent approach to this discussion, with nationally standardized information provision tools, should be prioritized to enable patient-centered decision making regarding LB, without prioritizing the incidentalome over the primary purpose of the test—refining and optimizing treatment options.

Neuroblastoma is a leading cause of pediatric cancer deaths. Accurate and early risk stratification is crucial; however, current approaches rely on invasive biopsies. MicroRNAs have shown promise as non-invasive biomarkers. PubMed and EMBASE were systematically searched for studies evaluating circulating microRNAs in patients with neuroblastoma until April 2025. Eligible studies were included in a narrative synthesis. If raw expression data were available, a standardized re-analysis was performed to assess diagnostic and prognostic performance using differential expression, receiver operating characteristic (ROC) curve, and Cox regression analyses. Nine studies comprising 385 patients were included. Narrative synthesis identified four microRNAs consistently associated with high-risk disease: miR-375, miR-124-3p, miR-218-5p, and miR-490-5p. In our re-analysis, the findings were largely driven by a single cohort of 185 patients, with limited contributions from smaller studies, and should therefore be interpreted as exploratory. We identified nine microRNAs that were associated with poor prognosis. MiR-9-3p demonstrated the most promising performance for identifying high-risk disease (sensitivity = 0.95; specificity = 0.89) and the strongest association with poor overall survival (hazard ratio [HR] = 14.80). In multivariable analyses, miR-9-3p remained an independent predictor of poor survival (HR = 5.48) after controlling for MYCN amplification (HR = 2.08), metastatic disease (HR = 1.62), and age ≥ 18 months at diagnosis (HR = 2.30). While miR-9-3p may serve as a promising candidate biomarker for high-risk disease, there are several limitations to this study. Prospective multicenter validation is required before clinical implementation.

Background: For localized germinoma, whole-ventricular radiotherapy (WVRT) is standard-of-care in Europe; 24 Gray (Gy) with 16 Gy boost following chemotherapy. The SIOP-CNS-GCT-II trial aimed to examine the effect of response-adapted radiotherapy on patient outcome. Methods: Patients with localized germinoma received 4 courses of “carboPEI” chemotherapy, then 24 Gy WVRT if in complete remission (CR), with an additional 16 Gy tumor boost if residual disease was present. Metastatic cases received 24 Gy craniospinal radiotherapy (CSRT) with 16 Gy boost to all sites. Results: Between 2012 and 2018, 227 fully-staged germinoma patients were treated according to protocol. Five-year event-free (EFS) and overall survival (OS) for 166 localized germinoma were 0.94 ± 0.02 and 0.98 ± 0.01, respectively. Sixty-five of 166 (39.2%) were in CR after chemotherapy, of whom 64/65 received 24 Gy WVRT, only 2 of whom (2/64; 3.1%) relapsed. Of 90/166 patients in partial remission (PR) after chemotherapy, 88/90 received 24 Gy WVRT + 16 Gy; 2 relapsed (2/88; 2.3%). Of the 11 remaining patients, 8 had stable disease (SD) after chemotherapy; 7/8 received 24 Gy WVRT + 16 Gy boost and 1 received 24 Gy WVRT + 30 Gy boost because of a teratoma component; none relapsed. Three remaining patients with progressive disease (PD) during chemotherapy had variable treatments, due to differing histologies and disease spread; 1 relapsed. Five-year EFS and OS for 61 metastatic germinoma were 0.98 ± 0.02 and 1.00 ± 0.00, respectively, with 55/61 (90.2%) in CR and 5/61 (11.5%) in PR after 24 Gy CSRT + 16 Gy; 1 relapsed. Conclusions: Response-adapted radiotherapy for localized germinoma led to excellent survival outcomes. Further treatment de-escalation should be considered in future strategies to reduce treatment burden and late effects.

Background: Renal tumours account for one in twenty paediatric cancers, with Wilms tumour (WT) the most common in young children and renal cell carcinoma (RCC) predominating in adolescents and young adults. Diagnostic work-up has traditionally focused on clinical features, radiology, and histology, with a limited role for molecular analysis. However, it is estimated that up to one-third of children with WT have underlying cancer predisposition, which could necessitate prolonged treatment and intensive follow-up. Methods: Here we describe five children and young adults treated at a single regional centre in England who had paired tumour and germline whole genome sequencing (WGS) as part of their routine diagnostic work-up. Results: One child diagnosed radiologically with a WT underwent pre-operative chemotherapy with good clinical and imaging response. Histological examination of the resection raised concerns over RCC; however, WGS was able to confirm that this was a WT with pathognomonic somatic WT changes. A young adult with upfront nephrectomy had a difficult-to-classify tumour; WGS revealed a novel ERC1-CCNY fusion as a likely novel driver event. Two further children, who did not meet clinical criteria for cancer predisposition testing, had predisposition syndromes identified via agnostic WGS. Finally, a child with piebaldism had a WT-associated REST deletion identified early through critical clinical thinking and expedited microarray. Conclusion: We highlight that molecular analysis, particularly agnostic WGS, has a key routine role in the care of children with renal tumours. It is likely that outcomes for these young people have been improved through more accurate diagnosis and early detection of cancer predisposition.

Small non-coding RNAs (sncRNA; < 200 nucleotide length) are of increasing research interest due to their key regulatory roles in a host of fundamental biological processes. For example, microRNAs (miRNAs), a specific class of sncRNAs, regulate gene expression through messenger RNA (mRNA) interactions, and their dysregulation is associated with disease. Classifying sncRNAs is an important bioinformatic task in small RNA-sequencing pipelines. Here we have developed an aligner called PymiRa, written in Python, to identify and quantify miRNAs from FASTA/FASTQ sequencing files. Unlike other approaches, PymiRa utilises a Burrows-Wheeler algorithm to align an input file against a reference hairpin precursor FASTA file derived from miRBase, the online miRNA registry, permitting up to two mismatches at the 3’ end of a read. Previous tools used either a Burrows-Wheeler genome alignment or dynamic programming alignment to precursors; we demonstrate that combining both approaches yields improved results and efficiency. Importantly, the PymiRa aligner accounts for 3’ post-transcriptional modifications to miRNAs that typically occur. PymiRa is a fast, accurate, and publicly accessible aligner available via GitHub and/or a webserver for sncRNA identification, including miRNAs, enabling accurate counts to be produced as part of a small RNA-sequencing pipeline. PymiRa will undergo relevant revisions over time e.g., with miRBase version updates. The PymiRa aligner will facilitate a deeper biological understanding of the landscape of sncRNA expression in normal physiological conditions and their dysregulation in disease states, including cancer.