Bastiaan B. J. Tops

DNA methylation-based classification is increasingly a cornerstone of CNS tumor diagnostics. Recently,nanopore sequencing enables very rapid methylation profiling to support surgical decision making. Time-constrained sequencing yields sparse methylation profiles, necessitating specialized classifiers for accurate predictions. All publicly available sparse-profile classifiers rely on a single reference set of only 2,801 samples published in 2018 and its accompanying classification system. Here we present Sturgeon V2, trained on nearly 10,000 methylation profiles assembled from a range of public repositories and collaborations. To harmonize differences in tumor class labels, we developed an iterative semi-supervised relabeling strategy with a two-tiered concordance filter. We demonstrate that this filtering is essential for optimal performance. Compared to Sturgeon V1, we further redesigned the classifier architecture to provide more stable, and consistent confidence estimates while accumulating sequencing data and we introduce an updated class system including new classes, such as meningioma subtypes, PATZ-fusion neuroepithelial tumors and PLAG1/2-amplified embryonal tumors. In a retrospective intraoperative cohort of 94 pediatric cases, Sturgeon V2 was confidently correct in 85 cases (Sturgeon V1: 78) with one confident misclassification, and in simulated sequencing runs (n=1880) reduced the error rate among confident calls from 7.1% to 2.8%, outperforming publicly available alternatives. Sturgeon V2 is the first openly available, locally executable classifier for sparse methylation profiles trained beyond the 2018 reference era.

Background: Central nervous system (CNS) tumors represent the leading cause of pediatric cancer-related deaths. Primary treatment often involves neurosurgical tumor resection. Accurate diagnosis is crucial to perform the best-suited extent of resection. Recently, we published Sturgeon, an AI-based validated intraoperative nanopore sequencing tool, delivering accurate methylation-based diagnoses during surgery. Although several other nanopore sequencing strategies have been described, clear evidence of clinical impact remains critically absent. Here, we report a care evaluation analyzing the impact of Sturgeon on real-time pediatric neurosurgical decision-making. Methods: Since May 2023, a dual diagnostic neuro-oncological workflow incorporating intraoperative frozen section analyses and ultra-fast nanopore sequencing using Sturgeon has been fully implemented as standard of care at the Princess Máxima Center for Pediatric Oncology in the Netherlands. This intraoperative diagnostic workflow has been applied to all pediatric patients (<19 years of age) with primary CNS tumors and the multidisciplinary indication for tumor resection surgery. Results: Sturgeon delivered correct diagnoses in 82 out of 94 consecutive patients (87.2%, <90 minutes), no diagnosis in 11.7% and one incorrect diagnosis (1.1%). The diagnosis obtained by Sturgeon supported the intended surgical strategy (85.7%) or changed the strategy (14.3%) toward a more aggressive or limited resection. This contributed to a low complication rate and less second-look surgeries. Earlier communication of the histomolecular diagnosis to patients and their families was observed. Conclusions: Intraoperative use of Sturgeon provides essential guidance toward the most optimal neurosurgical strategy for pediatric CNS tumors and thereby has great potential to contribute to better clinical outcome.

Hematologic malignancies are diagnosed through a fragmented, sequential workup of morphology, immunophenotyping, cytogenetics, and molecular testing that can take days to weeks and is unavailable at many centers. DNA methylation profiling has transformed central nervous system tumor diagnosis, yet hematologic classifiers have remained confined to narrow acute leukemia panels. Here we present Lamprey, a deep-learning methylation classifier spanning 86 hematologic malignancy entities, trained on a reference cohort of 8,544 patients and deployed directly from nanopore sequencing. A depth-aware training framework allows confident classification from the first minutes of a run. Against blinded integrated reference diagnoses across retrospective, external, and prospective cohorts, Lamprey exceeded 98% accuracy among classified cases. Lamprey reaches a confident call within minutes, and cost as little as $82 per sample. Lamprey consolidates a sequential diagnostic workup into a single, rapid, same-day molecular readout.

Whole genome sequencing (WGS) is increasingly accessible in clinical practice, enabling pharmacogenomics (PGx) integration, including in pediatric oncology. However, the lack of validated software to accurately annotate clinically actionable PGx variants from WGS limits routine implementation. We therefore aimed to identify and validate a PGx annotation tool suitable for clinical use in pediatric oncology. We evaluated several tools for technical performance and clinical integration. The Pharmacogenomics Clinical Annotation Tool (PharmCAT) was selected for its alignment with expert‐reviewed PGx guidelines, inclusion of relevant gene‐drug pairs and prescribing recommendations. PharmCAT was validated using an in silico dataset by introducing known diplotypes into the Genome in a Bottle (GIAB) reference sample, complemented by four clinically confirmed diplotypes from three patients. Diplotype and phenotype outputs from WGS were compared against the GIAB and patient reference data. We tested 71 diplotypes across seven genes ( TPMT , NUDT15 , CYP3A5 , CYP2C9 , CYP2C19 , DPYD , UGT1A1 ), using ≥ 95% sensitivity and specificity as validation criteria. CYP2D6 was excluded from this validation due to genotyping limitations from the input data used by PharmCAT. The tool was integrated into our WGS analysis pipeline using containerization for consistent, reproducible execution. Diplotype and phenotype results from PharmCAT fully matched the in silico GIAB set and patient samples, achieving 100% sensitivity and specificity. These findings confirm PharmCAT as a reliable tool for preemptive PGx annotation, supporting implementation in pediatric oncology. Its clinical integration supports individualized dosing, reducing adverse effects and improving efficacy. Further validation of additional gene‐drug pairs will broaden its clinical utility.

The high heterogeneity of pediatric cancers presents significant diagnostic challenges, underscoring the need for accurate classification. Although molecular profiling supports first-line diagnostics and guides treatment, it can delay final diagnosis. While Nanopore-based methylation analysis has enabled rapid CNS tumor diagnosis, its application to pediatric solid tumors and lymphomas has remained largely unexplored. We developed Tucan, a deep-learning classifier trained on 3,818 methylation array profiles representing 84 subtypes, designed to classify tumors from sparse Nanopore methylation data. In retrospective validation (n=514), Tucan generated confident predictions (CFT ≥ 0.7) within 30 minutes of sequencing in 385 cases, achieving 372 correct diagnoses (F1-score: 0.98). In prospective testing (n=74; 63 classifiable), 52 samples reached the confidence threshold with 96% accuracy, confirming the original diagnosis in 47 cases and correctly refining or revising it in three. Together, Tucan enables rapid, high-confidence molecular classification of pediatric solid tumors and lymphomas.

For pediatric patients with central nervous system (CNS) tumors, primary treatment often involves neurosurgical tumor resection. Accurate diagnosis is crucial to perform the best-suited extent of resection. Since two years, Sturgeon, an AI-based validated intraoperative nanopore sequencing tool, has been fully integrated with frozen section analyses as standard of care in our nationwide centralized pediatric oncology hospital. This care evaluation is the first to demonstrate the impact of Sturgeon on neurosurgical decision-making. Sturgeon delivered correct diagnoses in 82 out of 94 consecutive patients (87.2%, <90 minutes), no diagnosis in 11.7% and 1 incorrect diagnosis. The diagnosis obtained by Sturgeon supported the intended surgical strategy (85.7%) or changed the strategy (14.3%) toward a more aggressive or limited resection. This resulted in only 3.2% second-look surgeries. In conclusion, intraoperative use of Sturgeon provides essential guidance toward the most optimal neurosurgical strategy and thereby has great potential to contribute to better clinical outcome.

In pediatric oncology, pharmacogenetic guidelines are underutilized and the potential impact of pre‐emptive pharmacogenetic screening remains largely unexplored despite this field's need for individualized approaches. While comprehensive pharmacogenetic guidelines are not yet available for all anticancer drugs, evidence‐based recommendations exist for a subset of supportive care drugs and anticancer drugs, including thiopurines, irinotecan, capecitabine, and 5‐fluorouracil. In this study, we evaluate the potential impact of pre‐emptive pharmacogenetic screening by retrospectively identifying opportunities for dose or treatment adjustments within a national pediatric oncology cohort. Our analysis focused on ten genes and 28 drugs relevant to pediatric oncology, which are included in the Clinical Pharmacogenetics Implementation Consortium and the Dutch Pharmacogenetics Working Group guidelines. In a cohort of 1,151 pediatric oncology subjects, we identified that 16% of individuals could have benefited from altered drug dosing or treatment. These include dose and treatment recommendations for allopurinol, nonsteroidal anti‐inflammatory drugs, phenytoin, amitriptyline, proton pump inhibitors, voriconazole, tramadol, codeine, paroxetine, tacrolimus, rasburicase, and 6‐mercaptopurine. As genetic data increasingly becomes available through molecular diagnostics in pediatric oncology, there is a unique opportunity to re‐utilize this data for pre‐emptive pharmacogenetic screening. Leveraging genetic profiles to guide clinicians in drug selection and dose optimization can improve patient outcomes by enhancing the safety and efficacy of treatments. We therefore recommend incorporating pharmacogenetic screening into clinical workflows to advance personalized medicine in pediatric oncology.

PURPOSE: In high-grade osteosarcoma, prognostic factors at diagnosis are insufficient for stratifying patients into relevant subgroups. Recently, a transcriptomic study developed the G1/G2 gene expression signature, in which the G2 signature was associated with unfavorable survival. An orthogonal study identified MYC amplification as an unfavorable prognostic factor using targeted next-generation sequencing. The purpose of this study was to validate the independent prognostic value and to investigate the combined prognostic value of the G1/G2 signature with MYC amplification and/or MYC expression for survival prediction. MATERIALS AND METHODS: This study included pediatric and adolescent patients with high-grade osteosarcoma. RNA-seq was performed in 48 patients. Whole-exome sequencing was performed in 40 patients. Gene expression signature scores, MYC amplification (defined as >seven copies), and MYC expression levels were calculated. Multivariable Cox proportional hazards analysis was performed for event-free survival (EFS; primary end point) and overall survival (OS; secondary end point). RESULTS: In the full cohort, the 3-year EFS rate was 37%. In multivariable Cox regression analysis with metastatic disease stage (n = 21, 44%) as covariate, the G2 signature and MYC expression were independently associated with worse outcomes in terms of EFS (hazard ratio [HR], 3.32 [95% CI, 1.34 to 8.21] and HR, 3.38 [95% CI, 1.71 to 6.66], respectively) and OS (HR, 4.07 [95% CI, 1.19 to 13.9] and HR, 2.88 [95% CI, 1.22 to 6.76], respectively). MYC amplification was not associated with EFS or OS in univariable analysis (HR, 1.88 [95% CI, 0.74 to 4.77] and HR, 0.79 [95% CI, 0.21 to 3.05], respectively). CONCLUSION: The G2 gene expression signature and MYC expression were independently associated with unfavorable outcomes in a pediatric cohort of patients with high-grade osteosarcoma. The combined prognostic value warrants further prospective validation and could potentially serve as a stratification marker for future osteosarcoma treatment protocols.

Background: Pediatric high-grade central nervous system (pHG-CNS) tumors are the leading cause of childhood cancer-related deaths, partly due to poor response to standard treatments. B7-H3 is reportedly expressed in pHG-CNS tumors, making antigen-targeting therapies, including anti-B7-H3 chimeric antigen receptor T-cell (CAR-T) therapy, promising. However, given substantial inter-tumoral protein expression diversity in CNS tumors, it’s unclear which patients might benefit from these treatments. Therefore, we studied B7-H3 expression in a large set of pHG-CNS tumors. Methods: We retrospectively analyzed 136 pHG-CNS tumors (embryonal tumors (n = 44), high-grade neuroepithelial tumors (n = 4), ependymomas (n = 30),high-grade gliomas (HGGs, n = 58)) from the Princess Máxima Center for Pediatric Oncology. CD276 mRNA (encoding B7-H3) and immunohistochemical (IHC) protein expression of B7-H3 was measured and correlated to clinical-molecular data. Results: Large variability of B7-H3 mRNA and protein expression was observed both between and within tumor types. Many tumors expressed B7-H3, but 30% of diffuse midline glioma H3K27-altered and ependymomas posterior fossa type A showed no or minimal expression. This variability was unrelated to patient age, tumor location, epigenetic subclass, or molecular tumor driver. B7-H3 negative cases were high in tumor cells, ruling out low tumor cell percentage as an explanation for negative staining. Conclusions: Our study of B7-H3-expression in the largest pHG-CNS tumor set to date revealed significant interpatient variability and numerous negative cases. Our results urge for tumor tissue acquisition at enrollment in B7-H3 targeting therapeutic trials (including CAR-T cells) in order to thoroughly assess the value of IHC B7-H3 expression as biomarker and, ultimately, to allow for more tailored therapy.