Risk assessment in paediatric glioma—Time to move on from the binary classification

Longitudinal risk prediction for pediatric glioma with temporal deep learning

Pediatric glioma recurrence can cause morbidity and mortality; however, recurrence pattern and severity are heterogeneous and challenging to predict with established clinical and genomic markers. Resultingly, almost all children undergo frequent, long-term, magnetic resonance (MR) brain surveillance regardless of individual recurrence risk. Deep learning analysis of longitudinal MR may be an effective approach for improving individualized recurrence prediction in gliomas and other cancers but has thus far been infeasible with current frameworks. Here, we propose a self-supervised, deep learning approach to longitudinal medical imaging analysis, temporal learning, that models the spatiotemporal information from a patient's current and prior brain MRs to predict future recurrence. We apply temporal learning to pediatric glioma surveillance imaging for 715 patients (3,994 scans) from four distinct clinical settings. We find that longitudinal imaging analysis with temporal learning improves recurrence prediction performance by up to 41% compared to traditional approaches, with improvements in performance in both low- and high-grade glioma. We find that recurrence prediction accuracy increases incrementally with the number of historical scans available per patient. Temporal deep learning may enable point-of-care decision-support for pediatric brain tumors and be adaptable more broadly to patients with other cancers and chronic diseases undergoing surveillance imaging.

Clinicopathological analysis of BRAF and non-BRAF MAPK pathway-altered gliomas in paediatric and adult patients: a single-institution study of 40 patients

AIMS

Mitogen-activated protein kinase (MAPK) pathway alteration is a major oncogenic driver in paediatric low-grade gliomas (LGG) and some adult gliomas, encompassing BRAF (most common) and non-BRAF alterations. The aim was to determine the frequency, molecular spectrum and clinicopathological features of MAPK-altered gliomas in paediatric and adult patients at our neuropathology site in Kuwait.

METHODS

We retrospectively searched the data of molecularly sequenced gliomas between 2018 and 2023 for MAPK alterations, revised the pathology in view of the 2021 WHO classification and evaluated the clinicopathological data for possible correlations.

RESULTS

Of 272 gliomas, 40 (15%) harboured a MAPK pathway alteration in 19 paediatric (median 9.6 years; 1.2-17.6) and 21 adult patients (median 37 years; 18.9-89.2), comprising 42% and 9% of paediatric and adult cases, respectively. Pilocytic astrocytoma and glioblastoma were the most frequent diagnoses in children (47%) and adults (43%), respectively. BRAF V600E (n=17, 43%) showed a wide distribution across age groups, locations and pathological diagnoses while KIAA1549::BRAF fusion (n=8, 20%) was spatially and histologically restricted to cerebellar paediatric LGGs. Non-V600E variants and BRAF amplifications accompanied other molecular aberrations in high-grade tumours. Non-BRAF MAPK alterations (n=8) included mutations and gene fusions involving FGFR1, NTRK2, NF1, ROS1 and MYB. Fusions included KANK1::NTRK2, GOPC::ROS1 (both infant hemispheric gliomas), FGFR1::TACC1 (diffuse LGG), MYB::QKI (angiocentric glioma) and BCR::NTRK2 (glioblastoma). Paradoxical H3 K27M/MAPK co-mutations were observed in two LGGs.

CONCLUSION

The study provided insights into MAPK-altered gliomas in Kuwait highlighting the differences among paediatric and adult patients and providing a framework for planning therapeutic polices.

An institutional audit of the use of novel drugs in pediatric oncology

BACKGROUND

Significant challenges persist in treating children with rare, relapsed, or refractory malignancies. Novel molecularly targeted drugs promise improved outcomes for these children with reduced toxicity. However, there is often limited evidence to substantiate their clinical efficacy and guide their use. This raises issues for clinical decision-making, ethical concerns surrounding equity of access to these often-expensive agents, and the management of families' expectations for cure. This audit evaluated the off-label use of novel drugs and associated clinical outcomes in order to guide the development of future clinical and ethical guidelines.

AIM

To evaluate the patterns in the off-label use of novel drugs for treating childhood cancer and the associated clinical outcomes to guide prospective studies and inform ethical and clinical governance protocols for the use of these agents.

METHODS

A retrospective audit was performed for all patients who received novel drugs off-label as treatment for their malignancy at an Australian pediatric oncology center between 2010 and 2019.

RESULTS

One hundred patients with 32 unique diagnoses received 133 novel drugs across 124 regimens. Eighty-four patients received these drugs at the second line of treatment or greater. Novel drug median cost was $15 521 AUD (Range: $6.53 AUD to $258 339 AUD) and was primarily funded by the hospital (N = 60/133, 45.1%) or compassionate access from pharmaceutical companies (N = 52/133, 39.1%). Decision-making related to novel drugs was inconsistently documented. Ninety-one of 124 treatment regimens commenced between 2010 and 2019 resulted in objective responses (73.4%), but only 35 were still ongoing upon review in June 2020 (38.5%). Median response duration was 12.6 months (Range: 0-93.2 months).

CONCLUSIONS

While novel drugs were largely unable to definitively cure patients, most achieved objective responses. Prospective trials and more rigorous documentation are needed to fully inform the future use of these agents given the heterogeneity of their applications.

Loss of efficacy of subsequent nonsurgical therapy after primary treatment failure in pediatric low-grade glioma patients-Report from the German SIOP-LGG 2004 cohort

First-line treatment of pediatric low-grade glioma using surgery, radio- or chemotherapy fails in a relevant proportion of patients. We analyzed efficacy of subsequent surgical and nonsurgical therapies of the German cohort of the SIOP-LGG 2004 study (2004-2012, 1558 registered patients; median age at diagnosis 7.6 years, median observation time 9.2 years, overall survival 98%/96% at 5/10 years, 15% neurofibromatosis type 1 [NF1]). During follow-up, 1078/1558 patients remained observed without (n = 217), with 1 (n = 707), 2 (n = 124) or 3 to 6 (n = 30) tumor volume reductions; 480/1558 had 1 (n = 332), 2 (n = 80), 3 or more (n = 68) nonsurgical treatment-lines, accompanied by up to 4 tumor-reductive surgeries in 215/480; 265/480 patients never underwent any neurosurgical tumor volume reduction (163/265 optic pathway glioma). Patients with progressing tumors after first-line adjuvant treatment were at increased risk of suffering further progressions. Risk factors were young age (<1 year) at start of treatment, tumor dissemination or progression within 18 months after start of chemotherapy. Progression-free survival rates declined with subsequent treatment-lines, yet remaining higher for patients with NF1. In non-NF1-associated tumors, vinblastine monotherapy vs platinum-based chemotherapy was noticeably less effective when used as second-line treatment. Yet, for the entire cohort, results did not favor a certain sequence of specific treatment options. Rather, all can be aligned as a portfolio of choices which need careful balancing of risks and benefits. Future molecular data may predict long-term tumor biology.

Genetics of Common Pediatric Brain Tumors

Central nervous system tumors are the most common solid tumors in pediatrics and represent the largest cause of childhood cancer-related mortality. Improvements have occurred in the management of these patients leading to better survival, but significant morbidity persists. With the era of next generation sequencing, considerable advances have occurred in the understanding of these tumors both biologically and clinically. This information has impacted diagnosis and management. Subgroups have been identified, improving risk stratification. Novel therapeutic approaches, specifically targeting the biology of these tumors, are being investigated to improve overall survival and decrease treatment-related morbidity. The intent of this review is to discuss the genetics of common pediatric brain tumors and the clinical implications. This review will include known genetic disorders associated with central nervous system tumors, neurofibromatosis, tuberous sclerosis, Li-Fraumeni syndrome, Gorlin syndrome, and Turcot syndrome, as well as somatic mutations of glioma, medulloblastoma, and ependymoma.