RESUMO
Medulloblastoma and pilocytic astrocytoma are the two most common pediatric brain tumors with overlapping imaging features. In this proof-of-concept study, we investigated using a deep learning classifier trained on a multicenter data set to differentiate these tumor types. We developed a patch-based 3D-DenseNet classifier, utilizing automated tumor segmentation. Given the heterogeneity of imaging data (and available sequences), we used all individually available preoperative imaging sequences to make the model robust to varying input. We compared the classifier to diagnostic assessments by five readers with varying experience in pediatric brain tumors. Overall, we included 195 preoperative MRIs from children with medulloblastoma (n = 69) or pilocytic astrocytoma (n = 126) across six university hospitals. In the 64-patient test set, the DenseNet classifier achieved a high AUC of 0.986, correctly predicting 62/64 (97%) diagnoses. It misclassified one case of each tumor type. Human reader accuracy ranged from 100% (expert neuroradiologist) to 80% (resident). The classifier performed significantly better than relatively inexperienced readers (p < 0.05) and was on par with pediatric neuro-oncology experts. Our proof-of-concept study demonstrates a deep learning model based on automated tumor segmentation that can reliably preoperatively differentiate between medulloblastoma and pilocytic astrocytoma, even in heterogeneous data.
RESUMO
Extracranial malignant rhabdoid tumors (extracranial MRT) are rare, highly aggressive malignancies affecting mainly infants and children younger than 3 years. Common anatomic sites comprise the kidneys (RTK - rhabdoid tumor of kidney) and other soft tissues (eMRT - extracranial, extrarenal malignant rhabdoid tumor). The genetic origin of these diseases is linked to biallelic pathogenic variants in the genes SMARCB1, or rarely SMARCA4, encoding subunits of the SWI/SNF chromatin-remodeling complex. Even if extracranial MRT seem to be quite homogeneous, recent epigenome analyses reveal a certain degree of epigenetic heterogeneity. Use of intensified therapies has modestly improved survival for extracranial MRT. Patients at standard risk profit from conventional therapies; most high-risk patients still experience a dismal course and often therapy resistance. Discoveries of clinical and molecular hallmarks and the exploration of experimental therapeutic approaches open exciting perspectives for clinical and molecularly stratified experimental treatment approaches. To ultimately improve the outcome of patients with extracranial MRTs, they need to be characterized and stratified clinically and molecularly. High-risk patients need novel therapeutic approaches including selective experimental agents in phase I/II clinical trials.
RESUMO
BACKGROUND: Refined therapy has helped to improve survival rates in rhabdoid tumors (RT). Prognosis for patients with chemoresistant, recurrent, or progressive RT remains dismal. Although decitabine, an epigenetically active agent, has mainly been evaluated in the management of hematologic malignancies in adults, safety in children has also been demonstrated repeatedly. MATERIALS AND METHODS: A retrospective series of patients who received decitabine upon relapse or progression following therapy according to the EU-RHAB regimen is presented. Due to the retrospective nature of analyses, response was defined as measurable regression of at least one lesion on imaging. 850k methylation profiling was done whenever tumor tissue was available. RESULTS: A total of 22 patients with RT of any anatomical localization were included. Most patients (19/22) presented with metastases. All received low-dose decitabine with or preceding conventional chemotherapy. Patients received a median of two (1-6) courses of decitabine; 27.3% (6/22) demonstrated a radiological response. Molecular analyses revealed increased methylation levels in tumors from responders. No excessive toxicity was observed. Clinical benefits for responders included eligibility for early phase trials or local therapy. Responders showed prolonged time to progression and overall survival. Due to small sample size, statistical correction for survivorship bias demonstrated no significant effect on survival for responders. CONCLUSIONS: Patients with RT demonstrate promising signs of antitumor activity after multiagent relapse therapy including decitabine. Analyses of methylation data suggest a specific effect on an epigenetic level. We propose to consider decitabine and other epigenetic drugs as candidates for further clinical investigations in RT.