Deep learning (DL) is a widely applied mathematical modeling technique. Classically, DL models utilize large volumes of training data, which are not available in many healthcare contexts. For patients with brain tumors, non-invasive diagnosis would represent a substantial clinical advance, potentially sparing patients from the risks associated with surgical intervention on the brain. Such an approach will depend upon highly accurate models built using the limited datasets that are available. Herein, we present a novel genetic algorithm (GA) that identifies optimal architecture parameters using feature embeddings from state-of-the-art image classification networks to identify the pediatric brain tumor, adamantinomatous craniopharyngioma (ACP). We optimized classification models for preoperative Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and combined CT and MRI datasets with demonstrated test accuracies of 85.3%, 83.3%, and 87.8%, respectively. Notably, our GA improved baseline model performance by up to 38%. This work advances DL and its applications within healthcare by identifying optimized networks in small-scale data contexts. The proposed system is easily implementable and scalable for non-invasive computer-aided diagnosis, even for uncommon diseases.

Adamantinomatous craniopharyngioma (ACP) is a biologically benign but clinically aggressive lesion that has a significant impact on quality of life. The incidence of the disease has a bimodal distribution, with peaks occurring in children and older adults. Our group previously published the results of a transcriptome analysis of pediatric ACPs that identified several genes that were consistently overexpressed relative to other pediatric brain tumors and normal tissue. We now present the results of a transcriptome analysis comparing pediatric to adult ACP to identify biological differences between these groups that may provide novel therapeutic insights or support the assertion that potential therapies identified through the study of pediatric ACP may also have a role in adult ACP. Using our compiled transcriptome dataset of 27 pediatric and 9 adult ACPs, obtained through the Advancing Treatment for Pediatric Craniopharyngioma Consortium, we interrogated potential age-related transcriptional differences using several rigorous mathematical analyses. These included: canonical differential expression analysis; divisive, agglomerative, and probabilistic based hierarchical clustering; information theory based characterizations; and the deep learning approach, HD Spot. Our work indicates that there is no therapeutically relevant difference in ACP gene expression based on age. As such, potential therapeutic targets identified in pediatric ACP are also likely to have relvance for adult patients.