Category: Clinical Science
At the end of the session, participants will be able to:
- Understand potential limitations of DNA methylation profiling to distinguish piloid gliosis from pilocytic astrocytoma.
COI Disclosure:
None to disclose
Presenter
Andrew Valesano, MD, PhD is in his fourth year of the AP/CP pathology residency program at the University of Michigan Department of Pathology, where he serves as Chief Resident. He completed the Medical Scientist Training Program (MSTP) at the University of Michigan, earning his MD and PhD in 2023. He will begin a fellowship in Molecular Genetic Pathology at the Memorial Sloan Kettering Cancer Center in 2027. His research interests include molecular diagnostics and computational pathology approaches for diagnosis and prognosis of solid tumors, with a focus on CNS tumors and DNA methylation profiling.
Andrew L. Valesano1, Mark A. Rudolf1, Nicole Becker2, Ryan E. Landvater1, Sara L. Stone1, Pouya Jamshidi3, Robert C. Bell1, Kyle S. Conway1
1University of Michigan, Ann Arbor, MI, USA; 2University of Iowa, Iowa City, IA, USA; 3Northwestern University, Chicago, IL, USA
Target Audience:
Pathologists
Residents
CanMEDS:
Scholar
Piloid gliosis can mimic location-specific DNA methylation profiles of pilocytic astrocytoma
Abstract
Distinguishing non-neoplastic piloid gliosis from pilocytic astrocytoma (PA) has important implications for treatment and prognosis, but can be challenging given significant histologic overlap, particularly in small biopsies. Although DNA methylation profiling has become an integral part of neuropathology diagnosis, its ability to resolve this differential has not been systematically investigated. We identified cases exhibiting piloid gliosis adjacent to known non-glial CNS neoplasms, including cerebellar hemangioblastomas (n = 6) and sellar craniopharyngiomas (n = 5). We macrodissected the gliotic regions for DNA methylation profiling with the Illumina MethylationEPIC v2 array. Among hemangioblastoma cases, adjacent gliotic tissue matched to posterior fossa pilocytic astrocytoma by the NIH Bethesda v2 CNS classifier in 2/6 cases (score >0.9); an additional three cases had suggestive scores (0.5 – 0.9). By the DKFZ CNS classifier (v12.8.1), 3/6 cases suggested the low-grade glioma superfamily (scores 0.3–0.7). Among craniopharyngioma cases, adjacent gliotic tissue matched to the supratentorial pilocytic astrocytoma class by the NIH classifier in one case and the DKFZ classifier in another; four additional cases had suggestive scores across both classifiers (0.3 – 0.9). Unsupervised visualization demonstrated proximity of piloid gliosis to location-specific pilocytic astrocytoma references. These results indicate that piloid gliosis can be epigenetically similar to pilocytic astrocytoma, yielding potentially false-positive classifier results varying by model. These findings suggest that PA methylation profiles reflect regional astrocyte identity rather than purely tumor-specific epigenomic reprogramming, with implications for future classifier refinement. Caution is warranted when diagnosing pilocytic astrocytoma by methylation alone in specimens lacking a confirmed MAPK pathway alteration.