Atypical neurofibromatous lesions: Epigenetic cousins of other benign peripheral nerve sheath tumors?

Atypical neurofibromatous lesions (ANL) are considered a premalignant/early malignant disease stage that does not fulfill the diagnostic criteria for malignant peripheral nerve sheath tumor (MPNST) but are associated with increased risk of progression to MPNST particularly in the setting of neurofibromatosis type 1 (NF1).¹ It is currently recommended that ANL be called atypical neurofibroma if sporadic, or atypical neurofibromatous neoplasm of uncertain biological potential (ANNUBP) if associated with NF1.^2,3 ANL are defined as neurofibromas with at least 2 atypical histological features (cytological atypia with nuclear enlargement, hypercellularity, neurofibroma architectural loss, and increased mitotic activity). Furthermore, increased mitotic activity level in NF1-related ANL warrants an upgrade to low-grade MPNST.^2,3 ANL often harbor NF1 inactivation as well as CDKN2A/B homozygous deletions^4–6 with loss of p16 expression⁶ similar to low and high-grade MPNST. High-grade MPNST also shows inactivation of PRC2 core components (SUZ12, EED) with H3 p.K28me3 (K27me3) loss,^4,7,8TP53 mutations,^5,8,9 and complex copy-number profiles.⁴ Limited data suggest that ANL have an epigenetic signature overlapping with low-grade MPNST but distinct from other benign peripheral nerve sheath tumors and high-grade MPNST.⁴

The diagnosis of ANL based on histopathological criteria is susceptible to interobserver variability. Additionally, the similar molecular and methylation profile with low-grade MPNST makes ANL diagnostically challenging. Although an overlapping epigenetic signature has been suggested for ANL and low-grade MPNST, analysis of a larger number of cases using more sophisticated analytical approaches would be valuable. The retrospective multi-institutional study from Kresbach et al.¹⁰ expands on the epigenetic characterization of ANL and provides additional insights into the underlying molecular biology of this diagnostically and clinically challenging group of tumors.

In the current issue of Neuro-Oncology, Kresbach et al.¹⁰ performed methylation profiling of forty tumors with varying levels of atypical histological features and differential diagnosis of ANL. All cases with available data were NF1-associated tumors (27/40) and had worrisome radiological characteristics (12/40). Pediatric and adult age groups as well as both biological sexes were included in the study. Based on the 2021 WHO classification, these tumors corresponded to ANNUBP (24/40), plexiform neurofibroma (12/40), and low-grade MPNST (3/40). t-SNE analysis with methylation data of additional 134 cases (histological diagnosis of MPNST, dermal/plexiform neurofibroma, and schwannoma), using the v12.5 central nervous system (CNS) and v12.2 sarcoma classifiers showed that 36/40 cases grouped closely to benign peripheral nerve sheath tumor entities whereas 4/40 tumors (including all 3 low-grade MPNST and one ANUBBP by 2021 WHO criteria) clustered with MPNST. Consensus clustering of all 174 cases revealed a distinct “atypical neurofibroma (ANF)” methylation group with 21/40 ANL cases that also included four histologically defined schwannoma. ANL cases within this distinct ANF methylation group consisted of ANNUBP (13/21) or plexiform neurofibroma (8/21) based on the 2021 WHO classification, and often matched to the CNS classifier methylation class schwannoma and remained unclassifiable by the sarcoma classifier. The ANF methylation group had frequent CDKN2A/CDKN2B losses that were mainly a heterozygous (12/21) rather than homozygous deletion by the methylation array copy-number plot, and significantly higher tumor-infiltrating lymphocytes than MPNST, schwannoma, and neurofibromas. Available clinical data (10/21) suggested that tumors within ANF methylation cluster follow a benign clinical course; the sole case with malignant transformation arose in the vicinity of a recurrence of a prior MPNST in an NF1 patient.

This study supports the existence of a subset of histologically defined ANL with a distinct epigenetic signature that is closer to other benign peripheral nerve sheath tumor entities than to MPNST and enriched for CDKN2A/B losses and tumor-infiltrating lymphocytes. The limited clinical outcome data preclude assessment of the clinical significance of this epigenetic cluster but suggest a benign clinical course. The matching of all histologically defined low-grade MPNST cases to the MPNST epigenetic signature also supports the robustness of the 2021 WHO histological criteria to indicate biological aggressiveness. The lack of representation of ANF within the current reference sets of the CNS and sarcoma methylation classifiers resulting in misleading match to methylation class schwannoma or unclassifiable results should be considered as an opportunity to improve these classifiers and enable testing for matches to ANL in the diagnostic setting. While inclusion of well-characterized ANL cases in methylation classifier reference sets will be helpful, further study of the ANF epigenetic cluster is warranted. The ANF group seems to represent an epigenetic cousin of other benign peripheral nerve sheath tumors. However, the distinctive ANF epigenetic signature may represent an intermediate epigenetic stage of malignant progression as suggested by the associated worrisome histological and molecular features. Evaluation of additional well-characterized ANL cases with long-term clinical outcome data is crucial and will be facilitated as DNA methylation-based classification becomes increasingly available in the clinical diagnostic setting.