03/25/2025

Genetic and Clinical Landscape of NUTM1 Structural Variants

AACR 2025 PRESENTATION
Authors Sara R. Selitsky, Brandon A. Price, Michael Rollins, Adam J. Hockenberry, Ezra Cohen, Brian Piening, Alexander Rubinsteyn, Jennifer H. Choe, Kim A. Ely, Henry Kaplan

Introduction – NUT carcinoma is an aggressive cancer, originally described in midline structures of adolescents and young adults but now known to occur in a variety of anatomic sites with growing numbers of variant histologies being reported. As such, NUT immunoreactivity and/or demonstration of the NUTM1 fusion is required for definitive diagnosis.

Methods – We identified patients within the Tempus multimodal real-world database who had either a diagnosis of NUT carcinoma based on clinically abstracted diagnosis fields or via the presence of a NUTM1 fusion, as detected either by Tempus xT (targeted DNA panel) or Tempus xR (whole-exome capture RNA-seq).

Results – There were 59 patients in the Tempus database with NUT carcinoma annotated as their primary diagnosis. Tempus NGS testing confirmed the presence of a NUTM1 fusion for 48 (81%). Of the remaining 11 patients, none underwent RNA-seq testing at Tempus and only 2 had solid-tissue DNA-seq performed at Tempus. Additionally, there were 106 patients with a NUTM1 fusion based on NGS evidence that did not have an initial NUT diagnosis, including: non-small cell lung cancer (n=33), head and neck (n=10), cancer of unknown primary (n=10), thyroid cancer (n=11), soft tissue sarcoma (n=11), and several other cancer types. Eight out of 31 sweat gland adenocarcinomas/porocarcinomas in the cohort had a NUTM1 rearrangement. Several other NUTM1 fusion rearrangements with partners not previously associated with NUT carcinoma were identified, the clinical significance of which is unclear.Among the 154 patients where we detected a NUTM1 rearrangement, the most common fusion events were BRD4-NUTM1 (n = 55), NSD3-NUTM1 (n = 31), BRD3-NUTM1 (n = 17), SLC12A6-NUTM1 (n=11), YAP1-NUTM1 (n = 11), MXD4-NUTM1 (n=3), and CIC-NUTM1 (n=3)other fusion partners occurred in only one or two patients. The overall potential underdiagnosis rate (patients with the detected fusion but not an annotated formal NUT carcinoma diagnosis) was 69% (106/154); including BRD4-NUTM1 51% (28/55), BRD3-NUTM1 47% (8/17), NSD3-NUTM1 81% (25/31), and 91% for YAP1-NUTM1 (10/11).Some cancer types were enriched for specific fusion gene partners. Eleven of 12 patients (92%) that were diagnosed with thyroid cancer had a NSD3-NUTM1 fusion. Of the 8 sweat gland adenocarcinomas/porocarcinomas, 6 had a YAP1-NUTM1 fusion.The median overall survival was 5.29 months (n = 72) and the median age was 51 (n = 114) at diagnosis. The median age at diagnosis for patients annotated with NUT carcinoma was 41 years compared to 58.5 years for patients with a NUTM1 fusion but lacking the diagnosis.

Conclusions – NUT carcinomas may be underdiagnosed the majority of the time. Certain cancer types with a high enrichment of NUTM1 fusions, such as sweat gland adenocarcinomas/porocarcinomas, thyroid cancers, and sarcomas, may benefit from universal NGS sequencing to ensure correct diagnosis.

VIEW THE PUBLICATION