AI Epigenomic Model May Flag CSCC Risk Among Those With Hidradenitis Suppurativa

By: Lisa Astor
Posted: Thursday, August 13, 2026

Patients with hidradenitis suppurativa carry blood-based DNA methylation signatures that appear like early epigenomic alterations seen in cutaneous squamous cell carcinoma (CSCC), according to findings presented at the 2026 American Association for Cancer Research (AACR) Annual Meeting.

Hidradenitis suppurativa is a chronic inflammatory skin disease associated with an increased risk of developing CSCC. Epigenetic dysregulation is one potential early indicator of carcinogenic potential.

The team measured DNA methylation across the genome in blood from 24 patients with hidradenitis suppurativa and matched controls using the Illumina MethylationEPIC array. No patients in the study cohort had CSCC as of recruitment, soresearchers were able to look for precancerous epigenomic changes. CpG sites showing differential methylation were flagged through a bioinformatic workflow, then compared against existing CSCC datasets to see where they overlapped with cancer-related pathways.

To rank which sites carried the most signal for CSCC risk, the investigators applied artificial intelligence (AI) methods including deep learning, Cox elastic-net survival modeling, and random forest.

Thirty-two CpG sites spanning 32 genes showed meaningful methylation changes—24 hypomethylated and 8 hypermethylated—with a false detection rate of up to 0.05. Each gene had surfaced in earlier cSCC research, and together they mapped onto recognized cancer pathways such as EGFR/MAPK, p53, TERT, and NOTCH signaling, along with DNA repair and chromatin remodeling.

A Cox proportional hazards model trained on CSCC tumor data returned concordance index values of 0.78 to 0.84 in cross-validation. The resulting model produced a continuous CSCC Epigenetic Prognostic Score that sorted patients into low-, intermediate-, and high-risk methylation groups. SHapley Additive exPlanations analyses pointed to CpG sites in EGFR, DNMT1, TP53, NOTCH3, BRAF, and TERT as the strongest contributors to cancer risk.

“This AI-integrated framework provides the first evidence that these biomarkers can predict malignant transformation in hidradenitis suppurativa, positioning methylation profiling as a promising tool for early CSCC risk assessment,” Kuracha et al concluded.

Disclosure: The study authors reported no conflicts of interest.

AACR Annual Meeting (Abstract 4211)

---