DNA methylation analysis with nasal brushing for early diagnosis of sinonasal malignant tumours

TL;DR

This study explores using DNA methylation from nasal brushing to detect early-stage nasal tumors, showing promising accuracy in distinguishing malignant from benign and normal tissues.

Contribution

The study adapts a 13-gene DNA methylation assay from oral cavity cancer detection to nasal cavity tumors for early diagnosis.

Findings

The methylation assay showed 85.7% sensitivity and 85.6% specificity in detecting malignant tumors.

Most benign/borderline tumors and normal mucosa samples scored negative, indicating good discrimination.

Dimensionality reduction confirmed methylation profiles can distinguish between different nasal pathologies.

Abstract

Sinonasal tumours are rare entities presenting with non-specific symptoms, therefore being often misinterpreted. The present study aimed to evaluate if the 13-gene DNA Methylation assay for early cancer detection already assessed in the oral cavity, was also useful in nasal cavity tumours. The case series consisted of 93 patients (63 males/30 females), 49 with malignant tumours, 14 with benign/borderline tumours, 34 as control series with 33 inflammatory polyps and one fungus ball. We collected one flocked swab from the lesion and one from the contralateral nasal cavity. All sinonasal cancer cases were evaluated by bisulfite next generation DNA sequencing, investigating the following genes: ZAP70, ITGA4, KIF1A, PARP15, EPHX3, NTM, LRRTM1, FLI1, MIR193, LINC00599, MIR296, TERT, GP1BB. To evaluate the performance of the methylation assay, a specific methylation score was calculated for each sample using linear discriminant analysis, with a predefined positivity threshold of 1.0615547. The association between diagnosis and methylation score positivity was evaluated through Fisher’s exact test and calculation of risk ratios with 95% confidence intervals. Additionally, dimensionality reduction techniques were employed to explore the structure of the dataset and assess the ability of methylation profiles to distinguish between different pathological conditions. The 13-gene DNA Methylation scored positive in 42/49 malignant tumours; We included also 14 benign/borderline tumours of which 11 scored positive. Among 33 inflammatory polyps, 24 scored negative, as well as 64/70 normal contralateral mucosa and one fungus ball. Therefore, excluding benign/borderline tumours, the detected sensitivity was 85.7% and specificity 85.6% (AUC: 0.878).

The online version contains supplementary material available at 10.1007/s12672-026-04508-0.

Figures