Saliva-derived transcriptomic signature for gastric cancer detection using machine learning and leveraging publicly available datasets
Lopes et al.
The finding, in our words
A saliva-derived transcriptomic signature using support vector machine learning detected gastric cancer with an AUC of 0.87, sensitivity of 79% and specificity of 70%. Tissue-based gene expression models performed poorly when applied to saliva, indicating that saliva-specific biomarker panels are needed for effective non-invasive gastric cancer screening.
A paraphrase to the Library’s standard, never the abstract. The source is one link away and is always the authority.
This study used home self-collected saliva samples to successfully monitor cytokines and cortisol in mothers and infants, finding that a community singing intervention significantly reduced salivary interleukin-6 levels in mothers with postnatal depression.
Kirkpatrick et al., Brain, behavior, and immunity (paywalled) · source ↗
In people at high risk of nasopharyngeal carcinoma, saliva showed decreased EBV load and increased EBV DNA methylation compared with controls, and a stepwise strategy that adds saliva testing to serum EBV antibodies improved triage efficiency.
Salivary HPV DNA testing achieved 91.7% sensitivity and 100% specificity against oropharyngeal swab in patients with oral cavity cancer, demonstrating potential for non-invasive self-collection screening programmes and long-term monitoring of HPV-driven disease.
This paper describes a handheld electrochemical sensor for vitamin C measurement in saliva and urine, showing a linear range from 0.2 to 200 µm and detection limit of 0.03 µm. The device demonstrated satisfactory accuracy and recovery in human samples, suggesting potential for affordable home-based testing of vitamin C status.
Fu et al., Advanced healthcare materials (paywalled) · source ↗
A coffee-ring biosensor detected sepsis, COVID-19 and cancer proteins from self-collected saliva down to 3 pg/ml within 12 minutes, achieving sensitivity more than 100-fold higher than lateral flow assays. The system uses a smartphone camera and AI analysis, providing a rapid, affordable platform for decentralised protein diagnostics.