SeekIn's first OncoSeek study: an artificial intelligence model on seven routine protein tumour markers found about half of cancers while flagging about seven in a hundred people without cancer.
Luan and colleagues analysed 7,565 participants (954 with cancer) from SeekIn and Sun Yat-sen Memorial Hospital plus a Johns Hopkins validation cohort (1,005 with cancer, 812 without). Using one threshold per marker the seven markers had 56.9% specificity; the OncoSeek algorithm, combining the markers with age and sex, raised specificity to 92.9% (92.3 to 93.5) with overall sensitivity 51.7% (49.4 to 53.9). Sensitivities for nine common cancers ranged from 37.1% to 77.6% (pancreatic 77.6%), and tissue of origin was correct in 66.8% of true positives.
Cheap and runs on hospital analysers, but at about 93% specificity a screening programme would generate many false alarms for each cancer found.
Shares OncoSeek, SeekIn, Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Shares OncoSeek, SeekIn, Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Shares SeekIn, Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Shares Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Shares Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Shares Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Shares Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Shares Proteomics & phosphoproteomics, Multi-cancer early detection (MCED).
Every source this page cites, grouped by who published it. Grouping is read from the web address, not assigned by hand.
A source that has moved or gone is reported by the weekly link check. /audit/