Seeing the unseen: aI radiomics unmasking occult pancreatic cancer?
Would it be not attractive to have a scalable, imaging-based method at hand to predict pancreatic cancer before the patient becomes symptomatic and before a lesion becomes visible to the human eye—by ‘seeing the unseen’? For years, pancreatic ductal adenocarcinoma (PDAC) has remained one of the most...
Salvato in:
| Autori principali: | , |
|---|---|
| Natura: | Article (Journal) |
| Lingua: | inglese |
| Pubblicazione: |
8 May 2026
|
| In: |
Gut
Year: 2026, Pages: 1-2 |
| ISSN: | 1468-3288 |
| DOI: | 10.1136/gutjnl-2026-338712 |
| Accesso online: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1136/gutjnl-2026-338712 Verlag, lizenzpflichtig, Volltext: https://gut.bmj.com/content/early/2026/05/07/gutjnl-2026-338712 
|
| Note sull'autore: | Patrick Michl, Laura Roth |
| Riassunto: | Would it be not attractive to have a scalable, imaging-based method at hand to predict pancreatic cancer before the patient becomes symptomatic and before a lesion becomes visible to the human eye—by ‘seeing the unseen’? For years, pancreatic ductal adenocarcinoma (PDAC) has remained one of the most lethal malignancies, in large part because its diagnosis is still driven by late, symptom-based presentation. Most patients are diagnosed with an advanced, surgically incurable disease and incidentally detected PDAC at an asymptomatic stage is exceedingly rare. Shifting the diagnostic window so that a greater proportion of PDACs are being detected at an early resectable or even at a pre-invasive stage is the central unmet need in the field and represents a major determinant of improving survival in this dismal disease. The advent of artificial intelligence (AI) tools interrogating imaging data for signatures of malignancy long before anything is visible to the human eye opens intriguing new avenues in this context. - - In Gut , Mukherjee et al introduce a radiomics-based early detection model (REDMOD) designed precisely to enable earlier diagnosis of PDAC.1 The presented model is a fully automated AI framework that interrogated standard contrast-enhanced CT scans for subvisual … |
|---|---|
| Descrizione del documento: | Gesehen am 21.07.2026 |
| Descrizione fisica: | Online Resource |
| ISSN: | 1468-3288 |
| DOI: | 10.1136/gutjnl-2026-338712 |