Early Detection of Pancreatic Cancer in Type 2 Diabetes Mellitus Patients Based on 1H NMR Metabolomics.
Lenka MichálkováŠtěpán HorníkFebin KuriakoseLucie HabartováVladimír SetničkaBohuš BunganičPublished in: Journal of proteome research (2021)
The association of pancreatic cancer with type 2 diabetes mellitus was investigated by 1H NMR metabolomic analysis of blood plasma. Concentration data of 58 metabolites enabled discrimination of pancreatic cancer (PC) patients from healthy controls (HC) and long-term type 2 diabetes mellitus (T2DM) patients. A panel of eight metabolites was proposed and successfully tested for group discrimination. Furthermore, a prediction model for the identification of at-risk individuals for future development of pancreatic cancer was built and tested on recent-onset diabetes mellitus (RODM) patients. Six of 59 RODM samples were assessed as PC with an accuracy of more than 80%. The health condition of these individuals was re-examined, and in four cases, a correlation to the prediction was found. The current health condition can be retrospectively attributed to misdiagnosed pancreatogenic diabetes or to early-stage pancreatic cancer.
Keyphrases
- end stage renal disease
- ejection fraction
- early stage
- newly diagnosed
- chronic kidney disease
- healthcare
- prognostic factors
- public health
- type diabetes
- cardiovascular disease
- magnetic resonance
- high resolution
- adipose tissue
- machine learning
- patient reported outcomes
- neoadjuvant chemotherapy
- rectal cancer
- glycemic control
- cardiovascular risk factors
- patient reported
- artificial intelligence
- health information