From: Data-driven decision-making for precision diagnosis of digestive diseases
First author, year | Disease | n | Data source and specific task | ML method | Diagnostic performance | Refs. |
---|---|---|---|---|---|---|
Liu, 2020 | EC | 248 | MS-based proteomic and phosphoproteomic profiles of tumor and adjacent tissues; subtyping EC based on a protein signature | PCA/clustering/SVM | AUC: 0.976 | [119] |
Komor, 2021 | Colorectal adenomas | 281 | Stool proteomics data; classification based on a panel of protein biomarkers | LASSO | AUC: 0.711 | [120] |
Bhardwaj, 2020 | CRC | 259 | Quantitative data of 275 plasma proteins by PEA; classification based on selected protein features | LASSO | AUC: 0.920 | [121] |
Kalla, 2021 | IBD | 552 | Quantitative data of 460 serum proteins by PEA; classification based on six proteins with age and sex | LR | Accuracy: 79.8% | [122] |
Demirhan, 2023 | GC | 64 | N-glycomics data of tumor and adjacent tissues; classification by differentially expressed N-glycans | MLP | AUC: 0.980 | [123] |
Fan, 2022 | GC | 255 | Urine proteomics data; classification by 4 differentially expressed urine proteins | OPLS–DA | AUC: 0.810–0.920 | [124] |
Bergemalm, 2021 | UC | 451 | Quantitative data of 92 plasma proteins by PEA; preclinical prediction by a panel of up-regulated proteins | PCA/LR | AUC: 0.920 | [125] |
Zhao, 2020 | Acute appendicitis | 568 | Urinary proteomics data; classification based on a 10-protein signature | RF/SVM/Naive Bayes | Accuracy: 81.2–83.6% | [126] |
Song, 2020 | GC | 60 | Label-free global proteomics data of tumor and control tissues; classification based on a four-protein signature | RF | AUC: 0.886–0.996 | [127] |
Shen, 2019 | GC | 150 | Targeted proteomics data of serum by PEA; classification based on 19 proteins | Elastic-net regression | AUC: 0.990 | [128] |
Chatziioannou, 2018 | NEC | 86 | Serum proteomics profiles; classification based on two panels of three proteins | OPLS–DA | AUC: 0.999 | [129] |