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. |
---|---|---|---|---|---|---|
Ichikawa, 2017 | GC | 207 | Tumor tissue WGS data; actionable gene-based classification | Hierarchical clustering | – | [89] |
Imperiale, 2014 | CRC | 9989 | Multitarget stool DNA testing data; multimarker-based classification | LR | Sensitivity: 92.3% Specificity: 84.6% | [92] |
Luo, 2020 | CRC | 1822 | Circulating tumor DNA methylation markers; multimarker-based classification | LASSO/RF | AUC: 0.870 | [93] |
Romagnoni, 2019 | Crohn disease | 5277 | Genome-wide genotyping data; genetic variant-based classification | Penalized LR/GBT/ANN | AUC: 0.802 | [94] |
Chung, 2023 | CMMRD | 639 | Low-pass genomic instability characterization (LOGIC) assay; classification based on genomic microsatellite signature | LR | Sensitivity: 100% | [95] |
Zuo, 2022 | PEAC | 86 | Tumor tissue WES and targeted bisulfite sequencing data; DNA methylation-based classification | RF/LASSO/ SVM/ XGBoost | AUC: 0.900–1.000 | [96] |
Wan, 2019 | CRC | 817 | WGS data of plasma cfDNA; classification based on genetic features | PCA/SVM/ LR | AUC: 0.920 (95% CI, 0.910–0.930) | [97] |
Cakmak, 2023 | CRC | 115 | SNP profiles of immune phenotypes; prediction based on SNPs | LR/RF/SVM/KNN | AUC: 0.960 | [98] |
Guo, 2023 | CRC | 173 | Tissue RNA-seq data; WGCNA, classification based on key hub genes | LASSO | AUC: 0.821–1.000 | [99] |
Killcoyne, 2020 | EC | 412 | Shallow WGS data; classification based on genomic copy numbers | Elastic-net regression | Sensitivity: 72.0% Specificity: 82.0% | [100] |