From: Novel research and future prospects of artificial intelligence in cancer diagnosis and treatment
Type of cancer | Reference | Method | Year | Study population | Features and limitations | Performance |
---|---|---|---|---|---|---|
Breast cancer | Li et al. [208] | GBDT (XGBoost) | 2023 | SEER(2010–2019) | Focus on breast cancer brain metastases (BCBM) | 3-year survival AUC = 0.803 |
Li et al. [209] | SVM, CoxBoost | 2023 | TCGA | Establish robust and valid ROS signature (ROSig) to aid in assessing ROS levels | C-index: 0.736 for TCGA; 0.545 for Metabric | |
Verghese et al. [210] | FCN | 2023 | Hospital | Capture systemic immune features in lymph nodes | Dice coefficient of 0.86 and 0.74 for capturing GCs and sinuses, respectively | |
Li et al. [211] | RSF | 2022 | TCGA | Construct a novel hypoxia- and lactate metabolism-related gene signature | 5-year AUC: 0.638 | |
Wang et al. [212] | CNN | 2022 | ClinSeq, TCGA, SöS-BC, SCAN-B | Improve breast cancer histological grading | Hazard ratio [HR] 2.94, 95% confidence interval [CI] 1.24–6.97, p = 0.015 | |
Lung cancer | Ding et al. [213] | CNN(ResNet) | 2023 | Hospital | Assist pathologists in classifying histological patterns and prognosis stratification of LUAD patients | AUC: 0.93 |
She et al. [214] | Feed-forward deep neural network | 2020 | SEER | Explore the lack of studies on the performance of a deep learning survival neural network in non-small cell lung cancer (NSCLC) | C statistic = 0.739 vs. 0.706 | |
Hosny et al. [215] | CNN | 2018 | Hospital | Deep learning networks may be used for mortality risk stratification based on standard-of-care CT images from NSCLC patients | AUC: 0.70 in radiotherapy; AUC: 0.71 in surgery | |
Colorectal cancer | Finn et al. [216] | Multinomial logistic regression, elastic net regression, and random forest | 2023 | SEER-Medicare registry | Extend the ability of claims-based research to risk-adjust and stratify by stage | 95% CI, 0.43 to 0.46 |
Kleppe et al. [217] | MIL | 2022 | Hospital | Integrate DoMore-v1-CRC and pathological staging markers to provide a clinical decision-support system | 95% CI 6.39–17.93; p < 0.0001 | |
Bertsimas et al. [218] | RF, OPT | 2022 | Hospital | Provide a possible resolution to the long-standing debate on optimal margin width in CRLM | AUC: 0.76 | |
Kudo et al. [61] | ANN | 2021 | Hospital | Build a model to identify T1 colorectal tumors at risk for metastasis to lymph node and validate the model in a separate set of patients | AUC: 0.83 | |
Skrede et al. [219] | CNN | 2020 | Hospital | Develop a biomarker for patient outcome after primary colorectal cancer resection | 95% CI 2·72–5·43; p < 0·0001) | |
Prostate cancer | Deng et al. [220] | CNN | 2023 | Hospital | Predict Ki-67 expression in prostate cancer | AUC: 0.939–0.993 |
Saito et al. [221] | RSF, survival tree | 2023 | Hospital | Provide useful information for predicting the prognosis of metastatic prostate cancer | C-index: 0.64 | |
Lee et al. [222] | Cox proportional hazards, random survival forest, conditional inference survival forest, and DeepHit models | 2021 | SEER | Develop an improved prognostic model for predicting 10-year prostate cancer-specific mortality | C-index 0·829, 95% CI 0·820–0·838 | |
Pancreatic cancer | Nimgaonkar et al. [223] | CNN (HoVer-Net) | 2023 | TCGA | This imaging analysis pipeline has promise in the development of actionable markers in other clinical settings where few biomarkers currently exist | 95% CI [26.8, 63.9] |
Li et al. [224] | Random forest-based | 2023 | SEER | Explore two machine learning-based nomograms | 3-year OS: AUC, 0.792 (95% CI: 0.717–0.949) | |
Lee et al. [206] | Artificial neural network, logistic regression, random forest, gradient boosting, and support vector machine | 2022 | Hospital | Predict postoperative survival | 2-year OS: AUC, 0.67; p = 0.35 | |
Glioma | Voort et al. [106] | CNN | 2023 | Hospital | Noninvasively predicts multiple, clinically relevant features of glioma | AUC: 0.90 |
Skin cancer | Aung et al. [225] | ML | 2022 | TCGA and hospital | Evaluate the prognostic value of objective automated electronic TIL (eTIL) quantification | AUC: 0.793 |
Gastric cancer | Guan et al. [226] | SVM, RF | 2023 | Hospital | Evaluate and verify the predictive performance of computed tomography deep learning in gastric cancer | AUC: 0.9803 |
Oral cancer | Zhang et al. [227] | CNN | 2023 | Hospital | Identified OL patients with a high risk of OC development | HR = 4.52, 1.5–13.7 |
Singh et al. [228] | SVM, naïve Bayes, decision trees, multi-Layer perceptron, logistic regression, and K means (unsupervised) | 2022 | PIK3CA, KRAS, TP53 and Gingival | Reveal key candidate attributes for GBC prognosis | MLP accuracy: 63% |