Table 1 Features and challenges of cancer patient’s response prediction system using radiotherapy and risk.
Author [citation] | Methodology | Features | Challenges |
|---|---|---|---|
Hu et al.18 | XGBoost | It enhances predictions for individual patient risk It effectively improves the personalized treatment regimens and individual outcomes | It provides instability outcomes It suffers from overfitting issue |
Kim et al.21 | SLR | It gives high accuracy, specificity and sensitivity values It effectively identifies the robust gene signatures of clinical relevance | It does not support early diagnosis treatment It does not identify the younger patient’s symptoms |
Bernchou et al.19 | RM | It effectively identifies patients with low risk of symptomatic toxic It reduces the multi collinearity in the genome to effectively identify the affected genome | It suffers from minimal radiation issue It slightly decreases the survival rates |
Kim et al.23 | LR | It is useful for enhancing the treatment efficacy and it helps to avoid the ineffective drugs It increases the risk prediction score | It takes a lot of time to train the model It struggles to predict the time complexity |
Chen et al.24 | GSEA | It improves the prediction of prognosis and assists treatment stratification It effectively improves the treatment of cancer patients who have mild symptoms | It is not helpful for predicting the response of patients automatically It has a greater number of error rates |
Wilkins et al.25 | BCR | It effectively gives the warning of individual mortality risk It provides dose reductions to organs at risk | Training the model requires lot of power The implementation cost is high |
Cozzarini et al.22 | CNN | It maximizes the precision of delivery of the radiation dose It gives accurate prediction response for diseased patients | It struggles to train the model with a large dataset It is hard to predict the matched patient’s response |
Defraene et al.20 | ANN | It gives high dimensional dose distribution information in early stage It gives reliable and flexible outcomes | It increases the throughput and predictive power It is unstable and more complicated |
Pati et al.26 | DL | It enhances accuracy, ensuring reliable cancer diagnosis It is capable of diagnosing breast cancer in real-time | It may require significant resources for feature reduction and transfer learning Real-world deployment and adoption may face challenges |
Sahoo et al.27 | ML | It combines multiple ML models using various voting techniques for improved predictions It provides treatment planning by enhancing relapses and metastasis prediction accuracy | The employment of multiple models increases computational overhead and complexity |
