A review of computational modeling, machine learning and image analysis in cancer metastasis dynamics

Hirway, Shreyas U.; Weinberg, Seth H.

doi:10.1002/cso2.1044

Cited by 3 publications

(1 citation statement)

References 191 publications

(324 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is no Navier–Stokes equation rushing to the rescue and one can therefore naturally wonder whether hand-crafted models ( 33 , 34 ) could perhaps be usefully replaced by purely data-driven ones. This question is being actively investigated for a number of experimental cell motility systems ( 35 , 36 ) and of course is under intense study in many biomedical contexts, see e.g. work on digital pathology ( 37 ).…”

Section: The Spreading Of MLmentioning

confidence: 99%

Machine learning meets physics: A two-way street

Levine,

2024

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

Section: The Spreading Of MLmentioning

confidence: 99%

Machine learning meets physics: A two-way street

Levine,

2024

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

How much do we know about the metastatic process?

Rodriguez-Tirado,

Sosa

2024

Clin Exp Metastasis

View full text Add to dashboard Cite

Cancer cells can leave their primary sites and travel through the circulation to distant sites, where they lodge as disseminated cancer cells (DCCs), even during the early and asymptomatic stages of tumor progression. In experimental models and clinical samples, DCCs can be detected in a non-proliferative state, defined as cellular dormancy. This state can persist for extended periods until DCCs reawaken, usually in response to niche-derived reactivation signals. Therefore, their clinical detection in sites like lymph nodes and bone marrow is linked to poor survival. Current cancer therapy designs are based on the biology of the primary tumor and do not target the biology of the dormant DCC population and thus fail to eradicate the initial or subsequent waves of metastasis. In this brief review, we discuss the current methods for detecting DCCs and highlight new strategies that aim to target DCCs that constitute minimal residual disease to reduce or prevent metastasis formation. Furthermore, we present current evidence on the relevance of DCCs derived from early stages of tumor progression in metastatic disease and describe the animal models available for their study. We also discuss our current understanding of the dissemination mechanisms utilized by genetically less- and more-advanced cancer cells, which include the functional analysis of intermediate or hybrid states of epithelial–mesenchymal transition (EMT). Finally, we raise some intriguing questions regarding the clinical impact of studying the crosstalk between evolutionary waves of DCCs and the initiation of metastatic disease.

show abstract

Comparative Performance of Autoencoders and Traditional Machine Learning Algorithms in Clinical Data Analysis for Predicting Post-Staged GKRS Tumor Dynamics

Volovăț,

Popa,

Rusu

et al. 2024

Diagnostics

View full text Add to dashboard Cite

Introduction: Accurate prediction of tumor dynamics following Gamma Knife radiosurgery (GKRS) is critical for optimizing treatment strategies for patients with brain metastases (BMs). Traditional machine learning (ML) algorithms have been widely used for this purpose; however, recent advancements in deep learning, such as autoencoders, offer the potential to enhance predictive accuracy. This study aims to evaluate the efficacy of autoencoders compared to traditional ML models in predicting tumor progression or regression after GKRS. Objectives: The primary objective of this study is to assess whether integrating autoencoder-derived features into traditional ML models can improve their performance in predicting tumor dynamics three months post-GKRS in patients with brain metastases. Methods: This retrospective analysis utilized clinical data from 77 patients treated at the “Prof. Dr. Nicolae Oblu” Emergency Clinic Hospital-Iasi. Twelve variables, including socio-demographic, clinical, treatment, and radiosurgery-related factors, were considered. Tumor progression or regression within three months post-GKRS was the primary outcome, with 71 cases of regression and 6 cases of progression. Traditional ML models, such as Logistic Regression, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Extra Trees, Random Forest, and XGBoost, were trained and evaluated. The study further explored the impact of incorporating features derived from autoencoders, particularly focusing on the effect of compression in the bottleneck layer on model performance. Results: Traditional ML models achieved accuracy rates ranging from 0.91 (KNN) to 1.00 (Extra Trees). Integrating autoencoder-derived features generally enhanced model performance. Logistic Regression saw an accuracy increase from 0.91 to 0.94, and SVM improved from 0.85 to 0.96. XGBoost maintained consistent performance with an accuracy of 0.94 and an AUC of 0.98, regardless of the feature set used. These results demonstrate that hybrid models combining deep learning and traditional ML techniques can improve predictive accuracy. Conclusion: The study highlights the potential of hybrid models incorporating autoencoder-derived features to enhance the predictive accuracy and robustness of traditional ML models in forecasting tumor dynamics post-GKRS. These advancements could significantly contribute to personalized medicine, enabling more precise and individualized treatment planning based on refined predictive insights, ultimately improving patient outcomes.

show abstract

A review of computational modeling, machine learning and image analysis in cancer metastasis dynamics

Cited by 3 publications

References 191 publications

Machine learning meets physics: A two-way street

Machine learning meets physics: A two-way street

How much do we know about the metastatic process?

Comparative Performance of Autoencoders and Traditional Machine Learning Algorithms in Clinical Data Analysis for Predicting Post-Staged GKRS Tumor Dynamics

Contact Info

Product

Resources

About