Radiomics and artificial intelligence in prostate cancer: new tools for molecular hybrid imaging and theragnostics

Abstract

In prostate cancer (PCa), the use of new radiopharmaceuticals has improved the accuracy of diagnosis and staging, refined surveillance strategies, and introduced specific and personalized radioreceptor therapies. Nuclear medicine, therefore, holds great promise for improving the quality of life of PCa patients, through managing and processing a vast amount of molecular imaging data and beyond, using a multi-omics approach and improving patients' risk-stratification for tailored medicine. Artificial intelligence (AI) and radiomics may allow clinicians to improve the overall efficiency and accuracy of using these "big data" in both the diagnostic and theragnostic field: from technical aspects (such as semi-automatization of tumor segmentation, image reconstruction, and interpretation) to clinical outcomes, improving a deeper understanding of the molecular environment of PCa, refining personalized treatment strategies, and increasing the ability to predict the outcome. This systematic review aims to describe the current literature on AI and radiomics applied to molecular imaging of prostate cancer.

Keywords: Artificial intelligence; Positron emission tomography; Prostate cancer; Radiomics; Theragnostics.

Fig. 1

The workflow includes the steps required in a radiomic and artificial intelligence analysis in prostate cancer patients. The first step involves collecting clinical data on patient characteristics, histopathological data on tumor characteristics, and imaging data, with the extraction of radiomic features (such as shape, intensity, and texture features). Radiomic modeling involves three major aspects: feature selection, modeling methodology, and validation. The number of radiomic features that can be extracted from images is virtually unlimited. Once extracted, radiomic features must be selected; redundant or non-robust features against sources of variability must be identified and eliminated through dimensionality reduction techniques, to avoid overfitting problems. The choice of modeling methodology and the identification of optimal machine learning methods for radiomic applications are a crucial step in obtaining robust and clinically relevant results. The choice of a modeling methodology (supervised or unsupervised machine learning method) depends on the setting of the data, the characteristics of the analyzed population, and the experience of the researchers. The model chosen affects prediction and performance in radiomics, and hence, implementations of multiple modeling methodologies are highly desirable. Finally, validation techniques are useful tools for assessing model performance. An externally validated model has more credibility than an internally validated model because data obtained by the first approach are more independent. Validation is essential to verify the repeatability and reproducibility of the model, demonstrating statistical consistency between the training and validation datasets

Fig. 2

Schematic representation of the performed literature search and the review strategy