Table 3 Advantages and disadvantages of different approaches in terms of TME and cancer prognosis
Methods | Advantages | Disadvantages |
|---|---|---|
Radiogenomics | Investigating the biological significance of tumors, facilitating a comprehensive understanding of gene phenotypes Predicting macrolevel imaging biomarkers associated with the genome, enabling noninvasive molecular typing as well as prognostic evaluation and efficacy assessment of tumors | Small sample, single-center study Easy to overfit data in the study results Time-consuming image segmentation High cost of genomics data acquisition |
Radiotranscriptomics | Exhibiting correlations, enabling exploration of biological significance Predicting tumor molecular typing and prognostic assessment and facilitates the codevelopment of novel biomarkers | Small sample size and single-center design Between transcriptomic and imaging data can show correlation but cannot establish causation Transcriptome data from current studies represent the average of cell populations, possibly failing to effectively reflect tumor heterogeneity. Addressing this issue could be achieved with single-cell RNA sequencing technology |
Radiopathomics | Integrating macroradiomics and microscale pathomics enables improved characterization of tumor heterogeneity and the development of novel biomarkers This combination redefines the concept of “digital biopsy” | Image segmentation is time-consuming The biological interpretation between radiomics and pathomics features requires further investigation The combination of genomic and proteomic data is necessary to build a comprehensive predictive model for characterizing tumors from the macroscopic to microscopic level |
Pathogenomics | Pathomics provides a global view of the tumor, genomics responds to the biology of the tumor, and the histological context of the genomic data is important for a comprehensive understanding of tumor heterogeneity The association of the two provides biological interpretation, and the combination of the two allows for the development of novel biomarkers and assessment of tumor prognosis in terms of treatment response | Small-sample and single-center studies The association between genomic and pathohistological data will reveal correlations but cannot determine causality Pathology image segmentation is time-consuming |
Multiomics combinations based on radiomics, pathomics, and genomics | Understanding the underlying biological basis of specific quantitative imaging features Obtaining comprehensive information to visualize the spatial and molecular context of cancer Discovering new diagnostic and prognostic markers Establishing a holistic and comprehensive model and methodology to explore tumorigenesis, progression, and prognosis | High dimensionality and heterogeneity of multiomics data Difficulty in acquiring multiomics data The need for further combined proteomics, metabolomics, transcriptomics and other multiomics data for integrated and comprehensive research and exploration |