Utilization of Radiomics in Prognostication and Treatment Response

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

The relatively new field of radiomics seeks to leverage previously under/unutilized aspects of diagnostic imaging as a clinically meaningful tool with promise to change the face of how imaging may be used in future clinical practice. Through analysis of features in individual voxels and how those features interact with similar features of surrounding voxels as well as establishing patterns within individual features or classifications of features throughout an imaging set, a vast amount of data can be generated and mined. Further, through the current field of radiogenomics, researchers are seeking to gain a better understanding as to what these individual radiomic features and patterns truly represent at a cellular and subcellular level. Since its inception, the concept of radiomics has promised the potential to provide actionable information through multiple aspects of clinical care. To date, a majority of radiomic-related publications have focused on two aspects of potential use: Providing information that can act as an adjunct to clinical and pathological prognostic criteria for patients with newly diagnosed malignancies and producing radiomic feature signatures that can predict likelihood and degree of response to specific treatments. With these, the potential clinical utility is apparent. Providing patients with the most accurate prognostic information possible allows for a more thorough and realistic discussion regarding treatment options and will help to better optimize healthcare resources. Predicting treatment response in an accurate real-time manner allows for better decision-making regarding the need or lack thereof for individual therapies, determination of total duration of treatment, and help to further clarify which treatment or combination thereof provides the best likelihood of treatment response in each individual patient. Of increasing interest, however, are two additional potential avenues through which radiomics can provide aid to treating clinicians. Using radiographic signals beyond apparent tumor size reduction to track response to ongoing treatments can provide vital information to patients and providers regarding whether a current treatment should be continued or abandoned for other options, particularly in tumors that are difficult to objectively assess using standard imaging techniques or show a mixed response on follow-up imaging. Further, the potential utility of radiomic signatures to provide information regarding histology, mutational status, or targetable receptor status with a high degree of accuracy could open an avenue for patients and clinicians to gain information on primary tumors and metastatic disease without the need for biopsies, reducing the associated risk and morbidity to patients in the process. Through this chapter, we seek to offer a glimpse into the general state of knowledge gained from and current failures of modern radiomics with a focus on each of the four directions of radiomic research delineated above. Our desire is to provide a high-level overview of radiomics literature with commentary throughout, highlighting areas of consistent promise while additionally underscoring areas of inconsistency or shortcomings in the data. With this, we hope to illustrate the continued and growing excitement in this field and provide areas in need of further development, helping to guide readers in the consideration and design of the future research directions that will shape this field in the coming years. In keeping with the theme of this work as a whole, we will also offer commentary on how artificial intelligence has been incorporated into the development of radiomics signatures and the general themes that emerge when assessing this field in its totality. The primary uses of radiomics in cancer is summarized at the end of this chapter, in Table 1.

Original languageEnglish (US)
Title of host publicationArtificial Intelligence in Radiation Oncology
PublisherWorld Scientific Publishing Co.
Pages273-304
Number of pages32
ISBN (Electronic)9789811263545
ISBN (Print)9789811263538
DOIs
StatePublished - Jan 1 2022

ASJC Scopus subject areas

  • General Biochemistry, Genetics and Molecular Biology
  • General Medicine
  • General Computer Science

Fingerprint

Dive into the research topics of 'Utilization of Radiomics in Prognostication and Treatment Response'. Together they form a unique fingerprint.

Cite this