Advanced PET radiomics parameters have been recently proposed for histologic definition and prognosis prediction, evaluating the intra-lesion distribution of 18F-FDG. Possible radiomic quantitative analysis can be applied to oncological hematology, allowing further clinical applications such as survival prediction, assessment of bone marrow involvement, and differentiation of lymphoma from other malignancies. A specific systematic review of papers about PET radiomics in patients with lymphoma was already proposed in 2021. To update the timeframe 2021–2024, most recent papers were reported in this chapter. At the same time, artificial intelligence (AI) tools could be combined with radiomics parameters to improve the definition of treatment strategies such as tumor subtypes, survival time, and disease recurrence. These kinds of models can be established on multi-parametric sources of data to personalize the strategy of management of patients, alone or in combination with previously established clinical, biological, and laboratory markers.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Radiomics and Artificial Intelligence

  • Salvatore Annunziata,
  • Daniele Antonio Pizzuto,
  • Marco De Summa

摘要

Advanced PET radiomics parameters have been recently proposed for histologic definition and prognosis prediction, evaluating the intra-lesion distribution of 18F-FDG. Possible radiomic quantitative analysis can be applied to oncological hematology, allowing further clinical applications such as survival prediction, assessment of bone marrow involvement, and differentiation of lymphoma from other malignancies. A specific systematic review of papers about PET radiomics in patients with lymphoma was already proposed in 2021. To update the timeframe 2021–2024, most recent papers were reported in this chapter. At the same time, artificial intelligence (AI) tools could be combined with radiomics parameters to improve the definition of treatment strategies such as tumor subtypes, survival time, and disease recurrence. These kinds of models can be established on multi-parametric sources of data to personalize the strategy of management of patients, alone or in combination with previously established clinical, biological, and laboratory markers.