Applications of Artificial Intelligence for Pediatric Cancer Imaging.

Artificial intelligence (AI) is transforming medical imaging of adult patients. However, its utilization in pediatric oncology imaging remains constrained, in part due to the inherent data scarcity associated with childhood cancers. Pediatric cancers are rare, and imaging technologies are evolving rapidly, leading to insufficient data of a particular type to effectively train these algorithms. The small market size of pediatrics compared to adults could also contribute to this challenge, as market size is a driver of commercialization. This article provides an overview of the current state of AI applications for pediatric cancer imaging, including applications for medical image acquisition, processing, reconstruction, segmentation, diagnosis, staging, and treatment response monitoring. While current developments are promising, impediments due to diverse anatomies of growing children and nonstandardized imaging protocols have led to limited clinical translation thus far. Opportunities include leveraging reconstruction algorithms to achieve accelerated low-dose imaging and automating the generation of metric-based staging and treatment monitoring scores. Transfer-learning of adult-based AI models to pediatric cancers, multi-institutional data sharing, and ethical data privacy practices for pediatric patients with rare cancers will be keys to unlocking AI's full potential for clinical translation and improved outcomes for these young patients.

Comparison of diffusion-weighted MRI and [18F]FDG PET/MRI for treatment monitoring in pediatric Hodgkin and non-Hodgkin lymphoma.

OBJECTIVE: To compare tumor therapy response assessments with whole-body diffusion-weighted imaging (WB-DWI) and 18F-fluorodeoxyglucose ([18F]FDG) PET/MRI in pediatric patients with Hodgkin lymphoma and non-Hodgkin lymphoma. MATERIALS AND METHODS: In a retrospective, non-randomized single-center study, we reviewed serial simultaneous WB-DWI and [18F]FDG PET/MRI scans of 45 children and young adults (27 males; mean age, 13 years ± 5 [standard deviation]; age range, 1-21 years) with Hodgkin lymphoma (n = 20) and non-Hodgkin lymphoma (n = 25) between February 2018 and October 2022. We measured minimum tumor apparent diffusion coefficient (ADCmin) and maximum standardized uptake value (SUVmax) of up to six target lesions and assessed therapy response according to Lugano criteria and modified criteria for WB-DWI. We evaluated the agreement between WB-DWI- and [18F]FDG PET/MRI-based response classifications with Gwet's agreement coefficient (AC). RESULTS: After induction chemotherapy, 95% (19 of 20) of patients with Hodgkin lymphoma and 72% (18 of 25) of patients with non-Hodgkin lymphoma showed concordant response in tumor metabolism and proton diffusion. We found a high agreement between treatment response assessments on WB-DWI and [18F]FDG PET/MRI (Gwet's AC = 0.94; 95% confidence interval [CI]: 0.82, 1.00) in patients with Hodgkin lymphoma, and a lower agreement for patients with non-Hodgkin lymphoma (Gwet's AC = 0.66; 95% CI: 0.43, 0.90). After completion of therapy, there was an excellent agreement between WB-DWI and [18F]FDG PET/MRI response assessments (Gwet's AC = 0.97; 95% CI: 0.91, 1). CONCLUSION: Therapy response of Hodgkin lymphoma can be evaluated with either [18F]FDG PET or WB-DWI, whereas patients with non-Hodgkin lymphoma may benefit from a combined approach. CLINICAL RELEVANCE STATEMENT: Hodgkin lymphoma and non-Hodgkin lymphoma exhibit different patterns of tumor response to induction chemotherapy on diffusion-weighted MRI and PET/MRI. KEY POINTS: • Diffusion-weighted imaging has been proposed as an alternative imaging to assess tumor response without ionizing radiation. • After induction therapy, whole-body diffusion-weighted imaging and PET/MRI revealed a higher agreement in patients with Hodgkin lymphoma than in those with non-Hodgkin lymphoma. • At the end of therapy, whole-body diffusion-weighted imaging and PET/MRI revealed an excellent agreement for overall tumor therapy responses for all lymphoma types.

Radiomics; A Potential Next "Omics" in Psychiatric Disorders; An Introduction.

Psychiatric disorders remain one of the most debilitating conditions; however, most patients are never diagnosed and do not seek treatment. Despite its massive burden on modern society and the health system, many hurdles prevent proper diagnosis and management of these disorders. The diagnosis is primarily based on clinical symptoms, and efforts to find appropriate biomarkers have not been practical. Through the past years, researchers have put a tremendous effort into finding biomarkers in "omics" fields: genomics, transcriptomics, proteomics, metabolomics, and epigenomics. This article reviews the evolving field of radiomics and its role in diagnosing psychiatric disorders as the sixth potential "omics." The first section of this paper elaborates on the definition of radiomics and its potential to provide a detailed structural study of the brain. Following that, we have provided the latest promising results of this novel approach in a broad range of psychiatric disorders. Radiomics fits well within the concept of psychoradiology. Besides volumetric analysis, radiomics takes advantage of many other features. This technique may open a new field in psychiatry for diagnosing and classifying psychiatric disorders and treatment response prediction in the era of precision and personalized medicine. The initial results are encouraging, but radiomics in psychiatry is still in its infancy. Despite the extensive burden of psychiatric disorders, there are very few published studies in this field, with small patient populations. The lack of prospective multi-centric studies and heterogeneity of studies in design are the significant barriers against the clinical adaptation of radiomics in psychoradiology.