Radiomics and Artificial Intelligence for Outcome Prediction in Multiple Myeloma Patients Undergoing Autologous Transplantation: A Feasibility Study with CT Data

Abstract

Multiple myeloma is a plasma cell dyscrasia characterized by focal and non-focal bone lesions. Radiomic techniques extract morphological information from computerized tomography images and exploit them for stratification and risk prediction purposes. However, few papers so far have applied radiomics to multiple myeloma. A retrospective study approved by the institutional review board: n = 51 transplanted patients and n = 33 (64%) with focal lesion analyzed via an open-source toolbox that extracted 109 radiomics features. We also applied a dedicated tool for computing 24 features describing the whole skeleton asset. The redundancy reduction was realized via correlation and principal component analysis. Fuzzy clustering (FC) and Hough transform filtering (HTF) allowed for patient stratification, with effectiveness assessed by four skill scores. The highest sensitivity and critical success index (CSI) were obtained representing each patient, with 17 focal features selected via correlation with the 24 features describing the overall skeletal asset. These scores were higher than the ones associated with a standard cytogenetic classification. The Mann–Whitney U-test showed that three among the 17 imaging descriptors passed the null hypothesis. This AI-based interpretation of radiomics features stratified relapsed and non-relapsed MM patients, showing some potentiality for the determination of the prognostic image-based biomarkers in disease follow-up.

Current Status and Future of Artificial Intelligence in MM Imaging: A Systematic Review

“…Out of the nine articles and four abstracts included, 6/14 (43%) used CT images [23][24][25][26][27][28], 3/14 (21%) used PET-CT scans [29][30][31], and 5/14 (36%) used MRI as their imaging modality [17,[32][33][34]. A total of 5/14 (36%) studies used WB imaging or imaging from affected body parts, 1/14 (%) focused on the femur bone, 1/14 (7%) focused on the spine and pelvic bones, and 3/14 (%) focused on the spine alone.…”

A Review of Artificial Intelligence Applications in Hematology Management: Current Practices and Future Prospects

“…Values, n (%) Pathway stage [1,6,7,29,38,40,41,50,51,53,55,64,69,82,92,94,96,[99][100][101][102]105,117,118,136,143,144] 27 (20.6) Prediction [3,10,28,30,[32][33][34]36,[44][45][46]57,58,61,66,71,72,[76][77][78]80,83,85,89,91,…”

“…Normally, these blood elements mature and replenish depending on the body’s needs [ 5 ]. However, their growth can become disordered when certain hematologic malignancies are present [ 6 ]. For instance, because of the considerable increase in the number of abnormal WBCs, the ability of bone marrow to generate and support healthy RBCs and platelets in terms of oxygen and nutrition supply is impaired [ 2 , 7 ].…”

“…Motivated by the remarkable achievements of artificial intelligence (AI) in various fields, the applicability of such algorithms in solving critical problems related to oncology and hematology was recently investigated and proven efficient [ 6 ]. In particular, machine learning (ML) and deep learning (DL) methods were used to assist to classify various cancer types, facilitate faster diagnosis, and provide a basis for accurate clinical decisions for better health outcomes.…”

Segmentation accuracy and the reliability of radiomics features

“…Therefore, these imaging features can theoretically be used as cheaper and early surrogates of the genomic ones for prognostic purposes and as complementary information within a personalized diagnostic framework. Radiomics analysis was applied to several types of tumors, like breast [5], oropharyngeal [6] and pancreatic [7] cancer, as well as multiple myeloma [8] and brain tumors [9].…”