Radiomics analysis of bone marrow biopsy locations in [18F]FDG PET/CT images for measurable residual disease assessment in multiple myeloma

Abstract

The combination of visual assessment of whole body [18F]FDG PET images and evaluation of bone marrow samples by Multiparameter Flow Cytometry (MFC) or Next-Generation Sequencing (NGS) is currently the most common clinical practice for the detection of Measurable Residual Disease (MRD) in Multiple Myeloma (MM) patients. In this study, radiomic features extracted from the bone marrow biopsy locations are analyzed and compared to those extracted from the whole bone marrow in order to study the representativeness of these biopsy locations in the image-based MRD assessment. Whole body [18F]FDG PET of 39 patients with newly diagnosed MM were included in the database, and visually evaluated by experts in nuclear medicine. A methodology for the segmentation of biopsy sites from PET images, including sternum and posterior iliac crest, and their subsequent quantification is proposed. First, starting from the bone marrow segmentation, a segmentation of the biopsy sites is performed. Then, segmentations are quantified extracting SUV metrics and radiomic features from the [18F]FDG PET images and are evaluated by Mann–Whitney U-tests as valuable features differentiating PET+/PET− and MFC+ /MFC− groups. Moreover, correlation between whole bone marrow and biopsy sites is studied by Spearman ρ rank. Classification performance of the radiomics features is evaluated applying seven machine learning algorithms. Statistical analyses reveal that some images features are significant in PET+/PET− differentiation, such as SUVmax, Gray Level Non-Uniformity or Entropy, especially with a balanced database where 16 of the features show a p value < 0.001. Correlation analyses between whole bone marrow and biopsy sites results in significant and acceptable coefficients, with 11 of the variables reaching a correlation coefficient greater than 0.7, with a maximum of 0.853. Machine learning algorithms demonstrate high performances in PET+/PET− classification reaching a maximum AUC of 0.974, but not for MFC+/MFC− classification. The results demonstrate the representativeness of sample sites as well as the effectiveness of extracted features (SUV metrics and radiomic features) from the [18F]FDG PET images in MRD assessment in MM patients.

Positron Emission Tomography-Derived Radiomics and Artificial Intelligence in Multiple Myeloma: State-of-the-Art

“…The PRISMA flowchart of the articles included in our review is reported in Figu The literature search identified a total of 31 studies. According to the aforementio eligibility criteria, 10 studies were included [17][18][19][20][21][22][23][24][25][26]. Among these, eight studies w retrospective and two prospective.…”

“…R-Sofware [17,19,22] and Python [22,25,26] were the most used AI open source programming languages for the endpoint prediction and statistical analysis. In addition, Orange [18,21] was the most used toolkit to build the predictive model. The main data-mining algorithms used to perform ML strategies were random forests, including their extension in random survival forest [17,18,20,21,23,25] and k-nearest neighbors [18,20,21,23,26].…”

“…In addition, Orange [18,21] was the most used toolkit to build the predictive model. The main data-mining algorithms used to perform ML strategies were random forests, including their extension in random survival forest [17,18,20,21,23,25] and k-nearest neighbors [18,20,21,23,26]. The first two mentioned classifiers belong to the tree-based class.…”

“…Yet, they outperformed traditional ML methods, such as the random forest classifier (RF), k-nearest neighbors (k-NN), and the support vector machine (SVM). In addition, internal validation was performed in 8 out of 10 studies using a fold-cross validation procedure [17,18,[20][21][22][23]25,26]. Table 3 summarizes studies' data mining.…”

“…MM diagnosis is based on the so-called "CRAB criteria", which include hypercalcemia (C), renal damage (R), anemia (A), and lytic bone lesions (B), along with at least 10% of pathological infiltration on BM examination [3]. The NCCN guidelines recommend either whole-body (WB) low-dose computed tomography (CT) or [ 18 F]1fluorodeoxyglucose (FDG) positron emission tomography (PET)/CT for the initial diagnosis, staging, evaluation of treatment outcomes, and follow-up [4]. WB-magnetic resonance imaging (MRI) might be helpful in the case of diffuse BM infiltration and to evaluate the axial skeleton, particularly for localizing spinal cord and/or nerve root compression for surgery or radiotherapy [5].…”