Role of computed tomography texture analysis using dual-energy-based bone marrow imaging for multiple myeloma characterization: comparison with histology and established serologic parameters

MRI-based bone marrow radiomics for predicting cytogenetic abnormalities in multiple myeloma

AIM

To develop a radiomics signature applied to magnetic resonance imaging (MRI)-images to predict cytogenetic abnormalities in multiple myeloma (MM).

MATERIALS AND METHODS

Patients with newly diagnosed MM were enrolled retrospectively from March 2019 to September 2022. They were categorised into the high-risk cytogenetics (HRC) group and standard-risk cytogenetics (SRC) group. The patients were allocated randomly at a ratio of 7:3 into training and validation cohorts. Volumes of interest (VOI) was drawn manually on fat suppression T2-weighted imaging (FS-T2WI) and copied to the same location of the T1-weighted imaging (T1WI) sequence. Radiomics features were extracted from two sequences and selected by reproducibility and redundant analysis. The least absolute shrinkage selection operation (LASSO) algorithm was applied to build the radiomics signatures. The performance of the radiomics signatures to distinguish HRC with SRC was evaluated by ROC curves. The area under the curve (AUC), specificity, and sensitivity were also calculated.

RESULTS

A total of 105 MM patients were enrolled in this study. The four and 11 most significant and relevant features were selected separately from T1WI and FS-T2WI sequences to build the radiomics signatures based on the training cohort. Compared to the T1WI sequence, the radiomics signature based on the FS-T2WI sequence achieved better performance with AUCs of 0.896 and 0.729 in the training and validation cohorts respectively. A sensitivity of 0.833, specificity of 0.667, and Youden index of 0.500 were achieved for the FS-T2WI radiomics signature in the validation cohort.

CONCLUSIONS

The radiomics signature based on MRI provides a non-invasive and convenient tool to predict cytogenetic abnormalities in MM patients.

Quantitative assessment of bone marrow infiltration and characterization of tumor burden using dual-layer spectral CT in patients with multiple myeloma

BACKGROUND

The aim of the study was to evaluate whether virtual calcium subtraction (VNCa) image extracted from dual-layer spectral CT could estimate bone marrow (BM) infiltration with MRI as the reference standard and characterize tumor burden in patients with multiple myeloma (MM).

PATIENTS AND METHODS

Forty-seven patients with newly diagnosed MM were retrospectively enrolled. They had undergone whole-body low-dose dual-layer spectral CT (DLCT) and whole-body MRI within one week. VNCa images with calcium-suppressed (CaSupp) indices ranging from 25 to 95 at an interval of 10 and apparent diffusion coefficient (ADC) maps were quantitatively analyzed on vertebral bodies L1-L5 at the central slice of images. The optimal combination was selected by correlation analysis between CT numbers and ADC values. Then, it was used to characterize tumor burden by correlation analysis and receiver operating characteristic (ROC) curves analysis, including plasma cell infiltration rate (PCIR), high serum-free light chains (SFLC) ratio and the high-risk cytogenetic (HRC) status.

RESULTS

The most significant quantitative correlation between CT numbers of VNCa images and ADC values could be found at CaSupp index 85 for averaged L1-L5 (r = 0.612, p < 0.001). It allowed quantitative evaluation of PCIR (r = 0.835, p < 0.001). It could also anticipate high SFLC ratio and the HRC status with area under the curve (AUC) of 0.876 and 0.760, respectively.

CONCLUSIONS

The VNCa measurements of averaged L1-L5 showed the highest correlation with ADC at CaSupp index 85. It could therefore be used as additional imaging biomarker for non-invasive assessment of tumor burden if ADC is not feasible.

Imaging of Multiple Myeloma: Present and Future

Multiple myeloma (MM) is the second most common adult hematologic malignancy, and early intervention increases survival in asymptomatic high-risk patients. Imaging is crucial for the diagnosis and follow-up of MM, as the detection of bone and bone marrow lesions often dictates the decision to start treatment. Low-dose whole-body computed tomography (CT) is the modality of choice for the initial assessment, and dual-energy CT is a developing technique with the potential for detecting non-lytic marrow infiltration and evaluating the response to treatment. Magnetic resonance imaging (MRI) is more sensitive and specific than 18F-fluorodeoxyglucose positron emission tomography/computed tomography (FDG-PET/CT) for the detection of small focal lesions and diffuse marrow infiltration. However, FDG-PET/CT is recommended as the modality of choice for follow-up. Recently, diffusion-weighted MRI has become a new technique for the quantitative assessment of disease burden and therapy response. Although not widespread, we address current proposals for structured reporting to promote standardization and diminish variations. This review provides an up-to-date overview of MM imaging, indications, advantages, limitations, and recommended reporting of each technique. We also cover the main differential diagnosis and pitfalls and discuss the ongoing controversies and future directions, such as PET-MRI and artificial intelligence.

Discriminating minimal residual disease status in multiple myeloma based on MRI: utility of radiomics and comparison of machine-learning methods

AIM

To explore the possibility of discriminating minimal residual disease (MRD) status in multiple myeloma (MM) based on magnetic resonance imaging (MRI) and identify optimal machine-learning methods to optimise the clinical treatment regimen.

MATERIALS AND METHODS

A total of 83 patients were analysed retrospectively. They were divided randomly into training and validation cohorts. The regions of interest were segmented and radiomics features were extracted and analysed on two sequences, including T1-weighted imaging (WI) and fat saturated (FS)-T2WI, and then radiomics models were built in the training cohort and evaluated in the validation cohort. Clinical characteristics were calculated to build a traditional model. A combined model was also built using the clinical characteristics and radiomics features. Classification accuracy was assessed using area under the curve (AUC) and F1 score.

RESULTS

In the training cohort, only the bone marrow (BM) infiltrate ratio (p=0.005) was retained after univariate and multivariable logistic regression analysis. In T1WI, the linear support vector machine (SVM) achieved the best performance compared to other classifiers, with AUCs of 0.811 and 0.708 and F1 scores of 0.792 and 0.696 in the training and validation cohorts, respectively. Similarly, in FS-T2WI sequence, linear SVM achieved the best performance with AUCs of 0.833 and 0.800 and F1 score of 0.833 and 0.800. The combined model constructed by the FS-T2WI-linear SVM and BM infiltrate ratio outperformed the traditional model (p=0.050 and 0.012, Delong test), but showed no significant difference compared with the radiomics model (p=0.798 and 0.855).

CONCLUSION

The linear SVM-based machine-learning method can offer a non-invasive tool for discriminating MRD status in MM.

Lumbar MR-based radiomics nomogram for detecting minimal residual disease in patients with multiple myeloma

OBJECTIVES

Minimal residual disease (MRD) is a standard for assessing treatment response in multiple myeloma (MM). MRD negativity is considered to be the most powerful predictor of long-term good outcomes. This study aimed to develop and validate a radiomics nomogram based on magnetic resonance imaging (MRI) of the lumbar spine to detect MRD after MM treatment.

METHODS

A total of 130 MM patients (55 MRD negative and 75 MRD positive) who had undergone MRD testing through next-generation flow cytometry were divided into a training set (n = 90) and a test set (n = 40). Radiomics features were extracted from lumbar spinal MRI (T1-weighted images and fat-suppressed T2-weighted images) by means of the minimum redundancy maximum relevance method and the least absolute shrinkage and selection operator algorithm. A radiomics signature model was constructed. A clinical model was established using demographic features. A radiomics nomogram incorporating the radiomics signature and independent clinical factor was developed using multivariate logistic regression analysis.

RESULTS

Sixteen features were used to establish the radiomics signature. The radiomics nomogram included the radiomics signature and the independent clinical factor (free light chain ratio) and showed good performance in detecting the MRD status (area under the curve: 0.980 in the training set and 0.903 in the test set).

CONCLUSIONS

The lumbar MRI-based radiomics nomogram showed good performance in detecting MRD status in MM patients after treatment, and it is helpful for clinical decision-making.

KEY POINTS

• The presence or absence of minimal residual disease status has a strong predictive significance for the prognosis of patients with multiple myeloma. • A radiomics nomogram based on lumbar MRI is a potential and reliable tool for evaluating minimal residual disease status in MM.

Radiomics Analysis for Multiple Myeloma: A Systematic Review with Radiomics Quality Scoring

Multiple myeloma (MM) is one of the most common hematological malignancies affecting the bone marrow. Radiomics analysis has been employed in the literature in an attempt to evaluate the bone marrow of MM patients. This manuscript aimed to systematically review radiomics research on MM while employing a radiomics quality score (RQS) to accurately assess research quality in the field. A systematic search was performed on Web of Science, PubMed, and Scopus. The selected manuscripts were evaluated (data extraction and RQS scoring) by three independent readers (R1, R2, and R3) with experience in radiomics analysis. A total of 23 studies with 2682 patients were included, and the median RQS was 10 for R1 (IQR 5.5-12) and R3 (IQR 8.3-12) and 11 (IQR 7.5-12.5) for R2. RQS was not significantly correlated with any of the assessed bibliometric data (impact factor, quartile, year of publication, and imaging modality) (p > 0.05). Our results demonstrated the low quality of published radiomics research in MM, similarly to other fields of radiomics research, highlighting the need to tighten publication standards.

Combining Deep Learning and Radiomics for Automated, Objective, Comprehensive Bone Marrow Characterization From Whole-Body MRI: A Multicentric Feasibility Study

OBJECTIVES

Disseminated bone marrow (BM) involvement is frequent in multiple myeloma (MM). Whole-body magnetic resonance imaging (wb-MRI) enables to evaluate the whole BM. Reading of such whole-body scans is time-consuming, and yet radiologists can transfer only a small fraction of the information of the imaging data set to the report. This limits the influence that imaging can have on clinical decision-making and in research toward precision oncology. The objective of this feasibility study was to implement a concept for automatic, comprehensive characterization of the BM from wb-MRI, by automatic BM segmentation and subsequent radiomics analysis of 30 different BM spaces (BMS).

MATERIALS AND METHODS

This retrospective multicentric pilot study used a total of 106 wb-MRI from 102 patients with (smoldering) MM from 8 centers. Fifty wb-MRI from center 1 were used for training of segmentation algorithms (nnU-Nets) and radiomics algorithms. Fifty-six wb-MRI from 8 centers, acquired with a variety of different MRI scanners and protocols, were used for independent testing. Manual segmentations of 2700 BMS from 90 wb-MRI were performed for training and testing of the segmentation algorithms. For each BMS, 296 radiomics features were calculated individually. Dice score was used to assess similarity between automatic segmentations and manual reference segmentations.

RESULTS

The "multilabel nnU-Net" segmentation algorithm, which performs segmentation of 30 BMS and labels them individually, reached mean dice scores of 0.88 ± 0.06/0.87 ± 0.06/0.83 ± 0.11 in independent test sets from center 1/center 2/center 3-8 (interrater variability between radiologists, 0.88 ± 0.01). The subset from the multicenter, multivendor test set (center 3-8) that was of high imaging quality was segmented with high precision (mean dice score, 0.87), comparable to the internal test data from center 1. The radiomic BM phenotype consisting of 8880 descriptive parameters per patient, which result from calculation of 296 radiomics features for each of the 30 BMS, was calculated for all patients. Exemplary cases demonstrated connections between typical BM patterns in MM and radiomic signatures of the respective BMS. In plausibility tests, predicted size and weight based on radiomics models of the radiomic BM phenotype significantly correlated with patients' actual size and weight ( P = 0.002 and P = 0.003, respectively).

CONCLUSIONS

This pilot study demonstrates the feasibility of automatic, objective, comprehensive BM characterization from wb-MRI in multicentric data sets. This concept allows the extraction of high-dimensional phenotypes to capture the complexity of disseminated BM disorders from imaging. Further studies need to assess the clinical potential of this method for automatic staging, therapy response assessment, or prediction of biopsy results.

Utility of dual energy computed tomography in the evaluation of infiltrative skeletal lesions and metastasis: a literature review

Computed tomography (CT) is routinely used to diagnose and evaluate metastatic lesions in oncology. CT alone suffers from lack of sensitivity, especially for skeletal lesions in the bone marrow and lesions that have similar attenuation profiles to surrounding bone. Magnetic resonance imaging and nuclear medicine imaging remain the gold standard in evaluating skeletal lesions. However, compared to CT, these modalities are not as widely available or suitable for all patients. Dual energy computed tomography (DECT) exploits variations in linear attenuation coefficient of materials at different photon energy levels to reconstruct images based on material composition. DECT in musculoskeletal imaging is used in the imaging of crystal arthropathy and detecting subtle fractures, but it is not broadly utilized in evaluating infiltrative skeletal lesions. Malignant skeletal lesions have different tissue and molecular compositions compared to normal bone. DECT may exploit these physical differences to delineate infiltrative skeletal lesions from surrounding bone better than conventional monoenergetic CT. Studies so far have examined the utility of DECT in evaluating skeletal metastases, multiple myeloma lesions, pathologic fractures, and performing image-guided biopsies with promising results. These studies were mostly retrospective analyses and case reports containing small samples sizes. As DECT becomes more widely used clinically and more scientific studies evaluating the performance of DECT are published, DECT may eventually become an important modality in the work-up of infiltrative skeletal lesions. It may even challenge MRI and nuclear medicine because of relatively faster scanning times and ease of access.

Texture Features of Computed Tomography Image under the Artificial Intelligence Algorithm and Its Predictive Value for Colorectal Liver Metastasis

The aim of this research was to investigate the predictive role of texture features in computed tomography (CT) images based on artificial intelligence (AI) algorithms for colorectal liver metastases (CRLM). A total of 150 patients with colorectal cancer who were admitted to the hospital were selected as the research objects and randomly divided into three groups with 50 cases in each group. The patients who were found to suffer from the CRLM in the initial examination were included in group A. Patients who were found with CRLM in the follow-up were assigned to group B (B1: metastasis within 0.5 years, 16 cases; B2: metastasis within 0.5–1.0 years, 17 cases; and B3: metastasis within 1.0–2.0 years, 17 cases). Patients without liver metastases during the initial examination and subsequent follow-up were designated as group C. Image textures were analyzed for patients in each group. The prediction accuracy, sensitivity, and specificity of CRLM in patients with six classifiers were calculated, based on which the receiver operator characteristic (ROC) curves were drawn. The results showed that the logistic regression (LR) classifier had the highest prediction accuracy, sensitivity, and specificity, showing the best prediction effect, followed by the linear discriminant (LD) classifier. The prediction accuracy, sensitivity, and specificity of the LR classifier were higher in group B1 and group B3, and the prediction effect was better than that in group B2. The texture features of CT images based on the AI algorithms showed a good prediction effect on CRLM and had a guiding significance for the early diagnosis and treatment of CRLM. In addition, the LR classifier showed the best prediction effect and high clinical value and can be popularized and applied.

Correlation between Whole Skeleton Dual Energy CT Calcium-Subtracted Attenuation and Bone Marrow Infiltration in Multiple Myeloma

•

Quantification of whole skeleton calcium-subtracted attenuation with dual energy CT is feasible.

•

Whole skeleton calcium-subtracted attenuation correlates with the degree of marrow infiltration by plasma cells on bone marrow biopsy.

•

Whole skeleton calcium-subtracted attenuation provides complementary information to the detection of osteolytic bone lesions.

Objectives

Objective evaluation of the extent of skeletal marrow involvement in multiple myeloma remains a clinical gap for CT. We aimed to develop a quantitative segmentation pipeline for dual energy CT and to assess whether quantified whole skeleton calcium-subtracted attenuation values correlate with biopsy-derived bone marrow infiltration in multiple myeloma.

Methods

Consecutive prospective patients with suspected/established myeloma underwent dual source CT from the skull vertex to proximal tibia. Whole skeleton segmentation was performed for 120 kVp-equivalent images as follows: following Hounsfield unit (HU) thresholding, a Chan-Vese morphological operation was implemented to generate a whole skeleton segmentation mask. This mask was then applied to corresponding whole skeleton material decomposition calcium-subtracted maps, generating whole skeleton HU values. Associations with biopsy-derived bone marrow plasma cell infiltration percentage were assessed with Spearman’s rank correlation; significance was at 5%.

Results

21 patients (12 females; median (IQR) 67 (61, 73) years) were included; 16 patients had osteolytic bone lesions; 15 patients underwent bone marrow biopsy. Segmentation and quantification were feasible in all patients. Median (IQR) of the average skeletal calcium-subtracted attenuation was −59.9 HU (-66.3, −51.8HU). There was a positive correlation with bone marrow plasma cell infiltration percentage (Spearman’s rho: + 0.79, p < 0.001).

Conclusion

Whole skeleton calcium-subtracted attenuation is associated with the degree of bone marrow infiltration by plasma cells, providing an objective measure of marrow involvement with the potential to allow earlier detection of disease.

MRI-Based Bone Marrow Radiomics Nomogram for Prediction of Overall Survival in Patients With Multiple Myeloma

Purpose

To develop and validate a radiomics nomogram for predicting overall survival (OS) in multiple myeloma (MM) patients.

Material and Methods

A total of 121 MM patients was enrolled and divided into training (n=84) and validation (n=37) sets. The radiomics signature was established by the selected radiomics features from lumbar MRI. The radiomics signature and clinical risk factors were integrated in multivariate Cox regression model for constructing radiomics nomogram to predict MM OS. The predictive ability and accuracy of the nomogram were evaluated by the index of concordance (C-index) and calibration curves, and compared with other four models including the clinical model, radiomics signature model, the Durie-Salmon staging system (D-S) and the International Staging System (ISS). The potential association between the radiomics signature and progression-free survival (PFS) was also explored.

Results

The radiomics signature, 1q21 gain, del (17p), and β2-MG≥5.5 mg/L showed significant association with MM OS. The predictive ability of radiomics nomogram was better than the clinical model, radiomics signature model, the D-S and the ISS (C-index: 0.793 vs. 0.733 vs. 0.742 vs. 0.554 vs. 0.671 in training set, and 0.812 vs. 0.799 vs.0.717 vs. 0.512 vs. 0.761 in validation set). The radiomics signature lacked the predictive ability for PFS (log-rank P=0.001 in training set and log-rank P=0.103 in validation set), whereas the 1-, 2- and 3-year PFS rates all showed significant difference between the high and low risk groups (P ≤ 0.05).

Conclusion

The MRI-based bone marrow radiomics may be an additional useful tool for MM OS prediction.

Feasibility of artificial intelligence–supported assessment of bone marrow infiltration using dual-energy computed tomography in patients with evidence of monoclonal protein — a retrospective observational study

Objectives

To demonstrate the feasibility of an automated, non-invasive approach to estimate bone marrow (BM) infiltration of multiple myeloma (MM) by dual-energy computed tomography (DECT) after virtual non-calcium (VNCa) post-processing.

Methods

Individuals with MM and monoclonal gammopathy of unknown significance (MGUS) with concurrent DECT and BM biopsy between May 2018 and July 2020 were included in this retrospective observational study. Two pathologists and three radiologists reported BM infiltration and presence of osteolytic bone lesions, respectively. Bone mineral density (BMD) was quantified CT-based by a CE-certified software. Automated spine segmentation was implemented by a pre-trained convolutional neural network. The non-fatty portion of BM was defined as voxels > 0 HU in VNCa. For statistical assessment, multivariate regression and receiver operating characteristic (ROC) were conducted.

Results

Thirty-five patients (mean age 65 ± 12 years; 18 female) were evaluated. The non-fatty portion of BM significantly predicted BM infiltration after adjusting for the covariable BMD (p = 0.007, r = 0.46). A non-fatty portion of BM > 0.93% could anticipate osteolytic lesions and the clinical diagnosis of MM with an area under the ROC curve of 0.70 [0.49–0.90] and 0.71 [0.54–0.89], respectively. Our approach identified MM-patients without osteolytic lesions on conventional CT with a sensitivity and specificity of 0.63 and 0.71, respectively.

Conclusions

Automated, AI-supported attenuation assessment of the spine in DECT VNCa is feasible to predict BM infiltration in MM. Further, the proposed method might allow for pre-selecting patients with higher pre-test probability of osteolytic bone lesions and support the clinical diagnosis of MM without pathognomonic lesions on conventional CT.

Key Points

• The retrospective study provides an automated approach for quantification of the non-fatty portion of bone marrow, based on AI-supported spine segmentation and virtual non-calcium dual-energy CT data.

• An increasing non-fatty portion of bone marrow is associated with a higher infiltration determined by invasive biopsy after adjusting for bone mineral density as a control variable (p = 0.007, r = 0.46).

• The non-fatty portion of bone marrow might support the clinical diagnosis of multiple myeloma when conventional CT images are negative (sensitivity 0.63, specificity 0.71).

[Potential of radiomics and artificial intelligence in myeloma imaging : Development of automatic, comprehensive, objective skeletal analyses from whole-body imaging data]

CLINICAL/METHODICAL ISSUE

Multiple myeloma can affect the complete skeleton, which makes whole-body imaging necessary. With the current assessment of these complex datasets by radiologists, only a small part of the accessible information is assessed and reported.

STANDARD RADIOLOGICAL METHODS

Depending on the question and availability, computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET) is performed and the results are then visually examined by radiologists.

METHODOLOGICAL INNOVATIONS

A combination of automatic skeletal segmentation using artificial intelligence and subsequent radiomics analysis of each individual bone have the potential to provide automatic, comprehensive, and objective skeletal analyses.

PERFORMANCE

A few automatic skeletal segmentation algorithms for CT already show promising results. In addition, first studies indicate correlations between radiomics features of bone and bone marrow with established disease markers and therapy response.

ACHIEVEMENTS

Artificial intelligence (AI) and radiomics algorithms for automatic skeletal analysis from whole-body imaging are currently in an early phase of development.