Monoclonal Gammopathy of Clinical Significance (MGCS) and Related Disorders: A Review and the Role of Imaging.

The term monoclonal gammopathy of clinical significance (MGCS) refers to a group of symptomatic monoclonal gammopathies that do not meet the diagnostic criteria for malignant plasma cell disorders, such as multiple myeloma or Waldenström macroglobulinemia. These symptoms are attributable to the paraneoplastic effects of monoclonal immunoglobulins that occur through diverse mechanisms. The presence of symptoms distinguishes MGCS from monoclonal gammopathy of undetermined significance, which lacks significant symptomatic presentation. The presentations of MGCS are manifold, adding to the diagnostic challenge. Clinical suspicion is key for accurate and timely diagnosis. Radiologic imaging can provide pivotal information to guide the diagnosis. In this review, we discuss MGCS from a radiology perspective and highlight pertinent imaging features associated with the disorders.

Quantitative metrics of bone quality determined at the distal radius using photon-counting CT.

OBJECTIVE: To determine the accuracy of photon-counting-detector CT (PCD-CT) at deriving bone morphometric indices and demonstrate utility in vivo in the distal radius. METHODS: Ten cadaver wrists were scanned using PCD-CT and high-resolution peripheral quantitative CT (HRpQCT). Correlation between PCD-CT and HRpQCT morphometric indices was determined. Agreement was assessed by Lin's concordance correlation coefficient (Lin's CCC). Wrist PCD-CTs of patients between 02/2022 and 08/2023 were also evaluated for clinical utility. Morphometric indices of the in vivo distal radii were extracted and compared between patients with or without osteoporosis. RESULTS: In cadavers, strong correlation between PCD-CT and HRpQCT was observed for cortical thickness (Spearman correlation, ρ, 0.85), trabecular spacing (ρ = 0.98), and trabecular bone volume fraction (ρ = 0.68). Moderate negative correlation (ρ = - 0.49) was observed for trabecular thickness. PCD-CT shows good agreement to HRpQCT for cortical thickness, trabecular spacing, and trabecular bone volume fraction (Lin's CCC = 0.80, 0.94, and 0.86, respectively) but poor agreement (Lin's CCC = - 0.1) for trabecular thickness. In forty participants (31 adults and 9 pediatric), bone morphometrics indices for cortical thickness, trabecular thickness, trabecular spacing, and trabecular bone volume fraction were 0.99 mm (IQR, 0.89-1.06), 0.38 mm (IQR, 0.25-0.40), 0.82 mm (IQR, 0.72-1.05), and 0.28 (IQR, 0.25-0.33), respectively. Patients with osteoporosis had statistically significantly larger trabecular spacing (p = 0.025) and lower trabecular volumetric bone mineral density (p = 0.042). CONCLUSION: This study demonstrates the agreement of PCD-CT to HRpQCT in cadavers of most cortical and bone morphometrics examined and provide in vivo quantitative metrics of bone microarchitecture from routine clinical PCD-CT images of the distal radius.

Development of a deep learning model for the automated detection of green pixels indicative of gout on dual energy CT scan.

Background: Dual-energy CT (DECT) is a non-invasive way to determine the presence of monosodium urate (MSU) crystals in the workup of gout. Color-coding distinguishes MSU from calcium following material decomposition and post-processing. Most software labels MSU as green and calcium as blue. There are limitations in the current image processing methods of segmenting green-encoded pixels. Additionally, identifying green foci is tedious, and automated detection would improve workflow. This study aimed to determine the optimal deep learning (DL) algorithm for segmenting green-encoded pixels of MSU crystals on DECTs. Methods: DECT images of positive and negative gout cases were retrospectively collected. The dataset was split into train (N = 28) and held-out test (N = 30) sets. To perform cross-validation, the train set was split into seven folds. The images were presented to two musculoskeletal radiologists, who independently identified green-encoded voxels. Two 3D Unet-based DL models, Segresnet and SwinUNETR, were trained, and the Dice similarity coefficient (DSC), sensitivity, and specificity were reported as the segmentation metrics. Results: Segresnet showed superior performance, achieving a DSC of 0.9999 for the background pixels, 0.7868 for the green pixels, and an average DSC of 0.8934 for both types of pixels, respectively. According to the post-processed results, the Segresnet reached voxel-level sensitivity and specificity of 98.72 % and 99.98 %, respectively. Conclusion: In this study, we compared two DL-based segmentation approaches for detecting MSU deposits in a DECT dataset. The Segresnet resulted in superior performance metrics. The developed algorithm provides a potential fast, consistent, highly sensitive and specific computer-aided diagnosis tool. Ultimately, such an algorithm could be used by radiologists to streamline DECT workflow and improve accuracy in the detection of gout.

Whole-body low-dose computed tomography in patients with newly diagnosed multiple myeloma predicts cytogenetic risk: a deep learning radiogenomics study.

OBJECTIVE: To develop a whole-body low-dose CT (WBLDCT) deep learning model and determine its accuracy in predicting the presence of cytogenetic abnormalities in multiple myeloma (MM). MATERIALS AND METHODS: WBLDCTs of MM patients performed within a year of diagnosis were included. Cytogenetic assessments of clonal plasma cells via fluorescent in situ hybridization (FISH) were used to risk-stratify patients as high-risk (HR) or standard-risk (SR). Presence of any of del(17p), t(14;16), t(4;14), and t(14;20) on FISH was defined as HR. The dataset was evenly divided into five groups (folds) at the individual patient level for model training. Mean and standard deviation (SD) of the area under the receiver operating curve (AUROC) across the folds were recorded. RESULTS: One hundred fifty-one patients with MM were included in the study. The model performed best for t(4;14), mean (SD) AUROC of 0.874 (0.073). The lowest AUROC was observed for trisomies: AUROC of 0.717 (0.058). Two- and 5-year survival rates for HR cytogenetics were 87% and 71%, respectively, compared to 91% and 79% for SR cytogenetics. Survival predictions by the WBLDCT deep learning model revealed 2- and 5-year survival rates for patients with HR cytogenetics as 87% and 71%, respectively, compared to 92% and 81% for SR cytogenetics. CONCLUSION: A deep learning model trained on WBLDCT scans predicted the presence of cytogenetic abnormalities used for risk stratification in MM. Assessment of the model's performance revealed good to excellent classification of the various cytogenetic abnormalities.

Photon-counting CT in Thoracic Imaging: Early Clinical Evidence and Incorporation Into Clinical Practice.

Photon-counting CT (PCCT) is an emerging advanced CT technology that differs from conventional CT in its ability to directly convert incident x-ray photon energies into electrical signals. The detector design also permits substantial improvements in spatial resolution and radiation dose efficiency and allows for concurrent high-pitch and high-temporal-resolution multienergy imaging. This review summarizes (a) key differences in PCCT image acquisition and image reconstruction compared with conventional CT; (b) early evidence for the clinical benefit of PCCT for high-spatial-resolution diagnostic tasks in thoracic imaging, such as assessment of airway and parenchymal diseases, as well as benefits of high-pitch and multienergy scanning; (c) anticipated radiation dose reduction, depending on the diagnostic task, and increased utility for routine low-dose thoracic CT imaging; (d) adaptations for thoracic imaging in children; (e) potential for further quantitation of thoracic diseases; and (f) limitations and trade-offs. Moreover, important points for conducting and interpreting clinical studies examining the benefit of PCCT relative to conventional CT and integration of PCCT systems into multivendor, multispecialty radiology practices are discussed.

Deep-learning for automated detection of MSU deposits on DECT: evaluating impact on efficiency and reader confidence.

Introduction: Dual-energy CT (DECT) is a non-invasive way to determine the presence of monosodium urate (MSU) crystals in the workup of gout. Color-coding distinguishes MSU from calcium following material decomposition and post-processing. Manually identifying these foci (most commonly labeled green) is tedious, and an automated detection system could streamline the process. This study aims to evaluate the impact of a deep-learning (DL) algorithm developed for detecting green pixelations on DECT on reader time, accuracy, and confidence. Methods: We collected a sample of positive and negative DECTs, reviewed twice-once with and once without the DL tool-with a 2-week washout period. An attending musculoskeletal radiologist and a fellow separately reviewed the cases, simulating clinical workflow. Metrics such as time taken, confidence in diagnosis, and the tool's helpfulness were recorded and statistically analyzed. Results: We included thirty DECTs from different patients. The DL tool significantly reduced the reading time for the trainee radiologist (p = 0.02), but not for the attending radiologist (p = 0.15). Diagnostic confidence remained unchanged for both (p = 0.45). However, the DL model identified tiny MSU deposits that led to a change in diagnosis in two cases for the in-training radiologist and one case for the attending radiologist. In 3/3 of these cases, the diagnosis was correct when using DL. Conclusions: The implementation of the developed DL model slightly reduced reading time for our less experienced reader and led to improved diagnostic accuracy. There was no statistically significant difference in diagnostic confidence when studies were interpreted without and with the DL model.

Staging of breast cancer in the breast and regional lymph nodes using contrast-enhanced photon-counting detector CT: accuracy and potential impact on patient management.

OBJECTIVES: To describe the feasibility and evaluate the performance of multiphasic photon-counting detector (PCD) CT for detecting breast cancer and nodal metastases with correlative dynamic breast MRI and digital mammography as the reference standard. METHODS: Adult females with biopsy-proven breast cancer undergoing staging breast MRI were prospectively recruited to undergo a multiphasic PCD-CT using a 3-phase protocol: a non-contrast ultra-high-resolution (UHR) scan and 2 intravenous contrast-enhanced scans with 50 and 180 s delay. Three breast radiologists compared CT characteristics of the index malignancy, regional lymphadenopathy, and extramammary findings to MRI. RESULTS: Thirteen patients underwent both an MRI and PCD-CT (mean age: 53 years, range: 36-75 years). Eleven of thirteen cases demonstrated suspicious mass or non-mass enhancement on PCD-CT when compared to MRI. All cases with metastatic lymphadenopathy (3/3 cases) demonstrated early avid enhancement similar to the index malignancy. All cases with multifocal or multicentric disease on MRI were also identified on PCD-CT (3/3 cases), including a 4 mm suspicious satellite lesion. Four of five patients with residual suspicious post-biopsy calcifications on mammograms were detected on the UHR PCD-CT scan. Owing to increased field-of-view at PCD-CT, a 5 mm thoracic vertebral metastasis was identified at PCD-CT and not with the breast MRI. CONCLUSIONS: A 3-phase PCD-CT scan protocol shows initial promising results in characterizing breast cancer and regional lymphadenopathy similar to MRI and detects microcalcifications in 80% of cases. ADVANCES IN KNOWLEDGE: UHR and spectral capabilities of PCD-CT may allow for comprehensive characterization of breast cancer and may represent an alternative to breast MRI in select cases.

Musculoskeletal CT Imaging: State-of-the-Art Advancements and Future Directions.

CT is one of the most widely used modalities for musculoskeletal imaging. Recent advancements in the field include the introduction of four-dimensional CT, which captures a CT image during motion; cone-beam CT, which uses flat-panel detectors to capture the lower extremities in weight-bearing mode; and dual-energy CT, which operates at two different x-ray potentials to improve the contrast resolution to facilitate the assessment of tissue material compositions such as tophaceous gout deposits and bone marrow edema. Most recently, photon-counting CT (PCCT) has been introduced. PCCT is a technique that uses photon-counting detectors to produce an image with higher spatial and contrast resolution than conventional multidetector CT systems. In addition, postprocessing techniques such as three-dimensional printing and cinematic rendering have used CT data to improve the generation of both physical and digital anatomic models. Last, advancements in the application of artificial intelligence to CT imaging have enabled the automatic evaluation of musculoskeletal pathologies. In this review, the authors discuss the current state of the above CT technologies, their respective advantages and disadvantages, and their projected future directions for various musculoskeletal applications.

Quantifying Uncertainty in Deep Learning of Radiologic Images.

In recent years, deep learning (DL) has shown impressive performance in radiologic image analysis. However, for a DL model to be useful in a real-world setting, its confidence in a prediction must also be known. Each DL model's output has an estimated probability, and these estimated probabilities are not always reliable. Uncertainty represents the trustworthiness (validity) of estimated probabilities. The higher the uncertainty, the lower the validity. Uncertainty quantification (UQ) methods determine the uncertainty level of each prediction. Predictions made without UQ methods are generally not trustworthy. By implementing UQ in medical DL models, users can be alerted when a model does not have enough information to make a confident decision. Consequently, a medical expert could reevaluate the uncertain cases, which would eventually lead to gaining more trust when using a model. This review focuses on recent trends using UQ methods in DL radiologic image analysis within a conceptual framework. Also discussed in this review are potential applications, challenges, and future directions of UQ in DL radiologic image analysis.

Incidence of complications related to shoulder arthroplasties identified on computed tomography.

PURPOSE: Determine incidence of shoulder arthroplasty complications identified on computed tomography (CT). MATERIALS AND METHODS: Retrospective institutional database review of patients with shoulder arthroplasties who underwent CT between 01/2006-11/2021 at a tertiary academic referral center with subspecialized orthopedic shoulder surgeons. CT reports were reviewed for arthroplasty type and complication. Data were stratified and summarized. Associations between complications and arthroplasty types were determined with Chi-squared goodness of fit test. RESULTS: Eight hundred twelve CTs in 797 unique patients were included (438 (53.9%) females and 374 (46.1%) males; mean age 67 ± 11 years). There were 403 total shoulder arthroplasties (TSA), 317 reverse total shoulder arthroplasties (rTSA), and 92 hemiarthroplasties (HA). Complications were present in 527/812 (64.9%) and incidences were: loosening/aseptic osteolysis 36.9%, periprosthetic failure 21.6%, periprosthetic fracture 12.3%, periprosthetic dislocation 6.8%, joint/pseudocapsule effusion 5.9%, prosthetic failure 4.8%, infection 3.8%, and periprosthetic collection 2.1%. Complications per arthroplasty were: 305/403 (75.7%) TSAs, 176/317 (55.5%) rTSAs, and 46/92 (50%) HAs (p < 0.001). Periprosthetic fracture (20.8%), prosthetic dislocation (9.8%), and prosthetic failure (7.9%) were highest in rTSAs (p < 0.001, p < 0.013, p < 0.001, respectively). Loosening/aseptic osteolysis most frequent in TSAs (54.1%) (p < 0.001). Periprosthetic failure most frequent in HA (32.6%) (p < 0.001). Significant associations were identified with joint/pseudocapsule effusion and loosening/aseptic osteolysis (p = 0.04) and prosthetic dislocation (p < .001). CONCLUSION: In this single tertiary academic referral center cohort, the incidence of shoulder arthroplasty complication identified on CT was 64.9% and the most commonly occurring complication was loosening/aseptic osteolysis (36.9%). TSA had the highest incidence of complication (75.7%).

Clinical applications of photon counting detector CT.

The X-ray detector is a fundamental component of a CT system that determines the image quality and dose efficiency. Until the approval of the first clinical photon-counting-detector (PCD) system in 2021, all clinical CT scanners used scintillating detectors, which do not capture information about individual photons in the two-step detection process. In contrast, PCDs use a one-step process whereby X-ray energy is converted directly into an electrical signal. This preserves information about individual photons such that the numbers of X-ray in different energy ranges can be counted. Primary advantages of PCDs include the absence of electronic noise, improved radiation dose efficiency, increased iodine signal and the ability to use lower doses of iodinated contrast material, and better spatial resolution. PCDs with more than one energy threshold can sort the detected photons into two or more energy bins, making energy-resolved information available for all acquisitions. This allows for material classification or quantitation tasks to be performed in conjunction with high spatial resolution, and in the case of dual-source CT, high pitch, or high temporal resolution acquisitions. Some of the most promising applications of PCD-CT involve imaging of anatomy where exquisite spatial resolution adds clinical value. These include imaging of the inner ear, bones, small blood vessels, heart, and lung. This review describes the clinical benefits observed to date and future directions for this technical advance in CT imaging. KEY POINTS: • Beneficial characteristics of photon-counting detectors include the absence of electronic noise, increased iodine signal-to-noise ratio, improved spatial resolution, and full-time multi-energy imaging. • Promising applications of PCD-CT involve imaging of anatomy where exquisite spatial resolution adds clinical value and applications requiring multi-energy data simultaneous with high spatial and/or temporal resolution. • Future applications of PCD-CT technology may include extremely high spatial resolution tasks, such as the detection of breast micro-calcifications, and quantitative imaging of native tissue types and novel contrast agents.

Seeing More with Less: Clinical Benefits of Photon-counting Detector CT.

Photon-counting detector (PCD) CT is an emerging technology that has led to continued innovation and progress in diagnostic imaging after it was approved by the U.S. Food and Drug Administration for clinical use in September 2021. Conventional energy-integrating detector (EID) CT measures the total energy of x-rays by converting photons to visible light and subsequently using photodiodes to convert visible light to digital signals. In comparison, PCD CT directly records x-ray photons as electric signals, without intermediate conversion to visible light. The benefits of PCD CT systems include improved spatial resolution due to smaller detector pixels, higher iodine image contrast, increased geometric dose efficiency to allow high-resolution imaging, reduced radiation dose for all body parts, multienergy imaging capabilities, and reduced artifacts. To recognize these benefits, diagnostic applications of PCD CT in musculoskeletal, thoracic, neuroradiologic, cardiovascular, and abdominal imaging must be optimized and adapted for specific diagnostic tasks. The diagnostic benefits and clinical applications resulting from PCD CT in early studies have allowed improved visualization of key anatomic structures and radiologist confidence for some diagnostic tasks, which will continue as PCD CT evolves and clinical use and applications grow. ©RSNA, 2023 Quiz questions for this article are available in the supplemental material. See the invited commentary by Ananthakrishnan in this issue.

Feasibility of photon-counting CT for femoroacetabular impingement syndrome evaluation: lower radiation dose and improved diagnostic confidence.

OBJECTIVE: The feasibility of low-dose photon-counting detector (PCD) CT to measure alpha and acetabular version angles of femoroacetabular impingement (FAI). MATERIAL AND METHODS: FAI patients undergoing an energy-integrating detector (EID) CT underwent an IRB-approved prospective ultra-high-resolution (UHR) PCD-CT between 5/2021 and 12/2021. PCD-CT was dose-matched to the EID-CT or acquired at 50% dose. Simulated 50% dose EID-CT images were generated. Two radiologists evaluated randomized EID-CT and PCD-CT images and measured alpha and acetabular version angles on axial image slices. Image quality (noise, artifacts, and visualization of cortex) and confidence in non-FAI pathology were rated on a 4-point scale (3 = adequate). Preference tests of standard dose PCD-CT, 50% dose PCD-CT, and 50% dose EID-CT relative to standard dose EID-CT were performed using Wilcoxon Rank test. RESULTS: 20 patients underwent standard dose EID-CT (~ CTDIvol, 4.5 mGy); 10 patients, standard dose PCD-CT (4.0 mGy); 10 patients, 50% PCD-CT (2.6 mGy). Standard dose EID-CT images were scored as adequate for diagnostic task in all categories (range 2.8-3.0). Standard dose PCD-CT images scored higher than the reference in all categories (range 3.5-4, p < 0.0033). Half-dose PCD-CT images also scored higher for noise and cortex visualization (p < 0.0033) and equivalent for artifacts and visualization of non-FAI pathology. Finally, simulated 50% EID-CT images scored lower in all categories (range 1.8-2.4, p < 0.0033). CONCLUSIONS: Dose-matched PCD-CT is superior to EID-CT for alpha angle and acetabular version measurement in the work up of FAI. UHR-PCD-CT enables 50% radiation dose reduction compared to EID while remaining adequate for the imaging task.

Photon-Counting Detector CT for Musculoskeletal Imaging: A Clinical Perspective.

Photon-counting detector (PCD) CT has emerged as a novel imaging modality that represents a fundamental shift in the way that CT systems detect x-rays. After pre-clinical and clinical investigations showed benefits of PCD CT for a range of imaging tasks, the U.S. FDA in 2021 approved the first commercial PCD CT system for clinical use. The technologic features of PCD CT are particularly well suited for musculo-skeletal imaging applications. Advantages of PCD CT compared with conventional energy-integrating detector (EID) CT include smaller detector pixels and excellent geometric dose efficiency that enable imaging of large joints and central skeletal anatomy at ultrahigh spatial resolution; advanced multienergy spectral postprocessing that allows quantification of gout deposits and generation of virtual noncalcium images for visualization of bone edema; improved metal artifact reduction for imaging of orthopedic implants; and higher CNR and suppression of electronic noise. Given substantially improved cortical and trabecular detail, PCD CT images more clearly depict skeletal abnormalities, including fractures, lytic lesions, and mineralized tumor matrix. The purpose of this article is to review, by use of clinical examples comparing EID CT and PCD CT, the technical features of PCD CT and their associated impact on musculoskeletal imaging applications.

Photon-counting Detector CT with Deep Learning Noise Reduction to Detect Multiple Myeloma.

Background Photon-counting detector (PCD) CT and deep learning noise reduction may improve spatial resolution at lower radiation doses compared with energy-integrating detector (EID) CT. Purpose To demonstrate the diagnostic impact of improved spatial resolution in whole-body low-dose CT scans for viewing multiple myeloma by using PCD CT with deep learning denoising compared with conventional EID CT. Materials and Methods Between April and July 2021, adult participants who underwent a whole-body EID CT scan were prospectively enrolled and scanned with a PCD CT system in ultra-high-resolution mode at matched radiation dose (8 mSv for an average adult) at an academic medical center. EID CT and PCD CT images were reconstructed with Br44 and Br64 kernels at 2-mm section thickness. PCD CT images were also reconstructed with Br44 and Br76 kernels at 0.6-mm section thickness. The thinner PCD CT images were denoised by using a convolutional neural network. Image quality was objectively quantified in two phantoms and a randomly selected subset of participants (10 participants; median age, 63.5 years; five men). Two radiologists scored PCD CT images relative to EID CT by using a five-point Likert scale to detect findings reflecting multiple myeloma. The scoring for the matched reconstruction series was blinded to scanner type. Reader-averaged scores were tested with the null hypothesis of equivalent visualization between EID and PCD. Results Twenty-seven participants (median age, 68 years; IQR, 61-72 years; 16 men) were included. The blinded assessment of 2-mm images demonstrated improvement in viewing lytic lesions, intramedullary lesions, fatty metamorphosis, and pathologic fractures for PCD CT versus EID CT (P < .05 for all comparisons). The 0.6-mm PCD CT images with convolutional neural network denoising also demonstrated improvement in viewing all four pathologic abnormalities and detected one or more lytic lesions in 21 of 27 participants compared with the 2-mm EID CT images (P < .001). Conclusion Ultra-high-resolution photon-counting detector CT improved the visibility of multiple myeloma lesions relative to energy-integrating detector CT. © RSNA, 2022 Online supplemental material is available for this article.

Multimodality evaluation of transfascial muscle and other soft tissue herniations of the extremities.

This review illustrates the multimodality assessment of transfascial muscle and other soft tissue herniations of the extremities. Transfascial herniations of the extremities can develop from congenital or acquired disruptions of the deep fascia, resulting in herniation of the underlying muscle, nerve, or soft tissue tumor into the subcutaneous tissues. While most patients present with a painless subcutaneous nodule that may change in size with muscle activation, some may experience focal or diffuse extremity symptoms such as pain and paresthesias. Although the diagnosis may be clinically suspected, radiologic evaluation is useful for definitive diagnosis and characterization. Ultrasound is the preferred modality for initial workup through a focused and dynamic examination. Magnetic resonance imaging can be utilized for equivocal, complicated, and preoperative cases. Computed tomography is less useful in the evaluation of transfascial herniations in the extremities due to similarities in the attenuation between muscle and fascia, which can decrease the conspicuity of small defects.

A deep learning algorithm for detecting lytic bone lesions of multiple myeloma on CT.

BACKGROUND: Whole-body low-dose CT is the recommended initial imaging modality to evaluate bone destruction as a result of multiple myeloma. Accurate interpretation of these scans to detect small lytic bone lesions is time intensive. A functional deep learning) algorithm to detect lytic lesions on CTs could improve the value of these CTs for myeloma imaging. Our objectives were to develop a DL algorithm and determine its performance at detecting lytic lesions of multiple myeloma. METHODS: Axial slices (2-mm section thickness) from whole-body low-dose CT scans of subjects with biochemically confirmed plasma cell dyscrasias were included in the study. Data were split into train and test sets at the patient level targeting a 90%/10% split. Two musculoskeletal radiologists annotated lytic lesions on the images with bounding boxes. Subsequently, we developed a two-step deep learning model comprising bone segmentation followed by lesion detection. Unet and "You Look Only Once" (YOLO) models were used as bone segmentation and lesion detection algorithms, respectively. Diagnostic performance was determined using the area under the receiver operating characteristic curve (AUROC). RESULTS: Forty whole-body low-dose CTs from 40 subjects yielded 2193 image slices. A total of 5640 lytic lesions were annotated. The two-step model achieved a sensitivity of 91.6% and a specificity of 84.6%. Lesion detection AUROC was 90.4%. CONCLUSION: We developed a deep learning model that detects lytic bone lesions of multiple myeloma on whole-body low-dose CTs with high performance. External validation is required prior to widespread adoption in clinical practice.

Deep learning-based virtual noncalcium imaging in multiple myeloma using dual-energy CT.

BACKGROUND: Dual-energy CT with virtual noncalcium (VNCa) images allows the evaluation of focal intramedullary bone marrow involvement in patients with multiple myeloma. However, current commercial VNCa techniques suffer from excessive image noise and artifacts due to material decomposition used in synthesizing VNCa images. OBJECTIVES: In this work, we aim to improve VNCa image quality for the assessment of focal multiple myeloma, using an Artificial intelligence based Generalizable Algorithm for mulTi-Energy CT (AGATE) method. MATERIALS AND METHODS: AGATE method used a custom dual-task convolutional neural network (CNN) that concurrently carries out material classification and quantification. The material classification task provided an auxiliary regularization to the material quantification task. CNN parameters were optimized using custom loss functions that involved cross-entropy, physics-informed constraints, structural redundancy in spectral and material images, and texture information in spectral images. For training data, CT phantoms (diameters 30 to 45 cm) with tissue-mimicking inserts were scanned on a third generation dual-source CT system. Scans were performed at routine dose and half of the routine dose. Small image patches (i.e., 40 × 40 pixels) of tissue-mimicking inserts with known basis material densities were extracted for training samples. Numerically simulated insert materials with various shapes increased diversity of training samples. Generalizability of AGATE was evaluated using CT images from phantoms and patients. In phantoms, material decomposition accuracy was estimated using mean-absolute-percent-error (MAPE), using physical inserts that were not used during the training. Noise power spectrum (NPS) and modulation transfer function (MTF) were compared across phantom sizes and radiation dose levels. Five patients with multiple myeloma underwent dual-energy CT, with VNCa images generated using a commercial method and AGATE. Two fellowship-trained musculoskeletal radiologists reviewed the VNCa images (commercial and AGATE) side-by-side using a dual-monitor display, blinded to VNCa type, rating the image quality for focal multiple myeloma lesion visualization using a 5-level Likert comparison scale (-2 = worse visualization and diagnostic confidence, -1 = worse visualization but equivalent diagnostic confidence, 0 = equivalent visualization and diagnostic confidence, 1 = improved visualization but equivalent diagnostic confidence, 2 = improved visualization and diagnostic confidence). A post hoc assignment of comparison ratings was performed to rank AGATE images in comparison to commercial ones. RESULTS: AGATE demonstrated consistent material quantification accuracy across phantom sizes and radiation dose levels, with MAPE ranging from 0.7% to 4.4% across all testing materials. Compared to commercial VNCa images, the AGATE-synthesized VNCa images yielded considerably lower image noise (50-77% noise reduction) without compromising noise texture or spatial resolution across different phantom sizes and two radiation doses. AGATE VNCa images had markedly reduced area under NPS curves and maintained NPS peak frequency (0.7 lp/cm to 1.0 lp/cm), with similar MTF curves (50% MTF at 3.0 lp/cm). In patients, AGATE demonstrated reduced image noise and artifacts with improved delineation of focal multiple myeloma lesions (all readers comparison scores indicating improved overall diagnostic image quality [scores 1 or 2]). CONCLUSIONS: AGATE demonstrated reduced noise and artifacts in VNCa images and ability to improve visualization of bone marrow lesions for assessing multiple myeloma.