Search

Diffusion-weighted magnetic resonance imaging (MRI) is increasingly important in patients with multiple myeloma (MM). The objective of this study was to train and test an algorithm for automatic pelvic bone marrow analysis from whole-body apparent diffusion coefficient (ADC) maps in patients with MM, which automatically segments pelvic bones and subsequently extracts objective, representative ADC measurements from each bone.In this retrospective multicentric study, 180 MRIs from 54 patients were annotated (semi)manually and used to train an nnU-Net for automatic, individual segmentation of the right hip bone, the left hip bone, and the sacral bone. The quality of the automatic segmentation was evaluated on 15 manually segmented whole-body MRIs from 3 centers using the dice score. In 3 independent test sets from 3 centers, which comprised a total of 312 whole-body MRIs, agreement between automatically extracted mean ADC values from the nnU-Net segmentation and manual ADC measurements from 2 independent radiologists was evaluated. Bland-Altman plots were constructed, and absolute bias, relative bias to mean, limits of agreement, and coefficients of variation were calculated. In 56 patients with newly diagnosed MM who had undergone bone marrow biopsy, ADC measurements were correlated with biopsy results using Spearman correlation.The ADC-nnU-Net achieved automatic segmentations with mean dice scores of 0.92, 0.93, and 0.85 for the right pelvis, the left pelvis, and the sacral bone, whereas the interrater experiment gave mean dice scores of 0.86, 0.86, and 0.77, respectively. The agreement between radiologists' manual ADC measurements and automatic ADC measurements was as follows: the bias between the first reader and the automatic approach was 49 × 10-6 mm2/s, 7 × 10-6 mm2/s, and -58 × 10-6 mm2/s, and the bias between the second reader and the automatic approach was 12 × 10-6 mm2/s, 2 × 10-6 mm2/s, and -66 × 10-6 mm2/s for the right pelvis, the left pelvis, and the sacral bone, respectively. The bias between reader 1 and reader 2 was 40 × 10-6 mm2/s, 8 × 10-6 mm2/s, and 7 × 10-6 mm2/s, and the mean absolute difference between manual readers was 84 × 10-6 mm2/s, 65 × 10-6 mm2/s, and 75 × 10-6 mm2/s. Automatically extracted ADC values significantly correlated with bone marrow plasma cell infiltration (R = 0.36, P = 0.007).In this study, a nnU-Net was trained that can automatically segment pelvic bone marrow from whole-body ADC maps in multicentric data sets with a quality comparable to manual segmentations. This approach allows automatic, objective bone marrow ADC measurements, which agree well with manual ADC measurements and can help to overcome interrater variability or nonrepresentative measurements. Automatically extracted ADC values significantly correlate with bone marrow plasma cell infiltration and might be of value for automatic staging, risk stratification, or therapy response assessment.

Despite the extensive number of publications in the field of radiomics, radiomics algorithms barely enter large-scale clinical application. Supposedly, the low external generalizability of radiomics models is one of the main reasons, which hinders the translation from research to clinical application. The objectives of this study were to investigate reproducibility of radiomics features (RFs) in vivo under variation of patient positioning, magnetic resonance imaging (MRI) sequence, and MRI scanners, and to identify a subgroup of RFs that shows acceptable reproducibility across all different acquisition scenarios.Between November 30, 2020 and February 16, 2021, 55 patients with monoclonal plasma cell disorders were included in this prospective, bi-institutional, single-vendor study. Participants underwent one reference scan at a 1.5 T MRI scanner and several retest scans: once after simple repositioning, once with a second MRI protocol, once at another 1.5 T scanner, and once at a 3 T scanner. Radiomics feature from the bone marrow of the left hip bone were extracted, both from original scans and after different image normalizations. Intraclass correlation coefficient (ICC) was used to assess RF repeatability and reproducibility.Fifty-five participants (mean age, 59 ± 7 years; 36 men) were enrolled. For T1-weighted images after muscle normalization, in the simple test-retest experiment, 110 (37%) of 295 RFs showed an ICC ≥0.8: 54 (61%) of 89 first-order features (FOFs), 35 (95%) of 37 volume and shape features, and 21 (12%) of 169 texture features (TFs). When the retest was performed with different technical settings, even after muscle normalization, the number of FOF/TF with an ICC ≥0.8 declined to 58/13 for the second protocol, 29/7 for the second 1.5 T scanner, and 49/7 for the 3 T scanner, respectively. Twenty-five (28%) of the 89 FOFs and 6 (4%) of the 169 TFs from muscle-normalized T1-weighted images showed an ICC ≥0.8 throughout all repeatability and reproducibility experiments.In vivo, only few RFs are reproducible with different MRI sequences or different MRI scanners, even after application of a simple image normalization. Radiomics features selected by a repeatability experiment only are not necessarily suited to build radiomics models for multicenter clinical application. This study isolated a subset of RFs, which are robust to variations in MRI acquisition observed in scanners from 1 vendor, and therefore are candidates to build reproducible radiomics models for monoclonal plasma cell disorders for multicentric applications, at least when centers are equipped with scanners from this vendor.

Objective: To investigate the reproducibility of size measurements of focal bone marrow lesions (FL) in MRI in patients with monoclonal plasma cell disorders under variation of patient positioning and observer. Methods: A data set from a prospective test–retest study was used, in which 37 patients with a total of 140 FL had undergone 2 MRI scans with identical parameters after patient repositioning. Two readers measured long and short axis diameter on the initial scan in T1 weighted, T2 weighted short tau inversion recovery and diffusion-weighted imaging sequences. The first reader additionally measured FL on the retest-scan. The Bland–Altman method was used to assess limits of agreement (LoA), and the frequencies of absolute size changes were calculated. Results: In the simple test–retest experiment with one identical reader, a deviation of ≥1 mm / ≥2 mm / ≥3 mm for the long axis diameter in T1 weighted images was observed in 66% / 25% / 8% of cases. When comparing measurements of one reader on the first scan to the measurement of the other reader on the retest scan, a change of ≥1 mm / ≥3 mm / ≥5 mm for the long axis diameter in T1 weighted images was observed in 78% / 21% / 5% of cases. Conclusion: Small deviations in FL size are common and probably due to variation in patient positioning or inter-rater variability alone, without any actual biological change of the FL. Knowledge of the uncertainty associated with size measurements of FLs is critical for radiologists and oncologists when interpreting changes in FL size in clinical practice and in clinical trials. Advances in knowledge: According to the MY-RADs criteria, size measurements of focal lesions in MRI are now of relevance for response assessment in patients with monoclonal plasma cell disorders. Size changes of 1 or 2 mm are frequently observed due to uncertainty of the measurement only, while the actual focal lesion has not undergone any biological change. Size changes of at least 6 mm or more in T1 weighted or T2 weighted short tau inversion recovery sequences occur in only 5% or less of cases when the focal lesion has not undergone any biological change.

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).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.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).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.

Background/objectives Apparent diffusion coefficient (ADC) and signal intensity (SI) measurements play an increasing role in magnetic resonance imaging (MRI) of monoclonal plasma cell disorders. The purpose of this study was to assess interrater variability, repeatability, and reproducibility of ADC and SI measurements from bone marrow (BM) under variation of MRI protocols and scanners. Patients and methods Fifty-five patients with suspected or confirmed monoclonal plasma cell disorder were prospectively included in this institutional review board-approved study and underwent several measurements after the standard clinical whole-body MR scan, including repeated scan after repositioning, scan with a second MRI protocol, scan at a second 1.5 T scanner with a harmonized MRI protocol, and scan at a 3 T scanner. For T1-weighted, T2-weighted STIR, B800 images, and ADC maps, regions of interest were placed in the BM of the iliac crest and sacral bone, and in muscle tissue for image normalization. Bland-Altman plots were constructed, and absolute bias, relative bias to mean, limits of agreement, and coefficients of variation were calculated. Results Interrater variability and repeatability experiments showed a maximal relative bias of -0.077 and a maximal coefficient of variation of 16.2% for all sequences. Although the deviations at the second 1.5 T scanner with harmonized MRI protocol to the first 1.5 T scanner showed a maximal relative bias of 0.124 for all sequences, the variation of the MRI protocol and scan at the 3 T scanner led to large relative biases of up to -0.357 and -0.526, respectively. When comparing the 3 T scanner to the 1.5 T scanner, normalization to muscle reduced the bias of T1-weighted and T2-weighted sequences, but not of ADC maps. Conclusions The MRI scanners with identical field strength and harmonized MRI protocols can provide relatively stable quantitative measurements of BM ADC and SI. Deviations in MRI field strength and MRI protocol should be avoided when applying ADC cutoff values, which were established at other scanners or when performing multicentric imaging trials.

Novel oral anticoagulants are increasingly replacing vitamin K antagonists in the prophylaxis of thromboembolism as they are associated with lower incidence of spontaneous intracerebral hematomas and they do not require drug level monitoring. However, management dilemmas are apparent in patients on novel oral anticoagulants who have developed intracerebral hematomas after traumatic brain injury, since clinical experience with their reversal strategies is limited. We retrospectively studied 90 patients with traumatic intracerebral hematomas undergoing treatment at the surgical intensive care unit of the BG University Clinic Bergmannsheil in Bochum between 2015 and 2018. We analyzed potential prognostic factors for their radiological (expansion of intracerebral hematoma) and clinical (patients’ outcome) course, in particular the role of novel oral anticoagulants. 71.1% of patients were male; mean age was 67.3 years. Hematoma’s expansion occurred in 35.9% of our patients, whereas 62.2% of our cohort showed a favorable outcome, defined as Glasgow Outcome Scale 4 and 5. Intake of novel oral anticoagulants was associated with a higher rate of hematoma’s expansion compared to patients on vitamin K antagonists (p = 0.05) or to patients with normal coagulation status (p = 0.002). A younger age (p

Background and Study Aims Patients with large intracerebral hematomas (ICH) may demonstrate different demographics and underlying brain and systemic diseases, as well as different radiologic courses and distinct outcomes. It remains unclear whether their different behavior attributes to a different biology of the ICH or to the asymmetric characteristics of the two populations. To analyze and adjust for potential sources of selection and treatment bias, our study compared age-matched patients with traumatic and nontraumatic ICH in a single cohort diagnosed and treated in the same surgical department. Material and Methods We analyzed 135 consecutive patients with traumatic (n = 90) or spontaneous ICH (n = 45) undergoing treatment at a surgical intensive care unit of an urban university hospital. We documented their differences before and after adjustment for age in terms of demographics, the therapies applied, their radiologic (i.e., volume and rate of ICH expansion [HE]) and clinical (patients' outcome at 30 days) course, the length of hospital and ICU stay, as well as the hospital costs. Results Patients with traumatic ICH demonstrated more favorable clinical and radiologic characteristics at admission, that is, higher Glasgow Coma Scale score (p Conclusion Our data reveal a significant heterogeneity within the traumatic series. Whereas younger patients show an excellent outcome, the elderly population of the traumatic cases demonstrates a poor outcome similar to that of the nontraumatic cohort. HE under NOACs rather than under VKAs is more likely in the elderly after TBI. Larger prospective trials are warranted to elucidate the potential individual underlying molecular mechanisms for the development of an ICH and HE in these diseases.

Background Whole-body magnetic resonance imaging (wb-MRI) is an important diagnostic tool for staging, risk assessment and response evaluation in myeloma. Wb-MRIs contain approximately 110 million voxels per sequence, and only a limited amount of this information can be processed and reported by radiologists to date. Deep learning has brought striking advances in biomedical image segmentation in recent years. The goal of this work was to establish an automatic whole-body bone marrow (BM) segmentation algorithm for T1-weighted MRI sequence, and to use these segmentations for comprehensive MRI-phenotypic characterization of the BM by subsequent radiomics analysis, bone by bone. Methods For 66 patients with smoldering multiple myeloma (SMM), BM was manually segmented on T1-w images. Thirty different BM compartments were individually labelled: right and left humerus, second to seventh vertebral bodies of the cervical spine (C2-C7), all vertebral bodies of the thoracic (T1-T12) and lumbar (L1-L5) spine, sacrum, right and left hip bone and right and left femur. Data was split by date by a 3:1 ratio into training-set and independent test-set. A nnU-Net, which is state of the art deep learning framework for medical image segmentation, was trained on the 52 training cases for segmentation of BM compartments, and postprocessing was used to distinguish sides of paired bones (humeri, hip bones, femora). Mean Dice scores report accuracy of automatic segmentations on the test-cases and interrater variability between two radiologists. This study was approved by the institutional review board. Results The mean Dice scores of the nnU-Net segmentation on the 14 test-cases for BM of right and left humerus, C2-C7, T1-T12, L1-L5, sacrum, right and left hip bone, right and left femur were 0.95, 0.94, 0.87, 0.86, 0.84, 0.80, 0.82, 0.84, 0.86, 0.87, 0.88, 0.91, 0.85, 0.85, 0.83, 0.83, 0.87, 0.91, 0.93, 0.93, 0.89, 0.85, 0.85, 0.81, 0.76, 0.88, 0.93, 0.93, 0.97 and 0.97, respectively. The mean Dice scores between segmentations from 2 radiologists on 2 cases for these BM compartments in the same order were 0.94, 0.95, 0.79, 0.86, 0.88, 0.81, 0.78, 0.79, 0.84, 0.81, 0.91, 0.85, 0.89, 0.90, 0.91, 0.91, 0.92, 0.90, 0.90, 0.89, 0.88, 0.92, 0.93, 0.88, 0.88, 0.81, 0.89, 0.89, 0.95, 0.94, respectively. On a descriptive level, we found differences in radiomics signatures between vertebrae with physiological bone marrow, vertebrae with focal lesions and vertebrae with diffuse infiltration in exemplary cases. Conclusion We established automatic, bone by bone BM segmentation in SMM patients with accuracy only slightly worse compared to the interrater variability of radiologists, mostly due to lumbosacral transitional vertebra. In exemplary cases we found different radiomics-signatures between physiological BM and different pathologies, indicating that such BM segmentations can be used for in depth BM characterization from wb-MRI when combined with subsequent radiomics analysis.

Introduciton Diffusion weighed imaging (DWI) within whole-body magnetic resonance imaging (wb-MRI) is now recommended for imaging of patients with monoclonal plasma cell disorders, and quantitative ADC measurements are of high relevance for reading and reporting of wb-MRIs. However, the evidence on repeatability and especially on multicentric reproducibility of ADC measurements of the bone marrow (BM) is very limited. The goal of this study was to quantify interrater-variability, repeatability and reproducibility of ADC measurements under variation of MRI scanners and MRI sequences. Methods Fifty-five patients were prospectively included in this IRB-approved prospective study and underwent multiple scans in different settings within a maximum of 3 days between the scans. A standard scan (clinical 1.5T-MRI scanner, standard MRI-protocol) of the pelvic bone marrow was performed. Additionally, several measurements were performed: First, a retest after new positioning of the patient (same 1.5T scanner, same protocol, i.e., repeatability; n=37 paired scans). Second, retest at the same scanner but with different MRI protocol (n=37 paired scans). Third, retest at another 1.5T scanner with a harmonized MRI protocol to the initial setting (n=34 paired scans). Fourth, retest at another scanner with different field strength of 3T (n=40 paired scans). For measurement of ADC, one radiologist manually placed regions of interest (ROI) in a representative area of BM at the posterior iliac crest and a ROI in representative muscle tissue in the same slice. A second rater independently performed a second read of the standard scan. The Bland-Altman-approach was used to assess agreement of the different settings and mean bias, limits of agreement (LoA) and coefficients of variation (CoV) were calculated. Results Absolute bias / relative bias to mean / CoV of ADC measurements for the different retest measurements compared to the standard-scan were as follows: interrater setting -37.9mm2 s-1/-0.077/15.8%, repeatability -13.0mm2 s-1/-0.027/14.5%, second protocol -205.1mm2 s-1/-0.357/33.5%, second 1.5T scanner 5.8mm2 s-1/0.013/17.0%, 3T scanner 146.8mm2 s-1/0.371/41.3%. Normalization to muscle tissue did not improve the high relative bias of the second protocol or the 3T scanner to the standard setting. Conclusion In this study we quantified interrater-variability, repeatability and reproducibility of BM ADC measurements in repeated scans after patient repositioning, different protocols, different MRI scanners and different radiologists. Bias and CoV between two different 1.5T MRI scanners with harmonized protocols were in the range of interrater-variability and repeatability. On the contrary, variation of MRI protocols and field strength of the MRI scanner led to marked bias between ADC-measurements which could not be solved by normalization to muscle tissue. These findings are of high importance for planning of multicentric imaging studies in myeloma.

Background Advances in deep learning have made automatic biomedical image segmentation feasible. Additionally, radiomics analysis now allows computer based, in-depth tissue analysis from medical images. The goal of this work was to establish a full-automatic framework combining automatic pelvic bone marrow (BM) segmentation and radiomics analysis of the pelvic BM to predict BM plasma cell infiltration (PCI) directly and automatically from whole-body magnetic resonance imaging (wb-MRI). Methods A total of 541 MRIs acquired at 5 different MRI scanners from 270 patients with all stages of monoclonal plasma cell disorders were included. One-hundred fifty-eight patients who had received MRI at the standard clinical 1.5T MRI scanner and had information on PCI from concomitant BM biopsy at the iliac crest available were split by date into a training set (n=116) for both, nnU-Net and radiomics model, and an independent test set for the framework (n=42). All MRIs without biopsy data were used for training of the nnU-Net only, which is state of the art deep learning framework for medical image segmentation. Manual BM segmentations of the right and left pelvic bone on coronal T1tse images on the training cohort were used for training of the nnU-Net. A random forest classifier was trained on the pelvic radiomics features to predict PCI. The framework was then tested on the independent test set. Dice scores report accuracy of segmentations. Mean absolute error (MAE) in [%PCI] reports the accuracy of the PCI prediction. For comparison, two radiologists rated the diffuse infiltration according to 3 levels of severity (none-to-mild vs. moderate vs. severe). The mean PCI within each severity level from the training set was determined and assigned as a prediction to patients with the same level diffuse infiltration in the test set (the radiologists' PCI prediction). This study was approved by the institutional review board. Results The mean Dice scores of the nnU-Net segmentation for right / left pelvic BM on 8 cases of the test set (last 8 by acquisition date) were 0.94 and 0.94, respectively, the mean Dice scores between manual pelvic BM segmentation of two radiologists on these 8 cases were 0.87 and 0.88. The MAE of the prediction of PCI by the automated framework was 14.3 [%PCI]. The MAE of the radiologists PCI predictions were 16.1 [%PCI] (rater 1) and 16.7 [%PCI] (rater2). Conclusion We established automatic pelvic BM segmentation with radiologist level precision for all stages of monoclonal plasma cell disorders, which is a crucial step towards fully automated analysis of wb-MRI. Radiomics analysis of these segmentations can predict PCI with considerable accuracy. Further improvement of the PCI prediction model is necessary and is currently in progress, by adding additional MRI sequences into the model, enhancing the amount of training data by adding multi-institutional data, and learning about multi-scanner radiomics feature stability.

Purpose: To analyse the appearance of primary and recurrent malignant peripheral nerve sheath tumours (MPNSTs) in magnetic resonance imaging (MRI) with a focus on configuration, and to assess the occurrence of loco-regional post-treatment changes and metastases during post-treatment follow-up. Material and methods: Twenty patients with histologically proven MPNST underwent post-treatment 1.5 T MRI. Primary and recurrent MPNSTs were examined for configuration, contrast enhancement, extent, and signal intensity in MRI. Loco-regional post-treatment changes and information on metastases were extracted from the follow-up. Results: MPNSTs occurred most often in the extremities (p = 0.006). Twenty per cent (n = 4) of the patients developed recurrences, with a total of 24 lesions. Recurrent MPNSTs were significantly smaller than primary MPNSTs (p = 0.003). Primary MPNSTs mostly occurred unifocally as multilobulated or ovoid and heterogeneous lesions with mostly well-defined borders. Recurrent MPNSTs purely occurred multifocally as mostly nodular (p < 0.001), multilobulated, or ovoid lesions. 80%, 65% and 30% of the patients showed post-treatment subcutaneous oedema (p = 0.002 to 0.03), muscle oedema (p = 0.02), and seroma, respectively. Twenty-five per cent (n = 5) of patients presented metastases during follow-up. The relative risk in patients with recurrences to develop lung or lymph node metastases is eightfold (p = 0.056). Conclusions: While primary MPNSTs mostly appear unifocally as multilobulated or ovoid lesions, recurrent MPNSTs purely occur multifocally as mostly nodular lesions. Subcutaneous and muscle oedema are very common locoregional post-treatment changes. Patients with recurrences have a higher risk for lung and lymph node metastases.

Background One of the major causes of perioperative mortality of patients undergoing major hepatic resections is post-hepatectomy liver failure (PHLF). For preoperative appraisal of the risk of PHLF it is important to accurately predict resectate volume and future liver remnant volume (FLRV). The objective of our study is to prospectively evaluate the accuracy of hemihepatectomy resectate volumes that are determined by computed tomography volumetry (CTV) when compared with intraoperatively measured volumes and weights as gold standard in patients undergoing hemihepatectomy. Methods Twenty four patients (13 women, 11 men) scheduled for hemihepatectomy due to histologically proven primary or secondary hepatic malignancies were included in our study. CTV was performed using a semi-automated module (S, hereinafter) (syngo.CT Liver Analysis VA30, Siemens Healthcare, Germany). Conversion factors between CT volumes on the one side and intraoperative volumes and weights on the other side were calculated using the method of least squares. Absolute and relative disagreements between CT volumes and intraoperative volumes were determined. Results A conversion factor of c = 0.906 most precisely predicted intraoperative volumes of exsanguinated hemihepatectomy specimens from CT volumes in all patients with mean absolute and relative disagreements between CT volumes and intraoperative volumes of 57 ml and 6.3%. The use of operation-specific conversion factors yielded even better results. Conclusions CTV performed with S accurately predicts intraoperative volumes of hemihepatectomy specimens when applying conversion factors which compensate for exsanguination. This allows to precisely estimate the FLRV and thus minimize the risk of PHLF in patients undergoing major hepatic resections.