Slice Selection Research Articles

Brain Tumor Segmentation from Optimal MRI Slices Using a Lightweight U-Net

The timely detection and accurate localization of brain tumors is crucial in preserving people’s quality of life. Thankfully, intelligent computational systems have proven invaluable in addressing these challenges. In particular, the UNET model can extract essential pixel-level features to automatically identify the tumor’s location. However, known deep learning-based works usually directly feed the 3D volume into the model, which causes excessive computational complexity. This paper presents an approach to boost the UNET network, reducing computational workload while maintaining superior efficiency in locating brain tumors. This concept could benefit portable or embedded recognition systems with limited resources for operating in real time. This enhancement involves an automatic slice selection from the MRI T2 modality volumetric images containing the most relevant tumor information and implementing an adaptive learning rate to avoid local minima. Compared with the original model (7.7 M parameters), the proposed UNET model uses only 2 M parameters and was tested on the BraTS 2017, 2020, and 2021 datasets. Notably, the BraTS2021 dataset provided outstanding binary metric results: 0.7807 for the Intersection Over the Union (IoU), 0.860 for the Dice Similarity Coefficient (DSC), 0.656 for the Sensitivity, and 0.9964 for the Specificity compared to vanilla UNET.

Fluid-Attenuated Inversion Recovery Sequence in MRI in Distinguishing Fluid Cyst from Tumor in the Brain

Nuclear magnetic resonance (NMR) is a physical phenomenon used to investigate the characteristics of atomic nuclei. It is the absorption and emission of energy by nuclei in a magnetic field, which can offer extensive information on the structure, dynamics, reaction state, and chemical environment of molecules. Magnetic Resonance Imaging (MRI) is the safest imaging method. It gives a visual representation of human tissue without surgical intervention for clinical diagnosis. In this research, Fast Spin Echo (SE), Multi-Slice Spin Echo (MSSE), and Fluid Attenuated Inversion Recovery (FLAIR) sequences were used. The case under study is placed in an i-open permanent MRI system type in order to image a brain lesion. The obtained signals are encoded to fill k-space through the three stages, i.e., slice selection, phase encoding and frequency. The MRI-processed brain image is reconstructed by inverse Fourier transform in order to find the location of the cyst, and T1 weighted, T2 weighted, and FLAIR tests were performed to obtain three types of images. The images were compared with normal images in each type of applied sequence. Subsequently, the fluid signal was specifically attenuated in the targeted location to help distinguish between a tumor and fluid in the transverse relaxation time-weighted image. The signal was attenuated in the designated location, turned black, and appeared the color of free water, like cerebrospinal fluid, indicating that it was a fluid rather than a tumor. Consequently, a cyst fluid lesion is detected. Our study establishes a foundation for utilizing free fluid signal attenuation to differentiate between cyst fluid and tumors when other sequences fail to do the same.

Boost Your Immunity: VACCINE for Preventing a Novel Stealthy Slice Selection Attack in 5G and Beyond

G networks can offer network slices customized according to the demands of the services to enhance the quality of their users’ experience. The time for selecting an appropriate network slice to facilitate traffic flow between users and services by the core network functions in 5G networks is crucial for services such as emergency service and ultra-reliable low latency services. Therefore, we propose a distributed slice selection architecture for 5G and beyond networks to reduce the waiting time for starting services for users. The proposed architecture distributes slice selection function (SSF) to the edge of the network. The networks have to ensure stealthy slice selection attack (S3 attack) free operation, as moving the SSF to the edge increases attack surface. Attackers can launch S3 attack by manipulating the slice selection decisions of SSFs distributed in the network edge. The S3 attacker intentionally maps service requests from users to inappropriate network slices to damage cloud radio access network utilization and the quality of experience of users. In this article, we also present a countermeasure to tackle the S3 attack using a novel protocol called VACCINE ( v erifi a ble priva c y-preserving proto c ol for sl i ce selectio n in 5G and beyond n e tworks). VACCINE also ensures privacy-preserving slice selection by the SSFs to prevent traffic analysis attacks. We prove the chosen-ciphertext attack security strength of VACCINE and also compare the computational cost of VACCINE with other related protocols.

Bone marrow edema detection for diagnostic support of axial spondyloarthritis using MRI.

This study proposes a process for detecting slices with bone marrow edema (BME), a typical finding of axSpA, using MRI scans as the input. This process does not require manual input of ROIs and provides the results of the judgment of the presence or absence of BME on a slice and the location of edema as the rationale for the judgment. First, the signal intensity of the MRI scans of the sacroiliac joint was normalized to reduce the variation in signal values between scans. Next, slices containing synovial joints were extracted using a slice selection network. Finally, the BME slice detection network determines the presence or absence of the BME in each slice and outputs the location of the BME. The proposed method was applied to 86 MRI scans collected from 15 hospitals in Japan. The results showed that the average absolute error of the slice selection process was 1.49 slices for the misalignment between the upper and lower slices of the synovial joint range. The accuracy, sensitivity, and specificity of the BME slice detection network were 0.905, 0.532, and 0.974, respectively. This paper proposes a process to detect the slice with BME and its location as the rationale of the judgment from an MRI scan and shows its effectiveness using 86 MRI scans. In the future, we plan to develop a process for detecting other findings such as bone erosion from MR scans, followed by the development of a diagnostic support system.

Ultrashort echo time magnetic resonance elastography for quantification of the mechanical properties of short T2 tissues via optimal control-based radiofrequency pulses.

The aim of the current study is to demonstrate the feasibility of radiofrequency (RF) pulses generated via an optimal control (OC) algorithm to perform magnetic resonance elastography (MRE) and quantify the mechanical properties of materials with very short transverse relaxation times (T2 < 5 ms) for the first time. OC theory applied to MRE provides RF pulses that bring isochromats from the equilibrium state to a fixed target state, which corresponds to the phase pattern of a conventional MRE acquisition. Such RF pulses applied with a constant gradient allow to simultaneously perform slice selection and motion encoding in the slice direction. Unlike conventional MRE, no additional motion-encoding gradients (MEGs) are needed, enabling shorter echo times. OC pulses were implemented both in turbo spin echo (OC rapid acquisition with refocused echoes [RARE]) and ultrashort echo time (OC UTE) sequences to compare their motion-encoding efficiency with the conventional MEG encoding (classical MEG MRE). MRE experiments were carried out on agar phantoms with very short T2 values and on an ex vivo bovine tendon. Magnitude images, wave field images, phase-to-noise ratio (PNR), and shear storage modulus maps were compared between OC RARE, OC UTE, and classical MEG MRE in samples with different T2 values. Shear storage modulus values of the agar phantoms were in agreement with values found in the literature, and that of the bovine tendon was corroborated with rheometry measurements. Only the OC sequences could encode motion in very short T2 samples, and only OC UTE sequences yielded magnitude images enabling proper visualization of short T2 samples and tissues. The OC UTE sequence produced the best PNRs, demonstrating its ability to perform anatomical and mechanical characterization. Its success warrants in vivo confirmation in further studies.

Body Composition in Advanced Non-Small Cell Lung Cancer Treated With Immunotherapy

The association between body composition (BC) and cancer outcomes is complex and incompletely understood. Previous research in non-small-cell lung cancer (NSCLC) has been limited to small, single-institution studies and yielded promising, albeit heterogeneous, results. To evaluate the association of BC with oncologic outcomes in patients receiving immunotherapy for advanced or metastatic NSCLC. This comprehensive multicohort analysis included clinical data from cohorts receiving treatment at the Dana-Farber Brigham Cancer Center (DFBCC) who received immunotherapy given alone or in combination with chemotherapy and prospectively collected data from the phase 1/2 Study 1108 and the chemotherapy arm of the phase 3 MYSTIC trial. Baseline and follow-up computed tomography (CT) scans were collected and analyzed using deep neural networks for automatic L3 slice selection and body compartment segmentation (skeletal muscle [SM], subcutaneous adipose tissue [SAT], and visceral adipose tissue). Outcomes were compared based on baseline BC measures or their change at the first follow-up scan. The data were analyzed between July 2022 and April 2023. Hazard ratios (HRs) for the association of BC measurements with overall survival (OS) and progression-free survival (PFS). A total of 1791 patients (878 women [49%]) with NSCLC were analyzed, of whom 487 (27.2%) received chemoimmunotherapy at DFBCC (DFBCC-CIO), 825 (46.1%) received ICI monotherapy at DFBCC (DFBCC-IO), 222 (12.4%) were treated with durvalumab monotherapy on Study 1108, and 257 (14.3%) were treated with chemotherapy on MYSTIC; median (IQR) ages were 65 (58-74), 66 (57-71), 65 (26-87), and 63 (30-84) years, respectively. A loss in SM mass, as indicated by a change in the L3 SM area, was associated with worse oncologic outcome across patient groups (HR, 0.59 [95% CI, 0.43-0.81] and 0.61 [95% CI, 0.47-0.79] for OS and PFS, respectively, in DFBCC-CIO; HR, 0.74 [95% CI, 0.60-0.91] for OS in DFBCC-IO; HR, 0.46 [95% CI, 0.33-0.64] and 0.47 [95% CI, 0.34-0.64] for OS and PFS, respectively, in Study 1108; HR, 0.76 [95% CI, 0.61-0.96] for PFS in the MYSTIC trial). This association was most prominent among male patients, with a nonsignificant association among female patients in the MYSTIC trial and DFBCC-CIO cohorts on Kaplan-Meier analysis. An increase of more than 5% in SAT density, as quantified by the average CT attenuation in Hounsfield units of the SAT compartment, was associated with poorer OS in 3 patient cohorts (HR, 0.61 [95% CI, 0.43-0.86] for DFBCC-CIO; HR, 0.62 [95% CI, 0.49-0.79] for DFBCC-IO; and HR, 0.56 [95% CI, 0.40-0.77] for Study 1108). The change in SAT density was also associated with PFS for DFBCC-CIO (HR, 0.73; 95% CI, 0.54-0.97). This was primarily observed in female patients on Kaplan-Meier analysis. The results of this multicohort study suggest that loss in SM mass during systemic therapy for NSCLC is a marker of poor outcomes, especially in male patients. SAT density changes are also associated with prognosis, particularly in female patients. Automated CT-derived BC measurements should be considered in determining NSCLC prognosis.

Abstract 7650: An analysis of the prognostic role of sarcopenia derived with fully automated deep learning slice-based muscle estimation in patients with metastatic non-small cell lung cancer (NSCLC)

Abstract Introduction: Sarcopenia, progressive and generalized loss of skeletal muscle mass, is difficult and labor intensive to directly measure as it requires manual segmentations by radiologists from computed tomography (CT) scans. Sarcopenia has been shown to be a prognostic factor in some metastatic cancers. We evaluated the impact of sarcopenia derived via an automated segmentation algorithm on outcomes in the Phase 3 NSCLC clinical trial, CheckMate 227 (NCT02477826) Methods: The sarcopenia score assessment involves automated selection of the L3 vertebra level slice from abdominal or chest, abdomen, and pelvis (CAP) CT scans, followed by automated segmentation of the muscle area from that slice. Sarcopenia scores were derived from previously established definitions of sarcopenia, with values below 41 cm2/m2 for female patients (regardless of BMI) and below 43 cm2/m2 or 53 cm2/m2 for male (based on BMI &amp;lt; 25 kg/m2 or ≥ 25 kg/m2). Among the 967 patients with evaluable scans, sarcopenia at baseline was identified in 68.39% (n=329/481) patients receiving nivolumab (NIVO) with ipilimumab (IPI) and 70.2% (n=341/486) receiving chemotherapy (Chemo). Results: There was a trend of worse overall survival (OS) in patients with baseline sarcopenia across both treatment arms (minimum follow up 49.4 months). In patients with baseline sarcopenia vs ITT, median OS was 16.5 mo vs. 21.8 mo (p= 0.219) within the NIVO + IPI arm and 12.9 mo vs.15.0 mo (p= 0.273) within the Chemo arm respectively. Any increase in muscle mass from baseline to week 12 was associated with better OS for patients treated with NIVO+IPI (n=98) with sarcopenia at baseline compared to those without an increase in muscle mass (median OS 45.3 mo vs.19.5 mo, p= 0.012). This was not seen in those who received Chemo (n=87). Weight and muscle mass were moderately correlated at baseline (r=0.60). Conclusions: We found that in CheckMate 227, baseline sarcopenia is associated with numerically worse OS, regardless of treatment, leading to potential use as a negative prognostic biomarker. Patients with NSCLC receiving immunotherapy might benefit from changes in muscle mass being monitored during treatment. Validation of these findings in other studies and other tumor types is warranted. Citation Format: Wiem Safta, Rafael Santana-Davila, David Paulucci, William J. Geese, Diederik J. Grootendorst. An analysis of the prognostic role of sarcopenia derived with fully automated deep learning slice-based muscle estimation in patients with metastatic non-small cell lung cancer (NSCLC) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7650.

Efficient Slice Creation in Network Slicing using K-Prototype Clustering and Context-Aware Slice Selection for Service Provisioning

The advent of 5G technology has ushered in a new era of communication where the customization of network services is crucial to meet diverse user demands. Network slicing has emerged as a pivotal technology to achieve this customization. In this research, we present an innovative approach to optimize network slicing in 5G by employing K-Prototype Clustering for slice creation and Context-Aware Slice Selection for efficient resource allocation. In slice creation, we delve into the innovative application of the K-Prototype clustering algorithm. Recognizing that 5G networks encompass numerical and categorical attributes, the K-Prototype algorithm enables the creation of network slices that cater to diverse service requirements. By harnessing this clustering technique, our proposed method optimizes the creation of network slices, resulting in improved resource utilization and reduced network congestion. Furthermore, we introduce the concept of Context-Aware Slice Selection, which considers the dynamic and evolving nature of network demands. Context-awareness ensures that network slices are selected based on real-time contextual information, enabling a more adaptive and responsive network. This approach leads to the efficient allocation of resources and a higher quality of service for end-users. To evaluate the performance of our proposed methodology, we employ key performance metrics, including slice selection accuracy, slice selection delay, and radio link failure. Through comprehensive testing and analysis, our research demonstrates that our approach consistently outperforms existing methods in terms of these metrics.

Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients.

The importance of body composition and sarcopenia is well-recognized in cancer patient outcomes and treatment tolerance, yet routine evaluations are rare due to their time-intensive nature. While CT scans provide accurate measurements, they depend on manual processes. We developed and validated a deep learning algorithm to automatically select and segment abdominal muscles [SM], visceral fat [VAT], and subcutaneous fat [SAT] on CT scans. A total of 352 CT scans were collected from two cancer centers. The detection of the third lumbar vertebra and three different body tissues (SM, VAT, and SAT) were annotated manually. The 5-fold cross-validation method was used to develop the algorithm and validate its performance on the training cohort. The results were validated on an external, independent group of CT scans. The algorithm for automatic L3 slice selection had a mean absolute error of 4 mm for the internal validation dataset and 5.5 mm for the external validation dataset. The median DICE similarity coefficient for body composition was 0.94 for SM, 0.93 for VAT, and 0.86 for SAT in the internal validation dataset, whereas it was 0.93 for SM, 0.93 for VAT, and 0.85 for SAT in the external validation dataset. There were high correlation scores with sarcopenia metrics in both internal and external validation datasets. Our deep learning algorithm facilitates routine research use and could be integrated into electronic patient records, enhancing care through better monitoring and the incorporation of targeted supportive measures like exercise and nutrition.

MCE: Medical Cognition Embedded in 3D MRI feature extraction for advancing glioma staging.

In recent years, various data-driven algorithms have been applied to the classification and staging of brain glioma MRI detection. However, the restricted availability of brain glioma MRI data in purely data-driven deep learning algorithms has presented challenges in extracting high-quality features and capturing their complex patterns. Moreover, the analysis methods designed for 2D data necessitate the selection of ideal tumor image slices, which does not align with practical clinical scenarios. Our research proposes an novel brain glioma staging model, Medical Cognition Embedded (MCE) model for 3D data. This model embeds knowledge characteristics into data-driven approaches to enhance the quality of feature extraction. Approach includes the following key components: (1) Deep feature extraction, drawing upon the imaging technical characteristics of different MRI sequences, has led to the design of two methods at both the algorithmic and strategic levels to mimic the learning process of real image interpretation by medical professionals during film reading; (2) We conduct an extensive Radiomics feature extraction, capturing relevant features such as texture, morphology, and grayscale distribution; (3) By referencing key points in radiological diagnosis, Radiomics feature experimental results, and the imaging characteristics of various MRI sequences, we manually create diagnostic features (Diag-Features). The efficacy of proposed methodology is rigorously evaluated on the publicly available BraTS2018 and BraTS2020 datasets. Comparing it to most well-known purely data-driven models, our method achieved higher accuracy, recall, and precision, reaching 96.14%, 93.4%, 97.06%, and 97.57%, 92.80%, 95.96%, respectively.

Representative Slice Selection and Multi-View Projection Learning for Pulmonary Tuberculosis Infectiousness Identification Using CT Volume

Representative Slice Selection and Multi-View Projection Learning for Pulmonary Tuberculosis Infectiousness Identification Using CT Volume

A slice selection guided deep integrated pipeline for Alzheimer’s prediction from Structural Brain MRI

Alzheimer’s disease, a progressive form of dementia, has risen to become the fifth leading cause of death among individuals aged 65 and older. The diagnosis of Alzheimer’s is both time-consuming and costly, involving radiologists and clinical experts at multiple stages, which presents a significant challenge in the medical field. Moreover, cases of Alzheimer’s and dementia often go undiagnosed or misdiagnosed worldwide. To address this issue, medical experts meticulously analyze patients’ structural MRI (sMRI) scans to identify potential abnormalities linked to Alzheimer’s or other forms of dementia. Recognizing the devastating impact of this disease on people’s lives, Artificial Intelligence (AI) researchers have been dedicated to developing automated solutions for early-stage Alzheimer’s diagnosis in recent years, aiming to support medical practitioners in their efforts. Despite the application of various AI-driven solutions that use sMRI data for Alzheimer’s diagnosis, there are still research gaps that need attention. These gaps include the need for guided slice selection and the development of a simpler yet effective integrated pipeline where each stage of the process is fully automated, eliminating the need for medical practitioner intervention. In this study, we propose an integrated automated solution that incorporates a guided machine learning-based selection process using K-Means++ leading to a Gradient Boosting-based method for identifying the 16 most relevant 2-dimensional sMRI slices from 3-dimensional sMRI data. This step is crucial for accurate Alzheimer’s classification. Furthermore, we introduce a deep learning architecture that combines EfficientNetV2S-based transfer learning with densely-learned features in an optimized manner. To evaluate the effectiveness of our proposed deep-integrated architecture, we used two benchmark datasets from ADNI and OASIS, conducting rigorous experimental analysis and validation. The results demonstrated that our integrated architecture outperformed all other experimented architectures, achieving a 20-Fold Cross Validation Accuracy of 83.64% (CN vs AD), 82.69% (CN vs MCIc), and 71.40% (CN vs MCInc) on the ADNI dataset, and 91.54% (CN vs AD) on the OASIS dataset. This signifies the potential of our approach in improving Alzheimer’s diagnosis accuracy and offers hope for early detection and intervention in this debilitating disease.

AI-Derived CT Body Composition in Advanced Non-Small Cell Lung Cancer: A Multicohort Study

The relationship between body composition (BC) and cancer outcomes is complex and incompletely understood. Previous research in non-small cell lung cancer (NSCLC) has been limited to small single-institution studies, which have yielded inconsistent results. We conducted a comprehensive multicohort analysis to evaluate the impact of BC on overall survival (OS) in advanced NSCLC treated with systemic therapy. The analysis included data from the phase I/II CP1108 study (NSCLC Durvalumab cohort) and the chemotherapy arm of the phase III MYSTIC trial. We also analyzed data from Dana-Farber Cancer Institute (DFCI) cohorts receiving immunotherapy alone or in combination with chemotherapy. Baseline and follow-up (FU) CT scans were collected and analyzed using deep neural networks for automatic L3 slice selection and body compartment segmentation (skeletal muscle [SM], subcutaneous adipose tissue [SAT], and visceral adipose tissue [VAT]). We compared OS based on baseline BC measures or their change at the first FU scan. The impact of sarcopenia at baseline was evaluated in association with the delta metrics. A total of 1865 NSCLC patients were analyzed, of which 222 were treated on CP1108, 257 were treated on MYSTIC, 870 received IO monotherapy at DFCI (DFCI-IO), and 516 received chemoimmunotherapy at DFCI (DFCI-CIO). The median ages were 65, 63, 66, and 65, respectively. A loss in SM mass >5%, as indicated by a change in L3 SM area, was significantly associated with poorer OS across all patient groups (median [months]: 5 vs. 19; p<0.001 for CP1108, 11 vs. 14; p = 0.03 for MYSTIC, 11 vs. 17; p<0.001 for DFCI-IO, and 12 vs. 22; p<0.001 for DFCI-CIO). This effect was driven by male patients, with a non-significant association (p>0.5) among female patients in the MYSTIC and DFCI-CIO cohorts. An increase in SAT density >5%, as quantified by the average CT attenuation in HU of the SAT compartment, was significantly linked to poorer OS in three patient groups (median [months]: 4 vs. 19; p<0.001 for CP1108, 10 vs. 17; p<0.001 for DFCI-IO, and 12 vs. 20; p = 0.003 for DFCI-CIO). This was primarily observed among female patients, with a non-significant association (p>0.5) among male patients in the DFCI-CIO cohort. On subgroup analysis, loss in SM mass had an impact on OS in patients with baseline sarcopenia (median [months] 5 vs. 22; p<0.001 for CP1108, 5 vs. 12; p = 0.03 for MYSTIC, 11 vs. 17; p<0.001 for DFCI IO, and 9 vs. 17; p = 0.003 for DFCI-CIO). Conversely, no association was observed between change in SM mass and OS in patients without sarcopenia at baseline in the MYSTIC and DFCI-IO cohorts. Sarcopenia and loss in SM mass during systemic therapy for NSCLC are markers of poor outcome, especially in male patients. SAT density changes are also strongly associated with prognosis, particularly in female patients. Automated CT-derived BC measurements should be considered along with other risk factors in determining lung cancer prognosis and ability to tolerate oncologic treatments.

Automated quantification of glenoid bone loss in CT scans for shoulder dislocation surgery planning.

Estimation of glenoid bone loss in CT scans following shoulder dislocation is required to determine the type of surgery needed to restore shoulder stability. This paper presents a novel automatic method for the computation of glenoid bone loss in CT scans. The model-based method is a pipeline that consists of four steps: (1) computation of an oblique plane in the CT scan that best matches the glenoid face orientation; (2) selection of the glenoid oblique CT slice; (3) computation of the circle that best fits the posteroinferior glenoid contour; (4) quantification of the glenoid bone loss. The best-fit circle is computed with newly defined Glenoid Clock Circle Constraints. The pipeline and each of its steps were evaluated on 51 shoulder CT scans (44 patients). Ground truth oblique slice, best-fit circle, and glenoid bone loss measurements were obtained manually from three clinicians. The full pipeline yielded a mean absolute error (%) for the bone loss deficiency of 2.3 ± 2.9mm (4.67 ± 3.32%). The mean oblique CT slice selection difference was 1.42 ± 1.32 slices, above the observer variability of 1.74 ± 1.82 slices. The glenoid bone loss deficiency measure (%) on the ground truth oblique glenoid CT slice has a mean average error of 0.54 ± 1.03mm (4.76 ± 3.00%), close to the observer variability of 0.93 ± 1.40mm (2.98 ± 4.97%). Our pipeline is the first fully automatic method for the quantitative analysis of glenoid bone loss in CT scans. The computed glenoid bone loss report may assist orthopedists in selecting and planning surgical shoulder dislocation procedures.

Simultaneous T2 -weighted real-time MRI of two orthogonalslices.

MR guidance is used during therapy to detect and compensate for lesion motion. T2 -weighted MRI often has a superior lesion contrast in comparison to T1 -weighted real-time imaging. The purpose of this work was to design a fast T2 -weighted sequence capable of simultaneously acquiring two orthogonal slices, enabling real-time tracking of lesions. To generate a T2 contrast in two orthogonal slices simultaneously, a sequence (Ortho-SFFP-Echo) was designed that samples the T2 -weighted spin echo (S- ) signal in a TR-interleaved acquisition of two slices. Slice selection and phase-encoding directions are swapped between the slices, leading to a unique set of spin-echo signal conditions. To minimize motion-related signal dephasing, additional flow-compensation strategies are implemented. In both the abdominal breathing phantom and in vivo experiments, a time series was acquired using Ortho-SSFP-Echo. The centroid of the target was tracked in postprocessing steps. In the phantom, the lesion could be identified and delineated in the dynamic images. In the volunteer experiments, the kidney was visualized with a T2 contrast at a temporal resolution of 0.45 s under free-breathing conditions. A respiratory belt demonstrated a strong correlation with the time course of the kidney centroid in the head-foot direction. A hypointense saturation band at the slice overlap did not inhibit lesion tracking in the semi-automatic postprocessing steps. The Ortho-SFFP-Echo sequence delivers real-time images with a T2 -weighted contrast in two orthogonal slices. The sequence allows for simultaneous acquisition, which could be beneficial for real-time motion tracking in radiotherapy or interventional MRI.

Fully automated imaging protocol independent system for pituitary adenoma segmentation: a convolutional neural network—based model on sparsely annotated MRI

This study aims to develop a fully automated imaging protocol independent system for pituitary adenoma segmentation from magnetic resonance imaging (MRI) scans that can work without user interaction and evaluate its accuracy and utility for clinical applications. We trained two independent artificial neural networks on MRI scans of 394 patients. The scans were acquired according to various imaging protocols over the course of 11 years on 1.5T and 3T MRI systems. The segmentation model assigned a class label to each input pixel (pituitary adenoma, internal carotid artery, normal pituitary gland, background). The slice segmentation model classified slices as clinically relevant (structures of interest in slice) or irrelevant (anterior or posterior to sella turcica). We used MRI data of another 99 patients to evaluate the performance of the model during training. We validated the model on a prospective cohort of 28 patients, Dice coefficients of 0.910, 0.719, and 0.240 for tumour, internal carotid artery, and normal gland labels, respectively, were achieved. The slice selection model achieved 82.5% accuracy, 88.7% sensitivity, 76.7% specificity, and an AUC of 0.904. A human expert rated 71.4% of the segmentation results as accurate, 21.4% as slightly inaccurate, and 7.1% as coarsely inaccurate. Our model achieved good results comparable with recent works of other authors on the largest dataset to date and generalized well for various imaging protocols. We discussed future clinical applications, and their considerations. Models and frameworks for clinical use have yet to be developed and evaluated.

Analyzing the sensitivity of quantitative 3D MRI of longitudinal relaxation at very low field in Gd-doped phantoms.

Recently, new MRI systems working at magnetic field below 10 mT (Very and Ultra Low Field regime) have been developed, showing improved T1-contrast in projected 2D maps (i.e. images without slice selection). Moving from projected 2D to 3D maps is not trivial due to the low SNR of such devices. This work aimed to demonstrate the ability and the sensitivity of a VLF-MRI scanner operating at 8.9 mT in quantitatively obtaining 3D longitudinal relaxation rate (R1) maps and distinguishing between voxels intensities. We used phantoms consisting of vessels doped with different Gadolinium (Gd)-based Contrast Agent (CA) concentrations, providing a set of various R1 values. As CA, we used a commercial compound (MultiHance®, gadobenate dimeglumine) routinely used in clinical MRI. 3D R1 maps and T1-weighted MR images were analysed to identify each vessel. R1 maps were further processed by an automatic clustering analysis to evaluate the sensitivity at the single-voxel level. Results obtained at 8.9 mT were compared with commercial scanners operating at 0.2 T, 1.5 T, and 3 T. VLF R1 maps offered a higher sensitivity in distinguishing the different CA concentrations and an improved contrast compared to higher fields. Moreover, the high sensitivity of 3D quantitative VLF-MRI allowed an effective clustering of the 3D map values, assessing their reliability at the single voxel level. Conversely, in all fields, T1-weighted images were less reliable, even at higher CA concentrations. In summary, with few excitations and an isotropic voxel size of 3 mm, VLF-MRI 3D quantitative mapping showed a sensitivity better than 2.7 s-1 corresponding to a concentration difference of 0.17 mM of MultiHance in copper sulfate doped water, and improved contrast compared to higher fields. Based on these results, future studies should characterize R1 contrast at VLF, also with other CA, in the living tissues.

Deep learning based synthetic CT from cone beam CT generation for abdominal paediatric radiotherapy

Objective. Adaptive radiotherapy workflows require images with the quality of computed tomography (CT) for re-calculation and re-optimisation of radiation doses. In this work we aim to improve the quality of on-board cone beam CT (CBCT) images for dose calculation using deep learning. Approach. We propose a novel framework for CBCT-to-CT synthesis using cycle-consistent Generative Adversarial Networks (cycleGANs). The framework was tailored for paediatric abdominal patients, a challenging application due to the inter-fractional variability in bowel filling and small patient numbers. We introduced to the networks the concept of global residuals only learning and modified the cycleGAN loss function to explicitly promote structural consistency between source and synthetic images. Finally, to compensate for the anatomical variability and address the difficulties in collecting large datasets in the paediatric population, we applied a smart 2D slice selection based on the common field-of-view (abdomen) to our imaging dataset. This acted as a weakly paired data approach that allowed us to take advantage of scans from patients treated for a variety of malignancies (thoracic-abdominal-pelvic) for training purposes. We first optimised the proposed framework and benchmarked its performance on a development dataset. Later, a comprehensive quantitative evaluation was performed on an unseen dataset, which included calculating global image similarity metrics, segmentation-based measures and proton therapy-specific metrics. Main results. We found improved performance for our proposed method, compared to a baseline cycleGAN implementation, on image-similarity metrics such as Mean Absolute Error calculated for a matched virtual CT (55.0 ± 16.6 HU proposed versus 58.9 ± 16.8 HU baseline). There was also a higher level of structural agreement for gastrointestinal gas between source and synthetic images measured using the dice similarity coefficient (0.872 ± 0.053 proposed versus 0.846 ± 0.052 baseline). Differences found in water-equivalent thickness metrics were also smaller for our method (3.3 ± 2.4% proposed versus 3.7 ± 2.8% baseline). Significance. Our findings indicate that our innovations to the cycleGAN framework improved the quality and structure consistency of the synthetic CTs generated.

Modified Round Robin with Highest Response Ratio Next CPU Scheduling Algorithm using Dynamic Time Quantum

Background: The most popular time-sharing operating systems scheduling technique, whose efficiency heavily dependent on time slice selection, is the round robin CPU scheduling algorithm. The time slice works similar to the First-Come-First-Serve (FCFS) scheduling or processor sharing algorithm if it is large or extremely too small. Some of the existing research papers have an algorithm called Improved Round Robin with Highest Response Ratio Next (IRRHRRN) which made use of response ratio with a predefined time quantum of 10ms with the major aim of avoiding starvation. However, the IRRHRRN algorithm favors processes with shorter burst time than the ones with longer burst time, and gave no regard to the process arrival time, thus leading to starvation. Aim: This study tries to improve on the IRRHRRN algorithm by proposing the Modified Round Robin with Highest Response Ratio Next (MRRHRRN) CPU Scheduling Algorithm using Dynamic Time Quantum in order to reduce the problem of starvation. Method: Dynamic method of determining the time quantum was adopted. Results: The proposed algorithm was compared with four other existing algorisms such as Standard Round Robin (RR), Improved Round Robin (IRR), An Additional Improvement in Round Robin (AAIRR), and the Improved Round Robin with Highest Response Ratio Next (IRRHRRN) and it provided some promising results in terms of the Average Waiting Time of 35407.6 ms, Average Turnaround Time of 36117.6 ms, Average Response Time of 10894.8 ms and Number of Context Switch of 301for the Non-Zero Arrival Times Processes

Navigator-based slice tracking for kidney pCASL using spin-echo EPI acquisition.

To apply a navigator-based slice-tracking method to prospectively compensate respiratory motion for kidney pseudo-continuous arterial spin labeling (pCASL), using spin-echo (SE) EPI acquisition. A single gradient-echo slice selection and projection readout at the location of the diaphragm along the inferior-superior direction was applied as a navigator. Navigator acquisition and fat suppression were inserted before each transverse imaging slice of the readouts of a 2D-SE-EPI-based pCASL sequence. Motion information was calculated after exclusion of the signal saturation in the navigator signal caused by EPI excitations. The motion information was then used to directly adjust the slice positioning in real time. The respiratory motion from the navigator signal was calculated, and slice positioning was changed in real time based on the motion information. We could show that motion compensation reduces kidney movement, and that the coefficients of variation across renal perfusion values were significantly reduced when motion correction was applied. The average reduction of coefficients of variation was approximately 20%, resulting in a more accurate and detailed structure of the respective perfusion maps. This study demonstrates the feasibility of a navigator-based slice-tracking technique in kidney imaging with a SE-EPI readout pCASL sequence to reduce kidney motion.