Magnetic resonance biomarker assessment software (MR-BIAS): an automated open-source tool for the ISMRM/NIST system phantom

Objective.To provide an open-source software for repeatable and efficient quantification ofT₁andT₂relaxation times with the ISMRM/NIST system phantom. Quantitative magnetic resonance imaging (qMRI) biomarkers have the potential to improve disease detection, staging and monitoring of treatment response. Reference objects, such as the system phantom, play a major role in translating qMRI methods into the clinic. The currently available open-source software for ISMRM/NIST system phantom analysis, Phantom Viewer (PV), includes manual steps that are subject to variability.Approach.We developed the Magnetic Resonance BIomarker Assessment Software (MR-BIAS) to automatically extract system phantom relaxation times. The inter-observer variability (IOV) and time efficiency of MR-BIAS and PV was observed in six volunteers analysing three phantom datasets. The IOV was measured with the coefficient of variation (CV) of percent bias (%bias) inT₁andT₂with respect to NMR reference values. The accuracy of MR-BIAS was compared to a custom script from a published study of twelve phantom datasets. This included comparison of overall bias and %bias for variable inversion recovery (T₁^VIR), variable flip angle (T₁^VFA) and multiple spin-echo (T₂^MSE) relaxation models.Main results.MR-BIAS had a lower mean CV withT₁^VIR(0.03%) andT₂^MSE(0.05%) in comparison to PV withT₁^VIR(1.28%) andT₂^MSE(4.55%). The mean analysis duration was 9.7 times faster for MR-BIAS (0.8 min) than PV (7.6 min). There was no statistically significant difference in the overall bias, or the %bias for the majority of ROIs, as calculated by MR-BIAS or the custom script for all models.Significance.MR-BIAS has demonstrated repeatable and efficient analysis of the ISMRM/NIST system phantom, with comparable accuracy to previous studies. The software is freely available to the MRI community, providing a framework to automate required analysis tasks, with the flexibility to explore open questions and accelerate biomarker research.

Cascaded deep learning-based auto-segmentation for head and neck cancer patients: Organs at risk on T2-weighted magnetic resonance imaging

PURPOSE

To investigate multiple deep learning methods for automated segmentation (auto-segmentation) of the parotid glands, submandibular glands, and level II and level III lymph nodes on magnetic resonance imaging (MRI). Outlining radiosensitive organs on images used to assist radiation therapy (radiotherapy) of patients with head and neck cancer (HNC) is a time-consuming task, in which variability between observers may directly impact on patient treatment outcomes. Auto-segmentation on computed tomography imaging has been shown to result in significant time reductions and more consistent outlines of the organs at risk.

METHODS

Three convolutional neural network (CNN)-based auto-segmentation architectures were developed using manual segmentations and T2-weighted MRI images provided from the American Association of Physicists in Medicine (AAPM) radiotherapy MRI auto-contouring (RT-MAC) challenge dataset (n = 31). Auto-segmentation performance was evaluated with segmentation similarity and surface distance metrics on the RT-MAC dataset with institutional manual segmentations (n = 10). The generalizability of the auto-segmentation methods was assessed on an institutional MRI dataset (n = 10).

RESULTS

Auto-segmentation performance on the RT-MAC images with institutional segmentations was higher than previously reported MRI methods for the parotid glands (Dice: 0.860 ± 0.067, mean surface distance [MSD]: 1.33 ± 0.40 mm) and the first report of MRI performance for submandibular glands (Dice: 0.830 ± 0.032, MSD: 1.16 ± 0.47 mm). We demonstrate that high-resolution auto-segmentations with improved geometric accuracy can be generated for the parotid and submandibular glands by cascading a localizer CNN and a cropped high-resolution CNN. Improved MSDs were observed between automatic and manual segmentations of the submandibular glands when a low-resolution auto-segmentation was used as prior knowledge in the second-stage CNN. Reduced auto-segmentation performance was observed on our institutional MRI dataset when trained on external RT-MAC images; only the parotid gland auto-segmentations were considered clinically feasible for manual correction (Dice: 0.775 ± 0.105, MSD: 1.20 ± 0.60 mm).

CONCLUSIONS

This work demonstrates that CNNs are a suitable method to auto-segment the parotid and submandibular glands on MRI images of patients with HNC, and that cascaded CNNs can generate high-resolution segmentations with improved geometric accuracy. Deep learning methods may be suitable for auto-segmentation of the parotid glands on T2-weighted MRI images from different scanners, but further work is required to improve the performance and generalizability of these methods for auto-segmentation of the submandibular glands and lymph nodes.

Radiomics feature stability of open-source software evaluated on apparent diffusion coefficient maps in head and neck cancer

Radiomics is a promising technique for discovering image based biomarkers of therapy response in cancer. Reproducibility of radiomics features is a known issue that is addressed by the image biomarker standardisation initiative (IBSI), but it remains challenging to interpret previously published radiomics signatures. This study investigates the reproducibility of radiomics features calculated with two widely used radiomics software packages (IBEX, MaZda) in comparison to an IBSI compliant software package (PyRadiomics). Intensity histogram, shape and textural features were extracted from 334 diffusion weighted magnetic resonance images of 59 head and neck cancer (HNC) patients from the PREDICT-HN observational radiotherapy study. Based on name and linear correlation, PyRadiomics shares 83 features with IBEX and 49 features with MaZda, a sub-set of well correlated features are considered reproducible (IBEX: 15 features, MaZda: 18 features). We explore the impact of including non-reproducible radiomics features in a HNC radiotherapy response model. It is possible to classify equivalent patient groups using radiomic features from either software, but only when restricting the model to reliable features using a correlation threshold method. This is relevant for clinical biomarker validation trials as it provides a framework to assess the reproducibility of reported radiomic signatures from existing trials.

Abstract LB-023: Dynamic real time in vivo CAR T cell tracking: Clinical and preclinical studies using a novel dual PET-MR imaging agent

Objective: The aim is to demonstrate dynamic in-vivo tracking of CAR T cell therapy for treatment of solid tumors using Cu-64 labeled superparamagnetic iron oxide nanoparticles (SPION) as novel dual PET-MR imaging agent.

Methodology: Cu-64 SPION: Cu-64 radioisotope is bound to silica coated SPION using enhanced electrolysis plating techniques with tin and palladium seeding. Preclinical Model: Mouse splenic T cells were activated with anti-CD3, anti-CD28 & cultured with IL-2 and IL-7, prior to being transduced with second generation anti-Her-2 CAR (scFv-CD28-CD3ζ). 5 x 105 E0771-hHER2 breast tumor cells were implanted subcutaneously into male C57Bl/6-human HER2 transgenic mice. 107 labeled CAR T or control T cells (Cu-64 5-8 MBq) were injected into tail vein. Clinical Model: Activated T cells using antibody CD3 (OKT3) & IL-2 are transduced with retroviral vector constructs encoding for chimeric T-cell receptor specific for Lewis Y antigen. Modified T-cells are further expanded ex-vivo and reinfused. 3 x 108 CAR T cells were labeled with Cu-64 (200 - 300 MBq). Labeling of CAR T cells with Cu-64 SPION: Transfecting agent protamine sulphate facilitated cellular uptake of Cu-64 SPION within cells. Functional assays: 51Chromium release, cytometric bead array and cell viability showed that labeling process did not affect CAR T cell cytotoxicity, cytokine secretion (TNFα and IFN-γ) and viability. CAR T Cell Tracking: Scanning was performed using clinical grade dual PET-MR scanner.

Preliminary Data: In this clinical trial (HREC/16/PMCC/30) patients are being enrolled for first in human in vivo study to determine how many cells distribute to solid tumor sites within first few days of CAR T cell therapy. This is first data that demonstrates that CAR-T cells can be consistently and efficiently labeled (≤60%) with cell viability (≥85%) and at sensitivity comparable to detecting at least z cells at tumor site using clinical grade dual PET-MR scanner. SUVmean values provides insight into individual response to therapy. The observed increase in SUVmax over time specifies localization of CAR T cells at tumor sites.

Clinical data at early time point showed moderate uptake in lungs posterior basal segments without increased activity over next few days, thus suggesting transient process. Mild, diffuse bone marrow and relatively intense uptake in the liver and spleen suggests margination of cells to the reticulo-endothelial system. Distinct PET signal suggests localization of labeled cells in the secondary tumor sites. Little background uptake in important organs such as brain and heart indicate the safety profile of imaging agent. Absence of signal in bladder indicates hepatobiliary excretion, which may allow re-absorption from GI tract and re-circulation.

Distinct PET signal within tumor in preclinical studies confirms trafficking of CAR T cells to tumor site as compared to controls. A negative contrast in the liver on T2 weighted MRI in both the preclinical and clinical studies. Preliminary Conclusion:This is first in human in vivo study to show CAR T cell distribution in real time and provides insight into individual responses of tumors to therapy. CAR T cell functionality largely remain unchanged due to labeling process. The preliminary findings indicate that labeled cells traffic to tumor sites in first few hours of infusion and remain persistent for extended period.

Citation Format: Ritu Singla, Dominic Wall, Samuel Anderson, Nicholas Zia, James C. Korte, Lucy Kravets, Gerard McKiernan, Jeanne Butler, Amanda Gammilonghi, Jyoti Arora, Ben Solomon, Rodney Hicks, Timothy Cain, Phillip Darcy, Carleen Cullinane, Paul Neeson, Rajesh Ramanathan, Ravi Shukla, Vipul Bansal, Simon Harrison. Dynamic real time in vivo CAR T cell tracking: Clinical and preclinical studies using a novel dual PET-MR imaging agent [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr LB-023.

First in-human study of in vivo imaging of ex vivo labeled CAR T cells with dual PET-MR

3557

Background: This is a first in human in-vivo biodistribution of ex-vivo labelled CAR T cells assessed in a cohort of patients. Cells were labelled with novel Cu-64 labelled superparamagnetic iron oxide nanoparticles (SPION) and infused IV into patients with solid tumors & tracked using clinical dual PET-MR. The study validates the clinical translation of CAR T cell in-vivo tracking in real time. Methods: Cu-64 radioisotope was bound to silica coated SPION using electrolysis plating with tin & palladium seeding. Cellular uptake of Cu-64 SPION was facilitated with a transfecting agent. Functional assays including 51Chromium release, cytometric bead array demonstrated that labelling process did not affect cytotoxicity & cytokine secretion (TNFα & IFN-g). T cells were transduced with retroviral vector constructs encoding for second-generation chimeric T-cell receptor specific for carbohydrate Lewis Y antigen. Modified T-cells were expanded ex-vivo & were labelled with Cu-64 (~300 MBq) prior to re-infusion (3 x108 labelled cells). Scanning is performed with Siemens 3T dual PET-MR scanner. Results: In this first in human in-vivo study (HREC/16/PMCC/30) a cohort of patients received ex-vivo labelled CAR T cells to determine how many labelled cells distribute to solid tumor sites within 3-5 days. Our results demonstrate that cells can be efficiently labelled (≤60%) with high cell viability (≥85%) at a sensitivity sufficient to detect labelled cells at tumor site for up to 5 days. An observed trend in SUVmean & SUVmax provided insight into efficacy & individual response to therapy. Early time points showed moderate uptake of labelled cells in lungs posterior basal segments without increased activity over next few days, suggesting a transient process. Mild, diffuse bone marrow & relatively intense uptake of labelled cells in liver & spleen suggests margination of cells to reticulo-endothelial system. Distinct PET signal at some of the tumor sites at 24 h suggests antigen specific localization & time taken to reach these sites. Excretion via hepatobiliary indicated reabsorption from GI tract & re-circulation of labelled cells. Minimal uptake in brain & heart supported safety profile of labeling agent. Conclusions: This is first in human in-vivo study to provide highly valuable visual and dynamic data in real time and provides insight into individual responses to therapy. CAR T cell functionality largely remain unchanged due to labeling process. The findings indicate that labelled cells traffic to tumor sites at later time points & remain persistent for extended period of time.

Spin Lock Adiabatic Correction (SLAC) for B1-insensitive pulse design at 7T

A new framework for B₁ insensitive adiabatic pulse design is proposed, denoted Spin Lock Adiabatic Correction (SLAC), which counteracts deviations from ideal behaviour through inclusion of an additional correction component during pulse design. SLAC pulses are theoretically derived, then applied to the design of enhanced BIR-4 and hyperbolic secant pulses to demonstrate practical utility of the new pulses. At 7T, SLAC pulses are shown to improve the flip angle homogeneity compared to a standard adiabatic pulse with validation in both simulations and phantom experiments, under SAR equivalent experimental conditions. The SLAC framework can be applied to any arbitrary adiabatic pulse to deliver excitation with increased B₁ insensitivity.