Intrathecal delivery of nanoparticle PARP inhibitor to the cerebrospinal fluid for the treatment of metastatic medulloblastoma.

The morbidity associated with pediatric medulloblastoma, in particular in patients who develop leptomeningeal metastases, remains high in the absence of effective therapies. Administration of substances directly into the cerebrospinal fluid (CSF) is one approach to circumvent the blood-brain barrier and focus delivery of drugs to the site of tumor. However, high rates of CSF turnover prevent adequate drug accumulation and lead to rapid systemic clearance and toxicity. Here, we show that PLA-HPG nanoparticles, made with a single-emulsion, solvent evaporation process, can encapsulate talazoparib, a PARP inhibitor (BMN-673). These degradable polymer nanoparticles improve the therapeutic index when delivered intrathecally and lead to sustained drug retention in the tumor as measured with PET imaging and fluorescence microscopy. We demonstrate that administration of these particles into the CSF, alone or in combination with systemically administered temozolomide, is a highly effective therapy for tumor regression and prevention of leptomeningeal spread in xenograft mouse models of medulloblastoma. These results provide a rationale for harnessing nanoparticles for the delivery of drugs limited by brain penetration and therapeutic index and demonstrate important advantages in tolerability and efficacy for encapsulated drugs delivered locoregionally.

Deep-Learning-Based Few-Angle Cardiac SPECT Reconstruction Using Transformer.

Convolutional neural networks (CNNs) have been extremely successful in various medical imaging tasks. However, because the size of the convolutional kernel used in a CNN is much smaller than the image size, CNN has a strong spatial inductive bias and lacks a global understanding of the input images. Vision Transformer, a recently emerged network structure in computer vision, can potentially overcome the limitations of CNNs for image-reconstruction tasks. In this work, we proposed a slice-by-slice Transformer network (SSTrans-3D) to reconstruct cardiac SPECT images from 3D few-angle data. To be specific, the network reconstructs the whole 3D volume using a slice-by-slice scheme. By doing so, SSTrans-3D alleviates the memory burden required by 3D reconstructions using Transformer. The network can still obtain a global understanding of the image volume with the Transformer attention blocks. Lastly, already reconstructed slices are used as the input to the network so that SSTrans-3D can potentially obtain more informative features from these slices. Validated on porcine, phantom, and human studies acquired using a GE dedicated cardiac SPECT scanner, the proposed method produced images with clearer heart cavity, higher cardiac defect contrast, and more accurate quantitative measurements on the testing data as compared with a deep U-net.

Co-attention spatial transformer network for unsupervised motion tracking and cardiac strain analysis in 3D echocardiography.

Myocardial ischemia/infarction causes wall-motion abnormalities in the left ventricle. Therefore, reliable motion estimation and strain analysis using 3D+time echocardiography for localization and characterization of myocardial injury is valuable for early detection and targeted interventions. Previous unsupervised cardiac motion tracking methods rely on heavily-weighted regularization functions to smooth out the noisy displacement fields in echocardiography. In this work, we present a Co-Attention Spatial Transformer Network (STN) for improved motion tracking and strain analysis in 3D echocardiography. Co-Attention STN aims to extract inter-frame dependent features between frames to improve the motion tracking in otherwise noisy 3D echocardiography images. We also propose a novel temporal constraint to further regularize the motion field to produce smooth and realistic cardiac displacement paths over time without prior assumptions on cardiac motion. Our experimental results on both synthetic and in vivo 3D echocardiography datasets demonstrate that our Co-Attention STN provides superior performance compared to existing methods. Strain analysis from Co-Attention STNs also correspond well with the matched SPECT perfusion maps, demonstrating the clinical utility for using 3D echocardiography for infarct localization.

Increasing angular sampling through deep learning for stationary cardiac SPECT image reconstruction.

BACKGROUND: The GE Discovery NM (DNM) 530c/570c are dedicated cardiac SPECT scanners with 19 detector modules designed for stationary imaging. This study aims to incorporate additional projection angular sampling to improve reconstruction quality. A deep learning method is also proposed to generate synthetic dense-view image volumes from few-view counterparts. METHODS: By moving the detector array, a total of four projection angle sets were acquired and combined for image reconstructions. A deep neural network is proposed to generate synthetic four-angle images with 76 ([Formula: see text]) projections from corresponding one-angle images with 19 projections. Simulated data, pig, physical phantom, and human studies were used for network training and evaluation. Reconstruction results were quantitatively evaluated using representative image metrics. The myocardial perfusion defect size of different subjects was quantified using an FDA-cleared clinical software. RESULTS: Multi-angle reconstructions and network results have higher image resolution, improved uniformity on normal myocardium, more accurate defect quantification, and superior quantitative values on all the testing data. As validated against cardiac catheterization and diagnostic results, deep learning results showed improved image quality with better defect contrast on human studies. CONCLUSION: Increasing angular sampling can substantially improve image quality on DNM, and deep learning can be implemented to improve reconstruction quality in case of stationary imaging.

Reversal of synapse loss in Alzheimer mouse models by targeting mGluR5 to prevent synaptic tagging by C1Q.

Microglia-mediated synaptic loss contributes to the development of cognitive impairments in Alzheimer's disease (AD). However, the basis for this immune-mediated attack on synapses remains to be elucidated. Treatment with the metabotropic glutamate receptor 5 (mGluR5) silent allosteric modulator (SAM), BMS-984923, prevents ß-amyloid oligomer-induced aberrant synaptic signaling while preserving physiological glutamate response. Here, we show that oral BMS-984923 effectively occupies brain mGluR5 sites visualized by [18F]FPEB positron emission tomography (PET) at doses shown to be safe in rodents and nonhuman primates. In aged mouse models of AD (APPswe/PS1ΔE9 overexpressing transgenic and AppNL-G-F/hMapt double knock-in), SAM treatment fully restored synaptic density as measured by [18F]SynVesT-1 PET for SV2A and by histology, and the therapeutic benefit persisted after drug washout. Phospho-TAU accumulation in double knock-in mice was also reduced by SAM treatment. Single-nuclei transcriptomics demonstrated that SAM treatment in both models normalized expression patterns to a far greater extent in neurons than glia. Last, treatment prevented synaptic localization of the complement component C1Q and synaptic engulfment in AD mice. Thus, selective modulation of mGluR5 reversed neuronal gene expression changes to protect synapses from damage by microglial mediators in rodents.

[89Zr]ZrDFO-CR011 PET Correlates with Response to Glycoprotein Nonmetastatic Melanoma B-targeted Therapy in Triple-negative Breast Cancer.

There is a need for prognostic markers to select patients most likely to benefit from antibody-drug conjugate (ADC) therapy. We quantified the relationship between pretreatment PET imaging of glycoprotein nonmetastatic melanoma B (gpNMB) with 89Zr-labeled anti-gpNMB antibody ([89Zr]ZrDFO-CR011) and response to ADC therapy (CDX-011) in triple-negative breast cancer. First, we compared different PET imaging metrics and found that standardized uptake values (SUV) and tumor-to-heart SUV ratios were sufficient to delineate differences in radiotracer uptake in the tumor of four different cell- and patient-derived tumor models and achieved high standardized effect sizes. These tumor models with varying levels of gpNMB expression were imaged with [89Zr]ZrDFO-CR011 followed by treatment with a single bolus injection of CDX-011. The percent change in tumor volume relative to baseline (% CTV) was then correlated with SUVmean of [89Zr]ZrDFO-CR011 uptake in the tumor. All gpNMB-positive tumor models responded to CDX-011 over 6 weeks of treatment, except one patient-derived tumor regrew after 4 weeks of treatment. As expected, the gpNMB-negative tumor increased in volume by 130 ± 59% at endpoint. The magnitude of pretreatment SUV had the strongest inverse correlation with the % CTV at 2-4 weeks after treatment with CDX-011 (Spearman ρ = -0.8). However, pretreatment PET imaging with [89Zr]ZrDFO-CR011 did not inform on which tumor types will regrow over time. Other methods will be needed to predict resistance to treatment.

Development of [89Zr]ZrDFO-amivantamab bispecific to EGFR and c-MET for PET imaging of triple-negative breast cancer.

BACKGROUND: Amivantamab is a novel bispecific antibody that simultaneously targets the epidermal growth factor receptor (EGFR) and the hepatocyte growth factor receptor (HGFR/c-MET) that are overexpressed in several types of cancer including triple-negative breast cancer (TNBC). Targeting both receptors simultaneously can overcome resistance to mono-targeted therapy. The purpose of this study is to develop 89Zr-labeled amivantamab as a potential companion diagnostic imaging agent to amivantamab therapy using various preclinical models of TNBC for evaluation. METHODS: Amivantamab was conjugated to desferrioxamine (DFO) and radiolabeled with 89Zr to obtain [89Zr]ZrDFO-amivantamab. Binding of the bispecific [89Zr]ZrDFO-amivantamab as well as its mono-specific "single-arm" antibody controls were determined in vitro and in vivo. Biodistribution studies of [89Zr]ZrDFO-amivantamab were performed in MDA-MB-468 xenografts to determine the optimal imaging time point. PET/CT imaging with [89Zr]ZrDFO-amivantamab or its isotype control was performed in a panel of TNBC xenografts with varying levels of EGFR and c-MET expression. RESULTS: [89Zr]ZrDFO-amivantamab was synthesized with a specific activity of 148 MBq/mg and radiochemical yield of ≥ 95%. Radioligand binding studies and western blot confirmed the order of EGFR and c-MET expression levels: HCC827 lung cancer cell (positive control) > MDA-MB-468 > MDA-MB-231 > MDA-MB-453. [89Zr]ZrDFO-amivantamab demonstrated bispecific binding in cell lines co-expressed with EGFR and c-MET. PET/CT imaging with [89Zr]ZrDFO-amivantamab in TNBC xenografted mice showed standard uptake value (SUVmean) of 6.0 ± 1.1 in MDA-MB-468, 4.2 ± 1.4 in MDA-MB-231, and 1.5 ± 1.4 in MDA-MB-453 tumors, which are consistent with their receptors' expression levels on the cell surface. CONCLUSION: We have successfully prepared a radiolabeled bispecific antibody, [89Zr]ZrDFO-amivantamab, and evaluated its pharmacologic and imaging properties in comparison with its single-arm antibodies and non-specific isotype controls. [89Zr]ZrDFO-amivantamab demonstrated the greatest uptake in tumors co-expressing EGFR and c-MET.

Quantification of SV2A Binding in Rodent Brain Using [18F]SynVesT-1 and PET Imaging.

PURPOSE: Synapse loss is a hallmark of Alzheimer's disease (AD) and correlates with cognitive decline. The validation of a noninvasive in vivo imaging approach to quantify synapse would greatly facilitate our understanding of AD pathogenesis and assist drug developments for AD. As animal models of neurodegenerative and neuropsychiatric disorders play a critical role in the drug discovery and development process, a robust, objective, and translational method for quantifying therapeutic drug efficacy in animal models will facilitate the drug development process. In this study, we tested the quantification reliability of the SV2A PET tracer, [18F]SynVesT-1, in a mouse model of AD (APP/PS1) and wild-type controls, and developed a simplified quantification method to facilitate large cohort preclinical imaging studies. PROCEDURES: We generated nondisplaceable binding potential (BPND) and distribution volume ratio (DVR) values using the simplified reference tissue model (SRTM) on the 90-min dynamic PET imaging data, with brain stem and cerebellum as the reference region, respectively. Then, we correlated the standardized uptake value ratio (SUVR)-1 and SUVR averaged from different imaging windows with BPND and DVR, using brain stem and cerebellum as the reference region, respectively. We performed homologous competitive binding assay and autoradiographic saturation binding assay using [18F]SynVesT-1 to calculate the Bmax and Kd. RESULTS: Using brain stem as the reference region, the averaged SUVR-1 from 30 to 60 min postinjection correlated well with the BPND calculated using SRTM. Using cerebellum as the reference region, the averaged SUVR from 30 to 60 min postinjection correlated well with the SRTM DVR. From the homologous competitive binding assay and autoradiographic saturation binding assay, the calculated the Bmax and Kd were 4.5-18 pmol/mg protein and 9.8-19.6 nM, respectively, for rodent brain tissue. CONCLUSIONS: This simplified SUVR method provides reasonable SV2A measures in APP/PS1 mice and their littermate controls. Our data indicate that, in lieu of a full 90-min dynamic scan, a 30-min static PET scan (from 30 to 60 min postinjection) would be sufficient to provide quantification data on SV2A expression, equivalent to the data generated from kinetic modeling. The methods developed here are readily applicable to the evaluation of therapeutic effects of novel drugs in this rodent model using [18F]SynVesT-1 and small animal PET.

Imaging of HER2 with [89Zr]pertuzumab in Response to T-DM1 Therapy.

Background: The success of human epidermal growth factor receptor 2 (HER2)-targeted therapy depends on accurate characterization of HER2 expression, but current methods available have several limitations. This study aims to investigate the feasibility of [89Zr]pertuzumab imaging to monitor early response to Ado-trastuzumab emtansine (T-DM1) therapy in mice bearing xenografts of HER2-positive breast cancer (BCa). Materials and Methods: Pertuzumab was conjugated to DFO-Bz-NCS and labeled with 89Zr. Mice bearing BT-474 tumors were imaged with [89Zr]pertuzumab and [18F]FDG before and after T-DM1 therapy. Results: Pertuzumab was successfully labeled with 89Zr with a specific activity of 0.740 MBq/µg. Overall [18F]FDG images showed poor delineation of tumors. Using [18F]FDG-PET to measure tumor volume, the volume remained unchanged from 107.6 ± 20.7 mm3 before treatment to 89.87 ± 66.55 mm3 after treatment. In contrast, [89Zr]pertuzumab images showed good delineation of HER2-positive tumors, allowing accurate detection of changes in tumor volume (from 243.80 ± 40.91 mm3 before treatment to 78.4 ± 40.43 mm3 after treatment). Conclusion: [89Zr]pertuzumab may be an imaging probe for monitoring the response of HER2-positive BCa patients to T-DM1 therapy.

Preclinical PET imaging of glycoprotein non-metastatic melanoma B in triple negative breast cancer: feasibility of an antibody-based companion diagnostic agent.

High levels of expression of glycoprotein non-metastatic B (gpNMB) in triple negative breast cancer (TNBC) and its association with metastasis and recurrence make it an attractive target for therapy with the antibody drug conjugate, glembatumumab vedotin (CDX-011). This report describes the development of a companion PET-based diagnostic imaging agent using 89Zr-labeled glembatumumab ([89Zr]DFO-CR011) to potentially aid in the selection of patients most likely to respond to targeted treatment with CDX-011. [89Zr]DFO-CR011 was characterized for its pharmacologic properties in TNBC cell lines. Preclinical studies determined that [89Zr]DFO-CR011 binds specifically to gpNMB with high affinity (Kd = 25 ± 5 nM), immunoreactivity of 2.2-fold less than the native CR011, and its cellular uptake correlates with gpNMB expression (r = 0.95). In PET studies at the optimal imaging timepoint of 7 days p.i., the [89Zr]DFO-CR011 tumor uptake in gpNMB-expressing MDA-MB-468 xenografts had a mean SUV of 2.9, while significantly lower in gpNMB-negative MDA-MB-231 tumors with a mean SUV of 1.9. [89Zr]DFO-CR011 was also evaluated in patient-derived xenograft models of TNBC, where tumor uptake in vivo had a positive correlation with total gpNMB protein expression via ELISA (r = 0.79), despite the heterogeneity of gpNMB expression within the same group of PDX mice. Lastly, the radiation dosimetry calculated from biodistribution studies in MDA-MB-468 xenografts determined the effective dose for human use would be 0.54 mSv/MBq. Overall, these studies demonstrate that [89Zr]DFO-CR011 is a potential companion diagnostic imaging agent for CDX-011 which targets gpNMB, an emerging biomarker for TNBC.