Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet.

BACKGROUND: Magnetic resonance imaging (MRI) of the knee is the preferred method for diagnosing knee injuries. However, interpretation of knee MRI is time-intensive and subject to diagnostic error and variability. An automated system for interpreting knee MRI could prioritize high-risk patients and assist clinicians in making diagnoses. Deep learning methods, in being able to automatically learn layers of features, are well suited for modeling the complex relationships between medical images and their interpretations. In this study we developed a deep learning model for detecting general abnormalities and specific diagnoses (anterior cruciate ligament [ACL] tears and meniscal tears) on knee MRI exams. We then measured the effect of providing the model's predictions to clinical experts during interpretation. METHODS AND FINDINGS: Our dataset consisted of 1,370 knee MRI exams performed at Stanford University Medical Center between January 1, 2001, and December 31, 2012 (mean age 38.0 years; 569 [41.5%] female patients). The majority vote of 3 musculoskeletal radiologists established reference standard labels on an internal validation set of 120 exams. We developed MRNet, a convolutional neural network for classifying MRI series and combined predictions from 3 series per exam using logistic regression. In detecting abnormalities, ACL tears, and meniscal tears, this model achieved area under the receiver operating characteristic curve (AUC) values of 0.937 (95% CI 0.895, 0.980), 0.965 (95% CI 0.938, 0.993), and 0.847 (95% CI 0.780, 0.914), respectively, on the internal validation set. We also obtained a public dataset of 917 exams with sagittal T1-weighted series and labels for ACL injury from Clinical Hospital Centre Rijeka, Croatia. On the external validation set of 183 exams, the MRNet trained on Stanford sagittal T2-weighted series achieved an AUC of 0.824 (95% CI 0.757, 0.892) in the detection of ACL injuries with no additional training, while an MRNet trained on the rest of the external data achieved an AUC of 0.911 (95% CI 0.864, 0.958). We additionally measured the specificity, sensitivity, and accuracy of 9 clinical experts (7 board-certified general radiologists and 2 orthopedic surgeons) on the internal validation set both with and without model assistance. Using a 2-sided Pearson's chi-squared test with adjustment for multiple comparisons, we found no significant differences between the performance of the model and that of unassisted general radiologists in detecting abnormalities. General radiologists achieved significantly higher sensitivity in detecting ACL tears (p-value = 0.002; q-value = 0.019) and significantly higher specificity in detecting meniscal tears (p-value = 0.003; q-value = 0.019). Using a 1-tailed t test on the change in performance metrics, we found that providing model predictions significantly increased clinical experts' specificity in identifying ACL tears (p-value < 0.001; q-value = 0.006). The primary limitations of our study include lack of surgical ground truth and the small size of the panel of clinical experts. CONCLUSIONS: Our deep learning model can rapidly generate accurate clinical pathology classifications of knee MRI exams from both internal and external datasets. Moreover, our results support the assertion that deep learning models can improve the performance of clinical experts during medical imaging interpretation. Further research is needed to validate the model prospectively and to determine its utility in the clinical setting.

Deep learning for chest radiograph diagnosis: A retrospective comparison of the CheXNeXt algorithm to practicing radiologists.

BACKGROUND: Chest radiograph interpretation is critical for the detection of thoracic diseases, including tuberculosis and lung cancer, which affect millions of people worldwide each year. This time-consuming task typically requires expert radiologists to read the images, leading to fatigue-based diagnostic error and lack of diagnostic expertise in areas of the world where radiologists are not available. Recently, deep learning approaches have been able to achieve expert-level performance in medical image interpretation tasks, powered by large network architectures and fueled by the emergence of large labeled datasets. The purpose of this study is to investigate the performance of a deep learning algorithm on the detection of pathologies in chest radiographs compared with practicing radiologists. METHODS AND FINDINGS: We developed CheXNeXt, a convolutional neural network to concurrently detect the presence of 14 different pathologies, including pneumonia, pleural effusion, pulmonary masses, and nodules in frontal-view chest radiographs. CheXNeXt was trained and internally validated on the ChestX-ray8 dataset, with a held-out validation set consisting of 420 images, sampled to contain at least 50 cases of each of the original pathology labels. On this validation set, the majority vote of a panel of 3 board-certified cardiothoracic specialist radiologists served as reference standard. We compared CheXNeXt's discriminative performance on the validation set to the performance of 9 radiologists using the area under the receiver operating characteristic curve (AUC). The radiologists included 6 board-certified radiologists (average experience 12 years, range 4-28 years) and 3 senior radiology residents, from 3 academic institutions. We found that CheXNeXt achieved radiologist-level performance on 11 pathologies and did not achieve radiologist-level performance on 3 pathologies. The radiologists achieved statistically significantly higher AUC performance on cardiomegaly, emphysema, and hiatal hernia, with AUCs of 0.888 (95% confidence interval [CI] 0.863-0.910), 0.911 (95% CI 0.866-0.947), and 0.985 (95% CI 0.974-0.991), respectively, whereas CheXNeXt's AUCs were 0.831 (95% CI 0.790-0.870), 0.704 (95% CI 0.567-0.833), and 0.851 (95% CI 0.785-0.909), respectively. CheXNeXt performed better than radiologists in detecting atelectasis, with an AUC of 0.862 (95% CI 0.825-0.895), statistically significantly higher than radiologists' AUC of 0.808 (95% CI 0.777-0.838); there were no statistically significant differences in AUCs for the other 10 pathologies. The average time to interpret the 420 images in the validation set was substantially longer for the radiologists (240 minutes) than for CheXNeXt (1.5 minutes). The main limitations of our study are that neither CheXNeXt nor the radiologists were permitted to use patient history or review prior examinations and that evaluation was limited to a dataset from a single institution. CONCLUSIONS: In this study, we developed and validated a deep learning algorithm that classified clinically important abnormalities in chest radiographs at a performance level comparable to practicing radiologists. Once tested prospectively in clinical settings, the algorithm could have the potential to expand patient access to chest radiograph diagnostics.

Enhanced caspase activity contributes to aortic wall remodeling and early aneurysm development in a murine model of Marfan syndrome.

OBJECTIVE: Rupture and dissection of aortic root aneurysms remain the leading causes of death in patients with the Marfan syndrome, a hereditary connective tissue disorder that affects 1 in 5000 individuals worldwide. In the present study, we use a Marfan mouse model (Fbn1(C1039G/+)) to investigate the biological importance of apoptosis during aneurysm development in Marfan syndrome. APPROACH AND RESULTS: Using in vivo single-photon emission computed tomographic-imaging and ex vivo autoradiography for Tc99m-annexin, we discovered increased apoptosis in the Fbn1(C1039G/+) ascending aorta during early aneurysm development peaking at 4 weeks. Immunofluorescence colocalization studies identified smooth muscle cells (SMCs) as the apoptotic cell population. As biological proof of concept that early aortic wall apoptosis plays a role in aneurysm development in Marfan syndrome, Fbn1(C1039G/+) mice were treated daily from 2 to 6 weeks with either (1) a pan-caspase inhibitor, Q-VD-OPh (20 mg/kg), or (2) vehicle control intraperitoneally. Q-VD-OPh treatment led to a significant reduction in aneurysm size and decreased extracellular matrix degradation in the aortic wall compared with control mice. In vitro studies using Fbn1(C1039G/+) ascending SMCs showed that apoptotic SMCs have increased elastolytic potential compared with viable cells, mostly because of caspase activity. Moreover, in vitro (1) cell membrane isolation, (2) immunofluorescence staining, and (3) scanning electron microscopy studies illustrate that caspases are expressed on the exterior cell surface of apoptotic SMCs. CONCLUSIONS: Caspase inhibition attenuates aneurysm development in an Fbn1(C1039G/+) Marfan mouse model. Mechanistically, during apoptosis, caspases are expressed on the cell surface of SMCs and likely contribute to elastin degradation and aneurysm development in Marfan syndrome.

(99m)Tc-Annexin A5 quantification of apoptotic tumor response: a systematic review and meta-analysis of clinical imaging trials.

PURPOSE: (99m)Tc-Annexin A5 has been used as a molecular imaging probe for the visualization, characterization and measurement of apoptosis. In an effort to define the quantitative (99m)Tc-annexin A5 uptake criteria that best predict tumor response to treatment, we performed a systematic review and meta-analysis of the results of all clinical imaging trials found in the literature or publicly available databases. METHODS: Included in this review were 17 clinical trials investigating quantitative (99m)Tc-annexin A5 (qAnx5) imaging using different parameters in cancer patients before and after the first course of chemotherapy and/or radiation therapy. Qualitative assessment of the clinical studies for diagnostic accuracy was performed using the QUADAS-2 criteria. Of these studies, five prospective single-center clinical trials (92 patients in total) were included in the meta-analysis after exclusion of one multicenter clinical trial due to heterogeneity. Pooled positive predictive values (PPV) and pooled negative predictive values (NPV) (with 95% CI) were calculated using Meta-Disc software version 1.4. RESULTS: Absolute quantification and/or relative quantification of (99m)Tc-annexin A5 uptake were performed at baseline and after the start of treatment. Various quantitative parameters have been used for the calculation of (99m)Tc-annexin A5 tumor uptake and delta (Δ) tumor changes post-treatment compared to baseline including: tumor-to-background ratio (TBR), ΔTBR, tumor-to-noise ratio, relative tumor ratio (TR), ΔTR, standardized tumor uptake ratio (STU), ΔSTU, maximum count per pixel within the tumor volume (Cmax), Cmax%, absolute ΔU and percentage (ΔU%), maximum ΔU counts, semiquantitative visual scoring, percent injected dose (%ID) and %ID/cm(3). Clinical trials investigating qAnx5 imaging have included patients with lung cancer, lymphoma, breast cancer, head and neck cancer and other less common tumor types. In two phase I/II single-center clinical trials, an increase of ≥25% in uptake following treatment was considered a significant threshold for an apoptotic tumor response (partial response, complete response). In three other phase I/II clinical trials, increases of ≥28%, ≥42% and ≥47% in uptake following treatment were found to be the mean cut-off levels in responders. In a phase II/III multicenter clinical trial, an increase of ≥23% in uptake following treatment was found to be the minimum cut-off level for a tumor response. In one clinical trial, no significant difference in (99m)Tc-annexin A5 uptake in terms of %ID was found in healthy tissues after chemotherapy compared to baseline. In two other clinical trials, intraobserver and interobserver measurements of (99m)Tc-annexin A5 tumor uptake were found to be reproducible (mean difference <5%, kappa = 0.90 and 0.82, respectively) and to be highly correlated with treatment outcome (Spearman r = 0.99, p < 0.0001). The meta-analysis demonstrated a pooled positive PPV of 100% (95% CI 92 - 100%) and a pooled NPV of 70% (95% CI 55 - 82%) for prediction of a tumor response after the first course of chemotherapy and/or radiotherapy in terms of ΔU%. In a symmetric sROC analysis, the AUC was 0.919 and the Q* index was 85.21 %. CONCLUSION: Quantitative (99m)Tc-annexin A5 imaging has been investigated in clinical trials for the assessment of apoptotic tumor responses. This meta-analysis showed a high pooled PPV and a moderate pooled NPV with ΔU cut-off values ranging between 20% and 30%. Standardization of quantification and harmonization of results are required for high-quality clinical research. A standardized uptake value score (SUV, ΔSUV) using quantitative SPECT/CT imaging may be a promising approach to the simple, reproducible and semiquantitative assessment of apoptotic tumor changes.

Molecular imaging of VEGF receptors in angiogenic vasculature with single-chain VEGF-based probes.

We describe a new generation of protein-targeted contrast agents for multimodal imaging of the cell-surface receptors for vascular endothelial growth factor (VEGF). These receptors have a key role in angiogenesis and are important targets for drug development. Our probes are based on a single-chain recombinant VEGF expressed with a cysteine-containing tag that allows site-specific labeling with contrast agents for near-infrared fluorescence imaging, single-photon emission computed tomography or positron emission tomography. These probes retain VEGF activities in vitro and undergo selective and highly specific focal uptake into the vasculature of tumors and surrounding host tissue in vivo. The fluorescence contrast agent shows long-term persistence and co-localizes with endothelial cell markers, indicating that internalization is mediated by the receptors. We expect that multimodal imaging of VEGF receptors with these probes will be useful for clinical diagnosis and therapeutic monitoring, and will help to accelerate the development of new angiogenesis-directed drugs and treatments.

Hydrogels for Local and Sustained Delivery of Bacteriophages to Treat Wound Infections.

Bacteriophages (phages) are viruses that specifically target and kill bacteria, serving as a promising therapeutic to combat multidrug-resistant (MDR) pathogens such as Pseudomonas aeruginosa (Pa). However, delivering adequate concentrations of active phages directly to the infection site over sufficient times to eradicate infections remains an outstanding challenge to phage therapy (PT). Here we present "HydroPhage", a biocompatible hydrogel system for the sustained release of high-titre phages to effectively treat infections caused by MDR pathogens. We develop injectable hydrogels comprised of hyaluronic acid (HA) and polyethylene glycol (PEG) crosslinked through static covalent thioether bonds and hemithioacetal-based dynamic covalent crosslinks (DCC), which encapsulate phages at concentration up to 1011 PFU per mL gel, and achieve sustained release over a week with more than 60% total phage recovery. In a preclinical mouse model of extended wound infection, we demonstrate enhanced bacterial clearance compared to intravenous treatment. Thus, using hydrogels for local and sustained delivery of phage may represent an effective approach to eradicating MDR infections broadly.

The contribution of neutrophils to bacteriophage clearance and pharmacokinetics in vivo.

With the increasing prevalence of antimicrobial-resistant bacterial infections, there is great interest in using lytic bacteriophages (phages) to treat such infections. However, the factors that govern bacteriophage pharmacokinetics in vivo remain poorly understood. Here, we have examined the contribution of neutrophils, the most abundant phagocytes in the body, to the pharmacokinetics of intravenously administered bacteriophage in uninfected mice. A single dose of LPS-5, an antipseudomonal bacteriophage recently used in human clinical trials, was administered intravenously to both wild-type BALB/c and neutropenic ICR mice. Phage concentrations were assessed in peripheral blood and spleen at 0.5, 1, 2, 4, 8, 12, and 24 hours after administration by plaque assay and qPCR. We observed that the phage clearance is only minimally affected by neutropenia. Indeed, the half-life of phages in blood in BALB/c and ICR mice is 3.45 and 3.66 hours, respectively. These data suggest that neutrophil-mediated phagocytosis is not a major determinant of phage clearance. Conversely, we observed a substantial discrepancy in circulating phage levels over time when measured by qPCR versus plaque assay, suggesting that substantial functional inactivation of circulating phages occurs over time. These data indicate that circulating factors, but not neutrophils, inactivate intravenously administered phages.

Attenuation of apoptosis by telmisartan in atherosclerotic plaques of apolipoprotein E-/- mice: evaluation using technetium 99m-annexin A5.

Technetium 99m (99mTc)-annexin A5, a marker of ongoing apoptosis, is supposed to be useful in the detection of metabolically active atheroma. The aim of this study was to determine the potential of 99mTc-annexin A5 for evaluating the therapeutic effects of an angiotensin II receptor type 1 blocker (ARB) (telmisartan) on atherosclerosis. Male apolipoprotein E-/- mice were divided into telmisartan-treated (3 mg/kg/d, n  =  10) and control (n  =  10) groups. After 16 to 21 weeks of treatment, 99mTc-annexin A5 was injected and cryostat sections of aortic tissues (n  =  10-12/aorta) were prepared. The 99mTc-annexin A5 accumulation level in the plaques was evaluated by autoradiography. Serial sections of the plaques were histologically examined to identify the lesion phenotypes (normal vessels, early lesions, atheromatous lesions, and fibrotic lesions), plaque size, macrophage infiltration levels, and lipid deposition levels. Telmisartan treatment significantly decreased the plaque size (0.05 ± 0.05 vs 0.11 ± 0.08, mm2), macrophage infiltration level (0.02 ± 0.02 vs 0.03 ± 0.02, mm2), lipid deposition level (0.01 ± 0.01 vs 0.02 ± 0.02, mm2), and 99mTc-annexin A5 accumulation level (1.30 ± 1.09 vs 2.15 ± 1.91, × 10-6/g). 99mTc-annexin A5 accumulation levels in the plaques positively correlated with macrophage infiltration (r  =  .69, p < .05) and lipid deposition (r  =  .66, p < .05) levels. Apoptosis imaging with 99mTc-annexin A5 may be useful for evaluating the therapeutic effects of ARBs on atherosclerosis.

Rapid assessment of changes in phage bioactivity using dynamic light scattering.

Extensive efforts are underway to develop bacteriophages as therapies against antibiotic-resistant bacteria. However, these efforts are confounded by the instability of phage preparations and a lack of suitable tools to assess active phage concentrations over time. Here, we use Dynamic Light Scattering (DLS) to measure changes in phage physical state in response to environmental factors and time, finding that phages tend to decay and form aggregates and that the degree of aggregation can be used to predict phage bioactivity. We then use DLS to optimize phage storage conditions for phages from human clinical trials, predict bioactivity in 50-year-old archival stocks, and evaluate phage samples for use in a phage therapy/wound infection model. We also provide a web-application (Phage-ELF) to facilitate DLS studies of phages. We conclude that DLS provides a rapid, convenient, and non-destructive tool for quality control of phage preparations in academic and commercial settings.

Rapid assessment of changes in phage bioactivity using dynamic light scattering.

Extensive efforts are underway to develop bacteriophages as therapies against antibiotic-resistant bacteria. However, these efforts are confounded by the instability of phage preparations and a lack of suitable tools to assess active phage concentrations over time. In this study, we use dynamic light scattering (DLS) to measure changes in phage physical state in response to environmental factors and time, finding that phages tend to decay and form aggregates and that the degree of aggregation can be used to predict phage bioactivity. We then use DLS to optimize phage storage conditions for phages from human clinical trials, predict bioactivity in 50-y-old archival stocks, and evaluate phage samples for use in a phage therapy/wound infection model. We also provide a web application (Phage-Estimator of Lytic Function) to facilitate DLS studies of phages. We conclude that DLS provides a rapid, convenient, and nondestructive tool for quality control of phage preparations in academic and commercial settings.

Brain uptake of Tc99m-HMPAO correlates with clinical response to the novel redox modulating agent EPI-743 in patients with mitochondrial disease.

While decreased ATP production and redox imbalance are central to mitochondrial disease pathogenesis, efforts to develop effective treatments have been hampered by the lack of imaging markers of oxidative stress. In this study we wished to determine if Tc99m-HMPAO, a SPECT imaging marker of cerebral blood flow and glutathione/protein thiol content, could be used to monitor the effect(s) of EPI-743, an oral redox modulating, para-benzoquinone based therapeutic for mitochondrial disease. We hypothesized that treatment changes in HMPAO uptake would be inversely proportional to changes in oxidative stress within the brain and directly correlate to clinical response to EPI-743 therapy. Twenty-two patients with mitochondrial disease were treated with EPI-743. Each underwent baseline and 3-month Tc99m-HMPAO SPECT scanning along with clinical/neurologic evaluations. Diseases treated were: Leigh syndrome (n=7), polymerase γ deficiency (n=5), MELAS (n=5), Friedreich ataxia (n=2), Kearns-Sayre syndrome, Pearson syndrome, and mtDNA depletion syndrome. Neuro-anatomic uptake analyses of HMPAO were performed with NeuroGam™ (Segami Corp.) statistical software and clinical response was assessed by the Newcastle Paediatric Mitochondrial Disease Scale or Newcastle Mitochondrial Disease Adult Scale depending on patient age. For all 22 patients there was a significant linear correlation between the change in cerebellar uptake of HMPAO and the improvement in Newcastle score (r=0.623, **p=0.00161). The MELAS subgroup showed a significant relationship of whole brain uptake (n=5, r=0.917, *p=0.028) to improvement in Newcastle score. We conclude that Tc99m-HMPAO SPECT scanning has promise as a general marker of the oxidative state of the brain and its response to redox modulating therapies. Further studies will be needed to confirm these findings in a more homogenous study population.

Multimodality molecular imaging of apoptosis in oncology.

OBJECTIVE: The purposes of this review are to describe the signaling pathways of and the cellular changes that occur with apoptosis and other forms of cell death, summarize tracers and modalities used for imaging of apoptosis, delineate the relation between apoptosis and inhibition of protein translation, and describe spectroscopic technologies that entail high-frequency ultrasound and infrared and midinfrared light in characterizing the intracellular events of apoptosis. CONCLUSION: Apoptosis is a highly orchestrated set of biochemical and morphologic cellular events. These events present many potential targets for the imaging of apoptosis in vivo. Imaging of apoptosis can facilitate early assessment of anticancer treatment before tumor shrinkage, which may increase the effectiveness of delivery of chemotherapy and radiation therapy and speed drug development.

A novel roadmap connecting the 1H-MRS total choline resonance to all hallmarks of cancer following targeted therapy.

This review describes a cellular adaptive stress signalling roadmap connecting the 1H magnetic resonance spectroscopy (MRS) total choline peak at 3.2 ppm (tCho) to cancer response after targeted therapy (TT). Recent research on cell signalling, tCho metabolism, and TT of cancer has been retrospectively re-examined. Signalling research describes how the unfolded protein response (UPR), a major stress signalling network, transduces, regulates, and rewires the total membrane turnover in different cancer hallmarks after a TT stress. In particular, the UPR signalling maintains or increases total membrane turnover in all pro-survival hallmarks, whilst dramatically decreases turnover during apoptosis, a pro-death hallmark. Recent research depicts the TT-induced stress as a crucial event responsible for interrupting UPR pro-survival pathways, leading to an UPR-mediated cell death. The 1H-MRS tCho resonance represents the total mobile precursors and products during the enzymatic modification of phosphatidylcholine membrane abundance. The tCho profile represents a biomarker that noninvasively monitors TT-induced enzymatic changes in total membrane turnover in a wide variety of existing and new anticancer treatments targeting specific layers of the UPR signalling network. Our overview strongly suggests further evaluating and validating the 1H-MRS tCho peak as a powerful noninvasive imaging biomarker of cancer response in TT clinical trials.

Immunomodulation Therapy Using Tolerogenic Macrophages in a Rodent Model of Pulmonary Hypertension.

Inflammation plays a major role in the pathogenesis of pulmonary hypertension (PH). We sought to investigate the effects of a cell-based immunomodulation in a dysimmune model of PH. PH was induced in athymic nude rats using semaxinib (Su group, n = 6). Tolerogenic macrophages (toM) were generated from monocyte isolation and then injected either the day before semaxinib injection (Prevention group, n = 6) or 3 weeks after (Reversion group, n = 6). Six athymic nude rats were used as controls. In vivo trafficking of toM was investigated with bioluminescence imaging showing that toM were mainly located into the lungs until 48 h after injection. Right ventricular (RV) end-systolic pressure and RV systolic function were assessed at 4 weeks using echocardiography. Morphometric analysis and RNA sequencing of the lungs were realized at 4 weeks. Rats treated with toM (Prevention and Reversion groups) had a significantly lower RV end-systolic pressure at 4 weeks (respectively, 25 ± 8 and 30 ± 6 mmHg vs. 67 ± 9 mmHg, P < 0.001), while RV systolic dysfunction was observed in Su and Reversion groups. Mean medial wall thickness of small arterioles was lower in Prevention and Reversion groups compared with the Su group (respectively, 10.9% ± 0.8% and 16.4% ± 1.3% vs. 28.2% ± 2.1%, P < 0.001). Similarly, cardiomyocyte area was decreased in rats treated with toM (150 ± 18 and 160 ± 86 µm2 vs. 279 ± 50 µm2, P < 0.001). A trend toward upregulation of genes involved in pulmonary arterial hypertension pathobiology was found in Su rats, while KCNK3 was significantly downregulated (fold-change = 9.8, P < 0.001). Injection of toM was associated with a less severe phenotype of PH in rats exposed to angioproliferative stress. Preserved expression of KCNK3 may explain the protective effect of toM.

Characterization of a recombinant form of annexin VI for detection of apoptosis.

We developed a recombinant form of human annexin VI called annexin VI-601 (M(r) 76,224) with the N-terminal extension of Ala-Gly-Gly-Cys-Gly-His to allow ready attachment of fluorescent or radioactive labels. The protein was produced by expression in E. coli and was purified by calcium-dependent membrane binding, anion-exchange chromatography, and heparin-Sepharose affinity chromatography. The protein could be readily labeled with iodoacetamidofluorescein and with (99m)Tc. The protein bound with high affinity to PS-containing phospholipid vesicles and to erythrocytes with exposed phosphatidylserine. Fluorescent annexin VI-601 readily detected apoptosis of Jurkat cells by flow cytometry at much lower calcium concentrations than those required for equivalent detection by annexin V. In vivo administration of radiolabeled protein showed that blood clearance was much slower than annexin V. In conclusion, annexin VI may have advantages over annexin V in certain situations for both in vitro and in vivo detection of apoptosis and therapeutic targeting of PS due to its lower calcium requirement for membrane binding and its higher molecular weight.

Analysis of in situ and ex vivo vascular endothelial growth factor receptor expression during experimental aortic aneurysm progression.

OBJECTIVE: Mural inflammation and neovascularization are characteristic pathological features of abdominal aortic aneurysm (AAA) disease. Vascular endothelial growth factor receptor (VEGFR) expression may also mediate AAA growth and rupture. We examined VEGFR expression as a function of AAA disease progression in the Apolipoprotein E-deficient (Apo E(-/-)) murine AAA model. METHODS AND RESULTS: Apo E(-/-) mice maintained on a high-fat diet underwent continuous infusion with angiotensin II at 1000 ng/kg/min (Ang II) or vehicle (Control) via subcutaneous osmotic pump. Serial transabdominal ultrasound measurements of abdominal aortic diameter were recorded (n=16 mice, 3 to 4 time points per mouse) for up to 28 days. Near-infrared receptor fluorescent (NIRF) imaging was performed on Ang II mice (n=9) and Controls (n=5) with scVEGF/Cy, a single-chain VEGF homo-dimer labeled with Cy 5.5 fluorescent tracer (7 to 18 microg/mouse IV). NIRF with inactivated single chain VEGF/Cy tracer (scVEGF/In, 18 microg/mouse IV) was performed on 2 additional Ang II mice to control for nonreceptor-mediated tracer binding and uptake. After image acquisition and sacrifice, aortae were harvested for analysis. An additional AAA mouse cohort received either an oral angiogenesis inhibitor or suitable negative or positive controls to clarify the significance of angiogenesis in experimental aneurysm progression. Aneurysms developed in the suprarenal aortic segment of all Ang II mice. Significantly greater fluorescent signal was obtained from aneurysmal aorta as compared to remote, uninvolved aortic segments in Ang II scVEGF/Cy mice or AAA in scVEGF/In mice or suprarenal aortic segments in Control mice. Signal intensity increased in a diameter-dependent fashion in aneurysmal segments. Immunostaining confirmed mural VEGFR-2 expression in medial smooth muscle cells. Treatment with an angiogenesis inhibitor attenuated AAA formation while decreasing mural macrophage infiltration and CD-31(+) cell density. CONCLUSIONS: Mural VEGFR expression, as determined by scVEGF/Cy fluorescent imaging and VEGFR-2 immunostaining, increases in experimental AAAs in a diameter-dependent fashion. Angiogenesis inhibition limits AAA progression. Clinical VEGFR expression imaging strategies, if feasible, may improve real-time monitoring of AAA disease progression and response to suppressive strategies.

Molecular Imaging of Apoptosis in Atherosclerosis by Targeting Cell Membrane Phospholipid Asymmetry.

BACKGROUND: Apoptosis in atherosclerotic lesions contributes to plaque vulnerability by lipid core enlargement and fibrous cap attenuation. Apoptosis is associated with exteriorization of phosphatidylserine (PS) and phosphatidylethanolamine (PE) on the cell membrane. Although PS-avid radiolabeled annexin-V has been employed for molecular imaging of high-risk plaques, PE-targeted imaging in atherosclerosis has not been studied. OBJECTIVES: This study sought to evaluate the feasibility of molecular imaging with PE-avid radiolabeled duramycin in experimental atherosclerotic lesions in a rabbit model and compare duramycin targeting with radiolabeled annexin-V. METHODS: Of the 27 rabbits, 21 were fed high-cholesterol, high-fat diet for 16 weeks. Nine of the 21 rabbits received 99mTc-duramycin (test group), 6 received 99mTc-linear duramycin (duramycin without PE-binding capability, negative radiotracer control group), and 6 received 99mTc-annexin-V for radionuclide imaging. The remaining normal chow-fed 6 animals (disease control group) received 99mTc-duramycin. In vivo microSPECT/microCT imaging was performed, and the aortas were explanted for ex vivo imaging and for histological characterization of atherosclerosis. RESULTS: A significantly higher duramycin uptake was observed in the test group compared with that of disease control and negative radiotracer control animals; duramycin uptake was also significantly higher than the annexin-V uptake. Quantitative duramycin uptake, represented as the square root of percent injected dose per cm (√ID/cm) of abdominal aorta was >2-fold higher in atherosclerotic lesions in test group (0.08 ± 0.01%) than in comparable regions of disease control animals (0.039 ± 0.0061%, p = 3.70·10-8). Mean annexin uptake (0.060 ± 0.010%) was significantly lower than duramycin (p = 0.001). Duramycin uptake corresponded to the lesion severity and macrophage burden. The radiation burden to the kidneys was substantially lower with duramycin (0.49% ID/g) than annexin (5.48% ID/g; p = 4.00·10-4). CONCLUSIONS: Radiolabeled duramycin localizes in lipid-rich areas with high concentration of apoptotic macrophages in the experimental atherosclerosis model. Duramycin uptake in atherosclerotic lesions was significantly greater than annexin-V uptake and produced significantly lower radiation burden to nontarget organs.

The use of 99mTc-recombinant human annexin V imaging for differential diagnosis of aseptic loosening and low-grade infection in hip and knee prostheses.

UNLABELLED: The aim of this study was to assess the pattern of annexin V uptake in hip and knee prostheses suspected of being infected. METHODS: A total of 7 patients undergoing revision surgery for hip or knee prostheses were studied; 5 patients had total hip replacements, and 2 had total knee replacements. Infection was confirmed by pathology, culture results, laboratory evaluation, and clinical follow-up. All patients also underwent a bone scan before surgery. RESULTS: Annexin V scan findings were positive in 5 patients and negative in 2. Annexin V uptake was either focal (n = 4) or linear (n = 1). There were 4 true-positive, 2 true-negative, 1 false-positive, and no false-negative annexin V studies. Annexin V uptake was either more extensive or less extensive than, and usually was incongruent with, (99m)Tc-methylene diphosphonate uptake. CONCLUSION: Our findings suggest that annexin V imaging shows greater uptake with infection than with aseptic loosening and has a high negative predictive value for prosthetic infection.

SPECT and PET imaging of EGF receptors with site-specifically labeled EGF and dimeric EGF.

We describe a new generation of tracers for molecular imaging of the cell surface receptors for epidermal growth factor (EGF). These receptors play a key role in the progression of many tumors and are major drug development targets. Our tracers are based on a recombinant human EGF expressed with a cysteine-containing tag that enables facile site-specific radiolabeling with (99m)Tc for single photon emission computed tomography or site-specific conjugation of (64)Cu PEGylated chelators for positron emission tomography. These tracers retain EGF activities in vitro and display selective and highly specific focal uptake in tumors in vivo. We expect that nuclear imaging of EGF receptors with these tracers will be useful for clinical diagnosis, therapeutic monitoring, and development of new drugs and treatment regimens.

In vivo detection of apoptosis.

After several decades of debate, it is now widely acknowledged that apoptosis, also known as programmed cell death, is central to homoeostasis and normal development and physiology in all multicellular organisms, including humans. The dysregulation of apoptosis can lead to the destruction of normal tissues in a variety of disorders, including autoimmune and neurodegenerative diseases (too much apoptosis) or the growth of tumors (too little apoptosis). In addition, effective therapy of tumors requires the iatrogenic induction of programmed cell death by radiation, chemotherapy, or both. Given the central role of apoptosis, it would be desirable to have a noninvasive imaging method to serially detect and monitor this process in cancer patients undergoing conventional radiation and chemotherapy treatments as well as for the development and testing of new drugs. In this article, the latest modalities and contrast agents described in the literature for the imaging of apoptosis in vivo are reviewed. First, the most recent developments in the biochemical characterization of the many intracellular pathways involved in this complex process are discussed. Next, a variety of new radionuclide tracers, including radiolabeled annexin V and caspase inhibitors for PET and SPECT, are described. Finally, the use of MRI, MR spectroscopy, and ultrasound as possible alternative imaging modalities for the imaging of apoptosis is addressed.