In situ validation of VEGFR-2 and α v ß 3 integrin as targets for breast lesion characterization.

Vascular endothelial growth factor receptor 2 (VEGFR-2) and α v ß 3 integrin are the most frequently addressed targets in molecular imaging of tumor angiogenesis. In preclinical studies, molecular imaging of angiogenesis has shown potential to detect and differentiate benign and malignant lesions of the breast. Thus, in this retrospective clinical study employing patient tissues, the diagnostic value of VEGFR-2, α v ß 3 integrin and vascular area fraction for the diagnosis and differentiation of breast neoplasia was evaluated. To this end, tissue sections of breast cancer (n = 40), pre-invasive ductal carcinoma in situ (DCIS; n = 8), fibroadenoma (n = 40), radial scar (n = 6) and normal breast tissue (n = 40) were used to quantify (1) endothelial VEGFR-2, (2) endothelial α v ß 3 integrin and (3) total α v ß 3 integrin expression, as well as (4) the vascular area fraction. Sensitivity and specificity to differentiate benign from malignant lesions were calculated for each marker by receiver operating characteristics (ROC) analyses. Whereas vessel density, as commonly used, did not significantly differ between benign and malignant lesions (AUROC: 0.54), VEGFR-2 and α v ß 3 integrin levels were gradually up-regulated in carcinoma versus fibroadenoma versus healthy tissue. The highest diagnostic accuracy for differentiating carcinoma from fibroadenoma was found for total α v ß 3 integrin expression (AUROC: 0.76), followed by VEGFR-2 (AUROC: 0.71) and endothelial α v ß 3 integrin expression (AUROC: 0.68). In conclusion, total α v ß 3 integrin expression is the best discriminator between breast cancer, fibroadenoma and normal breast tissue. With respect to vascular targeting and molecular imaging of angiogenesis, endothelial VEGFR-2 appeared to be slightly superior to endothelial α v ß 3 for differentiating benign from cancerous lesions.

The high angiogenic activity in very early breast cancer enables reliable imaging with VEGFR2-targeted microbubbles (BR55).

OBJECTIVES: Tumour xenografts of well-discernible sizes can be examined well by molecular ultrasound. Here, we investigated whether very early breast carcinomas express sufficient levels of VEGFR2 for reliable molecular ultrasound imaging with targeted microbubbles. METHODS: MCF-7 breast cancer xenografts were orthotopically implanted in nude mice (n = 26). Tumours measuring from 4 mm(3) (2 mm diameter) up to 65 mm(3) (5 mm diameter) were examined with automated 3D molecular ultrasound using clinically translatable VEGFR2-targeted microbubbles (BR55). Additionally, the relative tumour blood volume was assessed with non-targeted microbubbles (BR38). In vivo ultrasound data were validated by quantitative immunohistochemistry. RESULTS: Very small lesions 2 mm in diameter showed the highest binding of VEGFR2-specific microbubbles. In larger tumours significantly less BR55 accumulated (p = 0.023). Nonetheless, binding of VEGFR2-targeted microbubbles was still high enough for imaging. The relative blood volume was comparable at all tumour sizes. Both findings were confirmed by immunohistochemistry. Additionally, a significantly enhanced number of large and mature vessels were detected with increasing tumour size (p < 0.01), explaining the decrease in VEGFR2 expression during tumour growth. CONCLUSIONS: 3D molecular ultrasound using BR55 is very well suited to depicting the angiogenic activity in very small breast lesions, suggesting its potential for detecting and characterising these lesions.

Development and validation of an intrinsic landmark-based gating protocol applicable for functional and molecular ultrasound imaging.

OBJECTIVES: To implement a retrospective intrinsic landmark-based (ILB) gating protocol for contrast-enhanced ultrasound (CEUS) and to compare its efficiency to non-gated, manually gated and extrinsically gated CEUS. METHODS: CEUS of the liver was performed in healthy mice (n = 5) and in NEMO knockout mice with dysplastic livers (n = 5). In healthy animals, first-pass kinetics of non-specific microbubbles was recorded. Knockout mice were analysed regarding retention of VEGFR2-specific microbubbles. For retrospective gating, a landmark which showed respiratory movement was encircled as a region of interest (ROI). During inspiration, the signal intensity within the ROI altered, which served as gating signal. To evaluate the accuracy, non-gated, extrinsically gated and ILB-gated time-intensity curves were created. For each curve, descriptive parameters were calculated and compared to the gold standard (manual frame-by-frame gating). RESULTS: No significant differences in the variation of ILB- and extrinsically gated time-intensity curves from the gold standard were observed. Non-gated data showed significantly higher variations. Also the variation of molecular ultrasound data was significantly lower for ILB-gated compared to non-gated data. CONCLUSION: ILB gating is a robust and easy method to improve data accuracy in functional and molecular ultrasound liver imaging. This technique can presumably be translated to contrast-enhanced ultrasound examinations in humans. KEY POINTS: • Quantitative analysis of the uptake of contrast agents during ultrasound is complex. • Intrinsic landmark-based gating (ILB) offers a simple implementable method for motion correction. • Results using ILB-gating are comparable to extrinsic gating using external biomonitoring devices. • Functional and molecular imaging of mobile organs will benefit from ILB gating.

Platelet-derived growth factor-B normalizes micromorphology and vessel function in vascular endothelial growth factor-A-induced squamous cell carcinomas.

Vascular endothelial growth factor (VEGF), which is a key regulator of angiogenesis, often induces formation of immature vessels with increased permeability and reduced vessel functionality. Here, we demonstrate that de novo expression of murine (m)VEGF-164 induces malignant and invasive tumor growth of HaCaT keratinocytes. However, the mVEGF-164-induced tumors are ulcerated with a disorganized epithelium that is interrupted by lacunae with limited basement membrane and endothelial cell coverage. Vessel maturation is strongly impaired. Tumor and vessel micromorphology are markedly improved by the combined expression of human platelet-derived growth factor (hPDGF)-B and mVEGF-164. Although tumor size and malignancy are comparable with either mVEGF-164 alone or combined human PDGF-B and mVEGF-164 expression, combined hPDGF-B and mVEGF-164 expression leads to a more solid and compact tumor tissue with a mature functional tumor vasculature and a higher microvessel density, as demonstrated histologically and by dynamic contrast-enhanced magnetic resonance imaging. Treatment of the hPDGF-B- and mVEGF-164-expressing tumors with imatinib mesylate to block PDGF-B signaling reverses this effect. In addition, tumor cell invasion of mVEGF-164 transfectants and mVEGF-164 plus hPDGF-B transfectants in vivo is associated with a marked induction of tumor-derived matrix metalloproteinase-1 and stromal matrix metalloproteinase-9 and -13, as was confirmed in three-dimensional organotypic co-cultures with fibroblasts in vitro. These data clearly demonstrate the need for a concerted action of different growth factors in the establishment of solid tumors with functional vasculature and emphasize the need for a multifactorial therapy.

Molecular and functional ultrasound imaging in differently aggressive breast cancer xenografts using two novel ultrasound contrast agents (BR55 and BR38).

OBJECTIVES: To characterise clinically translatable long-circulating (BR38) and VEGFR2-targeted (BR55) microbubbles (MB) and to assess their ability to discriminate breast cancer models with different aggressiveness. METHODS: The circulation characteristics of BR38 and BR55 were investigated in healthy mice. The relative blood volume (rBV) of MDA-MB-231 (n = 5) or MCF-7 (n = 6) tumours was determined using BR38. In the same tumours in-vivo binding specificity of BR55 was tested and VEGFR2 expression assessed. Data validation included quantitative immunohistological analysis. RESULTS: BR38 had a longer blood half-life than BR55 (>600 s vs. 218 s). BR38-enhanced ultrasound showed greater vascularisation in MDA-MB-231 tumours (p = 0.022), which was in line with immunohistology (p = 0.033). In-vivo competitive binding experiments proved the specificity of BR55 to VEGFR2 (p = 0.027). Binding of BR55 was significantly higher in MDA-MB-231 than in MCF-7 tumours (p = 0.049), which corresponded with the VEGFR2 levels found histologically (p = 0.015). However, differences became smaller when normalising the levels of BR55 to the rBV. CONCLUSIONS: BR38 and BR55 are well suited to characterising and distinguishing breast cancers with different angiogenesis and aggressiveness. Long-circulating BR38 MB allow extensive 3-dimensional examinations of larger or several organs. BR55 accumulation faithfully reflects the VEGFR2 status in tumours and depicts even small differences in angiogenesis.

Molecular ultrasound imaging and its potential for paediatric radiology.

US is a low-cost, real-time imaging modality that is the most used diagnostic tool in paediatric radiology. Reasons include the improved US image quality in children as compared to adults and the demand for avoiding X-rays as much as possible because children are more sensitive to radiation than adults. Stabilized microbubbles have been approved as US contrast agents for adults and show great potential in improving the diagnostic accuracy for many diseases. Initial studies show that in paediatric radiology contrast-enhanced US could also be beneficial for more than just the diagnosis of vesicoureteral reflux, if US contrast agents were approved for children. Molecular US imaging utilizes microbubbles conjugated to biomolecules that target intravascular disease-specific molecules. Many preclinical studies show that molecular US imaging is a highly sensitive tool to detect neovascularisation, inflammation and cardiovascular diseases. Its main advantages are the higher informative value, the longer persistence of the label at the target lesion and the chance to work with lower contrast agent dosages. Now, clinical translation of molecular US appears at the horizon. This review article reports on the current status of molecular US imaging and discusses its potential for paediatric radiology.

Recent advances in molecular, multimodal and theranostic ultrasound imaging.

Ultrasound (US) imaging is an exquisite tool for the non-invasive and real-time diagnosis of many different diseases. In this context, US contrast agents can improve lesion delineation, characterization and therapy response evaluation. US contrast agents are usually micrometer-sized gas bubbles, stabilized with soft or hard shells. By conjugating antibodies to the microbubble (MB) surface, and by incorporating diagnostic agents, drugs or nucleic acids into or onto the MB shell, molecular, multimodal and theranostic MBs can be generated. We here summarize recent advances in molecular, multimodal and theranostic US imaging, and introduce concepts how such advanced MB can be generated, applied and imaged. Examples are given for their use to image and treat oncological, cardiovascular and neurological diseases. Furthermore, we discuss for which therapeutic entities incorporation into (or conjugation to) MB is meaningful, and how US-mediated MB destruction can increase their extravasation, penetration, internalization and efficacy.

Influence of repetitive contrast agent injections on functional and molecular ultrasound measurements.

Quantitative contrast-enhanced ultrasound plays an important role in tumor characterization and treatment assessment. Besides established functional ultrasound techniques, ultrasound molecular imaging using microbubbles targeted to disease-associated markers is increasingly being applied in pre-clinical studies. Often, repeated injections of non-targeted or targeted microbubbles during the same imaging session are administered. However, the influence of repeated injections on the accuracy of the quantitative data is unclear. Therefore, in tumor-bearing mice, we investigated the influence of multiple injections of non-targeted microbubbles (SonoVue) on time to peak and peak enhancement in liver and tumor tissue and of vascular endothelial growth factor receptor 2 (VEGFR2)-targeted contrast agents (MicroMarker) on specific tumor accumulation. We found significantly decreasing values for time to peak and a tendency for increased values for peak enhancement after multiple injections. Repeated injections of VEGFR2-targeted microbubbles led to significantly increased tumor accumulation, which may result from the exposure of additional binding sites at endothelial surfaces caused by mechanical forces from destroyed microbubbles.

Accuracy of a clinical PET/CT vs. a preclinical µPET system for monitoring treatment effects in tumour xenografts.

PURPOSE: Small animal imaging is of growing importance for preclinical research and drug development. Tumour xenografts implanted in mice can be visualized with a clinical PET/CT (cPET); however, it is unclear whether early treatment effects can be monitored. Thus, we investigated the accuracy of a cPET versus a preclinical µPET using (18)F-FDG for assessing early treatment effects. MATERIALS AND METHODS: The spatial resolution and the quantitative accuracy of a clinical and preclinical PET were evaluated in phantom experiments. To investigate the sensitivity for assessing treatment response, A431 tumour xenografts were implanted in nude mice. Glucose metabolism was measured in untreated controls and in two therapy groups (either one or four days of antiangiogenic treatment). Data was validated by γ-counting of explanted tissues. RESULTS: In phantom experiments, cPET enabled reliable separation of boreholes≥5mm whereas µPET visualized boreholes≥2mm. In animal studies, µPET provided significantly higher tumour-to-muscle ratios for untreated control tumours than cPET (3.41±0.87 vs. 1.60±.0.28, respectively; p<0.01). During treatment, cPET detected significant therapy effects at day 4 (p<0.05) whereas µPET revealed highly significant therapy effects even at day one (p<0.01). Correspondingly, γ-counting of explanted tumours indicated significant therapy effects at day one and highly significant treatment response at day 4. Correlation with γ-counting was good for cPET (r=0.74; p<0.01) and excellent for µPET (r=0.85; p<0.01). CONCLUSION: Clinical PET is suited to investigate tumour xenografts≥5mm at an advanced time-point of treatment. For imaging smaller tumours or for the sensitive assessment of very early therapy effects, µPET should be preferred.

Targeted ultrasound imaging of cancer: an emerging technology on its way to clinics.

Ultrasound is one of the workhorses in clinical cancer diagnosis. In particular, it is routinely used to characterize lesions in liver, urogenital tract, head and neck and soft tissues. During the last years image quality steadily improved, which, among others, can be attributed to the development of harmonic image analysis. Microbubbles were introduced as intravascular contrast agents and can be detected with superb sensitivity and specificity using contrast specific imaging modes. By aid of these unspecific contrast agents tissues can be characterised regarding their vascularity. Antibodies, peptides and other targeting moieties were bound to microbubbles to target sites of angiogenesis and inflammation intending to get more disease-specific information. Indeed, many preclinical studies proved the high potential of targeted ultrasound imaging to better characterize tumors and to more sensitively monitor therapy response. Recently, first targeted microbubbles had been developed that meet the pharmacological demands of a clinical contrast agent. This review articles gives an overview on the history and current status of targeted ultrasound imaging of cancer. Different imaging concepts and contrast agent designs are introduced ranging from the use of experimental nanodroplets to agents undergoing clinical evaluation. Although it is clear that targeted ultrasound imaging works reliably, its broad acceptance is hindered by the user dependency of ultrasound imaging in general. Automated 3D-scanning techniques-like being used for breast diagnosis - and novel 3D transducers will help to make this fascinating method clinical reality.

Evaluation of high frequency ultrasound methods and contrast agents for characterising tumor response to anti-angiogenic treatment.

PURPOSE: To compare non-enhanced and contrast-enhanced high-frequency 3D Doppler ultrasound with contrast-enhanced 2D and 3D B-mode imaging for assessing tumor vascularity during antiangiogenic treatment using soft-shell and hard-shell microbubbles. MATERIALS AND METHODS: Antiangiogenic therapy effects (SU11248) on vascularity of subcutaneous epidermoid-carcinoma xenografts (A431) in female CD1 nude mice were investigated longitudinally using non-enhanced and contrast-enhanced 3D Doppler at 25 MHz. Additionally, contrast-enhanced 2D and 3D B-mode scans were performed by injecting hard-shell (poly-butyl-cyanoacrylate-based) and soft-shell (phospholipid-based) microbubbles. Suitability of both contrast agents for high frequency imaging and the sensitivity of the different ultrasound methods to assess early antiangiogenic therapy effects were investigated. Ultrasound data were validated by immunohistology. RESULTS: Hard-shell microbubbles induced higher signal intensity changes in tumors than soft-shell microbubbles in 2D B-mode measurements (424 ± 7 vs. 169 ± 8 A.U.; p<0.01). In 3D measurements, signals of soft-shell microbubbles were hardly above the background (5.48 ± 4.57 vs. 3.86 ± 2.92 A.U.), while signals from hard-shell microbubbles were sufficiently high (30.5 ± 8.06 A.U). Using hard-shell microbubbles 2D and 3D B-mode imaging depicted a significant decrease in tumor vascularity during antiangiogenic therapy from day 1 on. Using soft-shell microbubbles significant therapy effects were observed at day 4 after therapy in 2D B-mode imaging but could not be detected in the 3D mode. With non-enhanced and contrast-enhanced Doppler imaging significant differences between treated and untreated tumors were found from day 2 on. CONCLUSION: Hard-shell microbubble-enhanced 2D and 3D B-mode ultrasound achieved highest sensitivity for assessing therapy effects on tumor vascularisation and were superior to B-mode ultrasound with soft-shell microbubbles and to Doppler imaging.

Comparison of conventional time-intensity curves vs. maximum intensity over time for post-processing of dynamic contrast-enhanced ultrasound.

Our aim was to prospectively compare two post-processing techniques for dynamic contrast-enhanced ultrasound and to evaluate their impact for monitoring antiangiogenic therapy. Thus, mice with epidermoid carcinoma xenografts were examined during administration of polybutylcyanoacrylate-microbubbles using a small animal ultrasound system (40 MHz). Cine loops were acquired and analyzed using time-intensity (TI) and maximum intensity over time (MIOT) curves. Influences of fast (50 microl/2s) vs. slow (50 microl/10s) injection of microbubbles on both types of curves were investigated. Sensitivities of both methods for assessing effects of antiangiogenic treatment (SU11248) were examined. Correlative histological analysis was performed for vessel-density. Mann-Whitney test was used for statistical analysis. Microbubble injection rates significantly influenced upslope, time-to-peak and peak enhancement of conventional TI curves (p<0.05) but had almost no impact on maximum enhancement of MIOT curves (representing relative blood volume). Additionally, maximum enhancement of MIOT curves captured antiangiogenic therapy effects more reliably and earlier (already after 1 day of therapy; p<0.05) than peak enhancement of TI curves. Immunohistochemistry validated the significantly (p<0.01) lower vessel densities in treated tumors and high correlation (R(2)=0.95) between vessel-density and maximum enhancement of MIOT curves was observed. In conclusion, MIOT is less susceptible to variations of the injection's speed. It enables to assess changes of the relative blood volume earlier and with lower standard deviations than conventional TI curves. It can easily be translated into clinical practice and thus may provide a promising tool for cancer therapy monitoring.