Evaluating different combination methods to analyse ultrasound and shear wave elastography images automatically through discriminative convolutional neural network in breast cancer imaging.

PURPOSE: Ultrasound (US) and Shear Wave Elastography (SWE) imaging are non-invasive methods used for breast lesion characterization. While US and SWE images provide both morphological information, SWE visualizes in addition the elasticity of tissue. In this study a Discriminative Convolutional Neural Network (DCNN) model is applied to US and SWE images and their combination to classify the breast lesions into malignant or benign cases. Furthermore, it is identified whether analysing only the region of the elastogram or including the surrounding B-mode image gives a superior performance. METHODS: The dataset used in this study consists of 746 images obtained from 207 patients comprising 486 malignant and 260 benign breast lesions. From each image the US and SWE image was extracted, once including only the region of the elastogram and once including also the surrounding B-mode image. These four datasets were applied individually to a DCNN to determine their predictive capability. Each the best US and SWE dataset were used to examine different combination methods with DCNN. The results were compared to the manual assessment by an expert radiologist. RESULTS: The combination of US and SWE images with the surrounding B-mode image using two ensembled DCNN models achieved best results with an accuracy of 93.53 %, sensitivity of 94.42 %, specificity of 90.75 % and area under the curve (AUC) of 96.55 %. CONCLUSION: This study showed that using the whole US and SWE images through DCNN was superior to methods, in which only the region of elastogram was used. Combining breast cancer US and SWE images with two ensembled DCNN models in parallel improved the results. The accuracy, sensitivity and AUC of the best combination method were significantly superior to the results of using a single dataset through DCNN and to the results of the expert radiologist.

Introduction of a Measurement Setup to Monitor the Pressure Applied During Handheld Ultrasound Elastography.

Shear-wave elastography may produce misleadingly high values if too much pressure is applied during the imaging process. However, in clinical routine there is presently no way to monitor the pressure applied during the measurements. In this work we introduce a novel measurement setup which can directly be attached to an ultrasonic imaging transducer and allows observation of the applied pressure in real time. The setup supports free-hand imaging according to the clinical standard. We tested the setup by carrying out shear-wave elastography under varying pressures on ex vivo animal tissue. The values increased with pressure, as was expected. Thus, the setup is a possible solution for measuring applied pressure in real time.

Wound healing in rabbit corneas after flapless refractive lenticule extraction with a 345 nm ultraviolet femtosecond laser.

PURPOSE: To characterize corneal wound healing in a rabbit model after flapless refractive lenticule extraction with a 345 nm ultraviolet femtosecond laser. SETTING: Departments of Ophthalmology and Anatomy II, University of Erlangen-Nürnberg and Wavelight GmbH, Erlangen, Germany. DESIGN: Experimental study. METHODS: Flapless refractive lenticule extraction was performed in 1 eye each of 20 New Zealand white rabbits (-5.0 diopters). Groups of 4 animals were euthanized after 48 hours, 1 week, 2 weeks, 4 weeks, and 3 months, respectively. Corneal samples were prepared for histology and fluorescence microscopy. To assess corneal cell death, proliferation, and myofibroblastic transdifferentiation, terminal uridine deoxynucleotidyl nick end-labeling (TUNEL) assay as well as immunostaining for Ki67 and α-smooth muscle actin (αSMA) were performed on sagittal cryosections. RESULTS: Histology revealed a zone of keratocyte depletion with a thickness of approximately 50 µm around the extraction site. At 48 hours, pronounced TUNEL staining of keratocytes was detected around the interface (159.9 cells/mm ±â€¯18.4 [SD]), which steadily decreased to 74.9 ±â€¯19.8 cells/mm at 1 week and 5.7 ±â€¯4.8 cells/mm at 2 weeks. Ki67 staining of keratocytes was evident at 48 hours (10.0 ±â€¯3.8 cells/mm), which then decreased at 1 week (5.2 ±â€¯1.7 cells/mm) and 2 weeks (0.4 ±â€¯0.5 cells/mm). From 4 weeks onward, no TUNEL or Ki67 staining was detected. The corneal stroma was αSMA-negative at all timepoints. CONCLUSION: Application of the 345 nm laser showed no signs of problematic repair processes in the cornea, which supports the initiation of the clinical phase.

First step to facilitate long-term and multi-centre studies of shear wave elastography in solid breast lesions using a computer-assisted algorithm.

PURPOSE: Shear wave elastography (SWE) visualises the elasticity of tissue. As malignant tissue is generally stiffer than benign tissue, SWE is helpful to diagnose solid breast lesions. Until now, quantitative measurements of elasticity parameters have been possible only, while the images were still saved on the ultrasound imaging device. This work aims to overcome this issue and introduces an algorithm allowing fast offline evaluation of SWE images. METHODS: The algorithm was applied to a commercial phantom comprising three lesions of various elasticities and 207 in vivo solid breast lesions. All images were saved in DICOM, JPG and QDE (quantitative data export; for research only) format and evaluated according to our clinical routine using a computer-aided diagnosis algorithm. The results were compared to the manual evaluation (experienced radiologist and trained engineer) regarding their numerical discrepancies and their diagnostic performance using ROC and ICC analysis. RESULTS: ICCs of the elasticity parameters in all formats were nearly perfect (0.861-0.990). AUC for all formats was nearly identical for [Formula: see text] and [Formula: see text] (0.863-0.888). The diagnostic performance of SD using DICOM or JPG estimations was lower than the manual or QDE estimation (AUC 0.673 vs. 0.844). CONCLUSIONS: The algorithm introduced in this study is suitable for the estimation of the elasticity parameters offline from the ultrasound system to include images taken at different times and sites. This facilitates the performance of long-term and multi-centre studies.

Anisotropy of Solid Breast Lesions in 2D Shear Wave Elastography is an Indicator of Malignancy.

RATIONALE AND OBJECTIVES: To investigate if anisotropy at two-dimensional shear wave elastography (SWE) suggests malignancy and whether it correlates with prognostic and predictive factors in breast cancer. MATERIALS AND METHODS: Study group A of 244 solid breast lesions was imaged with SWE between April 2013 and May 2014. Each lesion was imaged in radial and in antiradial planes, and the maximum elasticity, mean elasticity, and standard deviation were recorded and correlated with benign/malignant status, and if malignant, correlated with conventional predictive and prognostic factors. The results were compared to a study group B of 968 solid breast lesions, which were imaged in sagittal and in axial planes between 2010 and 2013. RESULTS: Neither benign nor malignant lesion anisotropy is plane dependent. However, malignant lesions are more anisotropic than benign lesions (P ≤ 0.001). Anisotropy correlates with increasing elasticity parameters, breast imaging-reporting and data system categories, core biopsy result, and tumor grade. Large cancers are significantly more anisotropic than small cancers (P ≤ 0.001). The optimal anisotropy cutoff threshold for benign/malignant differentiation of 150 kPa(2) achieves the best sensitivity (74%) with a reasonable specificity (63%). CONCLUSIONS: Anisotropy may be useful during benign/malignant differentiation of solid breast masses using SWE. Anisotropy also correlates with some prognostic factors in breast cancer.

Evaluation of a 345 nm Femtosecond Laser for Corneal Surgery with Respect to Intraocular Radiation Hazard.

PURPOSE: We report our findings from a preclinical safety study designed to assess potential side effects of corneal ultraviolet femtosecond laser treatment on lens and retina. METHODS: Refractive lenticules (-5 dpt) with a diameter of 6 mm were created in the right cornea of eight Dutch Belted rabbits. Radiant exposure was 0.5 J/cm² in two animals and 18 J/cm² in six animals. The presence of lens opacities was assessed prior to and up to six months following laser application using Scheimpflug images (Pentacam, Oculus) and backscatter analysis (Opacity Lensmeter 702, Interzeag). Ganzfeld flash and flicker electroretinogram (ERG) recordings were obtained from both eyes prior to and up to six weeks following laser application. At the study endpoint, retinas were examined by light microscopy. RESULTS: Independent of energy dose applied, no cataract formation could be observed clinically or with either of the two objective methods used. No changes in ERG recordings over time and no difference between treated and untreated eye were detected. Histologically, retinal morphology was preserved and retinal pigment epithelium as well as photoreceptor inner and outer segments appeared undamaged. Quantitative digital image analysis did not reveal cell loss in inner or outer nuclear layers. CONCLUSIONS: Our analysis confirms theoretical considerations suggesting that ultraviolet femtosecond laser treatment of the cornea is safe for intraocular tissues. Transmitted light including stray light induces no photochemical effects in lens or retina at energy levels much higher than required for the clinical purpose. These conclusions cannot be applied to eyes with pre-existing retinal damage, as these may be more vulnerable to light.

Corneal tissue interactions of a new 345 nm ultraviolet femtosecond laser.

PURPOSE: To assess the suitability of a new 345 nm ultraviolet (UV) femtosecond laser for refractive surgery. SETTING: Department of Ophthalmology, University of Erlangen-Nürnberg, Erlangen, Germany. DESIGN: Experimental study. METHODS: Twenty-five porcine corneas were used for stromal flap or lamellar bed creation (stromal depth, 150 µm) and 15 rabbit corneas for lamellar bed creation near the endothelium. Ultraviolet femtosecond laser cutting-line morphology, gas formation, and keratocyte death rate were evaluated using light and electron microscopy and compared with a standard infrared (IR) femtosecond laser. Endothelial cell survival was examined after application of a laser cut near the endothelium. RESULTS: Flaps created by the UV laser were lifted easily. Gas formation was reduced 4.2-fold compared with the IR laser (P = .001). The keratocyte death rate near the interface was almost doubled; however, the death zone was confined to a region within 38 µm ± 10 (SD) along the cutting line. Histologically and ultrastructurally, a distinct and continuous cutting line was not found after UV femtosecond laser application if flap lifting was omitted and standard energy parameters were used. Instead, a regular pattern of vertical striations, presumably representing self-focusing induced regions of optical tissue breakdown, were identified. Lamellar bed creation with standard energy parameters 50 µm from the endothelium rendered the endothelial cells intact and viable. CONCLUSION: The new 345 nm femtosecond laser is a candidate for pending in vivo trials and future high-precision flap creation, intrastromal lenticule extraction, and ultrathin Descemet-stripping endothelial keratoplasty. FINANCIAL DISCLOSURES: Mr. Klenke and Ms. Skerl were paid employees of Wavelight GmbH when the study was performed. Dr. Seiler is a scientific consultant to Wavelight GmbH. No other author has a financial or proprietary interest in any material or method mentioned.

Biomechanical efficacy of collagen crosslinking in porcine cornea using a femtosecond laser pocket.

PURPOSE: The aim of this study was to assess the biomechanical efficacy of transepithelial collagen crosslinking using the femtosecond laser pocket technique compared with that using the standard crosslinking (CXL) technique. METHODS: Forty ex vivo porcine eyes were divided into 4 groups with 10 samples each. Group 1 comprised the untreated controls. Group 2 was the standard CXL group with debridement, instillation of 0.1% riboflavin-dextran solution for 15 minutes before and every 5 minutes during the 30 minutes of irradiation with ultraviolet A (UVA) light of 370 nm and an irradiance of 3 mW/cm². Group 3 pertained to the femtolaser pocket control with an intrastromal pocket but without riboflavin/UVA. Group 4 was the femtolaser pocket CXL group with an intrastromal pocket of an 8-mm diameter at a 180-µm depth, riboflavin/dextran application for 15 minutes and subsequent exposure to UVA light for 30 minutes. Postoperatively, biomechanical stress-strain measurements were performed. RESULTS: In the standard CXL group, the stress at 10% strain was 207.8 ± 64.1 × 10 Pa (+79.45% vs. controls; P = 0.021) compared with 115.8 ± 20.8 × 10 Pa in the untreated control group; in the crosslinked femtolaser pocket group, it was 159.5 ± 30.4 × 10 Pa (+37.74%; P = 0.049), in the non-cross-linked femtolaser pocket group, it was 103.5 ± 17.3 × 10 Pa (-10.62%; P = 0.103). The Young modulus was 5.4 MPa (+100% vs. controls) in the standard CXL group, 3.7 MPa (+37.04%) in the crosslinked femtolaser pocket group, and 2.4 MPa (-11.12%) in the non-cross-linked femtolaser pocket group compared with 2.7 MPa in the untreated control group. CONCLUSIONS: The biomechanical effect of CXL using the femtolaser pocket technique is about 50% less pronounced than that after standard CXL. Future studies will show whether the efficacy of the technique can still be improved and whether the clinical effect is sufficient for stabilizing ectatic corneas.