Structurally-constrained optical-flow-guided adversarial generation of synthetic CT for MR-only radiotherapy treatment planning.

The rapid progress in image-to-image translation methods using deep neural networks has led to advancements in the generation of synthetic CT (sCT) in MR-only radiotherapy workflow. Replacement of CT with MR reduces unnecessary radiation exposure, financial cost and enables more accurate delineation of organs at risk. Previous generative adversarial networks (GANs) have been oriented towards MR to sCT generation. In this work, we have implemented multiple augmented cycle consistent GANs. The augmentation involves structural information constraint (StructCGAN), optical flow consistency constraint (FlowCGAN) and the combination of both the conditions (SFCGAN). The networks were trained and tested on a publicly available Gold Atlas project dataset, consisting of T2-weighted MR and CT volumes of 19 subjects from 3 different sites. The network was tested on 8 volumes acquired from the third site with a different scanner to assess the generalizability of the network on multicenter data. The results indicate that all the networks are robust to scanner variations. The best model, SFCGAN achieved an average ME of 0.9   5.9 HU, an average MAE of 40.4   4.7 HU and 57.2   1.4 dB PSNR outperforming previous research works. Moreover, the optical flow constraint between consecutive frames preserves the consistency across all views compared to 2D image-to-image translation methods. SFCGAN exploits the features of both StructCGAN and FlowCGAN by delivering structurally robust and 3D consistent sCT images. The research work serves as a benchmark for further research in MR-only radiotherapy.

Artificial intelligence for teleophthalmology-based diabetic retinopathy screening in a national programme: an economic analysis modelling study.

BACKGROUND: Deep learning is a novel machine learning technique that has been shown to be as effective as human graders in detecting diabetic retinopathy from fundus photographs. We used a cost-minimisation analysis to evaluate the potential savings of two deep learning approaches as compared with the current human assessment: a semi-automated deep learning model as a triage filter before secondary human assessment; and a fully automated deep learning model without human assessment. METHODS: In this economic analysis modelling study, using 39 006 consecutive patients with diabetes in a national diabetic retinopathy screening programme in Singapore in 2015, we used a decision tree model and TreeAge Pro to compare the actual cost of screening this cohort with human graders against the simulated cost for semi-automated and fully automated screening models. Model parameters included diabetic retinopathy prevalence rates, diabetic retinopathy screening costs under each screening model, cost of medical consultation, and diagnostic performance (ie, sensitivity and specificity). The primary outcome was total cost for each screening model. Deterministic sensitivity analyses were done to gauge the sensitivity of the results to key model assumptions. FINDINGS: From the health system perspective, the semi-automated screening model was the least expensive of the three models, at US$62 per patient per year. The fully automated model was $66 per patient per year, and the human assessment model was $77 per patient per year. The savings to the Singapore health system associated with switching to the semi-automated model are estimated to be $489 000, which is roughly 20% of the current annual screening cost. By 2050, Singapore is projected to have 1 million people with diabetes; at this time, the estimated annual savings would be $15 million. INTERPRETATION: This study provides a strong economic rationale for using deep learning systems as an assistive tool to screen for diabetic retinopathy. FUNDING: Ministry of Health, Singapore.

Recent advancement in the early detection of melanoma using computerized tools: An image analysis perspective.

BACKGROUND: The paper reviews the advancement of tools and current technologies for the detection of melanoma. We discussed several computational strategies from pre- to postprocessing image operations, descriptors, and popular classifiers to diagnose a suspected skin lesion based on its virtual similarity to the malignant lesion with known histopathology. We reviewed the current state of smart phone-based apps as diagnostic tools for screening. METHODS: A literature survey was conducted using a combination of keywords in the bibliographic databases: PubMed, AJCC, PH2, EDRA, and ISIC melanoma project. A number of melanoma detection apps were downloaded for two major mobile operating systems, iOS and Android; their important uses, key challenges, and various expert opinions were evaluated and also discussed. RESULTS: We have provided an overview of research on the computer-aided diagnosis methods to estimate melanoma risk and early screening. Dermoscopic images are the most viable option for the advent of new image processing technologies based on which many of the skin cancer detection apps are being developed recently. We have categorized and explored their potential uses, evaluation criteria, limitations, and other details. CONCLUSION: Such advancements are helpful in the sense they are raising awareness. Diagnostic accuracy is the major issue of smart phone-based apps and it cannot replace an adequate clinical experience and biopsy procedures.

QuickCount®: a novel automated software for rapid cell detection and quantification.

We describe a novel automated cell detection and counting software, QuickCount® (QC), designed for rapid quantification of cells. The Bland-Altman plot and intraclass correlation coefficient (ICC) analyses demonstrated strong agreement between cell counts from QC to manual counts (mean and SD: -3.3 ± 4.5; ICC = 0.95). QC has higher recall in comparison to ImageJauto, CellProfiler and CellC and the precision of QC, ImageJauto, CellProfiler and CellC are high and comparable. QC can precisely delineate and count single cells from images of different cell densities with precision and recall above 0.9. QC is unique as it is equipped with real-time preview while optimizing the parameters for accurate cell count and needs minimum hands-on time where hundreds of images can be analyzed automatically in a matter of milliseconds. In conclusion, QC offers a rapid, accurate and versatile solution for large-scale cell quantification and addresses the challenges often faced in cell biology research.

Quantitative nuclear histomorphometry predicts oncotype DX risk categories for early stage ER+ breast cancer.

BACKGROUND: Gene-expression companion diagnostic tests, such as the Oncotype DX test, assess the risk of early stage Estrogen receptor (ER) positive (+) breast cancers, and guide clinicians in the decision of whether or not to use chemotherapy. However, these tests are typically expensive, time consuming, and tissue-destructive. METHODS: In this paper, we evaluate the ability of computer-extracted nuclear morphology features from routine hematoxylin and eosin (H&E) stained images of 178 early stage ER+ breast cancer patients to predict corresponding risk categories derived using the Oncotype DX test. A total of 216 features corresponding to the nuclear shape and architecture categories from each of the pathologic images were extracted and four feature selection schemes: Ranksum, Principal Component Analysis with Variable Importance on Projection (PCA-VIP), Maximum-Relevance, Minimum Redundancy Mutual Information Difference (MRMR MID), and Maximum-Relevance, Minimum Redundancy - Mutual Information Quotient (MRMR MIQ), were employed to identify the most discriminating features. These features were employed to train 4 machine learning classifiers: Random Forest, Neural Network, Support Vector Machine, and Linear Discriminant Analysis, via 3-fold cross validation. RESULTS: The four sets of risk categories, and the top Area Under the receiver operating characteristic Curve (AUC) machine classifier performances were: 1) Low ODx and Low mBR grade vs. High ODx and High mBR grade (Low-Low vs. High-High) (AUC = 0.83), 2) Low ODx vs. High ODx (AUC = 0.72), 3) Low ODx vs. Intermediate and High ODx (AUC = 0.58), and 4) Low and Intermediate ODx vs. High ODx (AUC = 0.65). Trained models were tested independent validation set of 53 cases which comprised of Low and High ODx risk, and demonstrated per-patient accuracies ranging from 75 to 86%. CONCLUSION: Our results suggest that computerized image analysis of digitized H&E pathology images of early stage ER+ breast cancer might be able predict the corresponding Oncotype DX risk categories.

A low-cost multimodal head-mounted display system for neuroendoscopic surgery.

Background: With rapid advances in technology, wearable devices as head-mounted display (HMD) have been adopted for various uses in medical science, ranging from simply aiding in fitness to assisting surgery. We aimed to investigate the feasibility and practicability of a low-cost multimodal HMD system in neuroendoscopic surgery. Methods: A multimodal HMD system, mainly consisted of a HMD with two built-in displays, an action camera, and a laptop computer displaying reconstructed medical images, was developed to assist neuroendoscopic surgery. With this intensively integrated system, the neurosurgeon could freely switch between endoscopic image, three-dimensional (3D) reconstructed virtual endoscopy images, and surrounding environment images. Using a leap motion controller, the neurosurgeon could adjust or rotate the 3D virtual endoscopic images at a distance to better understand the positional relation between lesions and normal tissues at will. Results: A total of 21 consecutive patients with ventricular system diseases underwent neuroendoscopic surgery with the aid of this system. All operations were accomplished successfully, and no system-related complications occurred. The HMD was comfortable to wear and easy to operate. Screen resolution of the HMD was high enough for the neurosurgeon to operate carefully. With the system, the neurosurgeon might get a better comprehension on lesions by freely switching among images of different modalities. The system had a steep learning curve, which meant a quick increment of skill with it. Compared with commercially available surgical assistant instruments, this system was relatively low-cost. Conclusions: The multimodal HMD system is feasible, practical, helpful, and relatively cost efficient in neuroendoscopic surgery.

An observational study to assess if automated diabetic retinopathy image assessment software can replace one or more steps of manual imaging grading and to determine their cost-effectiveness.

BACKGROUND: Diabetic retinopathy screening in England involves labour-intensive manual grading of retinal images. Automated retinal image analysis systems (ARIASs) may offer an alternative to manual grading. OBJECTIVES: To determine the screening performance and cost-effectiveness of ARIASs to replace level 1 human graders or pre-screen with ARIASs in the NHS diabetic eye screening programme (DESP). To examine technical issues associated with implementation. DESIGN: Observational retrospective measurement comparison study with a real-time evaluation of technical issues and a decision-analytic model to evaluate cost-effectiveness. SETTING: A NHS DESP. PARTICIPANTS: Consecutive diabetic patients who attended a routine annual NHS DESP visit. INTERVENTIONS: Retinal images were manually graded and processed by three ARIASs: iGradingM (version 1.1; originally Medalytix Group Ltd, Manchester, UK, but purchased by Digital Healthcare, Cambridge, UK, at the initiation of the study, purchased in turn by EMIS Health, Leeds, UK, after conclusion of the study), Retmarker (version 0.8.2, Retmarker Ltd, Coimbra, Portugal) and EyeArt (Eyenuk Inc., Woodland Hills, CA, USA). The final manual grade was used as the reference standard. Arbitration on a subset of discrepancies between manual grading and the use of an ARIAS by a reading centre masked to all grading was used to create a reference standard manual grade modified by arbitration. MAIN OUTCOME MEASURES: Screening performance (sensitivity, specificity, false-positive rate and likelihood ratios) and diagnostic accuracy [95% confidence intervals (CIs)] of ARIASs. A secondary analysis explored the influence of camera type and patients' ethnicity, age and sex on screening performance. Economic analysis estimated the cost per appropriate screening outcome identified. RESULTS: A total of 20,258 patients with 102,856 images were entered into the study. The sensitivity point estimates of the ARIASs were as follows: EyeArt 94.7% (95% CI 94.2% to 95.2%) for any retinopathy, 93.8% (95% CI 92.9% to 94.6%) for referable retinopathy and 99.6% (95% CI 97.0% to 99.9%) for proliferative retinopathy; and Retmarker 73.0% (95% CI 72.0% to 74.0%) for any retinopathy, 85.0% (95% CI 83.6% to 86.2%) for referable retinopathy and 97.9% (95% CI 94.9 to 99.1%) for proliferative retinopathy. iGradingM classified all images as either 'disease' or 'ungradable', limiting further iGradingM analysis. The sensitivity and false-positive rates for EyeArt were not affected by ethnicity, sex or camera type but sensitivity declined marginally with increasing patient age. The screening performance of Retmarker appeared to vary with patient's age, ethnicity and camera type. Both EyeArt and Retmarker were cost saving relative to manual grading either as a replacement for level 1 human grading or used prior to level 1 human grading, although the latter was less cost-effective. A threshold analysis testing the highest ARIAS cost per patient before which ARIASs became more expensive per appropriate outcome than human grading, when used to replace level 1 grader, was Retmarker £3.82 and EyeArt £2.71 per patient. LIMITATIONS: The non-randomised study design limited the health economic analysis but the same retinal images were processed by all ARIASs in this measurement comparison study. CONCLUSIONS: Retmarker and EyeArt achieved acceptable sensitivity for referable retinopathy and false-positive rates (compared with human graders as reference standard) and appear to be cost-effective alternatives to a purely manual grading approach. Future work is required to develop technical specifications to optimise deployment and address potential governance issues. FUNDING: The National Institute for Health Research (NIHR) Health Technology Assessment programme, a Fight for Sight Grant (Hirsch grant award) and the Department of Health's NIHR Biomedical Research Centre for Ophthalmology at Moorfields Eye Hospital and the University College London Institute of Ophthalmology.

Fast retrieval of calcification from sequential intravascular ultrasound gray-scale images.

Intravascular ultrasound (IVUS)-based tissue characterization is invaluable for the computer-aided diagnosis and interventional treatment of cardiac vessel diseases. Although the analysis of raw backscattered signals allows more accurate plaque characterization than gray-scale images, its applications are limited due to its nature of electrocardiogram-gated acquisition. Images acquired by IVUS devices that do not allow the acquisition of raw signals cannot be characterized. To address these limitations, we developed a method for fast frame-by-frame retrieval and location of calcification according to the jump features of radial gray-level variation curves from sequential IVUS gray-scale images. The proposed method consists of three main steps: (1) radial gray-level variation curves are extracted from each filtered polar view, (2) sequential images are preliminarily queried according to the maximal slopes of radial gray-level variation curves, and finally, (3) key frames that include calcification are selected through checking the gray-level features of successive pixel columns in the preliminary results. Experimental results with clinically acquired in vivo data sets indicate key frames that include calcification can be retrieved with the advantages of simplicity, high efficiency, and accuracy. Recognition results correlate well with manual characterization results obtained by experienced physicians and through virtual histology.

[Femtosecond laser-assisted cataract surgery]. / Chirurgie de la cataracte assistée par laser.

Femtosecond laser-assisted cataract surgery is a major technological innovation. The femtosecond laser, during a pretreatment step, helps to prepare the patient's eye for the surgery proper by creating corneal incisions, anterior capsulotomy and lens fragmentation in an automated fashion. Thus, these steps can be performed with precision and reproducibility, and lens fragmentation reduces the amount of ultrasound required during surgery. Drawbacks of this technology are a longer operating time, a more demanding surgical procedure and a much higher cost for patients and surgical centers. New models of organization in the operating room, patient flow, and financial systems have to be designed to adapt this procedure to our practice. The benefits of this technology should make it an essential tool in the future, provided that cataract surgery can be reconsidered logistically and economically.

Digital diffraction analysis enables low-cost molecular diagnostics on a smartphone.

The widespread distribution of smartphones, with their integrated sensors and communication capabilities, makes them an ideal platform for point-of-care (POC) diagnosis, especially in resource-limited settings. Molecular diagnostics, however, have been difficult to implement in smartphones. We herein report a diffraction-based approach that enables molecular and cellular diagnostics. The D3 (digital diffraction diagnosis) system uses microbeads to generate unique diffraction patterns which can be acquired by smartphones and processed by a remote server. We applied the D3 platform to screen for precancerous or cancerous cells in cervical specimens and to detect human papillomavirus (HPV) DNA. The D3 assay generated readouts within 45 min and showed excellent agreement with gold-standard pathology or HPV testing, respectively. This approach could have favorable global health applications where medical access is limited or when pathology bottlenecks challenge prompt diagnostic readouts.

Assessing diagnostic complexity: An image feature-based strategy to reduce annotation costs.

Computer-aided diagnosis systems can play an important role in lowering the workload of clinical radiologists and reducing costs by automatically analyzing vast amounts of image data and providing meaningful and timely insights during the decision making process. In this paper, we present strategies on how to better manage the limited time of clinical radiologists in conjunction with predictive model diagnosis. We first introduce a metric for discriminating between the different categories of diagnostic complexity (such as easy versus hard) encountered when interpreting CT scans. Second, we propose to learn the diagnostic complexity using a classification approach based on low-level image features automatically extracted from pixel data. We then show how this classification can be used to decide how to best allocate additional radiologists to interpret a case based on its diagnosis category. Using a lung nodule image dataset, we determined that, by a simple division of cases into hard and easy to diagnose, the number of interpretations can be distributed to significantly lower the cost with limited loss in prediction accuracy. Furthermore, we show that with just a few low-level image features (18% of the original set) we are able to determine the easy from hard cases for a significant subset (66%) of the lung nodule image data.

Current role of cardiac and extra-cardiac pathologies in clinically indicated cardiac computed tomography with emphasis on status before pulmonary vein isolation.

PURPOSE: The aim of this study was to assess the incidence of cardiac and significant extra-cardiac findings in clinical computed tomography of the heart in patients with atrial fibrillation before pulmonary vein isolation (PVI). MATERIALS AND METHODS: 224 patients (64 ±â€Š10 years; male 63 %) with atrial fibrillation were examined by cardiac 64-slice multidetector CT before PVI. Extra-cardiac findings were classified as "significant" if they were recommended to additional diagnostics or therapy, and otherwise as "non-significant". Additionally, cardiac findings were documented in detail. RESULTS: A total of 724 cardiac findings were identified in 203 patients (91 % of patients). Additionally, a total of 619 extra-cardiac findings were identified in 179 patients (80 % of patients). Among these extra-cardiac findings 196 (32 %) were "significant", and 423 (68 %) were "non-significant". In 2 patients (1 %) a previously unknown malignancy was detected (esophageal cancer and lung cancer, local stage, no metastasis). 203 additional imaging diagnostics followed to clarify the "significant" findings (124 additional CT, costs 38 314.69 US dollars). Overall, there were 3.2 cardiac and 2.8 extra-cardiac findings per patient. Extra-cardiac findings appear significantly more frequently in patients over 60 years old, in smokers and in patients with a history of cardiac findings (p <0.05). CONCLUSION: Cardiac CT scans before PVI should be screened for extracardiac incidental findings that could have important clinical implications for each patient.

Computer vision-based analysis of foods: a non-destructive colour measurement tool to monitor quality and safety.

Computer vision-based image analysis has been widely used in food industry to monitor food quality. It allows low-cost and non-contact measurements of colour to be performed. In this paper, two computer vision-based image analysis approaches are discussed to extract mean colour or featured colour information from the digital images of foods. These types of information may be of particular importance as colour indicates certain chemical changes or physical properties in foods. As exemplified here, the mean CIE a* value or browning ratio determined by means of computer vision-based image analysis algorithms can be correlated with acrylamide content of potato chips or cookies. Or, porosity index as an important physical property of breadcrumb can be calculated easily. In this respect, computer vision-based image analysis provides a useful tool for automatic inspection of food products in a manufacturing line, and it can be actively involved in the decision-making process where rapid quality/safety evaluation is needed.

Optical systems for point-of-care diagnostic instrumentation: analysis of imaging performance and cost.

One of the key elements in point-of-care (POC) diagnostic test instrumentation is the optical system required for signal detection and/or imaging. Many tests which use fluorescence, absorbance, or colorimetric optical signals are under development for management of infectious diseases in resource limited settings, where the overall size and cost of the device is of critical importance. At present, high-performance lenses are expensive to fabricate and difficult to obtain commercially, presenting barriers for developers of in vitro POC tests or microscopic image-based diagnostics. We recently described a compact "hybrid" objective lens incorporating both glass and plastic optical elements, with a numerical aperture of 1.0 and field-of-view of 250 µm. This design concept may potentially enable mass-production of high-performance, low-cost optical systems which can be easily incorporated in the readout path of existing and emerging POC diagnostic assays. In this paper, we evaluate the biological imaging performance of these lens systems in three broad POC diagnostic application areas; (1) bright field microscopy of histopathology slides, (2) cytologic examination of blood smears, and (3) immunofluorescence imaging. We also break down the fabrication costs and draw comparisons with other miniature optical systems. The hybrid lenses provided images with quality comparable to conventional microscopy, enabling examination of neoplastic pathology and infectious parasites including malaria and cryptosporidium. We describe how these components can be produced at below $10 per unit in full-scale production quantities, making these systems well suited for use within POC diagnostic instrumentation.

Cost-effectiveness analysis for breast cancer screening: double reading versus single + CAD reading.

BACKGROUND: Computer-aided detection (CAD) increases breast cancer detection, but its cost-effectiveness is unknown for breast cancer screening in Japan. We aimed to determine whether screening mammography diagnosed by one physician using CAD is cost-effective when compared with the standard double reading by two physicians. METHODS: We established our model with a decision tree and Markov model concept based on feasible screening and clinical pathways, combined with prognosis of the health state transition of breast cancer. Cost-effectiveness analysis between double reading by two readers and single reading with CAD by one reader was performed from a social perspective in terms of the expected cost, life expectancy and incremental cost-effectiveness ratio (ICER). The hypothetical population comprised 50-year-old female breast cancer screening examinees. Only direct medical costs related to breast cancer screening and treatment were considered. One simulation cycle was 2 years, and the annual discount rate was 3 %. Sensitivity analysis was performed to evaluate the robustness of the model and input data. RESULTS: Single reading with CAD increased expected costs by 2,704 yen and extended life expectancy by 0.0087 years compared with double reading. The ICER was 310,805 yen per life year gained, which is below the threshold. Sensitivity analysis showed that the sensitivity and specificity of CAD and the number of breast cancer screening examinees greatly affected the results. CONCLUSIONS: Single reading using CAD in mammography screening is more cost-effective than double reading, although the results are highly sensitive to the sensitivity and specificity of CAD and the numbers of examinees.

A blueprint for cost-efficient localization microscopy.

Crystal clear: The authors introduce a miniaturized localization microscopy setup based on cost-effective components. They demonstrate its feasibility for subdiffraction resolution fluorescence imaging in resolving different cellular nanostructures. The setup can be used advantageously in practical courses for training students in super-resolution fluorescence microscopy.

The reading of components of diabetic retinopathy: an evolutionary approach for filtering normal digital fundus imaging in screening and population based studies.

In any diabetic retinopathy screening program, about two-thirds of patients have no retinopathy. However, on average, it takes a human expert about one and a half times longer to decide an image is normal than to recognize an abnormal case with obvious features. In this work, we present an automated system for filtering out normal cases to facilitate a more effective use of grading time. The key aim with any such tool is to achieve high sensitivity and specificity to ensure patients' safety and service efficiency. There are many challenges to overcome, given the variation of images and characteristics to identify. The system combines computed evidence obtained from various processing stages, including segmentation of candidate regions, classification and contextual analysis through Hidden Markov Models. Furthermore, evolutionary algorithms are employed to optimize the Hidden Markov Models, feature selection and heterogeneous ensemble classifiers. In order to evaluate its capability of identifying normal images across diverse populations, a population-oriented study was undertaken comparing the software's output to grading by humans. In addition, population based studies collect large numbers of images on subjects expected to have no abnormality. These studies expect timely and cost-effective grading. Altogether 9954 previously unseen images taken from various populations were tested. All test images were masked so the automated system had not been exposed to them before. This system was trained using image subregions taken from about 400 sample images. Sensitivities of 92.2% and specificities of 90.4% were achieved varying between populations and population clusters. Of all images the automated system decided to be normal, 98.2% were true normal when compared to the manual grading results. These results demonstrate scalability and strong potential of such an integrated computational intelligence system as an effective tool to assist a grading service.

Radio-guided sentinel lymph node identification by lymphoscintigraphy fused with an anatomical vector profile: clinical applications.

OBJECTIVE: To develop a method to fuse lymphoscintigraphic images with an adaptable anatomical vector profile and to evaluate its role in the clinical practice. METHODS: We used Adobe Illustrator CS6 to create different vector profiles, we fused those profiles, using Adobe Photoshop CS6, with lymphoscintigraphic images of the patient. We processed 197 lymphoscintigraphies performed in patients with cutaneous melanomas, breast cancer or delayed lymph drainage. RESULTS: Our models can be adapted to every patient attitude or position and contain different levels of anatomical details ranging from external body profiles to the internal anatomical structures like bones, muscles, vessels, and lymph nodes. If needed, more new anatomical details can be added and embedded in the profile without redrawing them, saving a lot of time. Details can also be easily hidden, allowing the physician to view only relevant information and structures. Fusion times are about 85 s. The diagnostic confidence of the observers increased significantly. The validation process showed a slight shift (mean 4.9 mm). CONCLUSIONS: We have created a new, practical, inexpensive digital technique based on commercial software for fusing lymphoscintigraphic images with built-in anatomical reference profiles. It is easily reproducible and does not alter the original scintigraphic image. Our method allows a more meaningful interpretation of lymphoscintigraphies, an easier recognition of the anatomical site and better lymph node dissection planning.