Deep Multimodal Fusion of Data With Heterogeneous Dimensionality via Projective Networks.

The use of multimodal imaging has led to significant improvements in the diagnosis and treatment of many diseases. Similar to clinical practice, some works have demonstrated the benefits of multimodal fusion for automatic segmentation and classification using deep learning-based methods. However, current segmentation methods are limited to fusion of modalities with the same dimensionality (e.g., 3D + 3D, 2D + 2D), which is not always possible, and the fusion strategies implemented by classification methods are incompatible with localization tasks. In this work, we propose a novel deep learning-based framework for the fusion of multimodal data with heterogeneous dimensionality (e.g., 3D + 2D) that is compatible with localization tasks. The proposed framework extracts the features of the different modalities and projects them into the common feature subspace. The projected features are then fused and further processed to obtain the final prediction. The framework was validated on the following tasks: segmentation of geographic atrophy (GA), a late-stage manifestation of age-related macular degeneration, and segmentation of retinal blood vessels (RBV) in multimodal retinal imaging. Our results show that the proposed method outperforms the state-of-the-art monomodal methods on GA and RBV segmentation by up to 3.10% and 4.64% Dice, respectively.

Weakly-supervised detection of AMD-related lesions in color fundus images using explainable deep learning.

BACKGROUND AND OBJECTIVES: Age-related macular degeneration (AMD) is a degenerative disorder affecting the macula, a key area of the retina for visual acuity. Nowadays, AMD is the most frequent cause of blindness in developed countries. Although some promising treatments have been proposed that effectively slow down its development, their effectiveness significantly diminishes in the advanced stages. This emphasizes the importance of large-scale screening programs for early detection. Nevertheless, implementing such programs for a disease like AMD is usually unfeasible, since the population at risk is large and the diagnosis is challenging. For the characterization of the disease, clinicians have to identify and localize certain retinal lesions. All this motivates the development of automatic diagnostic methods. In this sense, several works have achieved highly positive results for AMD detection using convolutional neural networks (CNNs). However, none of them incorporates explainability mechanisms linking the diagnosis to its related lesions to help clinicians to better understand the decisions of the models. This is specially relevant, since the absence of such mechanisms limits the application of automatic methods in the clinical practice. In that regard, we propose an explainable deep learning approach for the diagnosis of AMD via the joint identification of its associated retinal lesions. METHODS: In our proposal, a CNN with a custom architectural setting is trained end-to-end for the joint identification of AMD and its associated retinal lesions. With the proposed setting, the lesion identification is directly derived from independent lesion activation maps; then, the diagnosis is obtained from the identified lesions. The training is performed end-to-end using image-level labels. Thus, lesion-specific activation maps are learned in a weakly-supervised manner. The provided lesion information is of high clinical interest, as it allows clinicians to assess the developmental stage of the disease. Additionally, the proposed approach allows to explain the diagnosis obtained by the models directly from the identified lesions and their corresponding activation maps. The training data necessary for the approach can be obtained without much extra work on the part of clinicians, since the lesion information is habitually present in medical records. This is an important advantage over other methods, including fully-supervised lesion segmentation methods, which require pixel-level labels whose acquisition is arduous. RESULTS: The experiments conducted in 4 different datasets demonstrate that the proposed approach is able to identify AMD and its associated lesions with satisfactory performance. Moreover, the evaluation of the lesion activation maps shows that the models trained using the proposed approach are able to identify the pathological areas within the image and, in most cases, to correctly determine to which lesion they correspond. CONCLUSIONS: The proposed approach provides meaningful information-lesion identification and lesion activation maps-that conveniently explains and complements the diagnosis, and is of particular interest to clinicians for the diagnostic process. Moreover, the data needed to train the networks using the proposed approach is commonly easy to obtain, what represents an important advantage in fields with particularly scarce data, such as medical imaging.

Simultaneous segmentation and classification of the retinal arteries and veins from color fundus images.

BACKGROUND AND OBJECTIVES: The study of the retinal vasculature represents a fundamental stage in the screening and diagnosis of many high-incidence diseases, both systemic and ophthalmic. A complete retinal vascular analysis requires the segmentation of the vascular tree along with the classification of the blood vessels into arteries and veins. Early automatic methods approach these complementary segmentation and classification tasks in two sequential stages. However, currently, these two tasks are approached as a joint semantic segmentation, because the classification results highly depend on the effectiveness of the vessel segmentation. In that regard, we propose a novel approach for the simultaneous segmentation and classification of the retinal arteries and veins from eye fundus images. METHODS: We propose a novel method that, unlike previous approaches, and thanks to the proposal of a novel loss, decomposes the joint task into three segmentation problems targeting arteries, veins and the whole vascular tree. This configuration allows to handle vessel crossings intuitively and directly provides accurate segmentation masks of the different target vascular trees. RESULTS: The provided ablation study on the public Retinal Images vessel Tree Extraction (RITE) dataset demonstrates that the proposed method provides a satisfactory performance, particularly in the segmentation of the different structures. Furthermore, the comparison with the state of the art shows that our method achieves highly competitive results in the artery/vein classification, while significantly improving the vascular segmentation. CONCLUSIONS: The proposed multi-segmentation method allows to detect more vessels and better segment the different structures, while achieving a competitive classification performance. Also, in these terms, our approach outperforms the approaches of various reference works. Moreover, in contrast with previous approaches, the proposed method allows to directly detect the vessel crossings, as well as preserving the continuity of both arteries and veins at these complex locations.

Random Network Models to Predict the Long-Term Impact of HPV Vaccination on Genital Warts.

The Human papillomaviruses (HPV) vaccine induces a herd immunity effect in genital warts when a large number of the population is vaccinated. This aspect should be taken into account when devising new vaccine strategies, like vaccination at older ages or male vaccination. Therefore, it is important to develop mathematical models with good predictive capacities. We devised a sexual contact network that was calibrated to simulate the Spanish epidemiology of different HPV genotypes. Through this model, we simulated the scenario that occurred in Australia in 2007, where 12-13 year-old girls were vaccinated with a three-dose schedule of a vaccine containing genotypes 6 and 11, which protect against genital warts, and also a catch-up program in women up to 26 years of age. Vaccine coverage were 73 % in girls with three doses and with coverage rates decreasing with age until 52 % for 20-26 year-olds. A fast 59 % reduction in the genital warts diagnoses occurred in the model in the first years after the start of the program, similar to what was described in the literature.

Modelling influenza A(H1N1) 2009 epidemics using a random network in a distributed computing environment.

In this paper we propose the use of a random network model for simulating and understanding the epidemics of influenza A(H1N1). The proposed model is used to simulate the transmission process of influenza A(H1N1) in a community region of Venezuela using distributed computing in order to accomplish many realizations of the underlying random process. These large scale epidemic simulations have recently become an important application of high-performance computing. The network model proposed performs better than the traditional epidemic model based on ordinary differential equations since it adjusts better to the irregularity of the real world data. In addition, the network model allows the consideration of many possibilities regarding the spread of influenza at the population level. The results presented here show how well the SEIR model fits the data for the AH1N1 time series despite the irregularity of the data and returns parameter values that are in good agreement with the medical data regarding AH1N1 influenza virus. This versatile network model approach may be applied to the simulation of the transmission dynamics of several epidemics in human networks. In addition, the simulation can provide useful information for the understanding, prediction and control of the transmission of influenza A(H1N1) epidemics.

Optimizing strategies for meningococcal C disease vaccination in Valencia (Spain).

BACKGROUND: Meningococcal C (MenC) conjugate vaccines have controlled invasive diseases associated with this serogroup in countries where they are included in National Immunization Programs and also in an extensive catch-up program involving subjects up to 20 years of age. Catch-up was important, not only because it prevented disease in adolescents and young adults at risk, but also because it decreased transmission of the bacteria, since it was in this age group where the organism was circulating. Our objective is to develop a new vaccination schedule to achieve maximum seroprotection in these groups. METHODS: A recent study has provided detailed age-structured information on the seroprotection levels against MenC in Valencia (Spain), where vaccination is routinely scheduled at 2 months and 6 months, with a booster dose at 18 months of age. A complementary catch-up campaign was also carried out in n for children from 12 months to 19 years of age. Statistical analyses of these data have provided an accurate picture on the evolution of seroprotection in the last few years. RESULTS: An agent-based model has been developed to study the future evolution of the seroprotection histogram. We have shown that the optimum strategy for achieving high protection levels in all infants, toddlers and adolescents is a change to a 2 months, 12 months and 12 years of age vaccination pattern. If the new schedule were implemented in January 2014, high-risk subjects between 15-19 years of age would have very low seroprotection for the next 6 years, thereby threatening the program. CONCLUSIONS: High protection levels and a low incidence of meningococcal C disease can be achieved in the future by means of a cost-free change in vaccination program. However, we recommend a new catch-up program simultaneous to the change in regular vaccination program.

[Experimental model of ultrasound thermotherapy in rats inoculated with Walker-236 tumor]. / Modelo experimental de termoterapia ultrassônica em ratos inoculados comtumor de Walker-236.

PURPOSE: To develop a model to evaluate the effects of focal pulsed ultrasound (US) waves as a source of heat for treatment of murine subcutaneous implanted Walker tumor. METHODS: An experimental, controlled, comparative study was conducted. Twenty male Wistar rats (160-300 g) randomized in 2 equal groups (G-1: Control and G-2: Hyperthermia) were inoculated with Walker-256 carcinosarcoma tumor. After 5 days G-2 rats were submitted to 45ºC hyperthermia. Heat was delivered directly to the tumor by an ultrasound (US) equipment (3 MHz frequency, 1,5W/cm³). Tumor temperature reached 45º C in 3 minutes and was maintained at this level for 5 minutes. Tumor volume was measured on days 5, 8, 11, 14 e 17 post inoculation in both groups. Unpaired t-test was used for comparison. P<0.05 was considered significant. RESULTS: Tumor volume was significantly greater in day 5 and decreased in days 11, 14 and 17 in treated rats. Rats treated with hyperthermia survived longer than control animals. On the 29th day following tumor inoculation, 40% of control rats and 77.78% of hyperthermia-treated rats remained alive. CONCLUSION: The proposed model is quite simple and may be used in less sophisticated laboratory settings for studying the effects of focal hyperthermia in the treatment of malignant implanted tumours or in survival studies.

Experimental model of ultrasound thermotherapy in rats inoculated with Walker-236 tumor.

PURPOSE: To develop a model to evaluate the effects of focal pulsed ultrasound (US) waves as a source of heat for treatment of murine subcutaneous implanted Walker tumor. METHODS: An experimental, controlled, comparative study was conducted. Twenty male Wistar rats (160-300 g) randomized in 2 equal groups (G-1: Control and G-2: Hyperthermia) were inoculated with Walker-256 carcinosarcoma tumor. After 5 days G-2 rats were submitted to 45ºC hyperthermia. Heat was delivered directly to the tumor by an ultrasound (US) equipment (3 MHz frequency, 1,5W/cm³). Tumor temperature reached 45º C in 3 minutes and was maintained at this level for 5 minutes. Tumor volume was measured on days 5, 8, 11, 14 e 17 post inoculation in both groups. Unpaired t-test was used for comparison. P<0.05 was considered significant. RESULTS: Tumor volume was significantly greater in day 5 and decreased in days 11, 14 and 17 in treated rats. Rats treated with hyperthermia survived longer than control animals. On the 29th day following tumor inoculation, 40% of control rats and 77.78% of hyperthermia-treated rats remained alive. CONCLUSION: The proposed model is quite simple and may be used in less sophisticated laboratory settings for studying the effects of focal hyperthermia in the treatment of malignant implanted tumours or in survival studies.

Experimental model of ultrasound thermotherapy in rats inoculated with Walker-236 tumor / Modelo experimental de termoterapia ultrassônica em ratos inoculados com tumor de Walker-236

PURPOSE: To develop a model to evaluate the effects of focal pulsed ultrasound (US) waves as a source of heat for treatment of murine subcutaneous implanted Walker tumor. METHODS: An experimental, controlled, comparative study was conducted. Twenty male Wistar rats (160-300 g) randomized in 2 equal groups (G-1: Control and G-2: Hyperthermia) were inoculated with Walker-256 carcinosarcoma tumor. After 5 days G-2 rats were submitted to 45ºC hyperthermia. Heat was delivered directly to the tumor by an ultrasound (US) equipment (3 MHz frequency, 1,5W/cm³). Tumor temperature reached 45º C in 3 minutes and was maintained at this level for 5 minutes. Tumor volume was measured on days 5, 8, 11, 14 e 17 post inoculation in both groups. Unpaired t-test was used for comparison. P<0.05 was considered significant. RESULTS: Tumor volume was significantly greater in day 5 and decreased in days 11, 14 and 17 in treated rats. Rats treated with hyperthermia survived longer than control animals. On the 29th day following tumor inoculation, 40 percent of control rats and 77.78 percent of hyperthermia-treated rats remained alive. CONCLUSION: The proposed model is quite simple and may be used in less sophisticated laboratory settings for studying the effects of focal hyperthermia in the treatment of malignant implanted tumours or in survival studies.

OBJETIVO: Desenvolver um modelo para avaliar os efeitos do ultra-som focal pulsado como fonte de calor para o tratamento de tumores de Walker subcutâneos implantados em ratos. MÉTODOS: Um estudo experimental, controlado, comparativo foi realizado. Vinte ratos Wistar machos (160-300 g) divididos em dois grupos (G-1: Controle e G-2: hipertermia) foram inoculados com tumor de Walker carcinossarcoma-256. Após cinco dias os ratos do grupo G-2 ratos foram submetidos a hipertermia (45ºC). O calor foi aplicado diretamente no tumor por um equipamento de ultrassonografia (3 MHz, 1,5 W/cm³). A temperatura no tumor atingiu 45ºC em 3 minutos e foi mantida nesse nível por 5 minutos. O volume do tumor foi medido nos dias 5, 8, 11, 14 e 17 após a inoculação, em ambos os grupos. Teste t não pareado foi utilizado para comparação. P <0,05 foi considerado significante. RESULTADOS: O volume do tumor foi significativamente maior no 5º dia e diminuiu nos dias 11, 14 e 17 nos ratos tratados. Animais submetidos à hipertermia sobreviveram mais tempo que os animais do grupo controle. No 29º dia após a inoculação do tumor, 40 por cento dos ratos do grupo controle e 77,78 por cento dos ratos tratados com hipertermia permaneceram vivos. CONCLUSÃO: Os resultados obtidos mostram que o modelo proposto é bastante simples e pode ser utilizado em laboratórios menos sofisticados para estudar os efeitos da hipertermia focal no tratamento dos tumores malignos implantados ou em estudos de sobrevida.

Stochastic modeling of the transmission of respiratory syncytial virus (RSV) in the region of Valencia, Spain.

In this paper, we study the dynamics of the transmission of respiratory syncytial virus (RSV) in the population using stochastic models. The stochastic models are developed introducing stochastic perturbations on the demographic parameter as well as on the transmission rate of the RSV. Numerical simulations of the deterministic and stochastic models are performed in order to understand the effect of fluctuating birth rate and transmission rate of the RSV on the population dynamics. The numerical solutions of stochastic models are calculated using Euler-Maruyama and Milstein schemes, and confidence intervals for stochastic solutions are given using Monte-Carlo method. Analysis of the numerical results reveals that perturbations on the transmission rate are more decisive in the dynamics of RSV than perturbations on demographic parameters. In addition, the stochastic models show the advantage of reproducing more effectively the noisy RSV hospitalization data. It is concluded that these stochastic models are a viable option to provide a realistic modeling of the RSV dynamics on the population.

Estudo comparativo entre as suturas contínua e separada em anastomose cólica: trabalho experimental em ratos / Comparative study between continuous and interrupted sutures in colonic anastomoses: experimental study in rats

Os autores realizam um estudo experimental em intestino grosso de ratos, comparando dois tipos de suturas intestinais (contínua e separada em plano único), avaliando os aspectos macroscópicos e histopatológicos dos dois tipos de anastomose. Os animais foram divididos em seis grupos, sendo sacrificados no 7§, 14§ e 21§ dia de pós-operatório. Concluíram que os animais submetidos a anastomose com sutura contínua apresentaram maior ganho de peso, reaçäo inflamatória menos exuberante e uma regeneraçäo epitelial glandular mais precoce e mais significativa. A sutura contínua apresenta ainda a vantagem de ser mais econômica, pois o consumo de fio cirúrgico é sigificativamente menor