Using Knowledge-Guided Machine Learning To Assess Patterns of Areal Change in Waterbodies across the Contiguous United States

Lake and reservoir surface areas are an important proxy for freshwater availability. Advancements in machine learning (ML) techniques and increased accessibility of remote sensing data products have enabled the analysis of waterbody surface area dynamics on broad spatial scales. However, interpreting the ML results remains a challenge. While ML provides important tools for identifying patterns, the resultant models do not include mechanisms. Thus, the "black-box" nature of ML techniques often lacks ecological meaning. Using ML, we characterized temporal patterns in lake and reservoir surface area change from 1984 to 2016 for 103,930 waterbodies in the contiguous United States. We then employed knowledge-guided machine learning (KGML) to classify all waterbodies into seven ecologically interpretable groups representing distinct patterns of surface area change over time. Many waterbodies were classified as having "no change" (43%), whereas the remaining 57% of waterbodies fell into other groups representing both linear and nonlinear patterns. This analysis demonstrates the potential of KGML not only for identifying ecologically relevant patterns of change across time but also for unraveling complex processes that underpin those changes.

ARTIFICIAL INTELLIGENCE: CREATING NEW PARADIGMS IN THE MANAGEMENT OF NON-COMMUNICABLE DISEASES

The role of Artificial Intelligence (AI) in medical science is growing immensely. Since AI contains features that can address both preventive and therapeutic aspects of non-communicable diseases (NCDs), it can potentially lessen the massive burden of morbidity and mortality associated with NCDs. AI can help in various ways in NCDs including predicting disease occurrence, monitoring, ensuring treatment and follow-up of patients. Low- and middle-income countries can harness the benefit of AI for the management of chronic diseases and effectively address challenges like manpower shortage, accessibility to health care, etc. However, AI needs to be used responsibly and rationally in NCDs for its maximum benefit.

Artificial Intelligence for Modeling Complex Systems: Taming the Complexity of Expert Models to Improve Decision Making

Major societal and environmental challenges involve complex systems that have diverse multi-scale interacting processes. Consider, for example, how droughts and water reserves affect crop production and how agriculture and industrial needs affect water quality and availability. Preventive measures, such as delaying planting dates and adopting new agricultural practices in response to changing weather patterns, can reduce the damage caused by natural processes. Understanding how these natural and human processes affect one another allows forecasting the effects of undesirable situations and study interventions to take preventive measures. For many of these processes, there are expert models that incorporate state-of-the-art theories and knowledge to quantify a system's response to a diversity of conditions. A major challenge for efficient modeling is the diversity of modeling approaches across disciplines and the wide variety of data sources available only in formats that require complex conversions. Using expert models for particular problems requires integration of models with third-party data as well as integration of models across disciplines. Modelers face significant heterogeneity that requires resolving semantic, spatiotemporal, and execution mismatches, which are largely done by hand today and may take more than 2 years of effort.

We are developing a modeling framework that uses artificial intelligence (AI) techniques to reduce modeling effort while ensuring utility for decision making. Our work to date makes several innovative contributions: (1) an intelligent user interface that guides analysts to frame their modeling problem and assists them by suggesting relevant choices and automating steps along the way; (2) semantic metadata for models, including their modeling variables and constraints, that ensures model relevance and proper use for a given decision-making problem; and (3) semantic representations of datasets in terms of modeling variables that enable automated data selection and data transformations. This framework is implemented in the MINT (Model INTegration) framework, and currently includes data and models to analyze the interactions between natural and human systems involving climate, water availability, agricultural production, and markets. Our work to date demonstrates the utility of AI techniques to accelerate modeling to support decision-making and uncovers several challenging directions for future work.

Plantation Mapping in Southeast Asia

Plantation mapping is important for understanding deforestation and climate change. While most existing plantation products are created manually, in this paper we study an ensemble learning based framework for automatically mapping plantations in southern Kalimantan on a yearly scale using remote sensing data. We study the effectiveness of several components in this framework, including class aggregation, data sampling, learning model selection and post-processing, by comparing with multiple baselines. In addition, we analyze the quality of our plantation mapping product by visual examination of high resolution images. We also compare our method to existing manually labeled plantation datasets and show that our method can achieve a better balance of precision (i.e., user's accuracy) and recall (i.e., producer's accuracy).

Blunt Cerebrovascular Injuries: Advances in Screening, Imaging, and Management Trends

Blunt cerebrovascular injury is a relatively uncommon but sometimes life-threatening injury, particularly in patients presenting with ischemic symptoms in that vascular territory. The decision to pursue vascular imaging (generally CT angiography) is based on clinical and imaging findings. Several grading scales or screening criteria have been developed to guide the decision to pursue vascular imaging, as well as to recommend different treatment options for various injuries. The data supporting many of these guidelines and options are limited however. The purpose of this article is to review and compare these scales and criteria and the data supporting clinical efficacy and to make recommendations for future research in this area.

Patterns of metastasis and recurrence in thymic epithelial tumours: longitudinal imaging review in correlation with histological subtypes

AIM

To determine the patterns of metastasis and recurrence in thymic epithelial tumours based on longitudinal imaging studies, and to correlate the patterns with World Health Organization (WHO) histological classifications.

MATERIALS AND METHODS

Seventy-seven patients with histopathologically confirmed thymomas (n=62) and thymic carcinomas (n=15) who were followed with cross-sectional follow-up imaging after surgery were retrospectively studied. All cross-sectional imaging studies during the disease course were reviewed to identify metastasis or recurrence. The sites of involvement and the time of involvement measured from surgery were recorded.

RESULTS

Metastasis or recurrence was noted in 24 (31%) of the 77 patients. Patients with metastasis or recurrence were significantly younger than those without (median age: 46 versus 60, respectively; p=0.0005), and more commonly had thymic carcinomas than thymomas (p=0.002). The most common site of involvement was the pleura (17/24), followed by the lung (9/24), and thoracic nodes (9/24). Abdominopelvic involvement was noted in 12 patients, most frequently in the liver (n=8). Lung metastasis was more common in thymic carcinomas than thymomas (p=0.0005). Time from surgery to the development of metastasis or recurrence was shortest in thymic carcinoma, followed by high-risk thymomas, and was longest in low-risk thymoma (median time in months: 25.1, 68.8, and not reached, respectively; p=0.0015).

CONCLUSIONS

The patterns of metastasis and recurrence of thymic epithelial tumours differ significantly across histological subgroups, with thymic carcinomas more commonly having metastasis with shorter length of time after surgery. The knowledge of different patterns of tumour spread may contribute to further understanding of the biological and clinical behaviours of these tumours.

Radiologist's perspective for the Meckel's diverticulum and its complications

The Meckel's diverticulum is the commonest congenital anomaly of the gastrointestinal tract, often presenting with complications such as gastrointestinal bleeding, intussusception, bowel obstruction and diverticulitis, which are often misdiagnosed. Imaging plays an important role in the early diagnosis and characterization of these conditions and is very helpful in decision making. The Meckel's diverticulum and its complications have myriad presentations and appearances on various imaging modalities. Thus, sound knowledge of the anatomy, embryology, clinical presentation, imaging characteristics and complications is crucial to the practice of abdominal imaging. We present a review of the literature and current radiological practices in the diagnosis and management of the Meckel's diverticulum and its various complications with special emphasis on the imaging of various complications, mimickers and pathological correlation.

Ruptured Aortic Aneurysm Secondary to Psoas abscess after Intravesical Bacilli Calmette-Guérin

Bacillus Calmette-Guérin (BCG) is a live attenuated strain of Mycobacterium bovis that has been used as effective treatment for early-stage transitional carcinoma of the urinary bladder. We present a case of a 68 year old man who had an abdominal aortic aneurysm following BCG therapy for bladder cancer. Contrast enhanced computerized tomogram (CECT) of abdomen and pelvis revealed bilateral hypodense lesions suggestive of psoas abscesses. In addition, a saccular abdominal aortic aneurysm measuring 4x3.6 cm involving infrarenal aorta with surrounding hematoma was seen. At surgery, he was found to have a psoas abscess and hemorrhage. He underwent ligation of the aorta and an axillary-bifemoral bypass. He was given one year of anti-tubercular therapy to which he responded clinically.

Hydrogen adsorption in several types of carbon nanotubes

In this work, we aim to study the hydrogen adsorption in several kinds of carbon nanotubes grown under different process conditions and to correlate the findings with the morphological microstructure and physical properties of these materials. The growth conditions and the behaviour with respect to hydrogen interaction of various carbon nanotubes are discussed, to establish microstructure-process-property relationships. In particular, we have analyzed several types of carbon nanotubes, namely one single-walled and five multi-walled having different tube diameter (due to different deposition techniques and conditions), different defectiveness and submitted to different surface treatments. To better understand the differences among the various samples, they have been investigated using field emission scanning electron microscopy and high resolution transmission electron microscopy for the morphological and structural characteristics, thermo-gravimetric analysis for the sample purity and Brunauer-Emmett-Teller analysis for the surface area. The experimental measurements on the ability of the different types of carbon nanotubes to adsorb and/or releasing hydrogen have been performed at 77 K with a volumetric Sievert analytical tool. Our findings clearly demonstrate a direct correlation between the exposed surface area and adsorbed hydrogen capacity, which confirms their linear relationship observed previously. For instance, single-walled nanotubes with surface area density of approximately 800 m2/g have showed hydrogen storage of approximately 1.7 wt% at a pressure of 35 atm. Adsorption process seems to be perfectly reversible. The adsorption values have been compared with a simple model, in order to evaluate the potentialities for carbon-based nanomaterials in future hydrogen storage applications.

Role of ABA and ABI3 in desiccation tolerance

We show in bryophytes that abscisic acid (ABA) pretreatment of moss (Physcomitrella patens) cells confers desiccation tolerance. In angiosperms, both ABA and the transcriptional regulator ABSCISIC ACID INSENSITIVE 3 (ABI3) are required to protect the seed during desiccation. ABA was not able to protect moss cells in stable deletion lines of ABI3 (DeltaPpabi3). Hence, moss has the same functional link between ABA, ABI3, and the desiccation tolerance phenotype that is found in angiosperms. Furthermore, we identified 22 genes that were induced during ABA pretreatment in wild-type lines. When their expression was compared with that of DeltaPpabi3 during ABA pretreatment and immediately after desiccation, a new target of ABI3 action appears to be in the recovery period.

Hydrogen adsorption in several types of carbon nanotubes

In this work, we aim to study the hydrogen adsorption in several kinds of carbon nanotubes grown under different process conditions and to correlate the findings with the morphological microstructure and physical properties of these materials. The growth conditions and the behaviour with respect to hydrogen interaction of various carbon nanotubes are discussed, to establish microstructure-process-property relationships. In particular, we have analyzed several types of carbon nanotubes, namely one single-walled and five multi-walled having different tube diameter (due to different deposition techniques and conditions), different defectiveness and submitted to different surface treatments. To better understand the differences among the various samples, they have been investigated using field emission scanning electron microscopy and high resolution transmission electron microscopy for the morphological and structural characteristics, thermo-gravimetric analysis for the sample purity and Brunauer-Emmett-Teller analysis for the surface area. The experimental measurements on the ability of the different types of carbon nanotubes to adsorb and/or releasing hydrogen have been performed at 77 K with a volumetric Sievert analytical tool. Our findings clearly demonstrate a direct correlation between the exposed surface area and adsorbed hydrogen capacity, which confirms their linear relationship observed previously. For instance, single-walled nanotubes with surface area density of approximately 800 m2/g have showed hydrogen storage of approximately 1.7 wt% at a pressure of 35 atm. Adsorption process seems to be perfectly reversible. The adsorption values have been compared with a simple model, in order to evaluate the potentialities for carbon-based nanomaterials in future hydrogen storage applications.

Sperm characteristics and teratology in rats following vas deferens occlusion with RISUG and its reversal

The functional success of the reversal of vas occlusion by styrene maleic anhydride (RISUG), using the solvent vehicle, Dimethyl Sulphoxide (DMSO), has been investigated. Reversal with DMSO was carried out in Wistar albino rats 90 days after bilateral vas occlusion. The body weight, organ weight, sperm characteristics, fertility test and teratology, including skeletal morphology were evaluated in vas occlusion and reversal animals and in F(1) progenies to assess the functional success of the occlusion and reversal. Body weight, organ weight and the cauda epididymal sperm characteristics of vas occlusion and reversal animals and of F(1) progenies were comparable to control. Ejaculated spermatozoa in the vaginal smear showed detached head/tail, acrosomal damage, bent midpiece, bent tail and morphological aberrations in sperm head after vas occlusion, which returned to normal, 90 days after reversal. Monthly fertility test, post-injection showed 0% fertility, which improved gradually and 100% fertility was achieved 90 days after reversal. The fertility/pregnancy/implantation record and skeletal morphology of the offspring were comparable to control. The results suggest functional success and safety of vas occlusion reversal by DMSO.

Effects of novel retinoic acid metabolism blocking agent (VN/14-1) on N-methyl-N-nitrosourea (MNU)-induced mammary carcinoma and uterus in the rat model

Abstract #2133

Background: All-trans-retinoic acid (ATRA) and other retinoids play key roles in prevention and therapy of many proliferative diseases including cancers. VN/14-1 [4-(±)-(1H-Imidazol-1-yl)-(E)-retinoic acid], which is a new generation novel retinoic acid metabolism blocking agent (RAMBA), works by inhibiting the breakdown of ATRA. The inhibitory effects of VN/14-1 on the growth of human breast cancer cells and human breast tumors in the nude mouse model have been previously demonstrated. The purpose of this study was to evaluate the effects of VN/14-1 on the N-methyl-N-nitrosourea (MNU)-induced rat mammary carcinoma model, as well as on the uterus in immature ovariectomized (OVX) rats.
 Methods: (1) VN/14-1 (5, 10, and 20 mg/kg/d) was given by oral gavage to grouped female Sprague Dawley (SD) rats bearing MNU-induced mammary carcinoma for 8 weeks, after which tumor weight and volume, as well as histology were measured. (2) VN/14-1 (10 and 20 mg/kg/d) and b-estradiol (10 mg/kg/d) were given alone or in combination, by oral gavage (VN/14-1) and subcutaneous injection (b-estradiol), to immature OVX SD rats for 3 days, after which uterine weight and histology were measured.
 Results: (1) At the end of the treatment period, the administration of 5, 10 and 20 mg/kg/d VN/14-1 caused significant reductions of 19.1, 34.4 and 44.3%, respectively, in mean tumor weight compared with the control animals (all p < 0.05). The cumulative tumor growth was also significantly slower in groups receiving 5, 10 and 20 mg/kg/d compared to the control group in a dose-dependent manner. (2) Immature OVX rats given VN/14-1 at doses of 10 and 20 mg/kg, had reduction in uterine wet weight of up to 56% compared to OVX controls (P < 0.001). OVX rats given VN/14-1 of 10 and 20 mg/kg in combination with β-estradiol had reduction in uterine wet weight of up to 58% compared to the OVX rats given β-estradiol alone (P < 0.001). The adverse toxic effects such as fatigue and anorexia were occurred in the groups at high dose of 20 mg/kg.
 Conclusions: RAMBA VN/14-1 was able to inhibit the growth of tumors in the MNU-induced ER positive rat mammary tumor model and antagonized the stimulatory effect of β-estradiol on the uterus. The studies suggest VN/14-1 might be an effective novel therapy for ER positive breast cancer.

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 2133.

Novel retinoic acid metabolism blocking agents have potent inhibitory activities on human breast cancer cells and tumour growth

Antitumour effects of retinoids are attributed to their influence on cell proliferation, differentiation, apoptosis and angiogenesis. In our effort to develop useful agents for breast cancer therapy, we evaluated the effects of four representative retinoic acid metabolism blocking agents (RAMBAs, VN/14-1, VN/50-1, VN/66-1 and VN/69-1) on growth inhibition of oestrogen receptor positive (ER +ve, MCF-7 and T-47D) and oestrogen receptor negative (ER −ve, MDA-MB-231) human breast cancer cells. Additionally, we investigated the biological effects/molecular mechanism(s) underlying their growth inhibitory properties as well as their antitumour efficacies against MCF-7 and MCF-7Ca tumour xenografts in nude mice. We also assessed the effect of combining VN/14-1 and all-trans-retinoic acid (ATRA) on MCF-7 tumuor xenografts. The ER +ve cell lines were more sensitive (IC₅₀ values between 3.0 and 609 nM) to the RAMBAs than the ER −ve MDA-MB-231 cell line (IC₅₀=5.6–24.0 μM). Retinoic acid metabolism blocking agents induced cell differentiation as determined by increased expression of cytokeratin 8/18 and oestrogen receptor-α (ER-α). Similar to ATRA, they also induced apoptosis via activation of caspase 9. Cell cycle analysis indicated that RAMBAs arrested cells in the G1 and G2/M phases and caused significant downregulation (>80%) of cyclin D1 protein. In vivo, the growth of MCF-7 mammary tumours was dose-dependently and significantly inhibited (92.6%, P<0.0005) by VN/14-1. The combination of VN/14-1 and ATRA also inhibited MCF-7 breast tumour growth in vivo (up to 120%) as compared with single agents (P<0.025). VN/14-1 was also very effective in preventing the formation of MCF-7Ca tumours and it significantly inhibited the growth of established MCF-7Ca tumours, being as effective as the clinically used aromatase inhibitors, anastrozole and letrozole. Decrease in cyclin D1 and upregulation of cytokeratins, Bad and Bax with VN/14-1 may be responsible for the efficacy of this compound in inhibiting breast cancer cell growth in vitro and in vivo. Our results suggest that our RAMBAs, especially VN/14-1 may be useful novel therapy for breast cancer.

Role of multidetector CT virtual bronchoscopy in the evaluation of post-tracheostomy tracheal stenosis--a preliminary study

AIM

To study the technique and utility of virtual bronchoscopy (virtual reality endobronchial simulation, VRES) as a tool to evaluate post-tracheostomy tracheal stenoses and to correlate the findings of virtual and invasive bronchoscopy and to follow-up treated lesions or those currently under treatment that were initially diagnosed with VRES.

METHODOLOGY

This prospective study comprised nine patients in the age group 13 to 65 years presenting with breathlessness and stridor following one or multiple tracheostomies. They underwent plain CT using a multidetector CT (MDCT) scanner (Siemens Volume Zoom) using narrow (1 mm) collimation. These thin slice images were post-processed using an Irix-based workstation with a 'Fly-Through' endoscopy application. These patients also underwent a rigid (three patients) or fiberoptic (six patients) bronchoscopy.

RESULTS

Of the nine patients that underwent VRES, five were found to have stenoses, three had obstructing granulation tissue, one had an obstructing membrane and one had synechiae. The invasive bronchoscopic findings supported the VRES diagnosis in all but one case of stenosis, one of granulation tissue and the case with synechiae. Membranes and synechiae were relatively difficult to diagnose without the corresponding axial and multiplanar images. VRES achieved a higher sensitivity, while invasive bronchoscopy a higher specificity.

CONCLUSIONS

VRES proved to be comparable to invasive bronchoscopy in the depiction of post-tracheostomy tracheal stenoses, with a notable advantage in critical stenoses in that the airway distal to the stenosis could be assessed with VRES but not with invasive bronchoscopy. A preliminary VRES was found to be of assistance in the selection of patients for the more invasive therapeutic procedures such as laser ablation of granulation tissue and its follow-up.