Machine learning algorithms for spatial data. Case studies: Environmental pollution, natural hazards, renewable resources

This paper presents general problems and approaches for the spatial data analysis using machine learning algorithms. Machine learning is a very powerful approach to adaptive data analysis, modelling and visualisation. The key feature of the machine learning algorithms is that they learn from empirical data and can be used in cases when the modelled environmental phenomena are hidden, nonlinear, noisy and highly variable in space and in time. Most of the machines learning algorithms are universal and adaptive modelling tools developed to solve basic problems of learning from data: classification/pattern recognition, regression/mapping and probability density modelling. In the present report some of the widely used machine learning algorithms, namely artificial neural networks (ANN) of different architectures and Support Vector Machines (SVM), are adapted to the problems of the analysis and modelling of geo-spatial data. Machine learning algorithms have an important advantage over traditional models of spatial statistics when problems are considered in a high dimensional geo-feature spaces, when the dimension of space exceeds 5. Such features are usually generated, for example, from digital elevation models, remote sensing images, etc. An important extension of models concerns considering of real space constrains like geomorphology, networks, and other natural structures. Recent developments in semi-supervised learning can improve modelling of environmental phenomena taking into account on geo-manifolds. An important part of the study deals with the analysis of relevant variables and models' inputs. This problem is approached by using different feature selection/feature extraction nonlinear tools. To demonstrate the application of machine learning algorithms several interesting case studies are considered: digital soil mapping using SVM, automatic mapping of soil and water system pollution using ANN; natural hazards risk analysis (avalanches, landslides), assessments of renewable resources (wind fields) with SVM and ANN models, etc. The dimensionality of spaces considered varies from 2 to more than 30. Figures 1, 2, 3 demonstrate some results of the studies and their outputs. Finally, the results of environmental mapping are discussed and compared with traditional models of geostatistics.

Photopsies: symptome souvent méconnu et apport de l'électrorétinographie [Photopsia: an often unrecognized symptom and sensitivity of electroretinography]

BACKGROUND: Photopsias are unformed luminous spontaneous visual hallucinations, often described as flickering or wiggling lights, sometimes like a glare. Phosphenes are more intense and of shorter duration whereas migraine fortifications have a specific time course and succession of events. Recognition of this symptom is often poor, hence patients are wrongly investigated. PURPOSE: To describe the clinical presentation and electroretinographic characteristics of patients with photopsias. METHODS: 4 patients were worked-up with clinical, psychophysical, angiographic and electroretinographic examinations. RESULTS: Despite normal fundus and angiographic examinations, full-field electroretinogram was diagnostic in all cases. Retinal dysfunction involved either inner or outer retina. Paraneoplastic, and probable autoimmune/inflammatory retinopathies were found in our cases. CONCLUSION: Photopsias are often secondary to sick retinal cells. In the presence of photopsias, investigations should be directed towards the retina and electroretinography is the gold standard. Recognition of the symptom should prevent useless and potentially harmful investigations for the patient.

Ultrasound Biomicroscopy Evaluation of Anterior Extension of Retinoblastomas : A Clinico-Pathological Study

Purpose: Extension of retinoblastoma cells anterior to the vitreous surface is a criteria used to categorize retinoblastomas in Group E. In some cases, the assessment of anterior chamber invasion is readily visible by slit lamp examination, but in other cases, invasion of the posterior chamber is clinically difficult to determinate. Ultrasound biomicroscopy (UBM) allows high-resolution images of the anterior segment and structures normally hidden from clinical visualization such as ciliary body, zonule or posterior chamber. This prompted us to evaluate the value of UBM in the assessment of posterior and anterior chamber involvement in patients with peripheral extending retinoblastomas. Methods: We retrospectively reviewed all retinoblastoma cases that underwent enucleation in our institution from 1.1996 till 12.2009 in which UBM ( 35-MHz or 50-MHz) evaluation was available. The UBM results were compared with the histopathological data. Results: From 1.1996 till 12.2009, 146 enucleations were performed in our institution. UBM information was available in 18 cases. There were 8 males and 10 females. The mean age was 4 years old. UBM allowed correct assessement of posterior chamber invasion in 15 cases (13 cases with retinoblastoma in the posterior chamber and 2 cases without retinoblastoma in the posterior chamber). There was a significant correlation between the presence of retinoblastoma cells in the posterior chamber detected by UBM and the histopathological confirmation of posterior chamber involvement (p=0,0008). The sensitivity of UBM in the assessment of posterior chamber invasion by retinoblastoma was 81% and the specificity 100%. UBM allowed correct assessment of anterior chamber invasion in 13 cases. The sensitivity of UBM for this purpose was 50% and the specificity 60 %. Conclusions: In selected cases of advanced retinoblastoma, UBM appears to represent a valuable tool in the evaluation of the precise extension of the disease. Although our series encompasses only a limited number of cases, the sensitivity and specificity of UBM in the assessment of retinoblastoma anterior extension is interesting. Further prospective multi-centered clinical studies would be necessary to better delineate the utility of this method in the precise categorization of retinoblastoma anterior extension

Carcinoembryonic antigen (CEA): demonstration of a partial identity between CEA and a normal glycoprotein.

The carcinoembryonic antigen of the human digestive tract (CEA), described by Gold as a glycoprotein specific for digestive carcinomatous and foetal tissues, was found to have common antigenic determinants with a glycoprotein of smaller size extracted from normal adult tissues. This observation suggests that only a part of the CEA molecule carries the onco-foetal specificity. It also has practical implications regarding the radioimmunoassay for CEA used for the diagnosis of certain carcinomas.

Decreased local control following radiation therapy alone in early-stage glottic carcinoma with anterior commissure extension.

PURPOSE: To assess the patterns of failure in the treatment of early-stage squamous cell carcinoma of the glottic larynx. PATIENTS AND METHODS: Between 1983-2000, 122 consecutive patients treated for early laryngeal cancer (UICC T1N0 and T2N0) by radical radiation therapy (RT) were retrospectively studied. Male-to-female ratio was 106 : 16, and median age 62 years (35-92 years). There were 68 patients with T1a, 18 with T1b, and 36 with T2 tumors. Diagnosis was made by biopsy in 104 patients, and by laser vaporization or stripping in 18. Treatment planning consisted of three-dimensional (3-D) conformal RT in 49 (40%) patients including nine patients irradiated using arytenoid protection. A median dose of 70 Gy (60-74 Gy) was given (2 Gy/fraction) over a median period of 46 days (21-79 days). Median follow-up period was 85 months. RESULTS: The 5-year overall, cancer-specific, and disease-free survival amounted to 80%, 94%, and 70%, respectively. 5-year local control was 83%. Median time to local recurrence in 19 patients was 13 months (5-58 months). Salvage treatment consisted of surgery in 17 patients (one patient refused salvage and one was inoperable; total laryngectomy in eleven, and partial laryngectomy or cordectomy in six patients). Six patients died because of laryngeal cancer. Univariate analyses revealed that prognostic factors negatively influencing local control were anterior commissure extension, arytenoid protection, and total RT dose < 66 Gy. Among the factors analyzed, multivariate analysis (Cox model) demonstrated that anterior commissure extension, arytenoid protection, and male gender were the worst independent prognostic factors in terms of local control. CONCLUSION: For early-stage laryngeal cancer, outcome after RT is excellent. In case of anterior commissure extension, surgery or higher RT doses are warranted. Because of a high relapse risk, arytenoid protection should not be attempted.

Mast cells are dispensable for normal and activin-promoted wound healing and skin carcinogenesis.

The growth and differentiation factor activin A is a key regulator of tissue repair, inflammation, fibrosis, and tumorigenesis. However, the cellular targets, which mediate the different activin functions, are still largely unknown. In this study, we show that activin increases the number of mature mast cells in mouse skin in vivo. To determine the relevance of this finding for wound healing and skin carcinogenesis, we mated activin transgenic mice with CreMaster mice, which are characterized by Cre recombinase-mediated mast cell eradication. Using single- and double-mutant mice, we show that loss of mast cells neither affected the stimulatory effect of overexpressed activin on granulation tissue formation and reepithelialization of skin wounds nor its protumorigenic activity in a model of chemically induced skin carcinogenesis. Furthermore, mast cell deficiency did not alter wounding-induced inflammation and new tissue formation or chemically induced angiogenesis and tumorigenesis in mice with normal activin levels. These findings reveal that mast cells are not major targets of activin during wound healing and skin cancer development and also argue against nonredundant functions of mast cells in wound healing and skin carcinogenesis in general.

Five Years Of Denosumab Treatment In Women With Postmenopausal Osteoporosis: Preliminary Results From The Freedom Extension Study

Aims: The pivotal FREEDOM study evaluated the effi cacy and safety of 3 years' denosumab treatment in women with postmenopausal osteoporosis (PMO).1 Since osteoporosis is a chronic condition requiring long-term therapy, FREEDOM was extended to further elucidate the safety and effi cacy of long-term denosumab administration. We present data from the fi rst 2 years of this extension, representing up to 5 years' continuous exposure to denosumab.Methods: Patients who completed FREEDOM were eligible for the extension. Women continued to receive (long-term group), or started after 3 years' placebo (cross-over group), denosumab 60 mg sc every 6 months and daily calcium and vitamin D. These data refl ect 5 years' (long-term) or 2 years' (cross-over) continuous denosumab treatment. Effi cacy measures include changes in BMD from extension study baseline and bone turnover markers (BTM). P-values are descriptive.Results: Of the 83.0% of subjects who completed FREEDOM, 70.2% (N = 4550) agreed to participate in the extension (long-term: 2343; cross-over: 2207). In the long-term group, there were further signifi cant gains (P < 0.0001) in BMD in years 4 and 5: 1.9% and 1.7% at the lumbar spine to a total of 13.7% from FREEDOM baseline and 0.7% and 0.6% at the total hip to a total of 7.0%. During their fi rst 2 years' denosumab treatment, women in the cross-over group had signifi cant improvements in lumbar spine (7.9%) and total hip BMD (4.1%) (P < 0.0001). Serum C-telopeptide (CTX) was rapidly reduced following denosumab dosing in both groups, with the characteristic attenuation of CTX reduction observed at the end of the dosing interval. A low incidence of new vertebral and nonvertebral fractures was reported for both groups. The denosumab safety profi le did not change over time.Conclusions: Denosumab treatment for up to 5 years in women with PMO remains well tolerated, maintains reduction of BTMs and continues to significantly increase BMD.Reference1. Cummings. NEJM 2009;361:756.

Instrumental leadership: Measurement and extension of transformational-transactional leadership theory.

Leaders must scan the internal and external environment, chart strategic and task objectives, and provide performance feedback. These instrumental leadership (IL) functions go beyond the motivational and quid-pro quo leader behaviors that comprise the full-range-transformational, transactional, and laissez faire-leadership model. In four studies we examined the construct validity of IL. We found evidence for a four-factor IL model that was highly prototypical of good leadership. IL predicted top-level leader emergence controlling for the full-range factors, initiating structure, and consideration. It also explained unique variance in outcomes beyond the full-range factors; the effects of transformational leadership were vastly overstated when IL was omitted from the model. We discuss the importance of a "fuller full-range" leadership theory for theory and practice. We also showcase our methodological contributions regarding corrections for common method variance (i.e., endogeneity) bias using two-stage least squares (2SLS) regression and Monte Carlo split-sample designs.

Transfer factor for carbon monoxide: a glance behind the scene.

The transfer factor for carbon monoxide (TLCO) is widely used in pulmonary function laboratories because it represents a unique non-invasive window on pulmonary microcirculation. The TLCO is the product of two primary measurements, the alveolar volume (VA) and the CO transfer coefficient (KCO). This test is most informative when VA and KCO are examined, together with their product TLCO. In a normal lung, a low VA due to incomplete expansion is associated with an elevated KCO, resulting in a mildly reduced TLCO. Thus, in case of low VA, a seemingly "normal KCO" must be interpreted as an abnormal gas transfer. The most common clinical conditions associated with an abnormal TLCO are characterised by a limited number of patterns for VA and KCO: incomplete lung expansion, discrete loss of alveolar units, diffuse loss of alveolar units, emphysema, pulmonary vascular disorders, high pulmonary blood volume, alveolar haemorrhage.

Calculation of correction factors for ionization chamber measurements with small fields in low-density media.

The quantity of interest for high-energy photon beam therapy recommended by most dosimetric protocols is the absorbed dose to water. Thus, ionization chambers are calibrated in absorbed dose to water, which is the same quantity as what is calculated by most treatment planning systems (TPS). However, when measurements are performed in a low-density medium, the presence of the ionization chamber generates a perturbation at the level of the secondary particle range. Therefore, the measured quantity is close to the absorbed dose to a volume of water equivalent to the chamber volume. This quantity is not equivalent to the dose calculated by a TPS, which is the absorbed dose to an infinitesimally small volume of water. This phenomenon can lead to an overestimation of the absorbed dose measured with an ionization chamber of up to 40% in extreme cases. In this paper, we propose a method to calculate correction factors based on the Monte Carlo simulations. These correction factors are obtained by the ratio of the absorbed dose to water in a low-density medium □D(w,Q,V1)(low) averaged over a scoring volume V₁ for a geometry where V₁ is filled with the low-density medium and the absorbed dose to water □D(w,QV2)(low) averaged over a volume V₂ for a geometry where V₂ is filled with water. In the Monte Carlo simulations, □D(w,QV2)(low) is obtained by replacing the volume of the ionization chamber by an equivalent volume of water, according to the definition of the absorbed dose to water. The method is validated in two different configurations which allowed us to study the behavior of this correction factor as a function of depth in phantom, photon beam energy, phantom density and field size.

The Eukaryotic Promoter Database EPD: the impact of in silico primer extension.

The Eukaryotic Promoter Database (EPD) is an annotated non-redundant collection of eukaryotic POL II promoters, experimentally defined by a transcription start site (TSS). There may be multiple promoter entries for a single gene. The underlying experimental evidence comes from journal articles and, starting from release 73, from 5' ESTs of full-length cDNA clones used for so-called in silico primer extension. Access to promoter sequences is provided by pointers to TSS positions in nucleotide sequence entries. The annotation part of an EPD entry includes a description of the type and source of the initiation site mapping data, links to other biological databases and bibliographic references. EPD is structured in a way that facilitates dynamic extraction of biologically meaningful promoter subsets for comparative sequence analysis. Web-based interfaces have been developed that enable the user to view EPD entries in different formats, to select and extract promoter sequences according to a variety of criteria and to navigate to related databases exploiting different cross-references. Tools for analysing sequence motifs around TSSs defined in EPD are provided by the signal search analysis server. EPD can be accessed at http://www.epd. isb-sib.ch.

Acquired auditory agnosia in childhood and normal sleep electroencephalography subsequently diagnosed as Landau-Kleffner syndrome: a report of three cases.

Aim  We report three cases of Landau-Kleffner syndrome (LKS) in children (two females, one male) in whom diagnosis was delayed because the sleep electroencephalography (EEG) was initially normal. Method  Case histories including EEG, positron emission tomography findings, and long-term outcome were reviewed. Results  Auditory agnosia occurred between the age of 2 years and 3 years 6 months, after a period of normal language development. Initial awake and sleep EEG, recorded weeks to months after the onset of language regression, during a nap period in two cases and during a full night of sleep in the third case, was normal. Repeat EEG between 2 months and 2 years later showed epileptiform discharges during wakefulness and strongly activated by sleep, with a pattern of continuous spike-waves during slow-wave sleep in two patients. Patients were diagnosed with LKS and treated with various antiepileptic regimens, including corticosteroids. One patient in whom EEG became normal on hydrocortisone is making significant recovery. The other two patients did not exhibit a sustained response to treatment and remained severely impaired. Interpretation  Sleep EEG may be normal in the early phase of acquired auditory agnosia. EEG should be repeated frequently in individuals in whom a firm clinical diagnosis is made to facilitate early treatment.