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.

Cardiopulmonary support and extracorporeal membrane oxygenation for cardiac assist.

BACKGROUND: Use of cardiopulmonary bypass for emergency resuscitation is not new. In fact, John Gibbon proposed this concept for the treatment of severe pulmonary embolism in 1937. Significant progress has been made since, and two main concepts for cardiac assist based on cardiopulmonary bypass have emerged: cardiopulmonary support (CPS) and extracorporeal membrane oxygenation (ECMO). The objective of this review is to summarize the state of the art in these two technologies. METHODS: Configuration of CPS is now fairly standard. A mobile cart with relatively large wheels allowing for easy transportation carries a centrifugal pump, a back-up battery with a charger, an oxygen cylinder, and a small heating system. Percutaneous cannulation, pump-driven venous return, rapid availability, and transportability are the main characteristics of a CPS system. Cardiocirculatory arrest is a major predictor of mortality despite the use of CPS. In contrast, CPS appears to be a powerful tool for patients in cardiogenic shock before cardiocirculatory arrest, requiring some type of therapeutic procedures, especially repair of anatomically correctable problems or bridging to other mechanical circulatory support systems such as ventricular assist devices. CPS is in general not suitable for long-term applications because of the small-bore cannulas, resulting in significant pressure gradients and eventually hemolysis. RESULTS: In contrast, ECMO can be designed for longer-term circulatory support. This requires large-bore cannulas and specifically designed oxygenators. The latter are either plasma leakage resistent (true membranes) or relatively thrombo-resistant (heparin coated). Both technologies require oxygenator changeovers although the main reason for this is different (clotting for the former, plasma leakage for the latter). Likewise, the tubing within a roller pump has to be displaced and centrifugal pump heads have to be replaced over time. ECMO is certainly the first choice for a circulatory support system in the neonatal and pediatric age groups, where the other assist systems are too bulky. ECMO is also indicated for patients improving on CPS. Septic conditions are, in general, considered as contraindications for ECMO. CONCLUSIONS: Ease of availability and moderate cost of cardiopulmonary bypass-based cardiac support technologies have to be balanced against the significant immobilization of human resources, which is required to make them successful.

Do resources, justice administration practices and federalism have an impact on registered and sentenced crime prevalence ?

This contribution, based on a statistical approach, undertakes to link data on resources (personnel and financial means) and the working of the administration of penal justice (prosecution, sentencing) taking into account the nationality of those prosecuted. In order to be able to distinguish prosecution and sentencing practices of judicial authorities and possible processes of discrimination, diverse sources have been used such as data from court administrations, public finances and police forces, collected by the Swiss Federal Statistical Office and the Swiss Federal administration of finances. The authors discuss discrimination in prosecution and sentencing between Swiss residents and foreigners taking into account localization and resources regarding personnel and public finances.

Maladies inflammatoires de l'intestin [Main therapeutic advances in IBD in 2008]

A coimmunosuppression with azathioprine or methotrexate in addition to infliximab does not improve the therapeutic efficacy in Crohn's disease but increase the risks of infectious complications and neoplasia.

Comparison of HIV-infected patients' characteristics, healthcare resources use and cost between native and migrant patients

OBJECTIVES: To assess whether patients' characteristics and healthcare resources consumption and costs were different between native and migrant populations in Switzerland. METHODS: All adult patients followed-up in the Swiss HIV-cohort study in our institution during 2000-2003 were considered. Patients' characteristics were retrieved from the cohort database. Hospital and outpatient resource use were extracted from individual charts and valued with 2002 tariffs. RESULTS: The 66 migrants were younger (29 +/- 8 years versus 37 +/- 11, p < 0.001), less often of male gender (38 % versus 70 %, p < 0.001), predominantly infected via heterosexual contact (87 % versus 52 %, p < 0.01), with lower mean CD4 level at enrollment (326 +/- 235 versus 437 +/- 305, p = 0.002) than their 200 native counterparts. Migrants had fewer hospitalizations, more frequent outpatient visits, laboratory tests, and lower total cost of care per year of follow-up (<euro> 2'215 +/- 4'206 versus 4'155 +/- 12'304, p = 0.037). Resource use and costs were significantly higher in people with < 200 CD4 cell counts in both groups. CONCLUSIONS: Migrant population had more advanced disease, more outpatient visits but less hospitalizations, resulting in lower costs of care when compared with native population.

Therapeutic Drug Monitoring of the new targeted anticancer agents imatinib, nilotinib, dasatinib, sunitinib, sorafenib and lapatinib by LC tandem mass spectrometry.

The treatment of some cancer patients has shifted from traditional, non-specific cytotoxic chemotherapy to chronic treatment with molecular targeted therapies. Imatinib mesylate, a selective inhibitor of tyrosine kinases (TKIs) is the most prominent example of this new era and has opened the way to the development of several additional TKIs, including sunitinib, nilotinib, dasatinib, sorafenib and lapatinib, in the treatment of various hematological malignancies and solid tumors. All these agents are characterized by an important inter-individual pharmacokinetic variability, are at risk for drug interactions, and are not devoid of toxicity. Additionally, they are administered for prolonged periods, anticipating the careful monitoring of their plasma exposure via Therapeutic Drug Monitoring (TDM) to be an important component of patients' follow-up. We have developed a liquid chromatography-tandem mass spectrometry method (LC-MS/MS) requiring 100 microL of plasma for the simultaneous determination of the six major TKIs currently in use. Plasma is purified by protein precipitation and the supernatant is diluted in ammonium formate 20 mM (pH 4.0) 1:2. Reverse-phase chromatographic separation of TKIs is obtained using a gradient elution of 20 mM ammonium formate pH 2.2 and acetonitrile containing 1% formic acid, followed by rinsing and re-equilibration to the initial solvent composition up to 20 min. Analyte quantification, using matrix-matched calibration samples, is performed by electro-spray ionization-triple quadrupole mass spectrometry by selected reaction monitoring detection using the positive mode. The method was validated according to FDA recommendations, including assessment of extraction yield, matrix effects variability (<9.6%), overall process efficiency (87.1-104.2%), as well as TKIs short- and long-term stability in plasma. The method is precise (inter-day CV%: 1.3-9.4%), accurate (-9.2 to +9.9%) and sensitive (lower limits of quantification comprised between 1 and 10 ng/mL). This is the first broad-range LC-MS/MS assay covering the major currently in-use TKIs. It is an improvement over previous methods in terms of convenience (a single extraction procedure for six major TKIs, reducing significantly the analytical time), sensitivity, selectivity and throughput. It may contribute to filling the current knowledge gaps in the pharmacokinetics/pharmacodynamics relationships of the latest TKIs developed after imatinib and better define their therapeutic ranges in different patient populations in order to evaluate whether a systematic TDM-guided dose adjustment of these anticancer drugs could contribute to minimize the risk of major adverse reactions and to increase the probability of efficient, long lasting, therapeutic response.

Douleurs chroniques chez les personnes âgées: dimensions psychologiques [Chronic pain in elderly people: psychosocial dimension].

Chronic pain in elderly people requires to take into account somatic co-morbidities as well as its psychosocial dimensions. Chronic pain often represents a distress signal addressed to the environment and the care providers. Psychological suffering or mood disorders can be presented in the form of somatic complaints often associated with functional impairments, sometimes severe. Therapeutic care has to address functionality through an image-enhancing approach aiming to summon the patients' resources. The treatment of a concomitant depressive state necessitates a true commitment from the therapist. Its benefits are documented in elderly patients. Analgesic treatment as a whole will seek in particular to restore feelings of self-esteem and help the patient recover a good quality of life.

Therapeutic drug monitoring of imatinib: Bayesian and alternative methods to predict trough levels

Background: The imatinib trough plasma concentration (C(min)) correlates with clinical response in cancer patients. Therapeutic drug monitoring (TDM) of plasma C(min) is therefore suggested. In practice, however, blood sampling for TDM is often not performed at trough. The corresponding measurement is thus only remotely informative about C(min) exposure. Objectives: The objectives of this study were to improve the interpretation of randomly measured concentrations by using a Bayesian approach for the prediction of C(min), incorporating correlation between pharmacokinetic parameters, and to compare the predictive performance of this method with alternative approaches, by comparing predictions with actual measured trough levels, and with predictions obtained by a reference method, respectively. Methods: A Bayesian maximum a posteriori (MAP) estimation method accounting for correlation (MAP-ρ) between pharmacokinetic parameters was developed on the basis of a population pharmacokinetic model, which was validated on external data. Thirty-one paired random and trough levels, observed in gastrointestinal stromal tumour patients, were then used for the evaluation of the Bayesian MAP-ρ method: individual C(min) predictions, derived from single random observations, were compared with actual measured trough levels for assessment of predictive performance (accuracy and precision). The method was also compared with alternative approaches: classical Bayesian MAP estimation assuming uncorrelated pharmacokinetic parameters, linear extrapolation along the typical elimination constant of imatinib, and non-linear mixed-effects modelling (NONMEM) first-order conditional estimation (FOCE) with interaction. Predictions of all methods were finally compared with 'best-possible' predictions obtained by a reference method (NONMEM FOCE, using both random and trough observations for individual C(min) prediction). Results: The developed Bayesian MAP-ρ method accounting for correlation between pharmacokinetic parameters allowed non-biased prediction of imatinib C(min) with a precision of ±30.7%. This predictive performance was similar for the alternative methods that were applied. The range of relative prediction errors was, however, smallest for the Bayesian MAP-ρ method and largest for the linear extrapolation method. When compared with the reference method, predictive performance was comparable for all methods. The time interval between random and trough sampling did not influence the precision of Bayesian MAP-ρ predictions. Conclusion: Clinical interpretation of randomly measured imatinib plasma concentrations can be assisted by Bayesian TDM. Classical Bayesian MAP estimation can be applied even without consideration of the correlation between pharmacokinetic parameters. Individual C(min) predictions are expected to vary less through Bayesian TDM than linear extrapolation. Bayesian TDM could be developed in the future for other targeted anticancer drugs and for the prediction of other pharmacokinetic parameters that have been correlated with clinical outcomes.

Presence of JC virus-specific CTL in the cerebrospinal fluid of PML patients: rationale for immune-based therapeutic strategies.

OBJECTIVES: There is urgent need of a treatment for progressive multifocal leukoencephalopathy (PML), caused by the polyomavirus JC (JCV). To evaluate the rationale for immunotherapy of PML, we explored whether JCV-specific cytotoxic T lymphocytes (CTL) can penetrate the central nervous system (CNS). In addition, we studied the breadth of their T-cell receptor (TCR) repertoire, and sought to establish a reliable method to expand these cells in vitro. DESIGN AND METHODS: We enrolled 18 patients in this study, including 16 with proven or possible PML (15 HIV-positive and one HIV-negative), and two HIV-positive patients with other neurological diseases. Detection of JCV-specific CTL in the blood and the cerebrospinal fluid was performed by Cr release and tetramer staining assays in 15 patients. RESULTS: Of 11 PML patients with analyzable cerebrospinal fluid (CSF), two had no detectable JCV-specific CTL in the blood and CSF and died 3.7 and 7.2 months later. The nine remaining patients had an inactive course of PML and detectable JCV-specific CTL in the blood. In addition, four of them (44%) also had detectable JCV-specific CTL in the CSF. Both HIV-positive patients with OND had detectable JCV-specific CTL in the blood and one in the CSF. Using tetramer technology, we obtained highly enriched JCV-specific CTL lines that were able to kill target cells presenting JCV peptides. The breadth of the TCR repertoire was CTL epitope dependent. CONCLUSIONS: These results indicate that JCV-specific CTL are present in the CNS of PML patients and pave the way for an immune-based therapeutic approach.

Insulin Resistance as a Therapeutic Target for Chronic Kidney Disease.

Insulin resistance (IR) is a prevalent metabolic feature in chronic kidney disease (CKD). Postreceptor insulin-signaling defects have been observed in uremia. A decrease in the activity of phosphatidylinositol 3-kinase appears critical in the pathophysiology of CKD-associated IR. Lipotoxicity due to ectopic accumulation of lipid moieties has recently emerged as another mechanism by which CKD and/or associated metabolic disorders may lead to IR through impairment of various insulin-signaling molecules. Metabolic acidosis, anemia, excess of fat mass, inflammation, vitamin D deficiency, adipokine imbalance, physical inactivity, and the accumulation of nitrogenous compounds of uremia all contribute to CKD-associated IR. The clinical impacts of IR in this setting are numerous, including endothelial dysfunction, increased cardiovascular mortality, muscle wasting, and possibly initiation and progression of CKD. This is why IR may be a therapeutic target in the attempt to improve outcomes in CKD. General measures to improve IR are directed to counteract causal factors. The use of pharmaceutical agents such as inhibitors of the renin-angiotensin system may improve IR in hypertensive and CKD patients. Pioglitazone appears a safe and promising therapeutic agent to reduce IR and uremic-associated abnormalities. However, interventional studies are needed to test if the reduction and/or normalization of IR may actually improve outcomes in these patients.

Body temperature regulation and outcome after cardiac arrest and therapeutic hypothermia.

OBJECTIVE: Therapeutic temperature modulation is recommended after cardiac arrest (CA). However, body temperature (BT) regulation has not been extensively studied in this setting. We investigated BT variation in CA patients treated with therapeutic hypothermia (TH) and analyzed its impact on outcome. METHODS: A prospective cohort of comatose CA patients treated with TH (32-34°C, 24h) at the medical/surgical intensive care unit of the Lausanne University Hospital was studied. Spontaneous BT was recorded on hospital admission. The following variables were measured during and after TH: time to target temperature (TTT=time from hospital admission to induced BT target <34°C), cooling rate (spontaneous BT-induced BT target/TTT) and time of passive rewarming to normothermia. Associations of spontaneous and induced BT with in-hospital mortality were examined. RESULTS: A total of 177 patients (median age 61 years; median time to ROSC 25 min) were studied. Non-survivors (N=90, 51%) had lower spontaneous admission BT than survivors (median 34.5 [interquartile range 33.7-35.9]°C vs. 35.1 [34.4-35.8]°C, p=0.04). Accordingly, time to target temperature was shorter among non-survivors (200 [25-363]min vs. 270 [158-375]min, p=0.03); however, when adjusting for admission BT, cooling rates were comparable between the two outcome groups (0.4 [0.2-0.5]°C/h vs. 0.3 [0.2-0.4]°C/h, p=0.65). Longer duration of passive rewarming (600 [464-744]min vs. 479 [360-600]min, p<0.001) was associated with mortality. CONCLUSIONS: Lower spontaneous admission BT and longer time of passive rewarming were associated with in-hospital mortality after CA and TH. Impaired thermoregulation may be an important physiologic determinant of post-resuscitation disease and CA prognosis. When assessing the benefit of early cooling on outcome, future trials should adjust for patient admission temperature and use the cooling rate rather than the time to target temperature.