Qualitative Evaluation of Common Quantitative Metrics for Clinical Acceptance of Automatic Segmentation: a Case Study on Heart Contouring from CT Images by Deep Learning Algorithms.

Organs-at-risk contouring is time consuming and labour intensive. Automation by deep learning algorithms would decrease the workload of radiotherapists and technicians considerably. However, the variety of metrics used for the evaluation of deep learning algorithms make the results of many papers difficult to interpret and compare. In this paper, a qualitative evaluation is done on five established metrics to assess whether their values correlate with clinical usability. A total of 377 CT volumes with heart delineations were randomly selected for training and evaluation. A deep learning algorithm was used to predict the contours of the heart. A total of 101 CT slices from the validation set with the predicted contours were shown to three experienced radiologists. They examined each slice independently whether they would accept or adjust the prediction and if there were (small) mistakes. For each slice, the scores of this qualitative evaluation were then compared with the Sørensen-Dice coefficient (DC), the Hausdorff distance (HD), pixel-wise accuracy, sensitivity and precision. The statistical analysis of the qualitative evaluation and metrics showed a significant correlation. Of the slices with a DC over 0.96 (N = 20) or a 95% HD under 5 voxels (N = 25), no slices were rejected by the readers. Contours with lower DC or higher HD were seen in both rejected and accepted contours. Qualitative evaluation shows that it is difficult to use common quantification metrics as indicator for use in clinic. We might need to change the reporting of quantitative metrics to better reflect clinical acceptance.

Understanding groundwater salinization mechanisms to secure freshwater resources in the water-scarce city of Maputo, Mozambique.

In this study hydrochemical, isotopic and multivariate statistical tools are combined with a recharge analysis and existing geophysical data to improve understanding of major factors controlling freshwater occurrence and the origins of high salinities in the multi-layered coastal aquifer system of the Great Maputo area in Mozambique. Access to freshwater in this semi-arid area is limited by an inefficient public supply network, scarce surface waters, long droughts and an increasing population growth. Groundwater has a large potential to enhance water security, but its exploitation is threatened by both coastal and inland salinization mechanisms that are poorly understood. A GIS approach is utilized to classify potential recharge zones based on hydrogeological properties and land use/cover, whereas potential recharge rates are estimated through a root zone water balance method. In combination with water stable isotope data results reveal that extreme rainfall events provide the most relevant contributions to recharge, and interception and evaporation play an important role in the low recharge areas. Hierarchical clustering of hydrochemical and isotopic data allows the classification of six water groups, varying from fresh to brackish/salt waters. Corresponding scatter plots and PHREEQC modelling show evaporation and mixing with seawater (up to 5%) as major processes affecting salinity in the area. The co-occurrence of high alkalinity and Cl concentrations, in combination with piezometric and geo-electrical data, suggests that: 1) inland brackish/salt groundwater is caused by mixing with seawater trapped within clay layers; and 2) brackish/salt surface waters result from seepage of brackish groundwater into rivers and wetlands, followed by evaporation, hence increasing salinity and δ18O values. Mixing with small fractions of trapped seawater as main salinity source, rather than halite dissolution, is further corroborated by Br/Cl ratios of brackish/salt water samples near the ocean ratio. Cation exchange upon salinization is mainly observed in the semi-confined aquifer, while freshening takes place in the phreatic aquifer, particularly in areas presenting high recharge rates.

The distribution of arsenic in shallow alluvial groundwater under agricultural land in central Portugal: insights from multivariate geostatistical modeling.

In this study multivariate and geostatistical methods are jointly applied to model the spatial and temporal distribution of arsenic (As) concentrations in shallow groundwater as a function of physicochemical, hydrogeological and land use parameters, as well as to assess the related uncertainty. The study site is located in the Mondego River alluvial body in Central Portugal, where maize, rice and some vegetable crops dominate. In a first analysis scatter plots are used, followed by the application of principal component analysis to two different data matrices, of 112 and 200 samples, with the aim of detecting associations between As levels and other quantitative parameters. In the following phase explanatory models of As are created through factorial regression based on correspondence analysis, integrating both quantitative and qualitative parameters. Finally, these are combined with indicator-geostatistical techniques to create maps indicating the predicted probability of As concentrations in groundwater exceeding the current global drinking water guideline of 10 µg/l. These maps further allow assessing the uncertainty and representativeness of the monitoring network. A clear effect of the redox state on the presence of As is observed, and together with significant correlations with dissolved oxygen, nitrate, sulfate, iron, manganese and alkalinity, points towards the reductive dissolution of Fe (hydr)oxides as the essential mechanism of As release. The association of high As values with rice crop, known to promote reduced environments due to ponding, further corroborates this hypothesis. An additional source of As from fertilizers cannot be excluded, as the correlation with As is higher where rice is associated with vegetables, normally associated with higher fertilization rates. The best explanatory model of As occurrence integrates the parameters season, crop type, well and water depth, nitrate and Eh, though a model without the last two parameters also gives quite satisfactory results.

A groundwater perspective on the river basin management plan for central Portugal - developing a methodology to assess the potential impact of N fertilizers on groundwater bodies.

The Water Framework Directive establishes that the river basin management plans must have a summary of the pressures and impacts of human activities, such as agriculture, on the chemical and quantitative status of groundwater bodies. In order to identify those areas where a potential impact from agricultural activities on groundwater bodies exists, but currently lacking groundwater monitoring data, a methodology was developed that combines the use of gross nitrogen balance values with the results of a specific vulnerability assessment index. A farm management efficiency parameter is added, to identify the factors that contribute to nitrogen use efficiency and to assess the near-future scenarios. This methodology allows the identification of significant pressures that may be responsible for a groundwater body failing good status where there is no representative monitoring network.

Environmental impacts on soil and groundwater at airports: origin, contaminants of concern and environmental risks.

Environmental impacts of airports are similar to those of many industries, though their operations expand over a very large area. Most international impact assessment studies and environmental management programmes have been giving less focus on the impacts to soil and groundwater than desirable. This may be the result of the large attention given to air and noise pollution, relegating other environmental descriptors to a second role, even when the first are comparatively less relevant. One reason that contributes to such "biased" evaluation is the lack of systematic information about impacts to soil and groundwater from airport activities, something the present study intends to help correct. Results presented here include the review of over seven hundred documents and online databases, with the objective of obtaining the following information to support environmental studies: (i) which operations are responsible for chemical releases?; (ii) where are these releases located?; (iii) which contaminants of concern are released?; (iv) what are the associated environmental risks? Results showed that the main impacts occur as a result of fuel storage, stormwater runoff and drainage systems, fuel hydrant systems, fuel transport and refuelling, atmospheric deposition, rescue and fire fighting training areas, winter operations, electrical substations, storage of chemical products by airport owners or tenants, and maintenance of green areas. A new method for ranking environmental risks of organic substances, based on chemical properties, is proposed and applied. Results show that the contaminants with the highest risks are the perfluorochemicals, benzene, trichloroethylene and CCl(4). The obtained information provides a basis for establishing the planning and checking phases of environmental management systems, and may also help in the best design of pollution prevention measures in order to avoid or reduce significant environmental impacts from airports.

Major issues regarding the efficiency of monitoring programs for nitrate contaminated groundwater.

Major issues regarding the efficiency of moni toring programs for nitrate contaminated groundwater are analyzed in this paper: (i) representativeness of monitoring networks; (ii) correct interpretation of the monitoring data and resulting time series and trends; and (iii) differentiation among the different sources of nitrates in groundwater. Following an overview of the nitrate contamination problem and possible solutions, as well as some of the difficulties found, a relatively straightforward method for assessing monitoring network representativity is presented, namely interpolation standard error assessment. It is shown how nitrate-concentration time series resulting from periodic observations can be corrected with a conservative tracer, in order to avoid misinterpretation and confirm or correct apparent trends. Finally, coupled ¹5N and ¹8O isotope signatures of nitrate (NO3⁻) in groundwater are used to differentiate among nitrogen (N) sources, to ensure correct targeting of restoration measures. The case study regards a Nitrate Vulnerable Zone in the south of Portugal, designated in compliance with the European Nitrates Directive, where coastal discharge of nutrient-rich groundwater threatens the good qualitative and ecological status of the Ria Formosa coastal lagoon. Results show that mineral fertilizer is the main source of N in groundwater, and that increases in N load can be masked by dilution phenomena.