Clinical classification of scoliosis patients using machine learning and markerless 3D surface trunk data.

Markerless 3D surface topography for scoliosis diagnosis and brace treatment can avoid repeated radiation known from standard X-ray analysis and possible side effects. Combined with the method of torso asymmetry analysis, curve severity and progression can be evaluated with high reliability. In the current study, a machine learning approach was utilised to classify scoliosis patients based on their trunk surface asymmetry pattern. Frontal X-ray and 3D scanning analysis with a clinical classification based on Cobb angle and spinal curve pattern were performed with 50 patients. Similar as in a previous study, each patient's trunk 3D reconstruction was used for an elastic registration of a reference surface mesh with fixed number of vertices. Subsequently, an asymmetry distance map between original and reflected torso was calculated. A fully connected neural network was then utilised to classify patients regarding their Cobb angle (mild, moderate, severe) and an Augmented Lehnert-Schroth (ALS) classification based on their full torso asymmetry distance map. The results reveal a classification success rate of 90% (SE: 80%, SP: 100%) regarding the curve severity (mild vs moderate-severe) and 50-72% regarding the ALS group. Identifying patient curve severity and treatment group was reasonably possible allowing for a decision support during diagnosis and treatment planning. Graphical abstract.

Innovative decision support for scoliosis brace therapy based on statistical modelling of markerless 3D trunk surface data.

Recently markerless 3D scanning methods receive an increased interest for therapy planning and brace treatment of patients with scoliosis. This avoids repeated radiation known from standard X-Ray analysis. Several authors introduced the method of asymmetry distance maps in order to classify curve severity and progression. The current work extends this approach by statistical mean shape 3D models of the human trunk in order to classify patients. 50 patients were included in this study performing frontal X-ray and 3D scanning analysis. All patients were classified by a clinician according to their Cobb angle and spinal curve pattern (Augmented-Lehnert-Schroth ALS). 3D reconstructions of each patient trunk were processed in a way to elastically register a reference surface mesh with fixed number of data points. Mean 3D shape models were generated for each curve pattern. An asymmetry distance map was then calculated for each patient and mean shape model. Single patient 3D reconstructions were classified according to severity and ALS treatment group. Optimal sensitivity and specificity was 97%/39% thoracic and 87%/42% lumbar respectively for detecting mild and moderate-severe patients. Identifying a treatment group was possible for three combined groups allowing to support decisions during diagnosis and therapy planning.

Workflow of CAD / CAM Scoliosis Brace Adjustment in Preparation Using 3D Printing.

BACKGROUND: High correction bracing is the most effective conservative treatment for patients with scoliosis during growth. Still today braces for the treatment of scoliosis are made by casting patients while computer aided design (CAD) and computer aided manufacturing (CAM) is available with all possibilities to standardize pattern specific brace treatment and improve wearing comfort. OBJECTIVE: CAD / CAM brace production mainly relies on carving a polyurethane foam model which is the basis for vacuuming a polyethylene (PE) or polypropylene (PP) brace. Purpose of this short communication is to describe the workflow currently used and to outline future requirements with respect to 3D printing technology. METHOD: Description of the steps of virtual brace adjustment as available today are content of this paper as well as an outline of the great potential there is for the future 3D printing technology. RESULTS: For 3D printing of scoliosis braces it is necessary to establish easy to use software plug-ins in order to allow adding 3D printing technology to the current workflow of virtual CAD / CAM brace adjustment. Textures and structures can be added to the brace models at certain well defined locations offering the potential of more wearing comfort without losing in-brace correction. CONCLUSIONS: Advances have to be made in the field of CAD / CAM software tools with respect to design and generation of individually structured brace models based on currently well established and standardized scoliosis brace libraries.

A phenomenological approach shows a high coherence of warming patterns in dimictic aquatic systems across latitude.

To predict the coherence in local responses to large-scale climatic forcing among aquatic systems, we developed a generalized approach to compare long-term data of dimictic water bodies based on phenomenologically defined hydrographic events. These climate-sensitive phases (inverse stratification, spring overturn, early thermal stratification, summer stagnation) were classified in a dual code (cold/warm) based on threshold temperatures. Accounting for a latitudinal gradient in seasonal timing of phases derived from gradients in cumulative irradiation (2.2 days per degree latitude), we found a high spatial and temporal coherence in warm-cold patterns for six lakes (84 %) and the Baltic Sea (78 %), even when using the same thresholds for all sites. Similarity to CW-codes for the North Sea still was up to 72 %. The approach allows prediction of phase-specific warming trends and resulting instantaneous or time-delayed ecological responses. Exemplarily, we show that warming during early thermal stratification controls food-web-mediated effects on key species during summer.

Comparing in situ particle monitoring to microscopic counts of plankton in a drinking water reservoir.

In a one-year study, the multispecies assemblages of phytoplankton (picoplankton to microplankton) within a drinking water reservoir were counted, determined and evaluated in their size fractions using microscope enumeration (MC). The manual counts were compared with the size evaluation obtained by a light obscuring particle counter (PC) in order to evaluate its use for the monitoring practice of a drinking water reservoir. With this multispecies comparison we present a novel approach for the evaluation of automated counting systems. The picoplankton clearly remained uncounted by the PC even though its lower size limits imply an adequate match. The highest and most consistent count numbers of plankton (nano- and microplankton) and particles were obtained during the spring mass development. However, from the middle of the year onwards, the measured particle concentration surpassed the counted plankton abundances by two- to threefold indicating the rise of seston within the water column. This fraction would be missed if counted solely by MC. Further, the PC consistently undersized the biological counts, but not the minerogenic fraction represented by the manganese oxidising bacteria. Consequently, the rise and decline of Metallogenium bacteria was reliably monitored with the PC. The PC provides additional size information compared to other bulk optical sensors (turbidity, chlorophyll-a). The correlation of particles with probe measurement always exceeded the plankton coefficient, but all combinations of plankton, particle and probe measurement revealed significant linear regressions. However, the redundancy of the chlorophyll-a probes was also shown in order to explain plankton abundances. Our results indicate that background knowledge of the monitored system and cautious interpretation of data is required to allocate and understand automated particle counts. Therefore, only in combination with MC, the PC enables phytoplankton or minerogenic particle counts under frequent real-time monitoring conditions. As such it may serve as a helpful tool for example in critical situations in the management of drinking water reservoirs.

Identifying cardinal dates in phytoplankton time series to enable the analysis of long-term trends.

Phenology and seasonal succession in aquatic ecosystems are strongly dependent on physical factors. In order to promote investigations into this coupling, methods of characterising annual time series of phytoplankton were derived and applied to a 31-year data set from Saidenbach Reservoir (Saxony, Germany). Field data are often scarce and irregularly sampled, particularly in the transition period from winter to spring, so reliable methods of determining cardinal dates in the time series are necessary. The proposed methods were used to determine the beginning, maximum and end of the spring mass development of phytoplankton by estimating the inflexion points (A), fitting a Weibull-type function (B) and fitting linear segments to the logarithmic values (C). For the data set from Saidenbach Reservoir, all three methods proved to be relevant to the analysis of long-term trends. Differences between the maxima determined by the different methods seemed small, but there were deviations when the maximum was related to physical factors such as ice-out. The Weibull-type fit gave the most reliable and comprehensible results and is recommended for trend analyses. For all methods, long-term analysis of the duration of the spring mass development and the duration of the spring full circulation revealed a period of consistently low values (1975-1990) followed by a period of higher values (1990-2005). These periods were also identified for the date of ice-out, although in this case there was a period of high values followed by a period of low values. A sensitivity analysis that compared results from subsampled time series with increasing time intervals indicated that a minimum of one sample every three weeks is needed to obtain reliable results.

Response of drinking-water reservoir ecosystems to decreased acidic atmospheric deposition in SE Germany: trends of chemical reversal.

This study evaluates chemical trends of seven acidified reservoirs and 22 tributaries in the Erzgebirge from 1993 to 2003. About 85% of these waters showed significantly (p < 0.05) declining concentrations of protons (-69%), nitrate (-41%), sulfate (-27%), and reactive aluminum (-50% on average). This reversal is attributed to the intense reduction of industrial SO2 and NOx emissions from formerly high levels, which declined by 99% and 82% in the German-Czech border region between 1993 and 1999. The deposition rates of protons and sulfur decreased by 70-90%. Since 1993, the dry deposition of total inorganic nitrogen diminished to a minor degree, but the wet deposition remained unchanged. The surface waters reflect a substantial decrease in Al exchange processes, a release of sulfur previously stored in soils, and an uptake of nitrate by forest vegetation. The latter effect may be supported by soil protection liming which contributed to the chemical reversal in almost 20% of the study waters.