A Generative Adversarial Inpainting Network to Enhance Prediction of Periodontal Clinical Attachment Level

Deep learning algorithms has recently been used to determine clinical attachment levels (CAL) which aid in the diagnosis of periodontal disease. However, the limited field-of-view of dental bitewing x-rays poses a challenge for convolutional neural networks (CNN) because out-of-view anatomy cannot be directly considered. This study presents an inpainting algorithm using generative adversarial networks (GANs) coupled with partial convolutions to predict out-of-view anatomy to enhance CAL prediction accuracy. 80,326 images were used for training, 12,901 images were used for validation and 10,687 images were used to compare non-inpainted methods to inpainted methods for CAL predictions. Statistical analyses were conducted using mean bias error (MBE), mean absolute error (MAE) and Dunn’s pairwise test comparing CAL at p=0.05. Comparator p-values demonstrated statistically significant improvement in CAL prediction accuracy between corresponding inpainted and non-inpainted methods with a MAE of 1.04mm and 1.50mm respectively. The Dunn’s pairwise test indicated a statistically significant improvement in CAL prediction accuracy between both inpainted methods compared to their non-inpainted counterparts, with the best performing methods achieving a Dunn’s pairwise value of -63.89. This study demonstrates the superiority of using a generative adversarial inpainting network with partial convolutions to predict CAL from bitewing images.

Interspecific agonistic behaviour of Odontotermes javanicus and Microcerotermes sp. (Isoptera: Termitidae): Preliminary study

Termite interaction from the same or different colonies or species can display a variety of agonistic behaviours. The study aimed to observe the interspecific agonistic behaviour of Odontotermes javanicus and Microcerotermes sp. For this purpose, an assay was carried out using a test arena separated into two groups. Termite responses were tested against each other in the following pairwise test (worker versus worker, worker versus soldier, and soldier versus soldier) with a 1:1 ratio (ten individuals) from each colony. Specimens were collected at the area of Forestry Faculty, Universitas Hasanuddin, Indonesia. The observed agonistic behaviour of both species is dominated by examination (antennation) and aggression (attacking, grappling, and biting). There were three treatment combinations that classified as no aggression, except for the treatment (Microcerotermes sp. soldier versus O. javanicus worker) that classified in aggression level 3 (aggressive). Based on termite survival after five minutes of testing, O. javanicus is more competitive than Microcerotermes sp,; no more individual termites of both species survive after 24 hours of testing.

The vegetation of rich fens (Sphagno warnstorfii-Tomentypnion nitentis) at the southeastern margins of their European range

Question: Rich fens of the Sphagno warnstorfii-Tomentypnion nitentis alliance require a specific combination of base richness and climate to occur. Their rarity at the southeastern margins of their European range has previously prevented rigorous vegetation classification. We asked how many associations may be delimited here and whether some of them are restricted to the high Balkan Mountains showing high endemicity. Study area: Entire territories of Bulgaria and Romania. Methods: We compiled all available vegetation-plot records, including some hitherto unprocessed data. We classified them by both divisive (modified TWINSPAN) and agglomerative (beta-flexible clustering) numerical classification method, with OPTIMCLASS1 applied to set the number of clusters. A semi-supervised approach (k-means) was additionally applied to confirm the classification of Southern-Carpathian (Romania) rich fens, where some Balkan taxa occur. Differences in base richness and elevation were tested by one-way ANOVA with Tukey’s pairwise test. Results: Three associations were delimited and all three occur in Bulgaria, from where only one association had been previously reported. Two associations characterised by Sphagnum contortum and Balkan and Southern-European species occur in Bulgaria, but not in Romania, one at lower elevations around 1,200 m, and one at higher elevations around 2,000 m where pH is lower. One lower-elevation (around 1,300 m) association with S. warnstorfii and S. teres is shared between Romania, Bulgaria and Central Europe. Conclusions: We have described a new high-mountain association, with two subassociations that differ by successional stage and dominant peat moss species (S. contortum and S. warnstorfii, respectively). These subassociations could be reconsidered when more data from other Balkan countries are available. Rich fens in southeastern Europe are rare, have a diverse vegetation, and are deserving of the further attention of nature conservation authorities and vegetation scientists. Taxonomic reference: The nomenclature was harmonized following The Euro+Med PlantBase (Euro+Med 2021) for vascular plants and Hill et al. (2006) for bryophytes, except of Angelica pancicii that is accepted as a separate taxon in Bulgaria (Andreev et al. 1992; Delipavlov et al. 2003). Critical taxa, not always reliably differentiated in the field and in literary sources, were merged to aggregates: Alchemilla vulgaris agg. (all Alchemilla species), Anthoxanthum odoratum agg. (A. alpinum, A. odoratum), Molinia caerulea agg. (M. arundinacea subsp. arundinacea, M. arundinacea subsp. freyi, M. caerulea), Palustriella commutata agg. (P. commutata, P. falcata), Plagiomnium affine agg. (P. affine, P. elatum, P. ellipticum), Sphagnum palustre agg. (S. centrale, S. palustre). Syntaxonomic reference: Peterka et al. (2017) for alliances.

Statistical approach for …a test chemical is considered to be positive… in regulatory toxicology: Trend and pairwise tests

In regulatory toxicology an outcome is claimed positive when both a trend is significant and any pairwise test against control. Two statistical approaches are proposed: a joint Dunnett and Williams test (assuming the dose as a qualitative factor) and a joint test of the Tukey regression test and Dunnett test (assuming the dose as a quantitative covariate). Related R software is available.

Combination of Different Fecal Immunochemical Tests in Colorectal Cancer Screening: Any Gain in Diagnostic Performance?

A variety of fecal immunochemical tests (FITs) are used for colorectal cancer screening. FIT performance could be improved further. It is unclear, whether the combination of different FITs with different analytical characteristics (such as, different antibodies for the detection of fecal hemoglobin) can yield a better diagnostic performance. Fecal samples were obtained from 2042 participants of screening colonoscopy. All participants with advanced neoplasm (AN, colorectal cancer (n = 16) or advanced adenoma (n = 200)) and 300 randomly selected participants without AN were included. Nine quantitative FITs were evaluated simultaneously. Sensitivity and specificity was calculated for single tests (n = 9) and for their pairwise test combinations (n = 36) (requiring either both FITs (P++) or at least one FIT (P+) to be positive for defining a positive test result). Mean age of the participants (n = 516) was 63 (range: 50–79) years and 56% were men. At cutoffs yielding a specificity of 96.7% for single FITs, the median gain in specificity by P++ combination was +1.0%, whereas the median loss in sensitivity for AN was −4.2%. For P+ combination the median gain in sensitivity for AN was +2.8%, at a prize of median loss of −1.0% of specificity. Combinations of different FITs do not yield any relevant gain in diagnostic performance.