@article{BrandTroyaKrenzeretal.2022, author = {Brand, Markus and Troya, Joel and Krenzer, Adrian and Saßmannshausen, Zita and Zoller, Wolfram G. and Meining, Alexander and Lux, Thomas J. and Hann, Alexander}, title = {Development and evaluation of a deep learning model to improve the usability of polyp detection systems during interventions}, series = {United European Gastroenterology Journal}, volume = {10}, journal = {United European Gastroenterology Journal}, number = {5}, doi = {10.1002/ueg2.12235}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-312708}, pages = {477-484}, year = {2022}, abstract = {Background The efficiency of artificial intelligence as computer-aided detection (CADe) systems for colorectal polyps has been demonstrated in several randomized trials. However, CADe systems generate many distracting detections, especially during interventions such as polypectomies. Those distracting CADe detections are often induced by the introduction of snares or biopsy forceps as the systems have not been trained for such situations. In addition, there are a significant number of non-false but not relevant detections, since the polyp has already been previously detected. All these detections have the potential to disturb the examiner's work. Objectives Development and evaluation of a convolutional neuronal network that recognizes instruments in the endoscopic image, suppresses distracting CADe detections, and reliably detects endoscopic interventions. Methods A total of 580 different examination videos from 9 different centers using 4 different processor types were screened for instruments and represented the training dataset (519,856 images in total, 144,217 contained a visible instrument). The test dataset included 10 full-colonoscopy videos that were analyzed for the recognition of visible instruments and detections by a commercially available CADe system (GI Genius, Medtronic). Results The test dataset contained 153,623 images, 8.84\% of those presented visible instruments (12 interventions, 19 instruments used). The convolutional neuronal network reached an overall accuracy in the detection of visible instruments of 98.59\%. Sensitivity and specificity were 98.55\% and 98.92\%, respectively. A mean of 462.8 frames containing distracting CADe detections per colonoscopy were avoided using the convolutional neuronal network. This accounted for 95.6\% of all distracting CADe detections. Conclusions Detection of endoscopic instruments in colonoscopy using artificial intelligence technology is reliable and achieves high sensitivity and specificity. Accordingly, the new convolutional neuronal network could be used to reduce distracting CADe detections during endoscopic procedures. Thus, our study demonstrates the great potential of artificial intelligence technology beyond mucosal assessment.}, language = {en} } @article{LuxBanckSassmannshausenetal.2022, author = {Lux, Thomas J. and Banck, Michael and Saßmannshausen, Zita and Troya, Joel and Krenzer, Adrian and Fitting, Daniel and Sudarevic, Boban and Zoller, Wolfram G. and Puppe, Frank and Meining, Alexander and Hann, Alexander}, title = {Pilot study of a new freely available computer-aided polyp detection system in clinical practice}, series = {International Journal of Colorectal Disease}, volume = {37}, journal = {International Journal of Colorectal Disease}, number = {6}, doi = {10.1007/s00384-022-04178-8}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-324459}, pages = {1349-1354}, year = {2022}, abstract = {Purpose Computer-aided polyp detection (CADe) systems for colonoscopy are already presented to increase adenoma detection rate (ADR) in randomized clinical trials. Those commercially available closed systems often do not allow for data collection and algorithm optimization, for example regarding the usage of different endoscopy processors. Here, we present the first clinical experiences of a, for research purposes publicly available, CADe system. Methods We developed an end-to-end data acquisition and polyp detection system named EndoMind. Examiners of four centers utilizing four different endoscopy processors used EndoMind during their clinical routine. Detected polyps, ADR, time to first detection of a polyp (TFD), and system usability were evaluated (NCT05006092). Results During 41 colonoscopies, EndoMind detected 29 of 29 adenomas in 66 of 66 polyps resulting in an ADR of 41.5\%. Median TFD was 130 ms (95\%-CI, 80-200 ms) while maintaining a median false positive rate of 2.2\% (95\%-CI, 1.7-2.8\%). The four participating centers rated the system using the System Usability Scale with a median of 96.3 (95\%-CI, 70-100). Conclusion EndoMind's ability to acquire data, detect polyps in real-time, and high usability score indicate substantial practical value for research and clinical practice. Still, clinical benefit, measured by ADR, has to be determined in a prospective randomized controlled trial.}, language = {en} } @article{LuxHuBenKraiemetal.2019, author = {Lux, Thomas J. and Hu, Xiawei and Ben-Kraiem, Adel and Blum, Robert and Chen, Jeremy Tsung-Chieh and Rittner, Heike L.}, title = {Regional differences in tight junction protein expression in the blood-DRG barrier and their alterations after nerve traumatic injury in rats}, series = {International Journal of Molecular Sciences}, volume = {21}, journal = {International Journal of Molecular Sciences}, number = {1}, issn = {1422-0067}, doi = {10.3390/ijms21010270}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-285029}, year = {2019}, abstract = {The nervous system is shielded by special barriers. Nerve injury results in blood-nerve barrier breakdown with downregulation of certain tight junction proteins accompanying the painful neuropathic phenotype. The dorsal root ganglion (DRG) consists of a neuron-rich region (NRR, somata of somatosensory and nociceptive neurons) and a fibre-rich region (FRR), and their putative epi-/perineurium (EPN). Here, we analysed blood-DRG barrier (BDB) properties in these physiologically distinct regions in Wistar rats after chronic constriction injury (CCI). Cldn5, Cldn12, and Tjp1 (rats) mRNA were downregulated 1 week after traumatic nerve injury. Claudin-1 immunoreactivity (IR) found in the EPN, claudin-19-IR in the FRR, and ZO-1-IR in FRR-EPN were unaltered after CCI. However, laser-assisted, vessel specific qPCR, and IR studies confirmed a significant loss of claudin-5 in the NRR. The NRR was three-times more permeable compared to the FRR for high and low molecular weight markers. NRR permeability was not further increased 1-week after CCI, but significantly more CD68\(^+\) macrophages had migrated into the NRR. In summary, NRR and FRR are different in na{\"i}ve rats. Short-term traumatic nerve injury leaves the already highly permeable BDB in the NRR unaltered for small and large molecules. Claudin-5 is downregulated in the NRR. This could facilitate macrophage invasion, and thereby neuronal sensitisation and hyperalgesia. Targeting the stabilisation of claudin-5 in microvessels and the BDB barrier could be a future approach for neuropathic pain therapy.}, language = {en} } @article{ReinholdSchwabeLuxetal.2018, author = {Reinhold, Ann Kristin and Schwabe, Joachim and Lux, Thomas J. and Salvador, Ellaine and Rittner, Heike L.}, title = {Quantitative and Microstructural Changes of the Blood-Nerve Barrier in Peripheral Neuropathy}, series = {Frontiers in Neuroscience}, volume = {12}, journal = {Frontiers in Neuroscience}, doi = {10.3389/fnins.2018.00936}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-225179}, pages = {936, 1-9}, year = {2018}, abstract = {Peripheral neuropathy is accompanied by changes in the neuronal environment. The blood-nerve barrier (BNB) is crucial in protecting the neural homeostasis: Tight junctions (TJ) seal paracellular spaces and thus prevent external stimuli from entering. In different models of neuropathic pain, the BNB is impaired, thus contributing to local damage, immune cell invasion and, ultimately, the development of neuropathy with its symptoms. In this study, we examined changes in expression and microstructural localization of two key tight junction proteins (TJP), claudin-1 and the cytoplasmic anchoring ZO-1, in the sciatic nerve of mice subjected to chronic constriction injury (CCI). Via qPCR and analysis of fluorescence immunohistochemistry, a marked downregulation of mRNA as well as decreased fluorescence intensity were observed in the nerve for both proteins. Moreover, a distinct zig-zag structure for both proteins located at cell-cell contacts, indicative of the localization of TJs, was observed in the perineurial compartment of sham-operated animals. This microstructural location in cell-cell-contacts was lost in neuropathy as semiquantified via computational analysis, based on a novel algorithm. In summary, we provide evidence that peripheral neuropathy is not only associated with decrease in relevant TJPs but also exhibits alterations in TJP arrangement and loss in barrier tightness, presumably due to internalization. Specifically, semiquantification of TJP in cell-cell-contacts of microcompartments could be used in the future for routine clinical samples of patients with neuropathy.}, language = {en} }