@phdthesis{Straub2023, author = {Straub, Anton}, title = {Wertigkeit der Digitalen Volumentomographie, Computertomographie und Magnetresonanztomographie in der Diagnostik einer Knocheninfiltration beim oralen Plattenepithelkarzinom}, doi = {10.25972/OPUS-31931}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-319318}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2023}, abstract = {Das orale Plattenepithelkarzinom (OSCC) ist eine Krebserkrankung des Menschen mit insgesamt schlechter Prognose. Die h{\"a}ufigsten Risikofaktoren f{\"u}r das Entstehen eines OSCCs sind dabei Alkohol- und Nikotinabusus. Die Vorstellung der Patient*innen erfolgt h{\"a}ufig erst in fortgeschrittenen Stadien, was lokale Destruktionen wie die Infiltration von Kieferknochen (kAD) bedingt. F{\"u}r die Festlegung des Tumorausmaßes und der Operationsplanung kommt daher der Detektion solch einer kAD eine besondere Bedeutung im Rahmen der Staginguntersuchungen beim OSCC zu. Die G{\"u}te der Computertomographie (CT) und Magnetresonanztomographie (MRT) bez{\"u}glich der Detektion einer kAD sind gut untersucht. Anders sieht es bei der Digitalen Volumentomographie (CBCT) aus. Sie scheint aufgrund des hohen Aufl{\"o}sungsverm{\"o}gens prima vista gut geeignet f{\"u}r die Detektion einer kAD beim OSCC zu sein, bisher fehlen jedoch Studien, die den Nutzen der CBCT bei der Detektion einer kAD belegen. Es ist daher Ziel vorliegender Arbeit, zu zeigen, dass die CBCT der CT und MRT bei der Detektion einer kAD nicht unterlegen ist. Im Rahmen vorliegender Arbeit wurden daher die CBCT-, CT- und MRT-Aufnahmen von 153 Patient*innen, die an einem OSCC erkrankt sind, mit einem 3-Punkte System (0 = keine kAD, 1 = kn{\"o}cherne Arrosion und 2 = kn{\"o}cherne Destruktion) nachuntersucht. Zur {\"U}berpr{\"u}fung der bildgebenden Verfahren wurde die histologische Untersuchung als Goldstandard oder ein Follow-Up {\"u}ber mindestens sechs Monate herangezogen. Das Ergebnis ergab f{\"u}r die CBCT eine Sensitivit{\"a}t von 95,6 \%, eine Spezifit{\"a}t von 87,0 \% und eine Accuracy von 89,5 \%. F{\"u}r die CT lagen diese Werte bei 84,4 \%, 91,7 \% und 89,0 \% und f{\"u}r die MRT bei 88,9 \%, 91,7 \% und 90,8 \%. Es konnte damit gezeigt werden, dass alle drei bildgebenden Verfahren mit hoher Sensitivit{\"a}t und Spezifit{\"a}t eine kAD detektieren k{\"o}nnen. Kein Verfahren war einem anderen {\"u}ber- oder unterlegen. Die Haupthypothese vorliegender Arbeit, dass die CBCT weder der CT noch der MRT bei der Detektion einer kAD unterlegen ist, wurde best{\"a}tigt und angenommen. Die Daten sprechen weiter daf{\"u}r, dass die CBCT routinem{\"a}ßig zur Detektion einer kAD herangezogen werden kann. Dies ist besonders zu unterst{\"u}tzen, da die CBCT fr{\"u}hzeitig, im besten Fall bei Erstvorstellung der Patient*innen, durchgef{\"u}hrt werden und Aufschluss {\"u}ber das operative Ausmaß geben kann.}, subject = {Plattenepithelcarcinom}, language = {de} } @article{StraubBrandsBorgmannetal.2022, author = {Straub, Anton and Brands, Roman and Borgmann, Anna and Vollmer, Andreas and Hohm, Julian and Linz, Christian and M{\"u}ller-Richter, Urs and K{\"u}bler, Alexander C. and Hartmann, Stefan}, title = {Free skin grafting to reconstruct donor sites after radial forearm flap harvesting: a prospective study with platelet-rich fibrin (PRF)}, series = {Journal of Clinical Medicine}, volume = {11}, journal = {Journal of Clinical Medicine}, number = {12}, issn = {2077-0383}, doi = {10.3390/jcm11123506}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-278854}, year = {2022}, abstract = {Reconstruction of the donor site after radial forearm flap harvesting is a common procedure in maxillofacial plastic surgery. It is normally carried out with split-thickness or full-thickness free skin grafts. Unfortunately, free skin graft transplantation faces wound healing impairments such as necrosis, (partial) graft loss, or tendon exposure. Several studies have investigated methods to reduce these impairments and demonstrated improvements if the wound bed is optimised, for example, through negative-pressure wound therapy or vacuum-assisted closure. However, these methods are device-dependent, expansive, and time-consuming. Therefore, the application of platelet-rich fibrin (PRF) to the wound bed could be a simple, cost-effective, and device-independent method to optimise wound-bed conditions instead. In this study, PRF membranes were applied between the wound bed and skin graft. Results of this study indicate improvements in the PRF versus non-PRF group (93.44\% versus 86.96\% graft survival, p = 0.0292). PRF applied to the wound bed increases graft survival and reduces impairments. A possible explanation for this is the release of growth factors, which stimulate angiogenesis and fibroblast migration. Furthermore, the solid PRF membranes act as a mechanical barrier ("lubrication" layer) to protect the skin graft from tendon motion. The results of this study support the application of PRF in donor-site reconstruction with free skin grafts.}, language = {en} } @article{VollmerVollmerLangetal.2022, author = {Vollmer, Andreas and Vollmer, Michael and Lang, Gernot and Straub, Anton and K{\"u}bler, Alexander and Gubik, Sebastian and Brands, Roman C. and Hartmann, Stefan and Saravi, Babak}, title = {Performance analysis of supervised machine learning algorithms for automatized radiographical classification of maxillary third molar impaction}, series = {Applied Sciences}, volume = {12}, journal = {Applied Sciences}, number = {13}, issn = {2076-3417}, doi = {10.3390/app12136740}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-281662}, year = {2022}, abstract = {Background: Oro-antral communication (OAC) is a common complication following the extraction of upper molar teeth. The Archer and the Root Sinus (RS) systems can be used to classify impacted teeth in panoramic radiographs. The Archer classes B-D and the Root Sinus classes III, IV have been associated with an increased risk of OAC following tooth extraction in the upper molar region. In our previous study, we found that panoramic radiographs are not reliable for predicting OAC. This study aimed to (1) determine the feasibility of automating the classification (Archer/RS classes) of impacted teeth from panoramic radiographs, (2) determine the distribution of OAC stratified by classification system classes for the purposes of decision tree construction, and (3) determine the feasibility of automating the prediction of OAC utilizing the mentioned classification systems. Methods: We utilized multiple supervised pre-trained machine learning models (VGG16, ResNet50, Inceptionv3, EfficientNet, MobileNetV2), one custom-made convolutional neural network (CNN) model, and a Bag of Visual Words (BoVW) technique to evaluate the performance to predict the clinical classification systems RS and Archer from panoramic radiographs (Aim 1). We then used Chi-square Automatic Interaction Detectors (CHAID) to determine the distribution of OAC stratified by the Archer/RS classes to introduce a decision tree for simple use in clinics (Aim 2). Lastly, we tested the ability of a multilayer perceptron artificial neural network (MLP) and a radial basis function neural network (RBNN) to predict OAC based on the high-risk classes RS III, IV, and Archer B-D (Aim 3). Results: We achieved accuracies of up to 0.771 for EfficientNet and MobileNetV2 when examining the Archer classification. For the AUC, we obtained values of up to 0.902 for our custom-made CNN. In comparison, the detection of the RS classification achieved accuracies of up to 0.792 for the BoVW and an AUC of up to 0.716 for our custom-made CNN. Overall, the Archer classification was detected more reliably than the RS classification when considering all algorithms. CHAID predicted 77.4\% correctness for the Archer classification and 81.4\% for the RS classification. MLP (AUC: 0.590) and RBNN (AUC: 0.590) for the Archer classification as well as MLP 0.638) and RBNN (0.630) for the RS classification did not show sufficient predictive capability for OAC. Conclusions: The results reveal that impacted teeth can be classified using panoramic radiographs (best AUC: 0.902), and the classification systems can be stratified according to their relationship to OAC (81.4\% correct for RS classification). However, the Archer and RS classes did not achieve satisfactory AUCs for predicting OAC (best AUC: 0.638). Additional research is needed to validate the results externally and to develop a reliable risk stratification tool based on the present findings.}, language = {en} } @article{SchmitzKodererElMeseryetal.2021, author = {Schmitz, Werner and Koderer, Corinna and El-Mesery, Mohamed and Gobik, Sebastian and Sampers, Rene and Straub, Anton and K{\"u}bler, Alexander Christian and Seher, Axel}, title = {Metabolic fingerprinting of murine L929 fibroblasts as a cell-based tumour suppressor model system for methionine restriction}, series = {International Journal of Molecular Sciences}, volume = {22}, journal = {International Journal of Molecular Sciences}, number = {6}, issn = {1422-0067}, doi = {10.3390/ijms22063039}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-259198}, year = {2021}, abstract = {Since Otto Warburg reported in 1924 that cancer cells address their increased energy requirement through a massive intake of glucose, the cellular energy level has offered a therapeutic anticancer strategy. Methionine restriction (MetR) is one of the most effective approaches for inducing low-energy metabolism (LEM) due to the central position in metabolism of this amino acid. However, no simple in vitro system for the rapid analysis of MetR is currently available, and this study establishes the murine cell line L929 as such a model system. L929 cells react rapidly and efficiently to MetR, and the analysis of more than 150 different metabolites belonging to different classes (amino acids, urea and tricarboxylic acid cycle (TCA) cycles, carbohydrates, etc.) by liquid chromatography/mass spectrometry (LC/MS) defines a metabolic fingerprint and enables the identification of specific metabolites representing normal or MetR conditions. The system facilitates the rapid and efficient testing of potential cancer therapeutic metabolic targets. To date, MS studies of MetR have been performed using organisms and yeast, and the current LC/MS analysis of the intra- and extracellular metabolites in the murine cell line L929 over a period of 5 days thus provides new insights into the effects of MetR at the cellular metabolic level.}, language = {en} } @article{BoschertTeuschAljasemetal.2020, author = {Boschert, Verena and Teusch, Jonas and Aljasem, Anwar and Schmucker, Philipp and Klenk, Nicola and Straub, Anton and Bittrich, Max and Seher, Axel and Linz, Christian and M{\"u}ller-Richter, Urs D. A. and Hartmann, Stefan}, title = {HGF-induced PD-L1 expression in head and neck cancer: preclinical and clinical findings}, series = {International Journal of Molecular Sciences}, volume = {21}, journal = {International Journal of Molecular Sciences}, number = {20}, issn = {1422-0067}, doi = {10.3390/ijms21228770}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-236220}, year = {2020}, abstract = {Head and neck squamous cell carcinoma (HNSCC) is a widespread disease with a low survival rate and a high risk of recurrence. Nowadays, immune checkpoint inhibitor (ICI) treatment is approved for HNSCC as a first-line treatment in recurrent and metastatic disease. ICI treatment yields a clear survival benefit, but overall response rates are still unsatisfactory. As shown in different cancer models, hepatocyte growth factor/mesenchymal-epithelial transition (HGF/Met) signaling contributes to an immunosuppressive microenvironment. Therefore, we investigated the relationship between HGF and programmed cell death protein 1 (PD-L1) expression in HNSCC cell lines. The preclinical data show a robust PD-L1 induction upon HGF stimulation. Further analysis revealed that the HGF-mediated upregulation of PD-L1 is MAP kinase-dependent. We then hypothesized that serum levels of HGF and soluble programmed cell death protein 1 (sPD-L1) could be potential markers of ICI treatment failure. Thus, we determined serum levels of these proteins in 20 HNSCC patients before ICI treatment and correlated them with treatment outcomes. Importantly, the clinical data showed a positive correlation of both serum proteins (HGF and sPD-L1) in HNSCC patient's sera. Moreover, the serum concentration of sPD-L1 was significantly higher in ICI non-responsive patients. Our findings indicate a potential role for sPD-L1 as a prognostic marker for ICI treatment in HNSCC.}, language = {en} } @article{VollmerSaraviVollmeretal.2022, author = {Vollmer, Andreas and Saravi, Babak and Vollmer, Michael and Lang, Gernot Michael and Straub, Anton and Brands, Roman C. and K{\"u}bler, Alexander and Gubik, Sebastian and Hartmann, Stefan}, title = {Artificial intelligence-based prediction of oroantral communication after tooth extraction utilizing preoperative panoramic radiography}, series = {Diagnostics}, volume = {12}, journal = {Diagnostics}, number = {6}, issn = {2075-4418}, doi = {10.3390/diagnostics12061406}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-278814}, year = {2022}, abstract = {Oroantral communication (OAC) is a common complication after tooth extraction of upper molars. Profound preoperative panoramic radiography analysis might potentially help predict OAC following tooth extraction. In this exploratory study, we evaluated n = 300 consecutive cases (100 OAC and 200 controls) and trained five machine learning algorithms (VGG16, InceptionV3, MobileNetV2, EfficientNet, and ResNet50) to predict OAC versus non-OAC (binary classification task) from the input images. Further, four oral and maxillofacial experts evaluated the respective panoramic radiography and determined performance metrics (accuracy, area under the curve (AUC), precision, recall, F1-score, and receiver operating characteristics curve) of all diagnostic approaches. Cohen's kappa was used to evaluate the agreement between expert evaluations. The deep learning algorithms reached high specificity (highest specificity 100\% for InceptionV3) but low sensitivity (highest sensitivity 42.86\% for MobileNetV2). The AUCs from VGG16, InceptionV3, MobileNetV2, EfficientNet, and ResNet50 were 0.53, 0.60, 0.67, 0.51, and 0.56, respectively. Expert 1-4 reached an AUC of 0.550, 0.629, 0.500, and 0.579, respectively. The specificity of the expert evaluations ranged from 51.74\% to 95.02\%, whereas sensitivity ranged from 14.14\% to 59.60\%. Cohen's kappa revealed a poor agreement for the oral and maxillofacial expert evaluations (Cohen's kappa: 0.1285). Overall, present data indicate that OAC cannot be sufficiently predicted from preoperative panoramic radiography. The false-negative rate, i.e., the rate of positive cases (OAC) missed by the deep learning algorithms, ranged from 57.14\% to 95.24\%. Surgeons should not solely rely on panoramic radiography when evaluating the probability of OAC occurrence. Clinical testing of OAC is warranted after each upper-molar tooth extraction.}, language = {en} } @article{VollmerVollmerLangetal.2022, author = {Vollmer, Andreas and Vollmer, Michael and Lang, Gernot and Straub, Anton and Shavlokhova, Veronika and K{\"u}bler, Alexander and Gubik, Sebastian and Brands, Roman and Hartmann, Stefan and Saravi, Babak}, title = {Associations between periodontitis and COPD: An artificial intelligence-based analysis of NHANES III}, series = {Journal of Clinical Medicine}, volume = {11}, journal = {Journal of Clinical Medicine}, number = {23}, issn = {2077-0383}, doi = {10.3390/jcm11237210}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-312713}, year = {2022}, abstract = {A number of cross-sectional epidemiological studies suggest that poor oral health is associated with respiratory diseases. However, the number of cases within the studies was limited, and the studies had different measurement conditions. By analyzing data from the National Health and Nutrition Examination Survey III (NHANES III), this study aimed to investigate possible associations between chronic obstructive pulmonary disease (COPD) and periodontitis in the general population. COPD was diagnosed in cases where FEV (1)/FVC ratio was below 70\% (non-COPD versus COPD; binary classification task). We used unsupervised learning utilizing k-means clustering to identify clusters in the data. COPD classes were predicted with logistic regression, a random forest classifier, a stochastic gradient descent (SGD) classifier, k-nearest neighbors, a decision tree classifier, Gaussian naive Bayes (GaussianNB), support vector machines (SVM), a custom-made convolutional neural network (CNN), a multilayer perceptron artificial neural network (MLP), and a radial basis function neural network (RBNN) in Python. We calculated the accuracy of the prediction and the area under the curve (AUC). The most important predictors were determined using feature importance analysis. Results: Overall, 15,868 participants and 19 feature variables were included. Based on k-means clustering, the data were separated into two clusters that identified two risk characteristic groups of patients. The algorithms reached AUCs between 0.608 (DTC) and 0.953\% (CNN) for the classification of COPD classes. Feature importance analysis of deep learning algorithms indicated that age and mean attachment loss were the most important features in predicting COPD. Conclusions: Data analysis of a large population showed that machine learning and deep learning algorithms could predict COPD cases based on demographics and oral health feature variables. This study indicates that periodontitis might be an important predictor of COPD. Further prospective studies examining the association between periodontitis and COPD are warranted to validate the present results.}, language = {en} } @article{WinterSchulzSchmitteretal.2022, author = {Winter, Anna and Schulz, Stefan M. and Schmitter, Marc and Brands, Roman C. and Straub, Anton and K{\"u}bler, Alexander and Borgmann, Anna and Hartmann, Stefan}, title = {Oral-health-related quality of life in patients with medication-related osteonecrosis of the jaw: a prospective clinical study}, series = {International Journal of Environmental Research and Public Health}, volume = {19}, journal = {International Journal of Environmental Research and Public Health}, number = {18}, issn = {1660-4601}, doi = {10.3390/ijerph191811709}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-288141}, year = {2022}, abstract = {Medication-related osteonecrosis of the jaw (MRONJ) represents an adverse side effect of antiresorptive and antiangiogenic medications. It is associated with impaired quality of life, oral health, and oral function and can be classified into various stages. The purpose of this prospective clinical study is to evaluate the impact of stages I and II MRONJ on oral-health-related quality of life (OHRQoL) and related parameters. Patients' OHRQoL, satisfaction with life, oral discomfort, and oral health were assessed using the German version of the Oral Health Impact Profile (OHIP-G49), visual analog scales (VAS), and Satisfaction with Life Scale (SWLS) at baseline (T0), 10 days (T1), and 3 months after treatment (T2) in 36 patients. Data were analyzed using Kolmogorov-Smirnov test, two-way mixed ANOVAs, and follow-up Mann-Whitney U tests. The impact of treatment effects on the original seven OHIP domain structures and the recently introduced four-dimensional OHIP structure were evaluated using linear regression analysis. Thirty-six patients received surgical MRONJ treatment. Before treatment, patients' perceived OHRQoL, oral discomfort, oral health, and satisfaction with life were negatively affected by MRONJ. Surgical treatment significantly improved OHRQoL and related parameters (all p ≤ 0.012). This improvement was greater in patients with higher impairment at T0. OHRQoL and oral restrictions were still impaired after treatment in patients who needed prosthetic treatment. The four-dimensional structure revealed valuable information beyond the standard seven OHIP domains. Increased awareness of MRONJ risks and an interdisciplinary treatment approach for MRONJ patients are needed.}, language = {en} } @article{StraubStapfFischeretal.2022, author = {Straub, Anton and Stapf, Maximilian and Fischer, Markus and Vollmer, Andreas and Linz, Christian and L{\^a}m, Thi{\^e}n-Tr{\´i} and K{\"u}bler, Alexander and Brands, Roman C. and Scherf-Clavel, Oliver and Hartmann, Stefan}, title = {Bone concentration of ampicillin/sulbactam: a pilot study in patients with osteonecrosis of the jaw}, series = {International Journal of Environmental Research and Public Health}, volume = {19}, journal = {International Journal of Environmental Research and Public Health}, number = {22}, issn = {1660-4601}, doi = {10.3390/ijerph192214917}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-297413}, year = {2022}, abstract = {Osteonecrosis of the jaw (ONJ) occurs typically after irradiation of the head and neck area or after the intake of antiresorptive agents. Both interventions can lead to compromised bone perfusion and can ultimately result in infection and necrosis. Treatment usually consists of surgical necrosectomy and prolonged antibiotic therapy, usually through beta-lactams such as ampicillin/sulbactam. The poor blood supply in particular raises the question as to whether this form of antibiosis can achieve sufficient concentrations in the bone. Therefore, we investigated the antibiotic concentration in plasma and bone samples in a prospective study. Bone samples were collected from the necrosis core and in the vital surrounding bone. The measured concentrations in plasma for ampicillin and sulbactam were 126.3 ± 77.6 and 60.2 ± 35.0 µg/mL, respectively. In vital bone and necrotic bone samples, the ampicillin/sulbactam concentrations were 6.3 ± 7.8/1.8 ± 2.0 µg/g and 4.9 ± 7.0/1.7 ± 1.7 µg/g, respectively. These concentrations are substantially lower than described in the literature. However, the concentration seems sufficient to kill most bacteria, such as Streptococci and Staphylococci, which are mostly present in the biofilm of ONJ. We, therefore, conclude that intravenous administration of ampicillin/sulbactam remains a valuable treatment in the therapy of ONJ. Nevertheless, increasing resistance of Escherichia coli towards beta-lactam antibiotics have been reported and should be considered.}, language = {en} } @article{VollmerVollmerLangetal.2023, author = {Vollmer, Andreas and Vollmer, Michael and Lang, Gernot and Straub, Anton and K{\"u}bler, Alexander and Gubik, Sebastian and Brands, Roman C. and Hartmann, Stefan and Saravi, Babak}, title = {Automated assessment of radiographic bone loss in the posterior maxilla utilizing a multi-object detection artificial intelligence algorithm}, series = {Applied Sciences}, volume = {13}, journal = {Applied Sciences}, number = {3}, issn = {2076-3417}, doi = {10.3390/app13031858}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-305050}, year = {2023}, abstract = {Periodontitis is one of the most prevalent diseases worldwide. The degree of radiographic bone loss can be used to assess the course of therapy or the severity of the disease. Since automated bone loss detection has many benefits, our goal was to develop a multi-object detection algorithm based on artificial intelligence that would be able to detect and quantify radiographic bone loss using standard two-dimensional radiographic images in the maxillary posterior region. This study was conducted by combining three recent online databases and validating the results using an external validation dataset from our organization. There were 1414 images for training and testing and 341 for external validation in the final dataset. We applied a Keypoint RCNN with a ResNet-50-FPN backbone network for both boundary box and keypoint detection. The intersection over union (IoU) and the object keypoint similarity (OKS) were used for model evaluation. The evaluation of the boundary box metrics showed a moderate overlapping with the ground truth, revealing an average precision of up to 0.758. The average precision and recall over all five folds were 0.694 and 0.611, respectively. Mean average precision and recall for the keypoint detection were 0.632 and 0.579, respectively. Despite only using a small and heterogeneous set of images for training, our results indicate that the algorithm is able to learn the objects of interest, although without sufficient accuracy due to the limited number of images and a large amount of information available in panoramic radiographs. Considering the widespread availability of panoramic radiographs as well as the increasing use of online databases, the presented model can be further improved in the future to facilitate its implementation in clinics.}, language = {en} }