University Hospital Dresden Research Articles

Using machine learning algorithms to detect fear of falling in people with multiple sclerosis in standardized gait analysis

Introduction: Multiple sclerosis (MS) is the most common chronic inflammatory disease of the central nervous system. The progressive impairment of gait is one of the most important pathognomic symptoms which are associated with falls and fear of falling (FOF) in people with MS (pwMS). 60 % of pwMS show a FOF, which leads to restrictions in mobility as well as physical activity and reduces the quality of life in general. Therefore, early detection of FOF is crucial because it enables early implementation of rehabilitation strategies as well as clinical decision-making to reduce progression. Qualitative and quantitative evaluation of gait pattern is an essential aspect of disease assessment and can provide valuable insights for personalized treatment decisions in pwMS. Our objective was to identify the most appropriate clinical gait analysis methods to identify FOF in pwMS and to detect the optimal machine learning (ML) algorithms to predict FOF using the complex multidimensional data from gait analysis.Methods: Data of 1240 pwMS was recorded at the MS Centre of the University Hospital Dresden between November 2020 and September 2021. Patients performed a multidimensional gait analysis with pressure and motion sensors, as well as patient-reported outcomes (PROs), according to a standardized protocol. A feature selection ensemble (FS-Ensemble) was developed to improve the classification performance. The FS-Ensemble consisted of four filtering methods: Chi-square test, information gain, minimum redundancy maximum relevance and ReliefF. Gaussian Naive Bayes, Decision Tree, k-Nearest Neighbor, and Support Vector Machines (SVM) were used to identify FOF.Results: The descriptive analysis showed that 37 % of the 1240 pwMS had a FOF (n = 458; age: 51 ± 16 years, 76 % women, median EDSS: 4.0). The FS-Ensemble improved classification performance in most cases. The SVM showed the best performance of the four classification models in detecting FOF. The PROs showed the best F1 scores (Early Mobility Impairment Questionnaire F1 = 0.81 ± 0.00 and 12-item Multiple Sclerosis Scale F1 = 0.80 ± 0.00).Conclusion: FOF is an important psychological risk factor associated with an increased risk of falls. To integrate a functional early warning system for fall detection into MS management and progression monitoring, it is necessary to detect the relevant gait parameters as well as assessment methods. In this context, ML strategies allow the integration of gait parameters from clinical routine to support the initiation of early rehabilitation measures and adaptation of course-modifying therapeutics. The results of this study confirm that patients' self-assessments play an important role in disease management.

Survival implications and radiographic predictors of prognosis in patients with IDH-mutant astrocytoma with CDKN2A/b homozygous deletion: An international, multi-centered study.

e14035 Background: Isocitrate Dehydrogenase (IDH) 1 or 2 mutations are associated with improved prognosis in glioma patients. Despite their less aggressive clinical course when compared to IDH-wild type gliomas, molecularly-defined IDH-mutant glioma subtypes with poor prognosis have been identified. Among these, CDKN2A/B homozygous deletion represents the strongest negative genetic predictor of prognosis yet identified in IDH-mutant astrocytoma. With the WHO 2021 classification recognizing CDKN2A/B homozygous deletion as sufficient for the designation of astrocytoma, IDH-mutant, WHO Grade 4, the translational implications and potential for therapies targeting CDKN2A/B homozygous deletion are an area of active investigation. We sought to further characterize overall (OS) and progression-free survival (PFS) of CDKN2A/B deleted patients within a large, multi-centered, international cohort and identify clinical characteristics that predict survival for patients with CDKN2A/B deletion. Methods: Patients with IDH-mutant glioma with CDKN2A/B homozygous deletion were retrospectively identified across four institutions (Massachusetts General Hospital, USA; University Hospital Heidelberg, Germany; University Hospital Dresden, Germany, and Foudan-Shanghai, China). IDH-mutation status was determined by immunohistochemistry or next-generation sequencing. CDKN2A/B status was determined by hybridized DNA capture or FISH. Results: We identified 75 patients with IDH-mutant astrocytoma with CDKN2A/B homozygous deletion. The cohort had a median follow-up time of 5.0 years, was 56% male, and had an average age of 38.8 years. Patients were treated with chemotherapy and radiation following initial surgery in 77% of cases. Preoperative contrast enhancement and T2 FLAIR mismatch were identified in 56 and 44% of cases, respectively. Compared to a cohort of 51 patients with CDKN2A/B intact, grade 4, IDH-mutant astrocytoma, those with CDKN2A/B-deleted tumors had significantly shorter median OS and PFS (OS: 4 vs 13.5 years, HR = 6.8, HR 95% CI = 3.1, 14.6, p < 0.0001; PFS: 2.8 vs 5.9 years, HR = 3.0, HR 95% CI = 1.7, 5.2, p = 0.0001). Within the CDKN2A/B deleted cohort, preoperative contrast enhancement predicted worse OS, but not PFS (OS: HR = 2.7, 95% CI = 1.1, 7.4, p = 0.04; PFS: HR = 1.17, 95% CI = 0.7, 3.7, p = 0.24). Conclusions: These data reinforce the strong negative prognostic impact of CDKN2A/B homozygous deletion on OS and PFS and identify a novel association between preoperative contrast enhancement and shorter OS in patients with CDKN2A/Bdeleted, IDH-mutant astrocytoma.

Endovascular aortic repair with sac embolization for the prevention of type II endoleaks (the EVAR-SE study): study protocol for a randomized controlled multicentre study in Germany

BackgroundBeyond a certain threshold diameter, abdominal aortic aneurysms (AAA) are to be treated by open surgical or endovascular aortic aneurysm repair (EVAR). In a quarter of patients who undergo EVAR, inversion of blood flow in the inferior mesenteric artery or lumbar arteries may lead to type II endoleak (T2EL), which is associated with complications (e.g. AAA growth, secondary type I endoleak, rupture). As secondary interventions to treat T2EL often fail and may be highly invasive, prevention of T2EL is desirable. The present study aims to assess the efficacy of sac embolization (SE) with metal coils during EVAR to prevent T2EL in patients at high risk.MethodsOver a 24-month recruitment period, a total of 100 patients undergoing EVAR in four vascular centres (i.e. Klinikum rechts der Isar of the Technical University of Munich, University Hospital Augsburg, University Hospital Dresden, St. Joseph’s Hospital Wiesbaden) are to be included in the present study. Patients at high risk for T2EL (i.e. ≥ 5 efferent vessels covered by endograft or aneurysmal thrombus volume <40%) are randomized to one group receiving standard EVAR and another group receiving EVAR with SE. Follow-up assessments postoperatively, after 30 days, and 6 months involve contrast-enhanced ultrasound scans (CEUS) and after 12 months an additional computed tomography angiography (CTA) scan. The presence of T2EL detected by CEUS or CTA after 12 months is the primary endpoint. Secondary endpoints comprise quality of life (quantified by the SF-36 questionnaire), reintervention rate, occurrence of type I/III endoleak, aortic rupture, death, alteration of aneurysm volume, or diameter. Standardized evaluation of CTA scans happens through a core lab. The study will be terminated after the final follow-up visit of the ultimate patient.DiscussionAlthough preexisting studies repeatedly indicated a beneficial effect of SE on T2EL rates after EVAR, patient relevant outcomes have not been assessed until now. The present study is the first randomized controlled multicentre study to assess the impact of SE on quality of life. Further unique features include employment of easily assessable high-risk criteria, a contemporary follow-up protocol, and approval to use any commercially available coil material. Overcoming limitations of previous studies might help SE to be implemented in daily practice and to enhance patient safety.Trial registration ClinicalTrials.gov NCT05665101. Registered on 23 December 2022.

Development and Validation of a Model for Postpancreatectomy Hemorrhage Risk

Postpancreatectomy hemorrhage (PPH) due to postoperative pancreatic fistula (POPF) is a life-threatening complication after pancreatoduodenectomy. However, there is no prediction tool for early identification of patients at high risk of late PPH. To develop and validate a prediction model for PPH. This retrospective prognostic study included consecutive patients with clinically relevant POPF who underwent pancreatoduodenectomy from January 1, 2009, to May 20, 2023, at the University Hospital Mannheim (derivation cohort), and from January 1, 2012, to May 31, 2022, at the University Hospital Dresden (validation cohort). Data analysis was performed from May 30 to July 29, 2023. Clinical and radiologic features of PPH. Accuracy of a predictive risk score of PPH. A multivariate prediction model-the hemorrhage risk score (HRS)-was established in the derivation cohort (n = 139) and validated in the validation cohort (n = 154). A total of 293 patients (187 [64%] men; median age, 69 [IQR, 60-76] years) were included. The HRS comprised 4 variables with associations: sentinel bleeding (odds ratio [OR], 35.10; 95% CI, 5.58-221.00; P < .001), drain fluid culture positive for Candida species (OR, 14.40; 95% CI, 2.24-92.20; P < .001), and radiologic proof of rim enhancement of (OR, 12.00; 95% CI, 2.08-69.50; P = .006) or gas within (OR, 12.10; 95% CI, 2.22-65.50; P = .004) a peripancreatic fluid collection. Two risk categories were identified with patients at low risk (0-1 points) and high risk (≥2 points) to develop PPH. Patients with PPH were predicted accurately in the derivation cohort (C index, 0.97) and validation cohort (C index 0.83). The need for more invasive PPH management (74% vs 34%; P < .001) and severe complications (49% vs 23%; P < .001) were more frequent in high-risk patients compared with low-risk patients. In this retrospective prognostic study, a robust prediction model for PPH was developed and validated. This tool may facilitate early identification of patients at high risk for PPH.

How patients become pilots of their disease

‚Patients need to know what the data is for, what the data means for them personally and, of course, for other patients,‘ explained Prof. Tjalf Ziemssen, MD, Director of the Centre for Clinical Neurosciences and Deputy Clinic Director of the Clinic and Polyclinic for Neurology at the Centre for Clinical Neurosciences at the Neurological University Hospital Dresden at the Roche Pharma DKVF Symposium entitled ‚Data use as an end in itself or for the benefit of the patient?‘. According to Ziemssen, it has often been the case in the past that although attempts have been made to obtain more and more data from patients in order to conduct research and introduce innovations, this is usually done without the patient benefiting directly or being able to understand what the data means for them personally. As a negative example, Mr Ziemssen cited the electronic patient file, which would have no structure that the patient would not be able to deal with operationally. The concept of the ‚digital twin‘, which he propagates, is different (see interview page 22 ff).

„Der ,Digitale Zwilling‘ kann die MS-Versorgung wirklich verbessern“

Prof. Dr Tjalf Ziemssen sees himself as a digital pioneer with a lot of passion. Alongside and parallel to his day-to-day work at the Centre for Clinical Neurosciences at the Neurological University Hospital Dresden, treating patients with multiple sclerosis, he and his team are developing the concept of the ‚digital MS twin‘. This can be thought of as a virtual copy, a digital dashboard or, in modern parlance, an avatar of a patient that is able to quantitatively describe the individual disease. The beauty of this is that a combination of artificial intelligence, machine learning and data analysis enables digital twins to uncover problems before they occur and optimise processes.

Explainable End-to-End Supervised Learning Identifies Myelodysplastic Neoplasms in Bone Marrow Smears

Myelodysplastic neoplasms (MDS) encompass clonal myeloid malignancies that are characterized by ineffective hematopoiesis, cytopenia, myelodysplasia, and recurrent genetic events. Apart from genetic abnormalities and the assessment of peripheral blood for cytopenia, cytomorphologic evaluation of bone marrow morphology is key in initial diagnosis and response assessment as well as to detect disease transformation to acute myeloid leukemia (AML). However, the identification of aberrant morphologies in bone marrow cells is challenging and prone to inter-observer variability even for seasoned morphologists resulting in discrepancies between local site and central review. After the introduction of deep learning (DL) models for computer vision tasks, a wide variety of applications in imaging diagnostics has been addressed using convolutional neural networks including the assessment of bone marrow cell types and prediction of hematological malignancies. Bone marrow smear (BMS) images of manually selected regions of interest (ROI) from 483 MDS patients treated at the University Hospital Dresden were captured at 40-fold magnification with a resolution of 2560*1920 pixels and pre-processed. In parallel, BMS images from two control cohorts were taken: (i) 1226 AML patients registered in the German Study Alliance Leukemia (SAL) registry, and (ii) 236 healthy bone marrow donors. In addition, an external validation cohort of 50 MDS patients was obtained from the Munich Leukemia Laboratory (MLL). Image-level labels were provided denoting the corresponding diagnosis: “MDS”, “AML” and “healthy donor”. To balance different sample sizes between the datasets, simple image augmentation techniques such as random sized cropping, color shifting, and linear transformations were used. Training and testing were performed with 5-fold cross-validation, i.e. multiple train-test-splits of 80:20. Importantly, cell-level labels, e.g. specifically annotated cell types or features like distinct dysplastic morphologies were not used to train our model as we intended to abstract visual information exclusively through our neural networks instead of using expert-guided feature engineering. Six different recently introduced convolutional neural network (CNN) architectures were compared (ResNet-34/50/101/152, ResNeXt-50/101, ShuffleNet V2, DenseNet-121/169/201, Wide ResNet-50/101, and SqueezeNet 1.1). To evaluate individual model performance, we used accuracy, recall (=sensitivity), precision (=positive predictive value), and area-under-the-curve (AUC) of both the receiver-operating-characteristic (ROC) and precision-recall-curve (PRC). Binary classification tasks showed high accuracy of 0.956 and 0.981 for distinguishing MDS from healthy donors and AML, respectively (Table I). Results were validated in the external MDS cohort reaching an accuracy of 0.916 for MDS vs. donors and 0.968 for MDS vs. AML (Table I). For explainability, we used occlusion sensitivity mapping which iteratively blocks image areas from being evaluated by the CNNs, measures consecutive drops in prediction performance, and thus identifies image areas of high importance (highlighted in green). By reviewing images, we found the CNNs to target nuclei not only of dysplastic cells but rather of all available cells in a ROI (Figure 1). Using supervised end-to-end deep learning, we developed a software framework to distinguish between MDS, AML, and healthy controls with very high accuracy. Importantly, models do not require laborious feature engineering, i.e. manually drafted, cell-level labels alleviating the need for time-consuming and cost-ineffective manual labeling and thereby bypassing a key bottleneck for current standards of computer vision in microscopy.

Treosulfan-Based Conditioning Prior to Allogeneic Hematopoietic Stem Cell Transplantation (alloHSCT) for Patients with Myelodysplastic Syndrome (MDS): Comparative Analysis of a Randomized Controlled Trial and Real-World Data

Introduction MDS is a clonal myeloid disorder manifested by diverse genotypes and phenotypes, characterized by ineffective hematopoiesis. AlloHSCT is the only potential curative therapy for MDS. Myeloablative conditioning regimens are integral part of alloHSCT that allow engraftment and provide a therapeutic effect. Treosulfan has myeloablative and immunosuppressive properties and provides less transplantation related mortality than some of the traditionally used myeloablative conditioning regimens like busulfan-based treatments. In our analysis, we compared the results from the MDS subgroup of the randomized controlled trial (RCT) MC-FludT.14/L (NCT00822393) with real-world data (RWD) of patients treated at the University Hospital Münster and the University Hospital Dresden. Methods MDS patients conditioned with treosulfan (30 g/m²) and fludarabine in the RCT (June 2013 to January 2018) and the data obtained from 2 hospitals (RWD) (August 2017 to February 2023) were compared. Baseline characteristics, overall survival, relapse-free survival, relapse (or progression), non-relapse mortality, time to engraftment, and safety profiles between the RCT and RWD were compared. A sensitivity analysis was conducted using propensity score matching (PSM). Additionally, subgroup analyses evaluated the clinical effectiveness of treosulfan by baseline morphologic blast count and the revised International Prognostic Scoring System (IPSS-R). Results Overall, 195 MDS patients (RCT 84, RWD 111) with at least 12 months follow-up were analyzed. Median age was 62 years (range 39-76 years). Baseline characteristics only differed between the RCT and RWD with respect to age, IPSS-R categories, Hematopoietic Cell Transplantation Co-Morbidity Index total score (HCT-CI), and neutrophil count, indicating some heterogeneity between the RCT and RWD patients. In the PSM analysis, 106 (RCT 53, RWD 53) patients were matched to address the heterogeneity. The Kaplan-Meier estimates for overall survival (OS) at 2 years were 73% (95% CI 62-81%) and 72% (95% CI 62-81%) for the RCT and RWD (Figure 1), and for relapse-free survival (RFS) at 2 years were 68% (95% CI 57-77%) and 67% (95% CI 56-76%), respectively. Cumulative incidence of relapse at 2 years was 11% (95% CI 4-18%) and 15% (95% CI 7-23%), and for non-relapse mortality (NRM) was 21% (95% CI 12-30%) and 18% (95% CI 10-26%) for the RCT and RWD, respectively. In addition, cumulative incidence of neutrophil engraftment at 28 days was 93% (95% CI 87-98%) and 94% (95% CI 89-98%) for the RCT and RWD, while the median engraftment time was later in the RCT (19 days vs 15 days). Cumulative incidence of platelet engraftment at 28 days was 89% (95% CI 83-96%) and 93% (95% CI 88-98%) for the RCT and RWD, with the same median time 14 days. Further, the PSM sensitivity analysis showed comparable efficacy between the RCT and RWD. Interestingly, bone marrow blast count prior to conditioning and IPSS-R subgroups only had minor, statistically not significant impact on survival endpoints (Table 1). The number of deaths were 27 (32%) and 28 (25%) in the RCT and RWD, respectively. The most frequent cause of death was transplantation related (20% vs 14%). One patient experienced primary graft failure in the RCT and none in the RWD. More patients had acute GVHD with at least grade 2 in the RCT (27%) than in RWD (18%). Conclusions The efficacy and safety of treosulfan-based conditioning therapy prior to alloHSCT in post-authorization real world setting was comparable with the results obtained from the RCT. Our data suggest that bone marrow blast count is no strong predictor for survival after alloHSCT. Therefore, we conclude that the approved treosulfan-based conditioning therapy for patients with MDS, including those with increased blasts, has promising outcomes after alloHSCT.

Quantification of timeloss within complex tumor therapies.

126 Background: There are various studies on reducing the waiting time of tumor patients before therapy sessions. In this study, we investigated how much waiting time within therapy sessions occurs due to delayed administration of the next drug. In addition, digital documented infusion time of specific drugs were determined to calculate waiting time. Both metrics serve as an indication of the potential to improve the effectiveness and efficiency of the infusion process. Methods: The therapy sequence of preliminary 29 patients in the Outpatient Therapy Center of the University Hospital Dresden was investigated. PDF export of the prescriptions of all cycles of the patient with the documentation of the therapy sequences from BD Cato is performed. In python the extraction of all time stamps and the processing of all start and stop times of the medication administration took place. Based on this data, waiting time between infusions were calculated. In addition, analysis of the total administration time of the specific medication was performed. Results: This revealed a median of 23 minutes of patient waiting time between infusion sequences (n = 29 patients and 1154 measured infusion changes). We compared 7 specific medications to the intended infusion time, taking particular note of the high variance in duration. We postulate due to standard administration protocols that the infusion time variance indicates additional waiting time. Both metrics are only an indication, since other factors besides the waiting time can also influence the time between two infusions. Conclusions: Treatment time for the same therapy protocol varied widely between patients, both in terms of waiting time between infusions and in the duration of infusion compared to the intended infusion time. Both indicate the opportunity to improve efficiency, thus providing more hospital-free lifetime for patients and increase revenue and standardization for treatment centers.

Selective Decontamination of the Digestive Tract in Pancreatic Head Resections-A Propensity Score-Matched Analysis.

(1) Background: The postoperative morbidity rate after pancreatic head resection remains high, partly due to infectious complications. The primary aim of this study was to analyze the influence of selective decontamination of the digestive tract (SDD) on the postoperative infection rate after pancreatic surgery. (2) Methods: From January 2019, the standard of care for patients undergoing pancreatic head resections at the Department for Visceral, Thoracic, and Vascular Surgery, University Hospital Dresden was the preoperative oral administration of SDD. The influence of SDD was evaluated for patients operated on between January 2019 and June 2020 in comparison to a propensity score-matched cohort, extracted from an existing database including all pancreatic resections from 2012 to 2018. The primary endpoint of the study was the shift of the bacterial load on the intraoperative bile swab test. The secondary endpoint was the association of SDD with postoperative complications. (3) Results: In total, 200 patients either with SDD (n = 100; 50%) or without SDD (non-SDD, n = 100; 50%) were analyzed. In the patient group without a preoperative biliary stent, 44% (n = 11) of the non-SDD group displayed positive bacterial results, whereas that was the case for only 21.7% (n = 10) in the SDD group (p = 0.05). Particularly, Enterobacter species (spp.) were reduced from 41.2% (n = 14) (non-SDD group) to 23.5% (n = 12) (SDD group) (p = 0.08), and Citrobacter spp. were reduced by 13.7% (p = 0.09) from the non-SDD to the SDD cohort. In patients with a preoperative biliary stent, the Gram-negative Enterobacter spp. were significantly reduced from 52.2% (n = 12) in the non-SDD group to 26.8% (n = 11) in the SDD group (p = 0.04). Similarly, Citrobacter spp. decreased by 20.6% from 30.4% (n = 7) to 9.8% (n = 4) in the non-SDD compared to the SDD group (p = 0.04). In general, deep fluid collection and abscesses occurred more frequently in the non-SDD group (36%; n = 36 vs. 27%; n = 27; p = 0.17). (4) Conclusions: Adoption of SDD before pancreatic head surgery may reduce the bacterial load in bile fluid. SDD administration does not significantly affect the postoperative infectious complication rate after pancreatic head resections.

Do certain surgical steps increase postoperative morbidity after cytoreductive surgery and HIPEC- a retrospective analysis

IntroductionIt has been shown that cytoreductive surgery (CRS) in combination with hyperthermic intraperitoneal chemotherapy (HIPEC) is an effective treatment for patients suffering from peritoneal malignancies. Despite good results, there is an ongoing debate about this treatment due to perioperative morbidity.The aim of this study is to identify relevant risk factors for an unfavorable postoperative outcome after CRS and HIPEC. Materials and methodsA retrospective analysis of a prospectively recorded database of all patients undergoing CRS and HIPEC between 2013 and 2020 in the Department of Surgery of the University Hospital Dresden was performed with a special focus on certain surgical steps of multivisceral resection, one- or 2- stage CRS/HIPEC and underlying diagnosis as possible risk factors for worse postoperative course. ResultsN = 173 CRS and HIPEC procedures were performed for various diagnoses. Relevant postoperative morbidity was 24% and 30d-mortality 1.2%. Simultaneous liver resections, preoperative hypalbuminemia and 2-staged CRS/HIPEC were significant risk factors for a worse postoperative course in multivariable analysis. Assessment of the association of simultaneous anastomoses and morbidity and mortality was inconclusive. ConclusionCRS and HIPEC is a safe treatment without relevant intraoperative morbidity and mortality and acceptable postoperative outcome. One-stage CRS/HIPEC should be preferred.

Extreme obesity is a strong predictor for in-hospital mortality and the prevalence of long-COVID in severe COVID-19 patients with acute respiratory distress syndrome

Acute Respiratory Distress Syndrome (ARDS) is common in COVID-19 patients and is associated with high mortality. The aim of this observational study was to describe patients’ characteristics and outcome, identifying potential risk factors for in-hospital mortality and for developing Long-COVID symptoms. This retrospective study included all patients with COVID-19 associated ARDS (cARDS) in the period from March 2020 to March 2021 who were invasively ventilated at the intensive care unit (ICU) of the University Hospital Dresden, Germany. Between October 2021 and December 2021 patients discharged alive (at minimum 6 months after hospital discharge—midterm survival) were contacted and interviewed about persistent symptoms possibly associated with COVID-19 as well as the quality of their lives using the EQ-5D-5L-questionnaire. Long-COVID was defined as the occurrence of one of the symptoms at least 6 months after discharge. Risk factors for mortality were assessed with Cox regression models and risk factors for developing Long-COVID symptoms by using relative risk (RR) regression. 184 Patients were included in this study (male: n = 134 (73%), median age 67 (range 25–92). All patients were diagnosed with ARDS according to the Berlin Definition. 89% of patients (n = 164) had severe ARDS (Horovitz-index < 100 mmHg). In 27% (n = 49) extracorporeal membrane oxygenation was necessary to maintain gas exchange. The median length of in-hospital stay was 19 days (range 1–60). ICU mortality was 51%, hospital mortality 59%. Midterm survival (median 11 months) was 83% (n = 55) and 78% (n = 43) of these patients presented Long-COVID symptoms with fatigue as the most common symptom (70%). Extreme obesity (BMI > 40 kg/m2) was the strongest predictor for in-hospital mortality (hazard ratio: 3.147, confidence interval 1.000–9.897) and for developing Long-COVID symptoms (RR 1.61, confidence interval 1.26–2.06). In-hospital mortality in severe cARDS patients was high, but > 80% of patients discharged alive survived the midterm observation period. Nonetheless, most patients developed Long-COVID symptoms. Extreme obesity with BMI > 40 kg/m2 was identified as independent risk factor for in-hospital mortality and for developing Long-COVID symptoms.Trial registration DRKS-ID DRKS00027856.

Characteristics and outcomes of sepsis patients with and without COVID-19

BackgroundThe aim of this study was to describe and compare clinical characteristics and outcomes in critically ill septic patients with and without COVID-19. MethodsFrom February 2020 to March 2021, patients from surgical and medical ICUs at the University Hospital Dresden were screened for sepsis. Patient characteristics and outcomes were assessed descriptively. Patient survival was analyzed using the Kaplan-Meier estimator. Associations between in-hospital mortality and risk factors were modeled using robust Poisson regression, which facilitates derivation of adjusted relative risks. ResultsIn 177 ICU patients treated for sepsis, COVID-19 was diagnosed and compared to 191 septic ICU patients without COVID-19. Age and sex did not differ significantly between sepsis patients with and without COVID-19, but SOFA score at ICU admission was significantly higher in septic COVID-19 patients. In-hospital mortality was significantly higher in COVID-19 patients with 59% compared to 29% in Non-COVID patients. Statistical analysis resulted in an adjusted relative risk for in-hospital mortality of 1.74 (95%-CI=1.35–2–24) in the presence of COVID-19 compared to other septic patients. Age, procalcitonin maximum value over 2 ng/ml, need for renal replacement therapy, need for invasive ventilation and septic shock were identified as additional risk factors for in-hospital mortality. ConclusionCOVID-19 was identified as independent risk factor for higher in-hospital mortality in sepsis patients. The need for invasive ventilation and renal replacement therapy as well as the presence of septic shock and higher PCT should be considered to identify high-risk patients.

Burnout or Depression? - Field Experience from a University Outpatient Clinic

With the increasing presence of burnout symptoms in clinical and social everyday life, it is necessary to closely examine patients with symptoms of burnout and to classify their respective symptoms in the spectrum of mental disorders. The sample includes 83 patients who presented themselves in a burnout consultation of the university outpatient clinic of the University Hospital Dresden. As part of the diagnostic examination, the patients completed the Structured Clinical Interview (SKID) and a comprehensive questionnaire diagnostic (PHQ-D; MBI, BMI) as well as a clinical interview. The clinical interview led to a burnout diagnosis in 35% of the patients, although for 1/3 of these patients no SKID diagnosis could be affirmed. However, a large proportion of the patients presenting at the university outpatient clinic suffered from a mental disorder, with depressive disorders accounting for the majority of complaints. Patients with burnout differed in a few points from patients with depressive symptoms: They scored lower in the MBI questionnaire and in the scales of mental and physical stress symptoms (p=.031) and boreout (p=.037). In addition, patients diagnosed with burnout were significantly less likely to report previous mental disorders requiring treatment (p=.017). Previous standardized diagnostic instruments are less able to differentiate between burnout and depressive disorders. The present study supports the hypothesis that burnout and depression are part of the same spectrum of symptoms, with patients with burnout being less distressed than patients with depressive disorders. There appears to be a significant overlap between depression and symptoms of burnout making the differentiation difficult in everyday clinical practice and calling for a accurate diagnostic.

Short-term preoperative drainage is associated with improved postoperative outcomes compared to that of long-term biliary drainage in pancreatic surgery

PurposeThe treatment of choice for patients presenting with obstructive cholestasis due to periampullary carcinoma is oncologic resection without preoperative biliary drainage (PBD). However, resection without PBD becomes virtually impossible in patients with obstructive cholangitis or severely impaired liver cell function. The appropriate duration of drainage by PBD has not yet been defined for these patients.MethodsA retrospective analysis was conducted on 170 patients scheduled for pancreatic resection following biliary drainage between January 2012 and June 2018 at the University Hospital Dresden in Germany. All patients were deemed eligible for inclusion, regardless of the underlying disease entity. The primary endpoint analysis was defined as the overall morbidity (according to the Clavien-Dindo classification). Secondary endpoints were the in-hospital mortality and malignancy adjusted overall and recurrence-free survival rates.ResultsA total of 170 patients were included, of which 45 (26.5%) and 125 (73.5%) were assigned to the short-term (< 4 weeks) and long-term (≥ 4 weeks) preoperative drainage groups, respectively.Surgical complications (Clavien-Dindo classification > 2) occurred in 80 (47.1%) patients, with significantly fewer complications observed in the short-term drainage group (31.1% vs. 52%; p = 0.02). We found that long-term preoperative drainage (unadjusted OR, 3.386; 95% CI, 1.507–7.606; p < 0.01) and periampullary carcinoma (unadjusted OR, 5.519; 95% CI, 1.722–17.685; p-value < 0.01) were independent risk factors for postoperative morbidity, based on the results of a multivariate regression model. The adjusted overall and recurrence-free survival did not differ between the groups (p = 0.12).ConclusionPBD in patients scheduled for pancreatic surgery is associated with substantial perioperative morbidity. Our results indicate that patients who have undergone PBD should be operated on within 4 weeks after drainage.

Diagnostic Red Blood Cell Distribution Width As a Prognostic Factor in Acute Myeloid Leukemia

Introduction: A high red blood cell distribution width (RDW) mirrors dysregulated erythrocyte homeostasis and is associated with a higher mortality in several solid cancers and myelodysplastic syndrome (MDS). A high RDW may also identify individuals more likely to develop a MDS or an acute myeloid leukemia (AML). The prognostic relevance of the RDW in patients diagnosed with AML has not been previously evaluated.Methods: We analyzed 294 newly diagnosed AML patients (median age at diagnosis 61, range 14-77 years [y]) that underwent an allogeneic hematopoietic stem cell transplantation (HSCT) at the University Hospital Leipzig in complete remission (CR, 63%), CR with incomplete peripheral recovery (CRi, 18%) or with active disease (19%). The RDW was measured at diagnosis and a cut-off of 20.7% was determined using the R package “OptimalCutpoints” to define patients with a low (87%) or high (13%) diagnostic RDW. Disease risk was assessed according to European LeukemiaNet (ELN) 2017. Median follow up after HSCT was 3.0 y. To verify the results irrespective of the consolidating therapy, a validation set of 392 AML patients (median age at diagnosis 63, range 18-87 years) treated at the University Hospital Dresden was analyzed. Here, 46% of patients received chemotherapy alone and patients that underwent allogeneic HSCT were censored at the time of HSCT.Results: The diagnostic RDW was above the local upper limit of normal (> 15%) in the majority (73%) of AML patients. The RDW as continuous parameter was higher in patients with AML developing from an antecedent myeloid disorder (sAML, after MDS, MPN, or MDS/MPN) than in patients with de novo (P=.01) or treatment-related AML (tAML, after lymphatic or solid neoplasm, P=.02, Figure 1A). A higher RDW was also observed in patients harboring gene mutations previously linked to AML of secondary origin (i.e. mutated JAK2 [P=.03], ASXL1 [P=.004], or spliceosome mutation [P=.03, compromising SF3B1, SRSF2, U2AF1, and ZRSR2, Figure 1B], which was particularly driven by SRSF2 mutations [P=.002]). Adapting an optimal cut, a higher RDW (> 20.7%) was associated with a higher non-relapse mortality (NRM, P=.02), shorter overall survival (OS, P=.009, Figure 1C), but similar relapse incidence (P=.96). In multivariate analyses, the RDW retained its prognostic significance for a higher NRM after adjustment for age at diagnosis. A RDW > 20.7% was also associated with lower hemoglobin levels at diagnosis and a trend for lower incidence of NPM1 mutations (P=.06), which are enriched in de novo AML. In contrast, RDW levels were not associated with chromosomal abnormalities (P=.72) or ELN2017 disease risk (P=.55). Despite known association of a high RDW with a higher cardiovascular/inflammatory risk, neither the comorbidity index at HSCT (HCT-CI, P=.77) nor the incidence of an acute (P=1) or chronic (P=.63) graft-versus-host disease differed according to the diagnostic RDW. Comparably, in the validation set, higher RDW levels were observed in patients with sAML compared to de novo (P=.02) or tAML (P=.02). Introducing the established 20.7% cut into the validation set, patients with high RDW levels again had lower hemoglobin levels at diagnosis (P<.001), less NPM1 mutations (P=.005) and no different ELN2017 disease risk (P=.25). A high RDW in the validation set was associated with a significantly shorter OS (P<.001, Figure 1D) and shorter event-free survival (P=.05).Conclusions: Our study is the first to evaluate the diagnostic RDW in AML patients. Patients with a secondary AML or secondary AML-associated gene mutations had higher diagnostic RDW levels. The presence of a high diagnostic RDW identifies patients with worse outcomes, irrespective of the ELN2017 risk classification and the applied consolidation therapy. This cost-effective and fast to assess parameter may help to personalize AML patient treatment. [Display omitted] DisclosuresVucinic: Gilead: Honoraria, Other: Travel Sponsoring; Novartis: Honoraria; MSD: Honoraria; Abbvie: Honoraria, Other: Travel Sponsoring; Janssen: Honoraria, Other: Travel Sponsoring. Backhaus: Bayer: Other: Current Employment of Family Member. Franke: BMS: Honoraria; Pfizer: Honoraria; Novartis: Honoraria; MSD: Honoraria; Jazz Pharmaceuticals: Honoraria, Other: Travel Sponsoring; Gilead: Honoraria, Other: Travel Sponsoring. Platzbecker: Janssen: Honoraria; Celgene/BMS: Honoraria; AbbVie: Honoraria; Novartis: Honoraria; Takeda: Honoraria; Geron: Honoraria. Schwind: Novartis: Honoraria, Research Funding; Pfizer: Honoraria. Jentzsch: Astellas: Honoraria; Jazz Pharmaceuticals: Honoraria; Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees.

An Automated Electronic Screening Tool (DETECT) for the Detection of Potentially Irreversible Loss of Brain Function.

A major reason for the low number of organ donors in Germany is a deficit in the recognition of patients who may have impending irreversible loss of brain function (ILBF) in hospitals capable of organ retrieval. We used anonymized data from the German Organ Procurement Organization (Deutsche Stiftung Organtransplantation, DSO) to compare two 12-month periods (a reference period and an evaluation period) before and after the implementation of an electronic screening tool (DETECT) at the University Hospital Dresden (UKD) with four other university hospitals without tool implementation (comparative cohort). DETECT is intended to aid in the recognition of potentially impending ILBF. The study endpoints encompassed patients with potentially unrecognized ILBF, patients with recognized ILBF, organ donations performed, and reports to the DSO. Changes in absolute risk were compared with Breslow-Day tests. 309 patients who died with primary or secondary brain lesions were identified in the UKD in the reference and evaluation periods (164 and 145 patients, respectively), and 1060 (529, 531) in the comparative cohort. In the UKD, the number of unrecognized cases of possibly impending ILBF was 14/164 (8.54%) in the reference period and 1/145 (0.69%) in the evaluation period, yielding an absolute reduction of 7.85% (95% confidence interval [--3.36; --12.33]); by contrast, in the comparative cohort, there was a 0.55% absolute increase between the two periods ([--2.21; 3.30]; p = 0.002 for the comparison between the two cohorts). Only minor differences in absolute risk change were seen with regard to the probability of recognized ILBF (7.09% [0.29; 13.88] vs. 2.42% [1.18; 6.01]; p = 0.234), organ donation (4.70% [--0.89; 10.28] vs. 0.55% [--2.17; 3.26]; p = 0.214), or reporting to the DSO (4.17% [--1.77; 10.11] vs. 2.22% [--1.44; 5.89; p = 0.447); these changes may have arisen by chance. These findings suggest that the use of DETECT can help to reduce the deficit in the recognition of patients with impending or manifest ILBF.

Reduced diffusion in white matter after radiotherapy with photons and protons

Background and purposeRadio(chemo)therapy is standard in the adjuvant treatment of glioblastoma. Inevitably, brain tissue surrounding the target volume is also irradiated, potentially causing acute and late side-effects. Diffusion imaging has been shown to be a sensitive method to detect early changes in the cerebral white matter (WM) after radiation. The aim of this work was to assess possible changes in the mean diffusivity (MD) of WM after radio(chemo)therapy using Diffusion-weighted imaging (DWI) and to compare these effects between patients treated with proton and photon irradiation. Materials and methods70 patients with glioblastoma underwent adjuvant radio(chemo)therapy with protons (n = 20) or photons (n = 50) at the University Hospital Dresden. MRI follow-ups were performed at three-monthly intervals and in this study were evaluated until 33 months after the end of therapy. Relative white matter MD changes between baseline and all follow-up visits were calculated in different dose regions. ResultsWe observed a significant decrease of MD (p < 0.05) in WM regions receiving more than 20 Gy. MD reduction was progressive with dose and time after radio(chemo)therapy (maximum: −7.9 ± 1.2% after 24 months, ≥50 Gy). In patients treated with photons, significant reductions of MD in the entire WM (p < 0.05) were seen at all time points. Conversely, in proton patients, whole brain MD did not change significantly. ConclusionsIrradiation leads to measurable MD reduction in white matter, progressing with both increasing dose and time. Treatment with protons reduces this effect most likely due to a lower total dose in the surrounding white matter. Further investigations are needed to assess whether those MD changes correlate with known radiation induced side-effects.

How Often is Rare Really Rare? A Survey on the Frequency of Rare Diseases at a University Hospital

The prevalence of rare diseases in hospitals and in university hospitals is unknown. As the ICD-10 coding system does not adequately represent rare diagnoses, the prevalence of rare diseases cannot be estimated based on ICD-10 coded discharge diagnoses. The current hospital reimbursement system does not seem to be designed to capture performance-related higher expenditures in the treatment of rare diseases. The aim of this study was to help estimate the frequency of rare diseases among inpatients treated at a university hospital where documentation of rare diseases is obligatory by analyzing the case load of such diseases for a given year. Since 2017, rare diseases have been coded for all inpatients treated at the University Hospital Dresden. This coding is based on the Orpha identification number, which was implemented in the hospital information system ORBIS for this purpose. Result For illustrative purposes, cases in 2019 were evaluated. During this period, 19% of all 70 937 inpatients seen at the University Hospital Dresden were coded as having a rare disease. For the first time, a prospective and complete documentation of rare diseases was implemented at a German university hospital. The prevalence of rare diseases of 6 to 8% as defined by the European Union was exceeded several fold. Probably it underestimates the actual prevalence considerably, since the quality of the coding correlates on user compliance. Nevertheless, the results of this survey underline the special role of patients with rare diseases in the medical care at university hospitals.

Perioperative Blood Management of Preoperative Anemia Determines Long-Term Outcome in Patients with Pancreatic Surgery

Background/PurposeAnemia affects the postoperative course of patients undergoing a major surgical procedure. However, it remains unclear whether anemia has a different impact on the long-term outcome of patients with malignant or benign pancreatic disease. MethodsA retrospective analysis of patients undergoing pancreatic surgery for pancreatic malignancies or chronic pancreatitis was conducted between January 2012 and June 2018 at the University Hospital Dresden, Germany. The occurrence of preoperative anemia and the administration of pre-, intra-, and postoperative blood transfusions were correlated with postoperative complications and survival data by uni- and multivariate analysis. ResultsA total of 682 patients were included with 482 (70.7%) undergoing surgical procedures for pancreatic malignancies. Univariate regression analysis confirmed preoperative anemia as a risk factor for postoperative complications > grade 2 according to the Clavien-Dindo classification. Multivariate regression analyses indicated postoperative blood transfusion as an independent risk factor for postoperative complications in patients with a benign (OR 20.5; p value < 0.001) and a malignant pancreatic lesion (OR 4.7; p value < 0.01). Univariate and multivariate analysis revealed preoperative anemia and pre-, intra-, and postoperative blood transfusions as independent prognostic factors for shorter overall survival in benign and malignant patients (p value < 0.001–0.01). ConclusionPreoperative anemia is a prevalent, independent, and adjustable factor in pancreatic surgery, which poses a significant risk for postoperative complications irrespective of the entity of the underlying disease. It should therefore be understood as an adjustable factor rather than an indicator of underlying disease severity.