Few studies on traumatic brain injury (TBI) have investigated the stability of blood serum biomarkers after long-term storage at low temperatures. In the current feasibility study we analyzed acute phase serum samples from patients with mild TBI as well as patients with moderate and severe TBI that were collected more than 10 years ago (old samples). We were particularly interested in mild TBI, because injury effects are more subtle in this category as compared to moderate-severe TBI. Therefore, the primary objective was to find out whether several biomarkers were still detectable for these patients. Additionally, we examined whether biomarker levels varied as a function of injury severity. For comparison, we also analyzed samples from an ongoing mTBI cohort (new samples) and healthy controls. Samples were treated with care and were not being subjected to freeze-thaw cycles. We measured concentrations of interleukins (IL6 and 10) and brain specific markers (total tau, UCH-L1, GFAP, and NF-L). No significant differences in biomarker concentrations were found between old and new mild TBI samples. For IL6, IL10, and UCH-L1 higher concentrations were found in moderate and severe TBI as compared to mild TBI. In conclusion, our study shows that long-term storage does not rule out the detection of meaningful biomarker concentrations in patients with TBI, although further research by other laboratories is warranted.

Objectives To investigate the impact on efficiency and quality of preprostatectomy multidisciplinary therapy conferences (MDT) at Karolinska University Hospital related to the use of a digital solution compared with standard of care. Further, to explore whether gains in MDT efficiency and quality impact oncological or functional patient outcomes. Methods We conducted a prospective, observational study of preoperative prostate cancer MDT at Karolinska between February 2017 and March 2021, including 1329 patients. We compared efficiency and quality of the standard MDT and the MDT using the digital solution IntelliSpace Precision Medicine Multidisciplinary Team Orchestrator (ISPM) based on the previously used MDT-MODe approach. Clinical and patient-reported functional outcomes were derived from the medical records and the Swedish National Prostate Cancer Register. Results While ISPM was used during the MDT meeting, the time spent per patient was reduced by 24% (p<0.001) and most of the MDT-MODe items were scored significantly higher. There was a reduction in pelvic lymph-node dissection procedures in the ISPM cohort (p=0.001) and an increased proportion of unilateral nerve-sparing procedures (p=0.005), while all other outcome-related measures were not significantly different between the two patient groups. Discussion and conclusion To increase the value of the MDT, all data relevant for treatment decision need to be purposefully presented and compiled, which also enables secondary use of the data. The use of a digital solution during preoperative MDTs for prostate cancer decision making at Karolinska University Hospital improved the efficiency and quality of this multidisciplinary team meeting without impacting patient outcomes.

Traumatic brain injury is a multifaceted condition that encompasses a spectrum of injuries: contusions, axonal injuries in specific brain regions, edema, and hemorrhage. Brain injury determines a broad clinical and disability spectrum due to the implication of various cellular pathways, genetic phenotypes, and environmental factors. It is challenging to predict patient outcomes, to appropriately evaluate the patients, to determine a suitable treatment strategy and rehabilitation program, and to communicate with patient relatives. Biomarkers detected from body fluids are potential evaluation tools for traumatic brain injury patients. These may serve as internal indicators of cerebral damage, delivering valuable information about the dynamic cellular, biochemical, and molecular environments. The diagnostic and prognostic value of biomarkers tested both in animal models of traumatic brain injury is still under question, despite a considerable scientific literature. Recent publications emphasize that a more realistic approach involves combining multiple types of biomarkers with other investigative tools (imaging, outcome scales, and genetic polymorphisms). Additionally, there is increasing interest in the use of biomarkers as tools for treatment monitoring and as surrogate outcome variables to facilitate the design of distinct randomized controlled trials. This review highlights the latest available evidence regarding biomarkers in adults after traumatic brain injury and discusses new approaches in the evaluation of this patient group.

Seric biomarkers have been tested in a large number of studies on traumatic brain injuries (TBI) patients in order to predict severity, especially related to the short-term outcome. However, TBI patients have a high risk of developing long-term complications such as physical disability, cognitive impairment, psychiatric pathology, epilepsy, and others. The aim of this study was to assess the correlation between protein biomarkers S100 and neuron-specific enolase (NSE) and neurocognitive status at 10- and 90-days post-injury. Both biomarkers were tested in the first 4h and after 72h post-injury in 62 patients with moderate-severe TBI. The patients were evaluated by a series of neurocognitive tests: Early Rehabilitation Barthel Index (ERBI), Glasgow Outcome Scale-Extended (GOSE), The Mini-Mental State Examination (MMSE), Processing Speed Index (PSI), and Stroop Test, at 10 and 90 days post-injury and supplementary by the Hospital Anxiety and Depression Scale at 90 days. For evaluating the whole neurocognitive status instead of every scale separately, we used Structural Equation Modeling (SEM), while for anxiety and depressive symptoms, we used multiple regression analyses. SEM showed that NSE values at 4 hours were significant predictors of the cognitive status at 10 (p=0.034) and 90 days (p= 0.023). Also, there were found significant correlations between NSE at 4h and the anxiety level. This study demonstrated a significant correlation between NSE at 4h and short and medium-term neuropsychological outcomes, which recommends using this biomarker for selecting patients with a higher risk of cognitive dysfunction.

One of the most common types of cancer in men is prostate cancer (PCa). Biopsies guided by bi-parametric magnetic resonance imaging (MRI) can aid PCa diagnosis. Previous works have mostly focused on either detection or classification of PCa from MRI. In this work, however, we present a neural network that simultaneously detects and grades cancer tissue in an end-to-end fashion. This is more clinically relevant than the classification goal of the ProstateX-2 challenge. We used the dataset of this challenge for training and testing. We use a 2D U-Net with MRI slices as input and lesion segmentation maps that encode the Gleason Grade Group (GGG), a measure for cancer aggressiveness, as output. We propose a method for encoding the GGG in the model target that takes advantage of the fact that the classes are ordinal. Furthermore, we evaluate methods for incorporating prostate zone segmentations as prior information, and ensembling techniques. The model scored a voxel-wise weighted kappa of 0.446 ±0.082 and a Dice similarity coefficient for segmenting clinically significant cancer of 0.370 ±0.046, obtained using 5-fold cross-validation. The lesion-wise weighted kappa on the ProstateX-2 challenge test set was 0.13 ±0.27. We show that our proposed model target outperforms standard multiclass classification and multi-label ordinal regression. Additionally, we present a comparison of methods for further improvement of the model performance.

Although the biological characteristics of uterine fibroids (UF) have implications for therapy choice and effectiveness, there is limited MRI data about these characteristics. Currently, the Funaki classification and Scaled Signal Intensity (SSI) are used to predict treatment outcome but both screening-tools appear to be suboptimal. Therefore, multiparametric and quantitative MRI was studied to evaluate various biological characteristics of UF. 87 patients with UF underwent an MRI-examination. Differences between UF tissues and myometrium were investigated using T2-mapping, Apparent Diffusion Coefficient (ADC) maps with different b-value combinations, contrast-enhanced T1-weighted and T2-weighted imaging. Additionally, the Funaki classification and SSI were calculated. Significant differences between myometrium and UF tissue in T2-mapping (p = 0.001), long-TE ADC low b-values (p = 0.002), ADC all b-values (p < 0.001) and high b-values (p < 0.001) were found. Significant differences between Funaki type 3 versus type 1 and 2 were observed in SSI (p < 0.001) and T2-values (p < 0.001). Significant correlations were found between SSI and T2-mapping (p < 0.001; ρs = 0.82), ADC all b-values (p = 0.004; ρs = 0.31), ADC high b-values (p < 0.001; ρs = 0.44) and long-TE ADC low b-values (p = 0.004; ρs = 0.31). Quantitative MR-data allowed us to distinguish UF tissue from myometrium and to discriminate different UF tissue types and may, therefore, be a useful tool to predict treatment outcome/determine optimal treatment modality.

Data analytics methods in the clinical domain are challenging to put into practice. Unsupervised learning provides opportunity for giving the level of personalization in evidence based decision-making that can otherwise only be achieved through the use of prediction models, by helping doctors gaining insights from data. In this context, grouping of clinical subjects, in terms of biomedical information of patients, is an important task for patient cohort identification for comparative effectiveness studies and clinical decision-support applications. It allows the decision-making process to leverage not only on data but also on doctors’ domain knowledge. However, one of the issues that needs to be addressed for a focused and realist unsupervised clustering of clinical subjects, is the fact that in the majority of the cases patients datasets are heterogeneous, i.e. their data features belong to several different feature spaces, e.g. nominal, ordinal, interval or rational, with completely different variation ranges and statistical distributions, affecting clustering quality and performance. In order to use these data measurements properly in an unsupervised manner, their corresponding weights need to be modeled. In this paper, we present a method for learning feature weights on clinical data. We show that learning feature weights is necessary in order to generate meaningful separation of data in high dimensional space. The method is based on silhouette score and principal component analysis, demonstrating its performance on a clinical test dataset.

The goal of this work is to detect prostate cancer (PCa) from multi-parametric MRI (mpMRI) and to simultaneously predict the Gleason Grade Group (GGG) of the detected tumors. We used the ProstateX-2 dataset, for training, validation and testing. The challenge training set contains 99 patients and 112 lesions. The challenge test set contains 63 patients and 70 lesions. T2-weighted and apparent diffusion coefficient images were used as input for a U-Net model. For each tumor a GGG was assigned based on biopsy and pathologic analysis. Segmentation maps of the tumors that were multiplied with a scaled value of the GGG, were used as target of the network, turning the problem into a binary soft-label ordinal regression problem. Using 5-fold cross validation, a voxel-wise quadratic-weighted kappa score of 0.391 ± 0.062 and a DSC (GGG ≥ 2) of 0.321 ± 0.039 were achieved. In order to evaluate using the challenge test set, the voxel-wise predictions were converted into a single GGG prediction per lesion. Our method ranks higher than 30 out of 43 participants of the challenge with a lesion-wise quadratic weighted kappa score of 0.082 ± 0.272 when evaluating on the ProstateX-2 test set. Despite solving a more difficult, but more clinically relevant, problem than the original ProstateX-2 challenge, a relatively high score was achieved using an approach to both grade and detect prostate cancer directly from mpMRI.

Purpose. To further validate the prognostic power of the biomarker PDE4D7, we investigated the correlation of PDE4D7 scores adjusted for presurgical clinical variables with longitudinal postsurgical biological outcomes. Methods. RNA was extracted from biopsy punches of resected tumors (550 patients; RP cohort) and diagnostic needle biopsies (168 patients; DB cohort). Cox regression and survival were applied to correlate PDE4D7 scores with patient outcomes. Logistic regression was used to combine the clinical CAPRA score with PDE4D7. Results. In univariate analysis, the PDE4D7 score was significantly associated with PSA recurrence after prostatectomy in both studied patient cohorts' analysis (HR 0.53; 95% CI 0.41-0.67; p<1.0E-04 and HR 0.47; 95% CI 0.33-0.65; p<1.0E-04, respectively). After adjustment for the presurgical clinical variables preoperative PSA, PSA density, biopsy Gleason, clinical stage, percentage tumor in the biopsy (data only available for RP cohort), and percentage of positive biopsies, the HR was 0.49 (95% CI 0.38-0.64; p<1.0E-04) and 0.43 (95% CI 0.29-0.63; p<1.0E-04), respectively. The addition of the PDE4D7 to the clinical CAPRA score increased the AUC by 5% over the CAPRA score alone (0.82 versus 0.77; p=0.004). This combination model stratified 14.6% patients of the DB cohort to no risk of biochemical relapse (NPV 100%) over a follow-up period of up to 15 years. Conclusions. The PDE4D7 score provides independent risk information for pretreatment risk stratification. Combining CAPRA with PDE4D7 scores significantly improved the clinical risk stratification before surgery.

BACKGROUND:The clinical metrics used to date to assess the progression risk of newly diagnosed prostate cancer patients only partly represent the true biological aggressiveness of the underlying disease. OBJECTIVE:Validation of the prognostic biomarker phosphodiesterase-4D7 (PDE4D7) in predicting longitudinal biological outcomes in a historical surgery cohort to improve postsurgical risk stratification. DESIGN, PATIENTS, AND METHODS:RNA was extracted from biopsy punches of resected tumors from 550 patients. PDE4D7 was quantified using one-step quantitative reverse transcription-polymerase chain reaction. PDE4D7 scores were calculated by normalization of PDE4D7 to reference genes. Multivariate analyses were adjusted for clinical prognostic variables. Outcomes tested were: prostate-specific antigen relapse, start of salvage treatment, progression to metastases, overall mortality, and prostate cancer-specific mortality. The PDE4D7 score was combined with the clinical risk model Cancer of the Prostate Risk Assessment Postsurgical Score (CAPRA-S) using multivariate regression modeling; the combined score was tested in post-treatment progression free survival prediction. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS:Correlations with outcomes were analyzed using multivariate Cox regression and logistic regression statistics. RESULTS AND LIMITATIONS:The PDE4D7 score was significantly associated with time-to-prostate specific antigen failure after prostatectomy (hazard ratio [HR]: 0.53, 95% confidence interval [CI]: 0.41-0.67 for each unit increase, p<0.0001). After adjustment for postsurgical prognostic variables the HR was 0.56 (95% CI: 0.43-0.73, p<0.0001). The PDE4D7 score remained significant after adjusting the multi-variate analysis for the CAPRA-S model categories (HR=0.54, 95% CI=0.42-0.69, p<0.0001). Combination of the PDE4D7 score with the CAPRA-S demonstrated a significant incremental value of 4-6% in 2-yr (p=0.004) or 5-yr (p=0.003) prediction of progression free survival after surgery. The combined model of PDE4D7 and CAPRA-S improves patient selection with very high risk of fast disease relapse after primary intervention. CONCLUSIONS:The PDE4D7 score has the potential to provide independent risk information and to restratify patients with clinical intermediate- to high-risk characteristics to a very low-risk profile. PATIENT SUMMARY:In this report, we studied the potential of a novel biomarker to predict outcomes of a cohort of prostate cancer patients who underwent surgery more than 10 yr ago. We found that a gene called phosphodiesterase-4D7 added extra information to the available clinical data. We conclude that the measurement of this gene in tumor tissue may contribute to more effective treatment decisions.

We provide preliminary details and formulation of an optimization strategy under current development that is able to automatically tune the parameters of a Support Vector Machine over new datasets. The optimization strategy is a heuristic based on Iterated Local Search, a modification of classic hill climbing which iterates calls to a local search routine.

In this paper, we propose a method for optimization of the parameters of a Support Vector Machine which is more accurate than the usually applied grid search method. The method is based on Iterated Local Search, a classic metaheuristic that performs multiple local searches in different parts of the space domain. When the local search arrives at a local optimum, a perturbation step is performed to calculate the starting point of a new local search based on the previously found local optimum. In this way, exploration of the space domain is balanced against wasting time in areas that are not giving good results. We show a preliminary evaluation of our method on a radial-basis kernel and some sample data, showing that it is more accurate than an application of grid search on the same problem. The method is applicable to other kernels and future work should demonstrate to what extent our Iterated Local Search based method outperforms the standard grid search method over other heterogeneous datasets from different domains.

We provide preliminary details and formulation of an optimization strategy under current development that is able to automatically tune the parameters of a Support Vector Machine over new datasets. The optimization strategy is a heuristic based on Iterated Local Search, a modification of classic hill climbing which iterates calls to a local search routine.

5069 Background: In this study we present the retrospective validation of the prognostic prostate cancer biomarker PDE4D7 in predicting longitudinal biological outcomes in a historical cohort of radical prostatectomy patients. Methods: Biopsy punches from 550 patients were collected from a representative tumor area of FFPE surgical resections. RNA was extracted and PDE4D7 quantified by one-step RT-qPCR. PDE4D7 scores were calculated by normalization of PDE4D7 to the averaged expression of four reference genes. The independent prognostic value of the PDE4D7 scores were evaluated using uni- and multivariate Cox proportional hazard regression. Multivariate analyses were adjusted for clinical prognostic variables. Post-surgical outcomes tested were: PSA relapse, start of salvage treatment, progression to metastases, overall and prostate cancer specific mortality. Logistic regression was used to create a combined prognostic model of PDE4D7 with clinical risk and tested in outcome prediction. Results: The PDE4D7 score was significantly associated with time to PSA failure after prostatectomy (HR 0.53; 95% CI 0.41-0.67 for each unit increase; p < 1.0E-04). After adjustment for pathology Gleason, pT stage, surgical margin status, and seminal vesicle invasion the HR was 0.55 (95% CI 0.43-0.72; p < 1.0E-04). Patients with a high PDE4D7 score that were clinically classified as intermediate to high risk of progression were re-classified into a group with an average progression risk less than the average cohort risk of clinically very low risk patients. The maximum benefit, compared to Gleason score, was observed in the clinically intermediate favorable risk group. Combining clinical risk with PDE4D7 scores improved the overall risk stratification. Conclusions: The PDE4D7 score has potential to provide independent risk information and, in particular, to re-stratify patients with clinical intermediate to high risk characteristics to a very low risk profile.

Traumatic brain injury (TBI) is one of the leading causes of injury-related death. In the United States alone, an estimated 1.7 million people sustain a TBI each year, and approximately 5.3 million people live with a TBI-related disability. The direct medical costs and indirect costs such as lost productivity of TBIs totaled an estimated $76.5 billion in the U.S. in the year 2000. Improving the limited treatment options for this condition remains challenging. However, recent reports from interdisciplinary working groups (consisting primarily of neurologists, neurosurgeons, neuropsychologists, and biostatisticians) have stated that to improve TBI treatment, important methodological lessons from the past must be taken into account in future clinical research. An evaluation of the neuroprotection intervention studies conducted over the last 30 years has indicated that a limited understanding of the underlying biological concepts and methodological design flaws are the major reasons for the failure of pharmacological agents to demonstrate efficacy. Cerebrolysin is a parenterally-administered neuro-peptide preparation that acts in a manner similar to endogenous neurotrophic factors. Cerebrolysin has a favorable adverse effect profile, and several meta-analyses have suggested that Cerebrolysin is beneficial as a dementia treatment. CAPTAIN is a randomized, double-blind, placebo-controlled, multi-center, multinational trial of the effects of Cerebrolysin on neuroprotection and neurorecovery after TBI using a multidimensional ensemble of outcome scales. The CAPTAIN trial will be the first TBI trial with a 'true' multidimensional approach based on full outcome scales, while avoiding prior weaknesses, such as loss of information through "dichotomization," or unrealistic assumptions such as "normal distribution."

Increased blood-brain barrier permeability (BBBP) can result from ischemia. In this study the relation between stroke severity, patient characteristics and admission BBBP values measured with CT-perfusion (CTP) was investigated in acute ischemic stroke patients.

Noise levels in hospitals, especially in intensive care units (ICUs) are known to be high, potentially affecting not only the patients’ well-being but also their clinical outcomes. In an observational study, we made a long-term measurement of noise levels in an ICU, and investigated the influence of various factors on the noise level, including the acute physiology and chronic health evaluation II (APACHE II) score.

Delayed cerebral ischemia (DCI) after aneurysmal subarachnoid hemorrhage (aSAH) can be reversible or progress to cerebral infarction. In patients with a deterioration clinically diagnosed as DCI, we investigated whether CT perfusion (CTP) can distinguish between reversible ischemia and ischemia progressing to cerebral infarction.From a prospectively collected series of aSAH patients, we included those with DCI, CTP on the day of clinical deterioration, and follow-up imaging. In qualitative CTP analyses (visual assessment), we calculated positive and negative predictive value (PPV and NPV) with 95% confidence intervals (95%CI) of a perfusion deficit for infarction on follow-up imaging. In quantitative analyses, we compared perfusion values of the least perfused brain tissue between patients with and without infarction by using receiver-operator characteristic curves and calculated a threshold value with PPV and NPV for the perfusion parameter with the highest area under the curve.In qualitative analyses of 33 included patients, 15 of 17 patients (88%) with and 6 of 16 patients (38%) without infarction on follow-up imaging had a perfusion deficit during clinical deterioration (p = 0.002). Presence of a perfusion deficit had a PPV of 71% (95%CI: 48-89%) and NPV of 83% (95%CI: 52-98%) for infarction on follow-up. Quantitative analyses showed that an absolute minimal cerebral blood flow (CBF) threshold of 17.7 mL/100 g/min had a PPV of 63% (95%CI: 41-81%) and a NPV of 78% (95%CI: 40-97%) for infarction.CTP may differ between patients with DCI who develop infarction and those who do not. For this purpose, qualitative evaluation may perform marginally better than quantitative evaluation.