Abstract. Acute renal failure increases risk of death after cardiac surgery. However, it is not known whether more subtle changes in renal function might have an impact on outcome. Thus, the association between small serum creatinine changes after surgery and mortality, independent of other established perioperative risk indicators, was analyzed. In a prospective cohort study in 4118 patients who underwent cardiac and thoracic aortic surgery, the effect of changes in serum creatinine within 48 h postoperatively on 30-d mortality was analyzed. Cox regression was used to correct for various established demographic preoperative risk indicators, intraoperative parameters, and postoperative complications. In the 2441 patients in whom serum creatinine decreased, early mortality was 2.6% in contrast to 8.9% in patients with increased postoperative serum creatinine values. Patients with large decreases (⌬Crea ϽϪ0.3 mg/dl) showed a progressively increasing 30-d mortality (16 of 199 [8%]). Mortality was lowest (47 of 2195 [2.1%]) in patients in whom serum creatinine decreased to a maximum of Ϫ0.3 mg/dl; mortality increased to 6% in patients in whom serum creatinine remained unchanged or increased up to 0.5 mg/dl. Mortality (65 of 200 [32.5%]) was highest in patients in whom creatinine increased Ն0.5 mg/dl. For all groups, increases in mortality remained significant in multivariate analyses, including postoperative renal replacement therapy. After cardiac and thoracic aortic surgery, 30-d mortality was lowest in patients with a slight postoperative decrease in serum creatinine. Any even minimal increase or profound decrease of serum creatinine was associated with a substantial decrease in survival.Acute renal failure (ARF) develops in 5 to 30% of patients who undergo cardiac surgery and is associated with a more complicated clinical course and with an excessive mortality of up to 80% (1-4). Actually, development of ARF was identified as the strongest risk factor for death with an odds ratio of 7.9 in patients who undergo cardiac surgery (1). Certainly, ARF presents an indicator for the severity and/or complicated course of disease; thus, perioperative patients with renal dysfunction are at a higher risk of dying. However, recently, is was shown convincingly that ARF acts as a risk factor for a grim prognosis independent of the severity of the underlying disease: that patients do not die with but rather from ARF (5,6).Nevertheless, it remains unknown whether not only manifest ARF but also more subtle changes in postoperative renal function might predict outcome in surgical patients. In patients with contrast-induced nephropathy, renal impairment as defined by an increase of 25% to at least 2 mg/dl in serum creatinine was associated with an odds ratio of 5.5 for death (7). Thus, the aim of the present investigation was to determine the consequences of small changes in serum creatinine within 48 h after surgery on 30-d and late mortality, independent of other established perioperative risk indicators. Materials and MethodsBet...

Background Deep learning offers considerable promise for medical diagnostics. We aimed to evaluate the diagnostic accuracy of deep learning algorithms versus health-care professionals in classifying diseases using medical imaging.Methods In this systematic review and meta-analysis, we searched Ovid-MEDLINE, Embase, Science Citation Index, and Conference Proceedings Citation Index for studies published from Jan 1, 2012, to June 6, 2019. Studies comparing the diagnostic performance of deep learning models and health-care professionals based on medical imaging, for any disease, were included. We excluded studies that used medical waveform data graphics material or investigated the accuracy of image segmentation rather than disease classification. We extracted binary diagnostic accuracy data and constructed contingency tables to derive the outcomes of interest: sensitivity and specificity. Studies undertaking an out-of-sample external validation were included in a meta-analysis, using a unified hierarchical model. This study is registered with PROSPERO, CRD42018091176.Findings Our search identified 31 587 studies, of which 82 (describing 147 patient cohorts) were included. 69 studies provided enough data to construct contingency tables, enabling calculation of test accuracy, with sensitivity ranging from 9•7% to 100•0% (mean 79•1%, SD 0•2) and specificity ranging from 38•9% to 100•0% (mean 88•3%, SD 0•1). An out-of-sample external validation was done in 25 studies, of which 14 made the comparison between deep learning models and health-care professionals in the same sample. Comparison of the performance between health-care professionals in these 14 studies, when restricting the analysis to the contingency table for each study reporting the highest accuracy, found a pooled sensitivity of 87•0% (95% CI 83•0-90•2) for deep learning models and 86•4% (79•9-91•0) for health-care professionals, and a pooled specificity of 92•5% (95% CI 85•1-96•4) for deep learning models and 90•5% (80•6-95•7) for health-care professionals.Interpretation Our review found the diagnostic performance of deep learning models to be equivalent to that of health-care professionals. However, a major finding of the review is that few studies presented externally validated results or compared the performance of deep learning models and health-care professionals using the same sample. Additionally, poor reporting is prevalent in deep learning studies, which limits reliable interpretation of the reported diagnostic accuracy. New reporting standards that address specific challenges of deep learning could improve future studies, enabling greater confidence in the results of future evaluations of this promising technology.

Background: One of the most challenging practical and daily problems in intensive care medicine is the interpretation of the results from diagnostic tests. In neonatology and pediatric intensive care the early diagnosis of potentially life-threatening infections is a particularly important issue. Focus: A plethora of tests have been suggested to improve diagnostic decision making in the clinical setting of infection which is a clinical example used in this article. Several criteria that are critical to evidence-based appraisal of published data are often not adhered to during the study or in reporting. To enhance the critical appraisal on articles on diagnostic tests we discuss various measures of test accuracy: sensitivity, specificity, receiver operating characteristic curves, positive and negative predictive values, likelihood ratios, pretest probability, posttest probability, and diagnostic odds ratio. Conclusions: We suggest the following minimal requirements for reporting on the diagnostic accuracy of tests: a plot of the raw data, multilevel likelihood ratios, the area under the receiver operating characteristic curve, and the cutoff yielding the highest discriminative ability. For critical appraisal it is mandatory to report confidence intervals for each of these measures. Moreover, to allow comparison to the readers' patient population authors should provide data on study population characteristics, in particular on the spectrum of diseases and illness severity.

Prostate biopsy schemes consisting of 12 cores that add laterally directed cores to the standard sextant scheme strike the balance between the cancer detection rate and adverse events. Taking more than 12 cores added no significant benefit.

No abstract