ABSTRACT
BACKGROUND: The application of artificial intelligence is a relatively new option to enable improved patient treatment in modern medicine and is therefore currently the focus of many research projects. In the clinical practice the application of artificial intelligence so far seems to be confined to the analysis of medical imaging. OBJECTIVE: In which form is the use of artificial intelligence possible in routine daily work in thoracic surgery and is already being practiced? MATERIAL AND METHODS: A search of the currently available literature was performed. RESULTS: Under current conditions artificial intelligence can best be used as part of diagnostics and treatment planning; however, in order to enable a comprehensive use standardization and evaluation of the centralized data collection are necessary. CONCLUSION: At the present time promising study results are available but the implementation into the surgical routine has so far been very difficult.
Subject(s)
Artificial Intelligence , Thoracic Surgery , Humans , Machine LearningABSTRACT
OBJECTIVE: To evaluate the diagnostic accuracy of anticarbamylated protein antibodies (CarP), alone and in combination with traditional biomarkers (rheumatoid factor [RF] and anticitrullinated peptide antibodies [ACPA]), in established rheumatoid arthritis (RA). METHODS: A commercially available enzyme-linked immunosorbent assay (ELISA) kit was used to assess CarP concentrations in serum samples of 200 established RA and 206 controls (115 healthy donors and 55 patients with other rheumatic diseases). Main outcome measures were sensitivity, specificity, and area under the curve (AUC; 95% confidence interval [CI]). Difference in accuracy was evaluated by comparison of the respective AUCs. RESULTS: A serum CarP cut-off of 1.47 ng/ml or more differentiated patients with RA from controls with 30% sensitivity, 97.1% specificity, and good accuracy (AUC[95%CI] = 0.83[0.79-0.86], P < 0.0001). However, it showed moderate diagnostic accuracy in seronegative RA patients: sensitivity 17.9%, specificity 96.9%, and AUC (95% CI) = 0.69 (0.63-0.75). The diagnostic accuracy of CarP_ACPA and CarP_RF combinations was significantly superior to that of ACPA and RF alone (P < 0.0001 and P = 0.015, respectively), but not to that of ACPA_RF combination (P = 0.089) In addition, the CarP_ACPA_RF combination did not improve the diagnostic accuracy of the ACPA_RF combination (AUC mean difference [95% CI] = 0.006 [-0.001 to 0.015], P = 0.10). The number of positive autoantibodies (0, 1, 2, or 3) was not significantly associated with moderate-severe disease (Disease Activity Score-28 [DAS-28] > 3.2) in adjusted multiple regression analysis. CONCLUSION: CarP has good diagnostic accuracy in established RA but not in seronegative RA. The addition of CarP to ACPA and RF alone or in combination does not significantly enhance the diagnostic accuracy of ACPA_RF combination.
ABSTRACT
BACKGROUND: Percutaneous dilatational tracheostomy (PDT) and surgical tracheostomy (ST) are widely accepted techniques and frequently performed in clinical practice. We compared PDT with ST tracheostomies in view of the benefits and drawbacks, time of duration, indication and complication rate of the respective procedures. METHODS: The evaluation was based on data from 378 tracheostomies. 209 of these tracheostomies were performed at bedside as PDT in the intensive care unit. These were compared to 169 ST tracheostomies performed in the operating room. All interventions were performed by the same team of surgeons or intensivists, however, at different training levels. RESULTS: The mean duration of the operation was shorter for PDT than for ST (18.2 ± 10 min versus 38.2 ± 14.2 min, p = < 0.001). The PDT was a simpler procedure and performed predominantly by physicians in postgraduate training. The rate of complications was low in both groups (8.6 % PDT, 8.3 % ST, p = 0.909). CONCLUSION: Although both interventions are safe and achieve comparable results, PDT can be applied in a shorter time. PDT is easier to perform and seems particularly suitable for physicians in postgraduate training.
