Supraorbital eyebrow approach and pterional approach in surgical treatment of ruptured and unruptured aneurysms: a propensity score-matched analysis.

The aim of this study is to reveal efficacy and efficiency of the supraorbital eyebrow approach (SEA) in clipping ruptured and unruptured aneurysms where both SEA and pterional approach (PA) are considered feasible by comparing SEA and PA using propensity score matching. A total of 229 patients who underwent surgical clipping of an aneurysm from 2013 to 2022 and met inclusion criteria were recruited in the study. A study group of 67 patients treated via the SEA and a comparison group of 162 patients treated via the PA were formed. Then, the subgroups of 66 patients each, with balanced incoming factors, were analyzed using the propensity score technique. The safety and efficacy endpoints were compared. Successful clipping was achieved in all cases of both groups. There were no patients in the SEA group who required conversion to the standard craniotomy. There were no procedure-related deaths in this series. No patient experienced early or late rebleeding in either group. Mean blood loss volume in the SEA group was lower than in the PA group by approximately 77.5 ml (p < 0.001). There were favorable differences in the SEA group regarding postoperative neurological deficit (p = 0.016), postoperative epileptic seizure rate (p = 0.013), ischemic and hemorrhagic complications (p = 0.028 and 0.0009, respectively), and outcomes (p < 0.001). Patients' satisfaction with cosmetic results measured by the visual analog scale was rated highly in both groups (p = 0.081). For patients where SEA provides adequate exposure, it results in safety and cosmetic outcomes not inferior to the PA.

Data Quality Estimation Via Model Performance: Machine Learning as a Validation Tool.

In our recent study, the attempt to classify neurosurgical operative reports into routinely used expert-derived classes exhibited an F-score not exceeding 0.74. This study aimed to test how improving the classifier (target variable) affected the short text classification with deep learning on real-world data. We redesigned the target variable based on three strict principles when applicable: pathology, localization, and manipulation type. The deep learning significantly improved with the best result of operative report classification into 13 classes (accuracy = 0.995, F1 = 0.990). Reasonable text classification with machine learning should be a two-way process: the model performance must be ensured by the unambiguous textual representation reflected in corresponding target variables. At the same time, the validity of human-generated codification can be inspected via machine learning.