Feasibility and safety of the posterior retroperitoneoscopic approach in the resection of aortocaval and infrarenal paraganglioma: a single-center experience

Predicting surgical outcome in posterior retroperitoneoscopic adrenalectomy with the aid of a preoperative nomogram

BACKGROUND

Posterior retroperitoneoscopic adrenalectomy (PRA) has several advantages over transperitoneal laparoscopic adrenalectomy (TLA) regarding operative time, blood loss, postoperative pain, and recovery. However, it can be a technically challenging procedure. To improve patient selection for PRA, we developed a preoperative nomogram to predict operative time.

METHODS

All consecutive patients with tumors of ≤ 7 cm and a body mass index (BMI) of < 35 kg/m2 undergoing unilateral PRA between February 2011 and March 2020 were included in the study. The primary outcome was operative time as surrogate endpoint for surgical complexity. Using ten patient variables, an optimal prediction model was created, with a best subsets regression analysis to find the best one-variable up to the best seven-variable model.

RESULTS

In total 215 patients were included, with a mean age of 52 years and mean tumor size of 2.4 cm. After best subsets regression analysis, a four-variable nomogram was selected and calibrated. This model included sex, pheochromocytoma, BMI, and perinephric fat, which were all individually significant predictors. This model showed an ideal balance between predictive power and applicability, with an R2 of 38.6.

CONCLUSIONS

A four-variable nomogram was developed to predict operative time in PRA, which can aid the surgeon to preoperatively identify suitable patients for PRA. If the nomogram predicts longer operative time and therefore a more complex operation, TLA should be considered as an alternative approach since it provides a larger working space. Also, the nomogram can be used for training purposes to select patients with favorable characteristics when learning this surgical approach.

Deep learning-based endoscopic anatomy classification: an accelerated approach for data preparation and model validation

BACKGROUND

Photodocumentation during endoscopy procedures is one of the indicators for endoscopy performance quality; however, this indicator is difficult to measure and audit in the endoscopy unit. Emerging artificial intelligence technology may solve this problem, which requires a large amount of material for model development. We developed a deep learning-based endoscopic anatomy classification system through convolutional neural networks with an accelerated data preparation approach.

PATIENTS AND METHODS

We retrospectively collected 8,041 images from esophagogastroduodenoscopy (EGD) procedures and labeled them using two experts for nine anatomical locations of the upper gastrointestinal tract. A base model for EGD image multiclass classification was first developed, and an additional 6,091 images were enrolled and classified by the base model. A total of 5,963 images were manually confirmed and added to develop the subsequent enhanced model. Additional internal and external endoscopy image datasets were used to test the model performance.

RESULTS

The base model achieved total accuracy of 96.29%. For the enhanced model, the total accuracy was 96.64%. The overall accuracy improved with the enhanced model compared with the base model for the internal test dataset without narrowband images (93.05% vs. 91.25%, p < 0.01) or with narrowband images (92.74% vs. 90.46%, p < 0.01). The total accuracy was 92.56% of the enhanced model on the external test dataset.

CONCLUSIONS

We constructed a deep learning-based model with an accelerated approach that can be used for quality control in endoscopy units. The model was also validated with both internal and external datasets with high accuracy.