Use of the KSVM-based system for the definition, validation and identification of the incisional hernia recurrence risk factors

BACKGROUND

Incisional hernia is one of the most common complications after abdominal surgery with an incidence rate of 11 to 20% post laparotomy. Many different factors can be considered as risk factors of incisional hernia recurrence. The aim of this study is to confirm and to validate the incisional hernia recurrence risk factors and to identify and to validate new ones.

METHODS

In the period from July 2007 to July 2017, 154 patients were selected and subjected to incisional hernia repair. The surgical operations were conducted under general anaesthesia. Patients received antibiotic prophylaxis when indicated, according to the hospital prophylaxis scheme. Inclusion criteria of the study were single operator case studies and open laparotomy for incisional hernia repair. The statistical analysis proposed to identify and to verify the risk factors for recurrence of incisional hernia is the Support Vector Machine (SVM). The analysis was conducted verifying 34 risk factors.

RESULTS

The data analysis confirmed the known correlations showed in the international literature with a greater incidence of comorbidities such as diabetes 37%, dyslipidaemia and hypercholesterolemia with a cumulative incidence of 16%; tobacco smoke - by combining categories smokers and ex-smokers - reach 46%, COPD 16% and hypertension 51%.

CONCLUSIONS

The analysis of the data therefore confirmed the correlations showed in the international literature. A KSVM-based system to classify incisional hernia recurrence has been presented. The type of prosthesis and the site of its implant also play a significant role in the development of the recurrence. Sensitivity (86,25%), Specificity (87,14%), Negative Predictive Value (84,72%), Precision (88,46%), Accuracy (86,67%), and Error (13,33%) scores obtained using the proposed technique highlight the validity for the relapse's classification methodology.

Multislice human organ extraction based on GVF

Segmentation techniques based on active contours algorithm are widely used in medical imaging. Unfortunately, they require a lot of parameters to be used and this can represent an issue for those physicians with not much informatics skills. This paper proposes a software tool which allows to segment multiple slice can differ organ extraction setting a small number of parameters. Moreover, the tool offers the functionality to perform a multiple segmentation the same time, so that an entire volume composed by hundreds slices can be segmented.