Retrospective ct-analysis using machine learning and feature selection for optimizing preoperative preparation of patients with large and giant postoperative ventral hernias
https://doi.org/10.17238/2072-3180-2025-1-56-63
Abstract
Introduction. Purpose: this study aimed to develop a machine learning-based approach for identifying key CT abdominometry features to assist in the preoperative planning of patients with large and giant postoperative ventral hernias.
Materials and methods of research. A retrospective analysis was conducted on 90 patients examined between 2020 and 2024. Patients were divided into three groups: those without postoperative ventral hernias (n=30), patients with W2 hernias (n=30), and patients with large or giant hernias who underwent preoperative botulinum therapy (n=30). CT scans of the abdomen, retroperitoneum, and pelvis were performed, and parameters such as anterior abdominal wall muscle characteristics, abdominal circumference, hernia gate width, defect ratio, and component separation index were measured. Statistical analyses included ANOVA, the Kruskal-Wallis test, Student’s t-test, and the Mann-Whitney U test. Feature selection was performed using a random forest algorithm, logistic regression was utilized for predicting the need for botulinum therapy, and clustering methods (K-Means, DBSCAN) were applied to identify distinct patient subgroups.
Treatment results. Logistic regression achieved an accuracy of 86 % (AUC-ROC = 0,95). 'The defect ratio, component separation index, and right muscle area were identified as significant predictors for botulinum therapy. K-Means clustering delineated a subgroup in which 83 % of patients underwent botulinum therapy, underscoring the objective differentiation based on CT features. An optimal predictive threshold of 0.6 was established to minimize false negatives.
Conclusion. The application of machine learning techniques facilitates objective and personalized preoperative planning, thereby enhancing surgical decision-making in the management of postoperative ventral hernias.
About the Authors
A. V. ProtasovRussian Federation
Protasov Andrey Vitalievich, Doctor of Medical Sciences, Professor, Head of the Department of Operative Surgery and Clinical Anatomy named after I.D. Kirpatovsky
117198, Miklukho-Maklaya St. 8, Moscow
L. B. Kanakhina
Russian Federation
Kanakhina Liya Beketaevna, Postgraduate student of the Department of Operative Surgery and Clinical Anatomy named after I.D. Kirpatovsky
117198, Miklukho-Maklaya St. 8, Moscow
O. I. Mazurova
Russian Federation
Mazurova Olga Igorevna, Candidate of Medical Sciences, Assistant Professor of the Department of Operative Surgery and Clinical Anatomy named after I.D. Kirpatovsky
117198, Miklukho-Maklaya St. 8, Moscow
A. P. Chorbadzhi
Russian Federation
Chorbadzhi Andrey Pavlovich, Resident of the Department of Operative Surgery and Clinical Anatomy named after I.D. Kirpatovsky
117198, Miklukho-Maklaya St. 8, Moscow
References
1. Le Huu Nho R., Mege D., Ouaïssi M., Sielezneff I., Sastre B. Incidence and prevention of ventral incisional hernia. Journal of Vascular Surgery, 2012, vol. 149, pp. e3–e14. https://doi.org/10.1016/j.jviscsurg.2012.05.004
2. Fink C., Baumann P., Wente M.N., Knebel P., Bruckner T., Ulrich A., Werner J., Büchler M.W., Diener M.K. Incisional hernia rate 3 years after midline laparotomy. British Journal of Surgery, 2014, № 101, pp. 51–54. https://doi.org/10.1002/bjs.9364
3. Ministry of Health of the Russian Federation. All-Russian public organization “Russian Society of Surgeons”, All-Russian public organization “Society of Herniologists”. Clinical recommendations «Posleoperacionnaja ventralnaja gryzha» [Postoperative ventral hernia]. (In Russ.). Available at: https://cr.minzdrav.gov.ru/preview-cr/685_2. (accessed: 23–02–2025)
4. Sulaymanova N., Rakhmonov Sh., Makhmudov F., Rakhimov A. Monitoring of intra-abdominal pressure and prevention of abdominal compression syndrome in hernioplasty of large and giant median hernias. Vestnik vracha, 2014, № 1, pp. 171–174. (In Russ.) Available at: https://inlibrary.uz/index.php/doctors_herald/article/view/4568 (accessed: 23–02–2025)
5. Samartsev V.A., Gavrilov V.A., Pushkarev B.S. Intraabdominal hypertension syndrome: the current state of the problem. Khirurgicheskaya praktika, 2020, № 2, pp. 35–42. (In Russ.). https://doi.org/10.38181/2223–2427–2020–2–35–42
6. Vasiliadis K., Knaebel H.P., Djakovic N., Nyarangi–Dix J., Schmidt J., Büchler M. Challenging surgical management of a giant inguinoscrotal hernia: Report of a case. Surgery Today, 2010, № 40, pp. 684–687. https://doi.org/10.1007/s00595–009–4125–3
7. El–Dessouki N.I. Preperitoneal mesh hernioplasty in giant inguinoscrotal hernias: A new technique with dual benefit in repair and abdominal rooming. Hernia, 2001, vol. 5, pp. 177–181. https://doi.org/10.1007/s10029–001–0030–4
8. Topchiev A.M., Protasov A.V., Fedoseev A.V., Topchiev M.A., Parshin D.S., Samsonov A.V. Possibilities of intraoperative stretching of the musculofascial structures of the anterior abdominal wall as a stage of their preparation for plastic surgery in postoperative hernias W3. Sovremennye problem nauki i obrazovanija, 2022, № 5, pp. 140–140. (In Russ.). https://doi.org/10.17513/spno.32137
9. Crist D.W., Gadacz T.R. Complications of laparoscopic surgery. Surgical clinics of North America, 1993, № 73, pp. 265–289. https://doi.org/10.1016/S0039–6109(16)45981–5
10. Huerta S. et al. Botulinum Toxin A as an Adjunct for the Repair of Giant Inguinal Hernias: Case Reports and a Review of the Literature. Journal of Clinical Medicine, 2024, vol. 13, № 7, pp. 1879. https://doi.org/10.3390/jcm13071879
Review
For citations:
Protasov A.V., Kanakhina L.B., Mazurova O.I., Chorbadzhi A.P. Retrospective ct-analysis using machine learning and feature selection for optimizing preoperative preparation of patients with large and giant postoperative ventral hernias. Moscow Surgical Journal. 2025;(1):56-63. (In Russ.) https://doi.org/10.17238/2072-3180-2025-1-56-63