Prediction of Clavien Dindo Classification ≥ Grade III Complications After Epithelial Ovarian Cancer Surgery Using Machine Learning Methods
Abstract
1. Introduction
2. Materials and Methods
Data Handling and Machine Learning Analysis
3. Statistical Methods
4. Results
5. Discussion
5.1. Summary of Main Results
5.2. Conclusions in the Context of Published Literature
5.3. Strengths and Weaknesses
6. Implications for Practice and Future Research
7. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
Abbreviations
ACS-SRC | American College of Surgeons Surgical Risk Calculator |
ASA | American Society of Anaesthesiologists |
CDC | Clavien Dindo Classification |
DM | Diabetes Mellitus |
ECOG | Eastern Cooperative Oncology Group |
KNN | K-nearest neighbour |
ML | Machine Learning |
NACT | Neoadjuvant Chemotherapy |
PARP | Poly ADP Ribose Polimerase İnhibitors |
POSSUM | Operative Severity Score for the Enumeration of Mortality and Morbidity |
RFE | Recursive Feature Elimination |
References
- PDQ Adult Treatment Editorial Board. PDQ Cancer Information Summaries; Mar 14 OE, Fallopian Tube, and Primary Peritoneal Cancer Treatment (PDQ®): Health Professional Version; National Cancer Institute (US): Bethesda, MD, USA, 2024.
- Modugno, F.; Edwards, R.P. Ovarian Cancer: Prevention, Detection, and Treatment of the Disease and its Recurrence. Molecular Mechanisms and Personalized Medicine Meeting Report. Int. J. Gynecol. Cancer 2012, 22, S45–S57. [Google Scholar] [CrossRef] [PubMed]
- Mancari, R.; Cutillo, G.; Bruno, V.; Vincenzoni, C.; Mancini, E.; Baiocco, E.; Bruni, S.; Vocaturo, G.; Chiofalo, B.; Vizza, E. Development of new medical treatment for epithelial ovarian cancer recurrence. Gland Surg. 2020, 9, 1149. [Google Scholar] [CrossRef] [PubMed]
- Eisenkop, S.M.; Spirtos, N.M.; Friedman, R.L.; Lin, W.-C.M.; Pisani, A.L.; Perticucci, S. Relative influences of tumor volume before surgery and the cytoreductive outcome on survival for patients with advanced ovarian cancer: A prospective study. Gynecol. Oncol. 2003, 90, 390–396. [Google Scholar] [CrossRef] [PubMed]
- Zou, R.; Jiang, Q.; Luo, X.; Chen, M.; Yuan, L.; Yao, L. Cytoreductive surgery is feasible in patients with limited regional platinum-resistant recurrent ovarian cancer. World J. Surg. Oncol. 2023, 21, 375. [Google Scholar] [CrossRef]
- Bristow, R.E.; Tomacruz, R.S.; Armstrong, D.K.; Trimble, E.L.; Montz, F. Survival effect of maximal cytoreductive surgery for advanced ovarian carcinoma during the platinum era: A meta-analysis. J. Clin. Oncol. 2002, 20, 1248–1259. [Google Scholar] [CrossRef]
- Dindo, D.; Demartines, N.; Clavien, P.-A. Classification of surgical complications: A new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann. Surg. 2004, 240, 205–213. [Google Scholar] [CrossRef]
- Seagle, B.-L.L.; Butler, S.K.; Strohl, A.E.; Nieves-Neira, W.; Shahabi, S. Chemotherapy delay after primary debulking surgery for ovarian cancer. Gynecol. Oncol. 2017, 144, 260–265. [Google Scholar] [CrossRef]
- Kengsakul, M.; Nieuwenhuyzen-de Boer, G.M.; Udomkarnjananun, S.; Kerr, S.J.; van Doorn, H.C.; van Beekhuizen, H.J. Factors Predicting 30-Day Grade IIIa–V Clavien–Dindo Classification Complications and Delayed Chemotherapy Initiation after Cytoreductive Surgery for Advanced-Stage Ovarian Cancer: A Prospective Cohort Study. Cancers 2022, 14, 4181. [Google Scholar] [CrossRef]
- Straubhar, A.M.; Wolf, J.L.; Zhou, Q.C.; Iasonos, A.; Cham, S.; Wright, J.D.; Roche, K.L.; Chi, D.S.; Zivanovic, O. Advanced ovarian cancer and cytoreductive surgery: Independent validation of a risk-calculator for perioperative adverse events. Gynecol. Oncol. 2021, 160, 438–444. [Google Scholar] [CrossRef]
- Copeland, G.; Jones, D.; Walters, M. POSSUM: A scoring system for surgical audit. Br. J. Surg. 1991, 78, 355–360. [Google Scholar] [CrossRef]
- Gawande, A.A.; Kwaan, M.R.; Regenbogen, S.E.; Lipsitz, S.A.; Zinner, M.J. An Apgar score for surgery. J. Am. Coll. Surg. 2007, 204, 201–208. [Google Scholar] [CrossRef] [PubMed]
- Bilimoria, K.Y.; Liu, Y.; Paruch, J.L.; Zhou, L.; Kmiecik, T.E.; Ko, C.Y.; Cohen, M.E. Development and evaluation of the universal ACS NSQIP surgical risk calculator: A decision aid and informed consent tool for patients and surgeons. J. Am. Coll. Surg. 2013, 217, 833–842.e3. [Google Scholar] [CrossRef] [PubMed]
- Cruz, J.A.; Wishart, D.S. Applications of machine learning in cancer prediction and prognosis. Cancer Inform. 2006, 2, 117693510600200030. [Google Scholar] [CrossRef]
- Miller, R.A.; Pople, H.E., Jr.; Myers, J.D. Internist-I, an experimental computer-based diagnostic consultant for general internal medicine. In Computer-Assisted Medical Decision Making; Springer: New York, NY, USA, 1985; pp. 139–158. [Google Scholar]
- Turing, A.M. Computing Machinery and Intelligence; Springer: Berlin, Germany, 2009. [Google Scholar]
- Alpaydin, E. Introduction to Machine Learning; MIT Press: Cambridge, MA, USA, 2020. [Google Scholar]
- Liang, Z.; Zhang, G.; Huang, J.X.; Hu, Q.V. Deep learning for healthcare decision making with EMRs. In Proceedings of the 2014 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Belfast, UK, 2–5 November 2014. [Google Scholar]
- Silver, D.L. Machine lifelong learning: Challenges and benefits for artificial general intelligence. In Proceedings of the Artificial General Intelligence: 4th International Conference, AGI 2011, Mountain View, CA, USA, 3–6 August 2011. [Google Scholar]
- Rajkomar, A.; Dean, J.; Kohane, I. Machine learning in medicine. N. Engl. J. Med. 2019, 380, 1347–1358. [Google Scholar] [CrossRef]
- Fagotti, A.; Ferrandina, M.G.; Vizzielli, G.; Pasciuto, T.; Fanfani, F.; Gallotta, V.; Margariti, P.A.; Chiantera, V.; Costantini, B.; Alletti, S.G.; et al. Randomized trial of primary debulking surgery versus neoadjuvant chemotherapy for advanced epithelial ovarian cancer (SCORPION-NCT01461850). Int. J. Gynecol. Cancer 2020, 30, 1657–1664. [Google Scholar] [CrossRef]
- Van Meurs, H.S.; Tajik, P.; Hof, M.H.; Vergote, I.; Kenter, G.G.; Mol, B.W.J.; Buist, M.R.; Bossuyt, P.M. Which patients benefit most from primary surgery or neoadjuvant chemotherapy in stage IIIC or IV ovarian cancer? An exploratory analysis of the European Organisation for Research and Treatment of Cancer 55971 randomised trial. Eur. J. Cancer 2013, 49, 3191–3201. [Google Scholar] [CrossRef]
- Berek, J.S.; Renz, M.; Kehoe, S.; Kumar, L.; Friedlander, M. Cancer of the ovary, fallopian tube, and peritoneum: 2021 update. Int. J. Gynecol. Obstet. 2021, 155, 61–85. [Google Scholar] [CrossRef]
- Kumar, A.; Janco, J.M.; Mariani, A.; Bakkum-Gamez, J.N.; Langstraat, C.L.; Weaver, A.L.; McGree, M.E.; Cliby, W.A. Risk-prediction model of severe postoperative complications after primary debulking surgery for advanced ovarian cancer. Gynecol. Oncol. 2016, 140, 15–21. [Google Scholar] [CrossRef]
- Panici, P.B.; Di Donato, V.; Fischetti, M.; Casorelli, A.; Perniola, G.; Musella, A.; Marchetti, C.; Palaia, I.; Berloco, P.; Muzii, L. Predictors of postoperative morbidity after cytoreduction for advanced ovarian cancer: Analysis and management of complications in upper abdominal surgery. Gynecol. Oncol. 2015, 137, 406–411. [Google Scholar] [CrossRef]
- Inci, M.G.; Rasch, J.; Woopen, H.; Mueller, K.; Richter, R.; Sehouli, J. ECOG and BMI as preoperative risk factors for severe postoperative complications in ovarian cancer patients: Results of a prospective study (RISC-GYN—Trial). Arch. Gynecol. Obstet. 2021, 304, 1323–1333. [Google Scholar] [CrossRef]
- Cham, S.; Chen, L.; Clair, C.M.S.; Hou, J.Y.; Tergas, A.I.; Melamed, A.; Ananth, C.V.; Neugut, A.I.; Hershman, D.L.; Wright, J.D. Development and validation of a risk-calculator for adverse perioperative outcomes for women with ovarian cancer. Am. J. Obstet. Gynecol. 2019, 220, 571.e1–571.e8. [Google Scholar] [CrossRef] [PubMed]
- Rausei, S.; Uccella, S.; D’Alessandro, V.; Gisone, B.; Frattini, F.; Lianos, G.; Rovera, F.; Boni, L.; Dionigi, G.; Ghezzi, F. Aggressive surgery for advanced ovarian cancer performed by a multidisciplinary team: A retrospective analysis on a large series of patients. Surg. Open Sci. 2019, 1, 43–47. [Google Scholar] [CrossRef] [PubMed]
- Yalcin, Y.; Tatar, B.; Erdemoglu, E.; Erdemoglu, E. The parameters to estimate postoperative severe complications classified through Clavien-Dindo after upper abdominal surgery in patients with primary and recurrent ovarian cancer. Ginekol. Pol. 2019, 90, 557–564. [Google Scholar] [CrossRef] [PubMed]
- Hernandez, M.C.; Chen, C.; Nguyen, A.; Choong, K.; Carlin, C.; Nelson, R.A.; Rossi, L.A.; Seth, N.; McNeese, K.; Yuh, B.; et al. Explainable Machine Learning Model to Preoperatively Predict Postoperative Complications in Inpatients With Cancer Undergoing Major Operations. JCO Clin. Cancer Inform. 2024, 8, e2300247. [Google Scholar] [CrossRef]
- Deng, H.; Eftekhari, Z.; Carlin, C.; Veerapong, J.; Fournier, K.F.; Johnston, F.M.; Dineen, S.P.; Powers, B.D.; Hendrix, R.; Lambert, L.A.; et al. Development and validation of an explainable machine learning model for major complications after cytoreductive surgery. JAMA Netw. Open 2022, 5, e2212930. [Google Scholar] [CrossRef]
- Yang, Z.; Zhang, Y.; Zhuo, L.; Sun, K.; Meng, F.; Zhou, M.; Sun, J. Prediction of prognosis and treatment response in ovarian cancer patients from histopathology images using graph deep learning: A multicenter retrospective study. Eur. J. Cancer. 2024, 199, 113532. [Google Scholar] [CrossRef]
- Sp, S.; C, R.K.; Kr, A.; Y, R.; N, A.K.; Fernandez, A.M.; Ahuja, V. Delay in Time to Adjuvant Chemotherapy and its Impact on Oncological Outcomes in Patients Undergoing Optimal Cytoreductive Surgery for Advanced Ovarian Cancer: Analysis of 1480 Cases From the Indian HIPEC Registry. J. Surg. Oncol. 2024, 130, 1358–1363. [Google Scholar] [CrossRef]
Parameter | N (%) | Parameter | N (%) | Parameter | N (%) | |||
---|---|---|---|---|---|---|---|---|
ECOG PS | 0–1 | 139 (77.65%) | Small bowel mesentery involvemenT | No | 119 (66.85%) | Pelvic peritonectomy | No | 84 (46.93%) |
≥2 | 40 (22.35%) | Localised foci | 22 (12.36%) | Yes | 95 (53.07%) | |||
Major Cardiac Comorbidity | No | 135 (75.42%) | Diffuse, military | 37 (20.79%) | Paracolic peritonectomy | No | 123 (68.72%) | |
Yes | 44 (24.58%) | Large bowel serosal involvement | No | 103 (57.54%) | Yes | 56 (31.28%) | ||
Major Pulmonary Comorbidity | No | 164 (91.62%) | Localised foci | 44 (24.58%) | Peritonectomy (partial or total) | No | 81 (45.25%) | |
Yes | 15 (8.38%) | Diffuse, military | 32 (17.88%) | Yes | 98 (54.75%) | |||
Diabetes mellitus | No | 149 (83.24%) | Large bowel mesentery involvement | No | 81 (45.25%) | Diaphragm peritonectomy | No | 150 (83.8%) |
Yes | 30 (16.76%) | Localised foci | 58 (32.4%) | Yes | 29 (16.2%) | |||
Neoadjuvant chemotherapy | No | 116 (64.8%) | Diffuse, military | 40 (22.35%) | Appendectomy | No | 111 (62.01%) | |
3 cycles | 25 (13.97%) | Spleen metastasis | No | 159 (88.83%) | Yes | 68 (37.99%) | ||
4 cycles | 25 (13.97%) | Yes | 20 (11.17%) | Splenectomy and/or distal pancreatectomy | No | 156 (87.15%) | ||
≥6 cycles | 13 (7.26%) | Liver metastasis | No | 161 (89.94%) | Yes | 23 (12.85%) | ||
Debulking surgery | Primary | 116 (64.8%) | Any surfacce lesion | 12 (6.7%) | Lymphadenectomy | No | 59 (32.96%) | |
Interval | 63 (35.2%) | Parencyhmal | 6 (3.35%) | Selective LN debulking | 5 (2.79%) | |||
Ascites | No | 113 (63.13%) | Pleural effusion | No | 140 (78.21%) | Systemic | 115 (64.25%) | |
Small volume | 31 (17.32%) | Yes | 39 (21.79%) | Intraoperative need for blood transfusion | No | 83 (46.37%) | ||
Large volume | 35 (19.55%) | Extraabdominal LN (+) | No | 147 (82.58%) | Yes | 96 (53.63%) | ||
Omental cake | No | 113 (63.13%) | Yes | 31 (17.42%) | Need for intensive care unit | No | 72 (40.22%) | |
Yes | 66 (36.87%) | Cytoreduction | Maximal (no visible) | 128 (71.51%) | Yes | 107 (59.78%) | ||
Peritoneal carcinomatosis | No | 72 (40.22%) | Optimal (<1 cm) | 38 (21.23%) | Tumour histotype | High grade | 137 (76.54%) | |
Localised foci | 36 (20.11%) | Suboptimal (≥1 cm) | 13 (7.26%) | Others | 42 (23.46%) | |||
Diffuse, military | 71 (39.66%) | Intestinal resection | No | 138 (77.09%) | FIGO Stage | I | 34 (18.99%) | |
Diaphragmatic disease | No | 130 (72.63%) | Yes | 41 (22.91%) | II | 17 (9.5%) | ||
Localised foci | 12 (6.7%) | Small bowel resection | No | 170 (94.97%) | III | 66 (36.87%) | ||
Diffuse, military | 37 (20.67%) | Yes | 9 (5.03%) | IV | 62 (34.64%) | |||
Small bowel serosal involvement | No | 132 (74.16%) | Colorectal anastomosis | No | 148 (82.68%) | Lymph node involvement | No | 75 (41.9%) |
Localised foci | 22 (12.36%) | Yes | 31 (17.32%) | Yes | 43 (24.02%) | |||
Diffuse, military | 24 (13.48%) | İleocolic anastomosis | No | 174 (97.21%) | Unknown or no LND | 61 (34.08%) | ||
Yes | 5 (2.79%) |
Parameter | Min | Max | Median | Mean | SD |
---|---|---|---|---|---|
Age | 22 | 82 | 57 | 58 | 11 |
Serum Ca-125 level prior to debulking surgery | 3 | 19,574 | 119 | 604 | 1913 |
Serum Albumin level prior to debulking surgery | 2 | 5 | 4 | 4 | 1 |
Operative time, minutes | 120 | 610 | 300 | 322 | 97 |
Length of intensive care unit stay, days | 0 | 69 | 1 | 3 | 8 |
Postoperative length of hospital stay, days | 1 | 77 | 11 | 14 | 9 |
Day 1, albumin, g/dL | 1 | 43 | 3 | 13 | 13 |
Day 1, CRP, mg/L | 1 | 317 | 63 | 99 | 85 |
Day 2, CRP, mg/L | 13 | 435 | 210 | 209 | 91 |
Day 3, CRP, mg/L | 30 | 499 | 218 | 219 | 95 |
Day 4, CRP, mg/L | 21 | 443 | 160 | 181 | 101 |
Day 5, CRP, mg/L | 16 | 417 | 102 | 128 | 92 |
Day 6, CRP, mg/L | 8 | 455 | 68 | 98 | 84 |
Day 7, CRP, mg/L | 0 | 439 | 71 | 90 | 75 |
Training Set | Test Set | Overall | |
---|---|---|---|
Group 1 | 107 (79.85%) | 32 (71.11%) | 139 (77.65%) |
Group 2 | 27 (20.15%) | 13 (28.89%) | 40 (22.35%) |
Overall | 134 (74.86%) | 45 (25.14%) | 179 (100.0%) |
z = −1.218 p * = 0.223 |
Evaluation of Model Success | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Model | Statistics—Predicting the Desired Group (CDC Grade 3-4-5) | |||||||||||
ML Algorithms * | Time Taken to Build Model | Mean Absolute Error | Root Mean Squared Error | Kappa (κ) | TP Rate | FP Rate | Precision | Recall | Overall Accuracy ¥ | F1 Score | MCC | ROC (AUC) |
J48 | <0.001 s | 0.2373 | 0.3849 | 0.498 | 0.462 | 0.031 | 0.857 | 0.462 | 0.822 | 0.600 | 0.538 | 0.776 |
Logistic | 0.08 s | 0.3555 | 0.5963 | 0.134 | 0.385 | 0.250 | 0.385 | 0.385 | 0.644 | 0.385 | 0.135 | 0.669 |
Simple Logistic Φ | 0.22 s | 0.2488 | 0.3526 | 0.642 | 0.615 | 0.031 | 0.889 | 0.615 | 0.867 | 0.727 | 0.662 | 0.880 |
Random Forest | 0.09 s | 0.2703 | 0.3762 | 0.299 | 0.231 | 0.0 | 1.00 | 0.231 | 0.778 | 0.375 | 0.419 | 0.885 |
BayesNet † | 0.03 s | 0.2066 | 0.4143 | 0.586 | 0.769 | 0.156 | 0.667 | 0.769 | 0.822 | 0.714 | 0.589 | 0.863 |
SMO (Support Vector Machine) | 0.03 s | 0.3333 | 0.5774 | 0.129 | 0.308 | 0.188 | 0.400 | 0.308 | 0.667 | 0.348 | 0.131 | 0.560 |
MultilayerPerceptron | 0.03 s | 0.3144 | 0.5331 | 0.084 | 0.231 | 0.156 | 0.375 | 0.231 | 0.667 | 0.286 | 0.088 | 0.690 |
Random Tree | 0.03 s | 0.3061 | 0.5018 | 0.073 | 0.154 | 0.094 | 0.400 | 0.154 | 0.689 | 0.222 | 0.087 | 0.606 |
Decision Stump | 0.02 s | 0.2866 | 0.4023 | 0.4774 | 0.538 | 0.094 | 0.700 | 0.538 | 0.800 | 0.609 | 0.485 | 0.722 |
Rep Tree | 0.03 s | 0.2411 | 0.3716 | 0.572 | 0.538 | 0.031 | 0.875 | 0.538 | 0.844 | 0.667 | 0.601 | 0.739 |
PART decision | 0.03 s | 0.2091 | 0.3888 | 0.5673 | 0.692 | 0.125 | 0.692 | 0.692 | 0.822 | 0.692 | 0.567 | 0.810 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Published by MDPI on behalf of the Lithuanian University of Health Sciences. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Alci, A.; Ikiz, F.; Yalcin, N.; Gokkaya, M.; Sari, G.E.; Ureyen, I.; Toptas, T. Prediction of Clavien Dindo Classification ≥ Grade III Complications After Epithelial Ovarian Cancer Surgery Using Machine Learning Methods. Medicina 2025, 61, 695. https://doi.org/10.3390/medicina61040695
Alci A, Ikiz F, Yalcin N, Gokkaya M, Sari GE, Ureyen I, Toptas T. Prediction of Clavien Dindo Classification ≥ Grade III Complications After Epithelial Ovarian Cancer Surgery Using Machine Learning Methods. Medicina. 2025; 61(4):695. https://doi.org/10.3390/medicina61040695
Chicago/Turabian StyleAlci, Aysun, Fatih Ikiz, Necim Yalcin, Mustafa Gokkaya, Gulsum Ekin Sari, Isin Ureyen, and Tayfun Toptas. 2025. "Prediction of Clavien Dindo Classification ≥ Grade III Complications After Epithelial Ovarian Cancer Surgery Using Machine Learning Methods" Medicina 61, no. 4: 695. https://doi.org/10.3390/medicina61040695
APA StyleAlci, A., Ikiz, F., Yalcin, N., Gokkaya, M., Sari, G. E., Ureyen, I., & Toptas, T. (2025). Prediction of Clavien Dindo Classification ≥ Grade III Complications After Epithelial Ovarian Cancer Surgery Using Machine Learning Methods. Medicina, 61(4), 695. https://doi.org/10.3390/medicina61040695