1
|
Klontzas ME, Ri M, Koltsakis E, Stenqvist E, Kalarakis G, Boström E, Kechagias A, Schizas D, Rouvelas I, Tzortzakakis A. Prediction of Anastomotic Leakage in Esophageal Cancer Surgery: A Multimodal Machine Learning Model Integrating Imaging and Clinical Data. Acad Radiol 2024:S1076-6332(24)00385-4. [PMID: 38955594 DOI: 10.1016/j.acra.2024.06.026] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/24/2024] [Revised: 04/18/2024] [Accepted: 06/18/2024] [Indexed: 07/04/2024]
Abstract
RATIONALE AND OBJECTIVES Surgery in combination with chemo/radiotherapy is the standard treatment for locally advanced esophageal cancer. Even after the introduction of minimally invasive techniques, esophagectomy carries significant morbidity and mortality. One of the most common and feared complications of esophagectomy is anastomotic leakage (AL). Our work aimed to develop a multimodal machine-learning model combining CT-derived and clinical data for predicting AL following esophagectomy for esophageal cancer. MATERIAL AND METHODS A total of 471 patients were prospectively included (Jan 2010-Dec 2022). Preoperative computed tomography (CT) was used to evaluate celia trunk stenosis and vessel calcification. Clinical variables, including demographics, disease stage, operation details, postoperative CRP, and stage, were combined with CT data to build a model for AL prediction. Data was split into 80%:20% for training and testing, and an XGBoost model was developed with 10-fold cross-validation and early stopping. ROC curves and respective areas under the curve (AUC), sensitivity, specificity, PPV, NPV, and F1-scores were calculated. RESULTS A total of 117 patients (24.8%) exhibited post-operative AL. The XGboost model achieved an AUC of 79.2% (95%CI 69%-89.4%) with a specificity of 77.46%, a sensitivity of 65.22%, PPV of 48.39%, NPV of 87.3%, and F1-score of 56%. Shapley Additive exPlanation analysis showed the effect of individual variables on the result of the model. Decision curve analysis showed that the model was particularly beneficial for threshold probabilities between 15% and 48%. CONCLUSION A clinically relevant multimodal model can predict AL, which is especially valuable in cases with low clinical probability of AL.
Collapse
Affiliation(s)
- Michail E Klontzas
- Department for Clinical Science, Intervention and Technology (CLINTEC), Division of Radiology, Karolinska Institutet, Stockholm, Sweden; Department of Medical Imaging, University Hospital of Heraklion, Crete, Greece; Computational BioMedicine Laboratory, Institute of Computer Science, Foundation for Research and Technology (FORTH), Heraklion, Crete, Greece; Department of Radiology, School of Medicine, University of Crete, Voutes Campus, Heraklion, Greece
| | - Motonari Ri
- Department of Clinical Science, Intervention and Technology (CLINTEC), Division of Surgery and Oncology, Karolinska Institutet, Solna, Sweden; Department of Upper Abdominal Diseases, Karolinska University Hospital, Huddinge, Stockholm, Sweden
| | - Emmanouil Koltsakis
- Department for Clinical Science, Intervention and Technology (CLINTEC), Division of Radiology, Karolinska Institutet, Stockholm, Sweden; Department of Diagnostic Radiology, Karolinska University Hospital, Stockholm, Sweden
| | - Erik Stenqvist
- Department of Diagnostic Radiology, Karolinska University Hospital, Stockholm, Sweden
| | - Georgios Kalarakis
- Department for Clinical Science, Intervention and Technology (CLINTEC), Division of Radiology, Karolinska Institutet, Stockholm, Sweden; Department of Radiology, School of Medicine, University of Crete, Voutes Campus, Heraklion, Greece; Department of Diagnostic Radiology, Karolinska University Hospital, Stockholm, Sweden
| | - Erik Boström
- Department of Diagnostic Radiology, Karolinska University Hospital, Stockholm, Sweden
| | - Aristotelis Kechagias
- Department of Digestive Surgery, Kanta-Häme Central Hospital, Hämeenlinna 13530, Finland
| | - Dimitrios Schizas
- First Department of Surgery, National and Kapodistrian University of Athens, Laikon General Hospital, Athens 115-27, Greece
| | - Ioannis Rouvelas
- Department of Clinical Science, Intervention and Technology (CLINTEC), Division of Surgery and Oncology, Karolinska Institutet, Solna, Sweden; Department of Upper Abdominal Diseases, Karolinska University Hospital, Huddinge, Stockholm, Sweden
| | - Antonios Tzortzakakis
- Department for Clinical Science, Intervention and Technology (CLINTEC), Division of Radiology, Karolinska Institutet, Stockholm, Sweden; Medical Radiation Physics and Nuclear Medicine, Section for Nuclear Medicine, Karolinska University Hospital, Huddinge, Stockholm, Sweden.
| |
Collapse
|
2
|
Aiolfi A, Bona D, Bonitta G, Bonavina L. Short-term Outcomes of Different Techniques for Gastric Ischemic Preconditioning Before Esophagectomy: A Network Meta-analysis. Ann Surg 2024; 279:410-418. [PMID: 37830253 DOI: 10.1097/sla.0000000000006124] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/14/2023]
Abstract
BACKGROUND Ischemia at the anastomotic site plays a critical role determinant in the development of anastomosis-related complications after esophagectomy. Gastric ischemic conditioning (GIC) before esophagectomy has been described to improve the vascular perfusion at the tip of the gastric conduit with a potential effect on anastomotic leak (AL) and stenosis (AS) risk minimization. Laparoscopic (LapGIC) and angioembolization (AngioGIC) techniques have been reported. PURPOSE Compare short-term outcomes among different GIC techniques. MATERIALS AND METHODS Systematic review and network meta-analysis. One-step esophagectomy (noGIC), LapGIC, and AngioGIC were compared. Primary outcomes were AL, AS, and gastric conduit necrosis (GCN). Risk ratio (RR) and weighted mean difference (WMD) were used as pooled effect size measures, whereas 95% credible intervals (CrIs) were used to assess relative inference. RESULTS Overall, 1760 patients (14 studies) were included. Of those, 1028 patients (58.4%) underwent noGIC, 593 (33.6%) LapGIC, and 139 (8%) AngioGIC. AL was reduced for LapGIC versus noGIC (RR=0.68; 95% CrI 0.47-0.98) and AngioGIC versus noGIC (RR=0.52; 95% CrI 0.31-0.93). Similarly, AS was reduced for LapGIC versus noGIC (RR=0.32; 95% CrI 0.12-0.68) and AngioGIC versus noGIC (RR=1.30; 95% CrI 0.65-2.46). The indirect comparison, assessed with the network methodology, did not show any differences for LapGIC versus AngioGIC in terms of postoperative AL and AS risk. No differences were found for GCN, pulmonary complications, overall complications, hospital length of stay, and 30-day mortality among different treatments. CONCLUSIONS Compared to noGIC, both LapGIC and AngioGIC before esophagectomy seem equivalent and associated with a reduced risk for postoperative AL and AS.
Collapse
Affiliation(s)
- Alberto Aiolfi
- Department of Biomedical Science for Health, Division of General Surgery, I.R.C.C.S. Ospedale Galeazzi-Sant'Ambrogio, University of Milan, Italy
| | - Davide Bona
- Department of Biomedical Science for Health, Division of General Surgery, I.R.C.C.S. Ospedale Galeazzi-Sant'Ambrogio, University of Milan, Italy
| | - Gianluca Bonitta
- Department of Biomedical Science for Health, Division of General Surgery, I.R.C.C.S. Ospedale Galeazzi-Sant'Ambrogio, University of Milan, Italy
| | - Luigi Bonavina
- Department of Biomedical Sciences for Health, Division of General and Foregut Surgery, University of Milan, IRCCS Policlinico San Donato, Milan, Italy
| |
Collapse
|
3
|
Schiffmann LM, de Groot E, Albert MC, Quaas A, Pinto Dos Santos D, Babic B, Fuchs HF, Walczak H, Chon SH, Ruurda JP, Kashkar H, Bruns CJ, Schröder W, van Hillegersberg R. Laparoscopic ischemic conditioning of the stomach prior to esophagectomy induces gastric neo-angiogenesis. EUROPEAN JOURNAL OF SURGICAL ONCOLOGY 2023; 49:107096. [PMID: 37801834 DOI: 10.1016/j.ejso.2023.107096] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/04/2023] [Revised: 09/20/2023] [Accepted: 09/25/2023] [Indexed: 10/08/2023]
Abstract
BACKGROUND The risk of an anastomotic leakage (AL) following Ivor-Lewis esophagectomy is increased in patients with calcifications of the aorta or a stenosis of the celiac trunc. Ischemic conditioning (ISCON) of the gastric conduit prior to esophagectomy is supposed to improve gastric vascularization at the anastomotic site. The prospective ISCON trial was conducted to proof the safety and feasibility of this strategy with partial gastric devascularization 14 days before esophagectomy in esophageal cancer patients with a compromised vascular status. This work reports the results from a translational project of the ISCON trial aimed to investigate variables of neo-angiogenesis. METHODS Twenty esophageal cancer patients scheduled for esophagectomy were included in the ISCON trial. Serum samples (n = 11) were collected for measurement of biomarkers and biopsies (n = 12) of the gastric fundus were taken before and after ISCON of the gastric conduit. Serum samples were analyzed including 62 different cytokines. Vascularization of the gastric mucosa was assessed on paraffin-embedded sections stained against CD34 to detect the degree of microvascular density and vessel size. RESULTS Between November 2019 and January 2022 patients were included in the ISCON Trial. While serum samples showed no differences regarding cytokine levels before and after ISCON biopsies of the gastric mucosa demonstrated a significant increase in microvascular density after ISCON as compared to the corresponding gastric sample before the intervention. CONCLUSION The data prove that ISCON of the gastric conduit as esophageal substitute induces significant neo-angiogenesis in the gastric fundus which is considered as surrogate of an improved vascularization at the anastomotic site.
Collapse
Affiliation(s)
- L M Schiffmann
- Department of General, Visceral, Cancer and Transplantation Surgery, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany
| | - E de Groot
- Department of Surgery, University Medical Center Utrecht, POBOX 85500, 3508 GA, Utrecht, the Netherlands
| | - M C Albert
- Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases (CECAD), University of Cologne, Cologne, Germany; Center for Biochemistry, Faculty of Medicine and University Hospital of Cologne, Cologne, Germany
| | - A Quaas
- Institute of Pathology, Faculty of Medicine and University Hospital Cologne, University of Cologne, Germany
| | - D Pinto Dos Santos
- Institute of Diagnostic and Interventional Radiology, Medical Faculty and University Hospital Cologne, University of Cologne, Cologne, Germany
| | - B Babic
- Department of General, Visceral, Cancer and Transplantation Surgery, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany
| | - H F Fuchs
- Department of General, Visceral, Cancer and Transplantation Surgery, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany
| | - H Walczak
- Center for Biochemistry, Faculty of Medicine and University Hospital of Cologne, Cologne, Germany
| | - S-H Chon
- Department of General, Visceral, Cancer and Transplantation Surgery, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany
| | - J P Ruurda
- Department of Surgery, University Medical Center Utrecht, POBOX 85500, 3508 GA, Utrecht, the Netherlands
| | - H Kashkar
- Excellence Cluster on Cellular Stress Responses in Aging-Associated Diseases (CECAD), University of Cologne, Cologne, Germany; Institute for Molecular Immunology, Faculty of Medicine and University Hospital of Cologne, University of Cologne, Cologne, Germany
| | - C J Bruns
- Department of General, Visceral, Cancer and Transplantation Surgery, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany
| | - W Schröder
- Department of General, Visceral, Cancer and Transplantation Surgery, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany
| | - R van Hillegersberg
- Department of Surgery, University Medical Center Utrecht, POBOX 85500, 3508 GA, Utrecht, the Netherlands.
| |
Collapse
|
4
|
Liu XR, Liu F, Zhang W, Peng D. The aortic calcification is a risk factor for colorectal anastomotic leakage. Updates Surg 2023; 75:1857-1865. [PMID: 37594659 DOI: 10.1007/s13304-023-01630-4] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/26/2023] [Accepted: 08/12/2023] [Indexed: 08/19/2023]
Abstract
PURPOSE The current pooling up analysis aimed to evaluate whether aortic calcification (AC) was a potential risk factor for anastomotic leakage (AL) after colorectal surgery. METHODS In this study, we searched studies in three databases including PubMed, Embase, and the Cochrane Library on April 20, 2022. In order to investigate the association between AC and AL, the hazard ratios (HRs) and 95% confidence intervals (CIs) of AC were pooled up. Our study was performed with RevMan 5.3 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. RESULTS We finally enrolled eight studies involving 1955 patients for statistical analysis. As for all patients, we found that AC could significantly increase the risk of AL after surgery (HR = 2.31, I2 = 0%, 95%CI = 1.58 to 3.38, P < 0.01). In five studies including patients undergoing colorectal surgery (benign diseases and cancers), AC was also a risk factor for AL (HR = 3.30, I2 = 2%; 95%CI = 1.83 to 5.95, P < 0.01). In terms of the other three studies that only included CRC patients, there was still a correction between AC and AL (HR = 1.80, I2 = 0%, 95%CI = 1.10 to 2.96, P = 0.02). CONCLUSION Patients with AC were more likely to develop AL after colorectal surgery. Moreover, subgroup analysis suggested that AC was a predictor for AL after CRC surgery.
Collapse
Affiliation(s)
- Xu-Rui Liu
- Department of Gastrointestinal Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, 400016, China
| | - Fei Liu
- Department of Gastrointestinal Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, 400016, China
| | - Wei Zhang
- Department of Gastrointestinal Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, 400016, China
| | - Dong Peng
- Department of Gastrointestinal Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, 400016, China.
| |
Collapse
|
5
|
Aiolfi A, Bona D, Bonitta G, Bonavina L. Effect of gastric ischemic conditioning prior to esophagectomy: systematic review and meta-analysis. Updates Surg 2023; 75:1633-1643. [PMID: 37498484 DOI: 10.1007/s13304-023-01601-9] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/23/2023] [Accepted: 07/17/2023] [Indexed: 07/28/2023]
Abstract
Ischemia at the anastomotic site is thought to be a protagonist in the development of anastomosis-related complications while different strategies to overcome this problem have been reported. Gastric ischemic conditioning (GIC) prior to esophagectomy has been described with this intent. Evaluate the effect of GIC on anastomotic complications after esophagectomy. Scopus, Web of Science, MEDLINE, and PubMed were investigated up to March 31st, 2023. We considered articles that appraised short-term outcomes after GIC vs. no GIC in patients undergoing esophagectomy. Anastomotic leak (AL), anastomotic stricture (AS), and gastric conduit necrosis (GCN) were primary outcomes. Risk ratio (RR) and standardized mean difference (SMD) were used as pooled effect size measures, whereas 95% confidence intervals (95% CIs) were used to calculate related inference. Fourteen studies (1760 patients) were included. Of those, 732 (41.6%) underwent GIC, while 1028 (58.4%) underwent one-step esophagectomy. Compared with no GIC, GIC was related to a reduced RR for AL (R RR = 0.63; 95% CI 0.47-0.86; p < 0.01) and AS (RR = 0.51; 95% CI 0.29-0.91; p = 0.02), whereas no differences were found for GCN (RR = 0.56; 95% CI 0.19-1.61; p = 0.28). Postoperative pneumonia (RR = 1.09; p = 0.99), overall complications (RR = 0.87; p = 0.19), operative time (SMD - 0.58; p = 0.07), hospital stay (SMD 0.66; p = 0.09), and 30-day mortality (RR = 0.69; p = 0.22) were comparable. GIC prior to esophagectomy seems associated with a reduced risk for AL and AS. Further studies are necessary to identify the subset of patients who can benefit from this procedure, the optimal technique, and the timing of GIC prior to esophagectomy.
Collapse
Affiliation(s)
- Alberto Aiolfi
- I.R.C.C.S. Ospedale Galeazzi-Sant'Ambrogio, Division of General Surgery, Department of Biomedical Science for Health, University of Milan, Via C. Belgioioso, 173, 20157, Milan, Italy.
| | - Davide Bona
- I.R.C.C.S. Ospedale Galeazzi-Sant'Ambrogio, Division of General Surgery, Department of Biomedical Science for Health, University of Milan, Via C. Belgioioso, 173, 20157, Milan, Italy
| | - Gianluca Bonitta
- I.R.C.C.S. Ospedale Galeazzi-Sant'Ambrogio, Division of General Surgery, Department of Biomedical Science for Health, University of Milan, Via C. Belgioioso, 173, 20157, Milan, Italy
| | - Luigi Bonavina
- IRCCS Policlinico San Donato, Division of General and Foregut Surgery, Department of Biomedical Sciences for Health, University of Milan, Milan, Italy
| |
Collapse
|
6
|
Development of a Novel Scoring Model to Estimate the Severity Grade of Mesenteric Artery Stenosis. J Clin Med 2022; 11:jcm11247420. [PMID: 36556035 PMCID: PMC9785168 DOI: 10.3390/jcm11247420] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/02/2022] [Revised: 11/19/2022] [Accepted: 12/12/2022] [Indexed: 12/23/2022] Open
Abstract
Objective: This study aimed to derive a new scoring model from estimating the severity grade of mesenteric artery stenosis. We sought to analyze the relationship between the new scoring model and the development, treatment, and mortality of chronic mesenteric ischemia (CMI). Methods: This retrospective study included 242 patients (128 (53%) women and 114 (47%) men) with suspected CMI from January 2011 to December 2020. A weighted sum six-point score (CSI-score; the celiac artery is abbreviated by “C”, superior mesenteric artery by “S”, and inferior mesenteric artery by “I”) based on the number of affected vessels and the extent and grade of the stenosis or occlusion of the involved visceral arteries was derived by maximizing the area under the ROC curve. The calculated CSI-score ranged from 0 to 22. The patients were divided according to the best cut-off point into low-score (CSI-score < 8) and high-score (CSI-score ≥ 8) groups. Results: The area under the receiver operating characteristic curve (AUC) of the CSI-score was 0.86 (95% CI, 0.82−0.91). The best cut-off point of “8” represented the highest value of Youden’s index (0.58) with a sensitivity of 87% and specificity of 72%. The cohort was divided according to the cut-off point into a low-score group (n = 100 patients, 41%) and high-score group (n = 142 patients, 59%) and according to the clinical presentation into a CMI group (n = 109 patients, 45%) and non-CMI group (n = 133 patients, 55%). The median CSI-score for all patients was 10 (range: 0 -22). High-scoring patients showed statistically significant higher rates of coronary artery disease (54% vs. 36%, p = 0.007), chronic renal insufficiency (50% vs. 30%, p = 0.002), and peripheral arterial disease (57% vs. 16%, p < 0.001). A total of 109 (45%) patients underwent invasive treatment of the visceral arteries and were more often in the high-score group (69% vs. 11%, p < 0.001). Of those, 79 (72%) patients underwent primary endovascular treatment, and 44 (40%) patients underwent primary open surgery or open conversion after endovascular treatment. Sixteen (7%) patients died during the follow-up, with a statistically significant difference between high- and low-scoring patients (9% vs. 0%, p = 0.008). The score stratification showed that the percentage of patients treated with endovascular and open surgical methods, the recurrence of the stenosis or failure of the endovascular treatment, the need for a bypass procedure, and the mortality rates significantly increased in the subgroups. The CSI-score demonstrated an excellent ability to discriminate between patients who needed treatment and those who did not, with an AUC of 0.87 (95% CI, 0.82−0.91). Additionally, the CSI-score’s ability to predict the patients’ mortality was moderate, with an AUC of 0.73 (95% CI, 0.62−0.83). Conclusions: The new scoring model can estimate the severity grade of the stenosis of the mesenteric arteries. Our study showed a strong association of the score with the presence of chronic mesenteric ischemia, the need for treatment, the need for open surgery, and mortality.
Collapse
|