Artificial intelligence in gastrointestinal surgery: A systematic review

Abstract

BACKGROUND

Artificial intelligence (AI) is gaining widespread traction in surgical disciplines, particularly in gastrointestinal (GI) surgery, where it offers opportunities to enhance decision-making, improve accuracy, and optimize patient outcomes across the entire surgical continuum.

AIM

To comprehensively evaluate current AI applications in GI surgery, highlighting its role in preoperative planning, intraoperative guidance, postoperative monitoring, endoscopic diagnosis, and surgical education.

METHODS

This systematic review was conducted in accordance with PRISMA guidelines. We searched the Web of Science Core Collection through March 31, 2025 using the terms “artificial intelligence” AND “gastrointestinal surgery”. Inclusion criteria: Original, English-language, full-text articles indexed under the “Surgery” category reporting quantitative AI performance metrics in GI surgery. Exclusion criteria: Reviews, editorials, letters, conference abstracts, non-English publications, ESCI/SSCI/Index Chemicus-only papers, studies without full text, and articles outside the surgical domain. Full texts of potentially eligible studies were assessed, yielding 45 studies from an initial 955 records for qualitative and quantitative synthesis.

RESULTS

The included studies demonstrated that AI has superior performance compared to traditional clinical tools in areas such as risk prediction, lesion detection, nerve identification, and complication forecasting. Notably, convolutional neural networks, random forests, support vector machines, and reinforcement learning models were commonly used. AI-enhanced systems improved diagnostic accuracy, procedural safety, documentation quality, and educational feedback. However, there are several limitations, such as lack of external validation, dataset standardization, and explainability.

CONCLUSION

AI is transforming GI surgery from preoperative risk assessment to postoperative care and training. While many tools now match or exceed expert-level performance, successful clinical adoption requires transparent, validated models that seamlessly integrate into surgical workflows. With continued multidisciplinary collaboration, AI is positioned to become a trusted companion in surgical practice.

Key Words: Artificial intelligence; Gastrointestinal surgery; Convolutional neural networks; Random forests; Support vector machines; Reinforcement learning; Risk prediction

Core Tip: This systematic review highlights how artificial intelligence (AI) is rapidly reshaping gastrointestinal surgery across preoperative, intraoperative, and postoperative stages. By synthesizing findings from 45 original studies, this review identifies AI’s strengths in enhancing surgical precision, predicting complications, improving diagnostic accuracy, and supporting surgical education. It emphasizes that while many AI tools now match or exceed clinician performance, challenges remain in validation, integration, and interpretability. This review offers a comprehensive roadmap for clinicians and researchers seeking to responsibly implement AI in gastrointestinal surgical practice.

INTRODUCTION

Gastrointestinal (GI) surgery poses unique challenges due to the complex anatomy of the alimentary tract and the high variability of tumor location, vascular supply and tissue perfusion. Recent advances in artificial intelligence (AI) have begun to address these challenges by providing real-time image enhancement, predictive modeling, and decision support. AI-driven navigation systems that use indocyanine green angiography or hyperspectral imaging enhance intraoperative tissue delineation and have reduced complication rates in procedures like total mesorectal excision[1,2]. Machine learning algorithms such as support vector machines and random forest models facilitate personalized risk evaluation for complications prior to surgery, including anastomotic leakage following rectal cancer resection, and have shown enhanced precision compared to conventional clinical staging in forecasting lymph node involvement in early colorectal cancer[3,4]. Collectively, these advancements demonstrate that AI is poised to become an essential instrument in enhancing the safety and effectiveness of GI surgery[5]. AI facilitates patient care postoperatively by allowing for the prompt identification of issues such as surgery site infections or leakage. By continuously analyzing clinical data, AI systems can flag warning signs sooner, allowing for timely interventions and improved recovery outcomes[6,7]. Moreover, AI is beginning to reshape surgical education. Through simulation, performance analysis, and real-time feedback, it offers a new way to train and assess surgical skills, ultimately contributing to safer and more effective care. For technically demanding operations such as laparoscopic colorectal procedures, AI-powered simulation tools provide real-time feedback, helping trainees refine their skills more objectively[1,8]. Some systems even assess adherence to surgical safety steps, with promising results in shortening procedure times and improving safety during operations[9]. In the future, AI is expected to become even more embedded in surgical robotics. Advanced robotic systems such as Da Vinci and Hugo robot-assisted surgery are beginning to integrate AI algorithms that process imaging, patient data, and even surgical instrument motion to assist in decision making, enhancing surgical precision, safety, and efficiency[10,11]. However, challenges remain. The lack of standardized, high-quality datasets limits the generalizability of many models. Ethical questions, legal concerns, and patient trust are critical issues that need to be addressed before AI can be fully adopted in everyday surgical practice[12,13].

Therefore, we aimed to summarize current evidence regarding the performance, validation status, and clinical applicability of AI models in preoperative, intraoperative, and postoperative phases of GI surgery. Our aims were to: (1) Quantify AI performance metrics in GI surgical tasks; (2) Evaluate methodological quality, including external validation and dataset standardization; and (3) Identify barriers and facilitators to clinical integration. By addressing this gap, this systematic review will provide surgeons and researchers with a clear roadmap for translating AI innovations into safe and effective clinical practice.

MATERIALS AND METHODS

This systematic review was executed in compliance with the PRISMA criteria, which provide a recognized framework for guaranteeing transparency and rigor in systematic reviews. To present a comprehensive summary of the existing literature, an extensive search was conducted utilizing the Web of Science (WoS) Core Collection database, encompassing all accessible papers until March 2025. There were no limitations on publication year, enabling the tracking of the evolution of this topic throughout time. The search method included the terms “artificial intelligence” and “gastrointestinal” through the Boolean operator AND, facilitating the identification of a comprehensive and pertinent collection of papers for inclusion. The search was restricted to articles published in English. The inclusion criteria were as follows: (1) Manuscripts focused specifically on the application of AI in GI surgery; (2) Were original research articles; (3) Written in English; and (4) Indexed under the “Surgery” category in the WoS database. Review articles, conference papers, editorials, letters, abstracts, and any other publication types that did not present original research findings were excluded. Additionally, studies indexed in unrelated WoS categories (n = 366), as well as those found in ESCI (n = 80), SSCI (n = 5), and Index Chemicus (n = 1) were excluded. Articles without full-text availability (n = 1) or not written in English (n = 1) were also omitted. In total, 955 records were initially identified. After removing 456 based on publication type, 499 articles remained for title screening. Abstract screening excluded 89 additional studies based on the predefined criteria. Ultimately, 45 full-text articles were deemed eligible and included in the final synthesis.

From each included study, key information was systematically extracted, including: Author(s) and year of publication, country of origin, surgical area and purpose of AI implementation, type of AI model used, clinical context or target pathology, and reported performance metrics. Due to significant variation in study designs, AI models, and how results were reported, a meta-analysis was not feasible. Instead, a qualitative, thematic approach was adopted to synthesize the findings. Results were grouped and analyzed according to thematic categories reflecting different stages and functions of AI application within GI surgery. To provide transparency and traceability, Figure 1 illustrates the full selection process from the initial database search to the final inclusion of studies along with specific reasons for exclusion at each stage, in accordance with PRISMA standards.

Figure 1 Methodological flow diagram of literature search and study inclusion process based on PRISMA guidelines. A total of 955 records were identified from Web of Science databases. After applying exclusion criteria based on publication type, indexation status, accessibility, language, and Web of Science categories, 45 studies were deemed eligible and included in the final systematic review.

RESULTS

The word cloud in Figure 2 highlights the dominant research themes in AI applications within GI surgery. Terms such as “Cancer”, “Endoscopy”, “Surgery”, “Detection”, “Computer”, and “Diagnosis” appeared most prominently, indicating a strong focus on cancer detection, preoperative planning, and surgical decision support using endoscopic imaging. Additionally, keywords like “Capsule”, “Colonoscopy”, and “Polyp” suggest a notable emphasis on non-invasive screening techniques. Overall, this visual representation provides a comprehensive overview of the key concepts and clinical areas addressed in the reviewed literature.

Figure 2 Keyword word cloud. This figure was generated using the keywords from the 48 studies included in this systematic review. The size of each word reflects its frequency across the selected articles.

As seen in Figure 3A, the number of studies on AI applications in GI surgery has significantly increased in recent years, peaking in 2022. This trend reflects the growing research interest and clinical adoption of AI-driven techniques in surgical disciplines. Figure 3B illustrates the distribution of the selected articles across journal quartiles, with a strong concentration in Q1 journals (n = 37), indicating the high impact and academic quality of the studies included in this study. Only a small number of studies appeared in Q2 (n = 3), Q3 (n = 4), and Q4 (n = 1) journals. This distribution supports the scientific relevance and credibility of the analyzed literature.

Figure 3 Temporal and quartile-based distribution of the included studies. A: The number of publications per year from 2019 to 2025; B: The quartile rankings (Q1-Q4) of the journals in which the articles were published.

The analysis of publication venues (Table 1) shows that research on AI applications in GI surgery is primarily concentrated in a few high impact journals. Surgical endoscopy and other interventional techniques (n = 14), endoscopy (n = 11), and digestive endoscopy (n = 7) collectively account for the majority of included studies, emphasizing the central role of endoscopy-driven research in this field. The remaining publications are distributed across various journals such as Annals of Surgery, Colorectal Disease, and International Journal of Surgery, each contributing a single study. This distribution reflects both the interdisciplinary nature and the growing clinical interest in AI-assisted techniques within minimally invasive and image-guided surgical practices.

Table 1 The 45 studies’ journals included in this systematic review and the number of publications per journal.

Journal title	Number of publications
Surgical Endoscopy and Other Interventional Techniques	14
Endoscopy	11
Digestive Endoscopy	7
Annals of Surgery	1
ANZ Journal of Surgery	1
Cirugía y Cirujanos	1
Colorectal Disease	1
Diseases of the Colon & Rectum	1
International Journal of Computer Assisted Radiology and Surgery	1
International Journal of Surgery	1
Journal of Gastrointestinal Surgery	1
Journal of Surgical Research	1
Obesity Surgery	1
Surgery	1
Surgical Laparoscopy Endoscopy & Percutaneous Techniques	1

Preoperative planning and risk prediction

Preoperative planning in GI surgery aims to reduce complications, improve operative precision, and optimize patient selection. Embedding AI in the preoperative phase has produced risk-prediction models that outperform conventional scoring systems in both diagnostic accuracy and clinical relevance. Several studies illustrate how AI-based algorithms accurately forecast choledocholithiasis, colorectal neoplasia, and postoperative complications, while also improving anatomical mapping to more precisely guide surgeons during operative planning. Table 2 shows that AI has quickly become an indispensable tool for preoperative planning and risk stratification in GI surgery. Across studies, AI models ranging from image-based deep learning networks to classic machine-learning risk scores consistently outperform traditional methods in speed, accuracy and personalization. They not only predict conditions such as choledocholithiasis or colorectal neoplasia and flag patients at high risk for complications, but also improve anatomical mapping to guide surgeons before the first incision. In sum, integrating AI into the preoperative workflow goes beyond a mere technical upgrade. It represents a strategic advance toward safer and more tailored surgical care.

Table 2 Overview of studies focusing on preoperative planning and risk prediction in gastrointestinal surgery.

Ref.	Focus area	AI method/model	Key outcome	Number of data points	Results
Galvis-García et al[13], 2023	Colorectal polyp detection and classification	Deep learning, CNN, CAD (CADe/CADx)	Increased adenoma and polyp detection rates using AI assisted colonoscopy in real time	1038 patients (RCT), 8641 images (CNN study), 466 polyps (CADx), 238 lesions (endocytoscopy)	Sensitivity up to 96.5%, specificity up to 93%, accuracy up to 96.4%, F1 score approximately 94%, NPV up to 99.6%
Zhang et al[14], 2023	Diagnosis of choledocholithiasis in gallstone patients	Machine learning (7 models), AI (ModelArts)	Developed and validated AI model with high diagnostic accuracy for CBD stones prediction	1199 patients (681 with CBD stones)	ModelArts AI: Accuracy 0.97, recall 097, precision 0.971, F1 score 097; machine learning AUCs: 0.77-0.81
Ahmad et al[15], 2022	Detection of subtle and advanced colorectal neoplasia	Deep learning (ResNet 101 CNN)	High sensitivity for detecting flat lesions, sessile serrated lesions, and advanced colorectal polyps	173 polyps, 35114 polyp positive frames, 634988 polyp negative frames across multiple datasets	Per polyp sensitivity: 100%, 98.9%, and 79.5% (in subtle set); F1 score: Up to 87.9%; CNN outperformed expert and trainee endoscopists in detection speed and accuracy
Lei et al[16], 2023	Polyp detection via colon capsule endoscopy	Deep learning (CNN, AiSPEED™)	Feasibility and diagnostic accuracy of AI assisted reading compared to clinician interpretation	Target: 674 patients (597 needed for power; both prospective and retrospective recruitment)	Sensitivity/specificity of AI to be compared to clinician standard; exact results pending study is ongoing
Eckhoff et al[17], 2023	Surgical phase recognition for Ivor-Lewis esophagectomy	CNN + LSTM (TEsoNet), transfer learning	Demonstrated feasibility of knowledge transfer from sleeve gastrectomy to esophagectomy with moderate accuracy	60 sleeve gastrectomy videos, 40 esophagectomy videos (used in combinations across 5 experiments)	Single procedure accuracy: 87.7% (sleeve). Transfer learning: 23.4% overall (4 overlapping phases: 58.6%). Co training max accuracy: 40.8%
Blum et al[18], 2024	Prediction of choledocholithiasis	Logistic regression, RF, XGBoost, KNN; ensemble model	Machine learning models can outperform ASGE guidelines in predicting choledocholithiasis risk using pre MRCP data	222 patients	AUROC: 0.83 (RF), accuracy: 0.81 (ensemble), sensitivity: Up to 0.94, F1 score: Up to 0.82
Axon[19], 2020	Evolution and future of digestive endoscopy	Conceptual AI (future prediction)	Highlights AI’s potential to surpass expert level diagnostic accuracy and support real time treatment decisions		Predicts that AI will revolutionize diagnosis and treatment in endoscopy with real time support tools
Hsu et al[20], 2023	Predicting postoperative GIB after bariatric surgery	RF, XGBoost, NNs	Machine learning models outperform logistic regression in predicting GIB, aiding clinical decision making	159959 patients (632 with GIB)	RF AUROC: 0.764, sensitivity: 75.4%, specificity: 70.0%; XGBoost AUROC: 0.746; NN AUROC: 0.741; logistic regression AUROC: 0.709
Athanasiadis et al[21], 2025	Accuracy of self-assessment in laparoscopic cholecystectomy (CVS quality)		Surgeons frequently overestimate CVS performance; self-assessment alone is insufficient	25 surgeons enrolled, 13 submitted 1 video, 4 submitted 2 videos	No surgeon achieved adequate CVS per expert review; significant discrepancy between self and expert ratings on Strasberg scale
Bisschops et al[22], 2019	Advanced endoscopic imaging for colorectal neoplasia		Guidance on when to use advanced imaging (HD WLE, CE, NBI, etc.) for detection and differentiation of colorectal lesions		AI suggested for future use if validated; various imaging techniques reviewed for ADR, miss rates, lesion detection
Han et al[23], 2025	Intraoperative recognition of PAN during total mesorectal excision	DeepLabv3+ with ResNet50 backbone	AI model (AINS) achieved real time neuro recognition of PAN, aiding nerve preservation during rectal cancer surgery	1780 images (1424 training, 356 validation)	Accuracy: 0.9609, precision: 0.7494, recall: 0.6587, F1 score: 0.7011; AI outperformed surgeons (F1: 0.4568) and operated faster (3 minutes vs 25 minutes)

AI: Artificial intelligence; CNN: Convolutional neural network; CAD: Computer-aided diagnosis; RCT: Randomized controlled trial; NPV: Negative predictive value; CBD: Common bile duct; AUC: Area under the curve; LSTM: Long short-term memory; RF: Random forest; XGBoost: Extreme gradient boosting; KNN: K-nearest neighbors; ASGE: American Society for Gastrointestinal Endoscopy; MRCP: Magnetic resonance cholangiopancreatography; AUROC: Area under the receiver operating characteristic curve; NN: Neural network; GIB: Gastrointestinal bleeding; CVS: Clinical vignette scoring; HD WLE: High-definition white light endoscopy; CE: Chromoendoscopy; NBI: Narrow band imaging; ADR: Adverse drug reaction; PAN: Pelvic autonomic nerves.

Intraoperative guidance systems

The integration of AI into the operating room and endoscopy suite represents a major advance in GI surgery. By providing real-time guidance for anatomical identification and lesion characterization, these tools reduce the cognitive load on surgeons in complex cases and enhance patient safety. Table 3 summarizes recent examples across laparoscopy, white light endoscopy, and endoscopic ultrasonography that employ convolutional neural networks, object-detection frameworks like YOLOv5/YOLOv8, and segmentation networks such as DeepLabv3+. Most of these systems are designed for real-time or near real-time use and have been evaluated against expert performance or clinical outcomes, reflecting their growing reliability in the surgical setting.

Table 3 Summary of studies related to artificial intelligence-based intraoperative guidance in gastrointestinal surgery.

Ref.	Surgical context	AI method/model	Guidance function	Data type/source	Real time capability	Performance metrics
Sato et al[24], 2022	Thoracoscopic esophagectomy	DeepLabv3+ (CNN based semantic segmentation)	Recurrent laryngeal nerve identification and navigation	3000 annotated intraoperative images from 20 videos (train/val) + 40 test images from 8 other videos	Yes	Dice: AI 0.58, expert 0.62, general surgeons 0.47
Niikura et al[25], 2022	Upper GI endoscopy	Single shot multibox detector (CNN)	Detection of gastric cancer in endoscopic images	23892 white light images from 500 patients (1:1 matched to expert endoscopists)	No	Per patient: 100%, per image: 99.87%, IOU: 0.842
Yang et al[26], 2022	EUS for subepithelial lesions	ResNet 50 (CNN based deep learning)	Differentiation of gastrointestinal stromal tumors and leiomyomas using EUS	10439 EUS images from 752 patients (multicenter); 132 prospective patients for clinical validation	Yes	AUC: 0.986 (internal), 0.642 (external), accuracy: 96.2%
Schnelldorfer et al[27], 2024	Staging laparoscopy	YOLOv5 (detection), ensemble ResNet18 CNN (classification)	Identification and classification of peritoneal surface metastases	4287 lesions from 132 patients (365 biopsied lesions; 3650 image patches)	No	AUC PR: 0.69, AUROC: 0.78, accuracy: 78%
Guo et al[28], 2021	GI endoscopy (multi lesion)	Deep CNN (ResNet 50 with TTA and transfer learning)	Detecting four lesion categories to support endoscopic diagnosis	327121 WLI images for training; 33959 images in validation from 1734 cases	No (0.05 seconds/image)	Sensitivity: 88.3%, specificity: 90.3%, accuracy: 89.7%
Houwen et al[29], 2022	Colonoscopy		Defines competence standards (SODA) for AI or endoscopists in optical diagnosis	Simulation studies + systematic literature review; ESGE Delphi consensus panel	Yes	Sensitivity ≥ 90%, specificity ≥ 80% (leave in situ); ≥ 80% both (resect and discard)
Tatar and Çubukçu[30], 2024	Colonoscopy	YOLOv8 based CNN (ColoNet)	Identification of neoplastic/premalignant/malignant lesions; biopsy decision support	1760 colonoscopy images (306 patients) for training/validation; 91 external images for real time testing	Yes	mAP50: 0.832, accuracy: 82.4%, sensitivity: 70.7%, specificity: 92.0%

AI: Artificial intelligence; CNN: Convolutional neural network; GI: Gastrointestinal; IOU: Intersection over union; EUS: Endoscopic ultrasonography; AUC: Area under the curve; AUC PR: Area under the precision-recall curve; AUROC: Area under the receiver operating characteristic curve; TTA: Test time augmentation; WLI: White light imaging; SODA: Simple optical diagnosis accuracy; ESGE: European Society of Gastrointestinal Endoscopy; mAP: Mean average precision.

The studies in Table 3 highlight the versatility and expanding role of AI in intraoperative GI procedures. Whether assisting with fine anatomical recognition (e.g., nerves, tumors) or automating detection tasks in endoscopy, AI systems are increasingly achieving or surpassing expert level accuracy. Nevertheless, challenges remain, particularly in ensuring real time compatibility, external validation, and cross platform generalizability. The consistent inclusion of performance metrics and rigorous validation, as demonstrated by these studies, is critical to fostering trust and regulatory approval in the surgical AI ecosystem.

Postoperative monitoring and complication detection

AI is swiftly becoming an essential instrument for the early detection and forecasting of postoperative problems, often before their clinical manifestation. Through the ongoing analysis of varied patient data, including electronic health records, postoperative imaging, and physiological signals, AI has the capacity to enhance recovery outcomes, reduce hospital durations, and assist doctors in making prompt and educated decisions. This section emphasizes recent research demonstrating the application of AI in postoperative monitoring within GI treatment. As summarized in Table 4, these approaches show strong potential to enhance early detection of complications, minimize overtreatment, and support more precise, personalized postoperative interventions. From hospital discharge prediction to lesion identification and treatment response modeling, AI systems demonstrate strong performance across varied clinical targets. However, challenges such as external validation, interpretability, and integration into clinical workflows remain critical for translating these systems into routine use. Nevertheless, these studies collectively illustrate how AI can serve as a powerful ally in the postoperative management for GI patients.

Table 4 Artificial intelligence applications for postoperative monitoring and complication prediction in gastrointestinal care.

Ref.	Complication targeted	AI method/model	Timing of prediction	Data source/type	Performance metrics	Clinical utility
van de Sande et al[31], 2022	Need for hospital specific interventions (e.g., reoperation, radiological intervention, IV antibiotics)	Random forest (4 variants tested)	After second postoperative day	EHR data from 3 non-academic hospitals; 18 perioperative variables (e.g., age, BMI, ASA, meds, surgery time)	AUROC: 0.83, sensitivity: 77.9%, specificity: 79.2%, PPV: 61.6%, NPV: 89.3%	Supports early safe discharge and capacity management
Choi et al[32], 2024	Missed small bowel lesions in negative capsule endoscopy	CNN	After initial human reading of SBCE videos	103 negative SBCE videos retrospectively reanalyzed; images from two academic hospitals	CNN detected additional lesions in 61.2% of cases; model had > 96% accuracy in prior study (AUROC = 0.9957)	Reduces diagnostic oversight; changed diagnosis in 10.3%
Blum et al[18], 2024	Choledocholithiasis	Logistic regression, random forest, XGBoost, KNN, ensemble	Before MRCP, using pre intervention data	Retrospective data from 222 patients (clinical, biochemical, imaging variables) from Royal Hobart Hospital	Ensemble & random forest model (accuracy: 0.81, AUROC: 0.83, sensitivity: 0.94, specificity: 0.69, F1: 0.82)	Avoids unnecessary MRCP, triages patients for ERCP
Haak et al[33], 2022	Incomplete response after CRT in rectal cancer	CNNs (EfficientNet B2, Xception, etc.); FFN for clinical data; combined model	After chemoradiation, pre surgery	226 patients; 731 endoscopic images; clinical features from single institute retrospective cohort	EfficientNet B2-AUC: 0.83, accuracy: 0.75, sensitivity: 0.77, specificity: 0.75, PPV: 0.74, NPV: 0.77	Identifies candidates for non-surgical follow up
Noar and Khan[34], 2023	GP	AI derived GMAT threshold using multivariate regression	Pre-treatment (based on GMA from EGG + WLST)	30 patients with GP; GMA via EGG; WLST; gastric emptying tests; symptom scores (GCSI DD, Leeds)	Sensitivity: 96%, specificity: 75%, accuracy: 93%, AUC: 0.95 for GMAT ≥ 0.59	Guides selection for balloon dilation; personalized therapy

AI: Artificial intelligence; IV: Intravenous; EHR: Electronic health record; BMI: Body mass index; ASA: American Society of Anesthesiologists; AUROC: Area under the receiver operating characteristic curve; PPV: Positive predictive value; NPV: Negative predictive value; CNN: Convolutional neural network; SBCE: Small bowel capsule endoscopy; XGBoost: Extreme gradient boosting; KNN: K-nearest neighbors; MRCP: Magnetic resonance cholangiopancreatography; ERCP: Endoscopic retrograde cholangiopancreatography; CRT: Chemoradiotherapy; FFN: Feed forward neural network; AUC: Area under the curve; GP: Gastroparesis; GMAT: Gastric myoelectric activity threshold; EGG: Electrogastrography; WLST: Water loaded stomach test; GCSI DD: Gastroparesis cardinal symptom index-daily diary.

AI in endoscopic applications for GI cancers

AI-enhanced endoscopic systems are rapidly transforming the landscape of GI cancer diagnosis and surveillance. By enabling real time lesion detection, precise localization, and classification, these systems enhance diagnostic accuracy and standardize quality control in upper and lower GI endoscopy. Table 5 highlights a wide range of studies that apply deep learning algorithms, object detection models, and reinforcement learning in various endoscopic modalities such as white light endoscopy, capsule endoscopy, and endoscopic ultrasonography to support the detection and management of GI neoplasia. From real time detection and documentation auditing to risk stratification and capsule reading automation, these systems demonstrate exceptional performance. Notably, many models now match or exceed expert level accuracy, particularly in detecting subtle lesions or enhancing procedural completeness. However, continued efforts in prospective validation, generalizability across populations, and clinical integration are necessary to ensure routine use. Nevertheless, AI-assisted endoscopy is poised to become a standard in GI cancer diagnosis and surveillance.

Table 5 Artificial intelligence applications in endoscopic diagnosis and quality control for gastrointestinal cancers.

Ref.	Cancer type/lesion	Endoscopic modality	AI method/model	Task/objective	Dataset size/source	Performance metrics	Clinical relevance/impact
Messmann et al[35], 2022	BERN	Upper GI endoscopy	AI assisted deep learning systems	Real time detection and localization of Barrett’s neoplasia	Meta analyzer and real time work (n > 1000 images)	Sensitivity: 83.7%-95.4%, accuracy: 88%-96%	Improves detection of subtle lesions; supports targeted biopsies over Seattle protocol
Choi et al[36], 2022		EGD	CNN (squeeze and excitation network)	Classification of anatomical landmarks and completeness of photo documentation	2599 images from 250 EGD procedures (Korea University Hospital)	Landmark classification: Accuracy 97.58%, sensitivity 97.42%, specificity 99.66%; completeness detection: Accuracy 89.20%, specificity 100%	Enhances quality control in EGD by verifying complete anatomical documentation automatically
Inaba et al[37], 2024		Colonoscopy (preparation phase)	MobileNetV3 based CNN (smartphone app)	AI based stool image classification to assess bowel preparation quality	1689 images from 121 patients; 106 patient prospective validation	Accuracy: 90.2% (grade 1), 65.0% (grade 2), 89.3% (grade 3); BBPS ≥ 6 in 99.0% of app users	Improved bowel prep monitoring; 100% cecal intubation; reduced burden on patients and nurses
Zhang et al[14], 2023	Suspected choledocholithiasis	Not applicable (pre-endoscopy prediction)	ModelArts AI platform (Huawei); 7 machine learning models also tested	Predictive classification of CBD stones before cholecystectomy	1199 patients with symptomatic gallstones; retrospective, single center	ModelArts AI: Accuracy 0.97, recall 097, precision 0.971, F1 score 0.97	May outperform guideline based risk stratification; reduces unnecessary ERCP
Wu et al[38], 2021	EGC	EGD	ENDOANGEL system (CNNs + deep reinforcement learning)	Real time monitoring of blind spots and detection of EGC	1050 patients in multicenter RCT; 196 gastric lesions biopsied	Accuracy: 84.7%, sensitivity: 100%, specificity: 84.3%	Reduced blind spots, improved EGD quality, potential for real time EGC detection in clinical setting
Rondonotti et al[39], 2023	DRSPs ≤ 5 mm	Colonoscopy with blue light imaging	CAD EYE (Fujifilm, Tokyo, Japan), CNN based real time system	Optical diagnosis to support “resect and discard” strategy	596 DRSPs in 389 patients, 4 center prospective study (Italy)	NPV: 91.0%, sensitivity: 88.6%, specificity: 88.1%, accuracy: 88.4%	Meets ASGE PIVI thresholds; may enable safe omission of histology in DRSPs, especially beneficial for nonexperts
Koh et al[40], 2023	Colonic adenomas including SSA	Colonoscopy	GI Genius™ (CADe system, Medtronic, MN, United States)	Real time detection of colonic polyps and ADR improvement	298 colonoscopies; 487 AI “hits”; 250 polyps removed	Post AI ADR: 30.4% vs baseline 243% (P = 0.02); SSA rate: 5.6%	Enhanced ADR even in experienced endoscopists; improved SSA detection; supports AI use in routine colonoscopy
Yuan et al[41], 2022	Gastric lesions (EGC, AGC, SMT, polyp, PU, erosion)	White light endoscopy	YOLO based DCNN model	Multiclass diagnosis of six gastric lesions + lesion free mucosa	31388 images (29809 train/1579 test) from 9443 patients	Overall accuracy: 85.7%; EGC: Sensitivity 59.2%, specificity 99.3%; AGC: Sensitivity 100%, specificity 98.1%	Comparable to senior endoscopists; improved diagnostic accuracy and efficiency; potential for real time support in diverse gastric lesion detection
Munir et al[42], 2024		Not applicable (survey based assessment)	ChatGPT	Evaluation of AI responses to perioperative GI surgery questions	1080 responses assessed by 45 surgeons	Majority graded “fair” or “good” (57.6%); highest “very good/excellent” rate for cholecystectomy (45.3%)	ChatGPT may aid in patient education, but only 20% deemed it accurate; limited utility in reducing message load
Sudarevic et al[43], 2023	Colorectal polyps	Colonoscopy	Poseidon system (EndoMind + waterjet based AI)	AI based in situ measurement of polyp size using waterjet as reference	28 polyps in silicone model + 29 polyps in routine colonoscopies	Median error: Poseidon 7.4% (model), 7.7% (clinical); visual: 25.1%/22.1%; forceps: 20.0%	Significantly improved sizing accuracy; does not require additional tools; useful for clinical polyp surveillance and resection decisions
Tsuboi et al[44], 2020	Small bowel angioectasia	Capsule endoscopy (PillCam SB2/SB3, Medtronic, MN, United States)	CNN (single shot multibox detector)	Automatic detection of angioectasia in CE images	2237 training images, 10488 validation images (488 angioectasia, 10000 normal)	AUC: 0.998; sensitivity: 98.8%, specificity: 98.4%, PPV: 75.4%, NPV: 99.9%	Enables high accuracy detection of angioectasia; may reduce oversight and physician workload during capsule reading
Chang et al[45], 2022		Upper GI endoscopy (EGD)	ResNeSt deep learning model	Evaluate photodocumentation completeness via anatomical classification	15305 training images; 15723 test images from 472 EGD cases	Accuracy: 96.64% (deep learning model), Photodocumentation rate: 78% (esophagus duodenum), 53.8% (pharynx duodenum)	Enables automated auditing of image completeness; higher completeness linked to higher ADR; applicable for routine EGD quality control
Hwang et al[46], 2021	Small bowel hemorrhagic and ulcerative lesions	CE	VGGNet based CNN + Grad CAM	Classification and localization of hemorrhagic vs ulcerative lesions	30224 abnormal + 30224 normal images (train); 5760 images (validation)	Combined model: Accuracy 96.83%, sensitivity 97.61%, specificity 96.04%, AUROC approximately 0.996	Enhanced lesion localization without manual annotation; Grad CAM improves interpretability; supports efficient clinical CE analysis
Jazi et al[47], 2023		Not applicable (survey + clinical scenarios)	ChatGPT 4 (LLM by OpenAI)	Assess alignment of ChatGPT 4 with expert opinions on bariatric surgery suitability and recommendations	10 patient scenarios; 30 international bariatric surgeons	Expert match: 30%; ChatGPT 4 inconsistency: 40%; recommended surgery in 60% vs experts 90%	ChatGPT 4 showed limited alignment and inconsistency; suitable for education, but not yet reliable for clinical decision making
Meinikheim et al[48], 2024	BERN	Upper GI endoscopy (video based)	DeepLabV3+ with ResNet50 backbone (clinical decision support system)	Evaluate add on effect of AI on endoscopist performance in BERN detection	96 videos from 72 patients; 51273 images (train); 22 endoscopists from 12 centers	AI alone: Sensitivity 92.2%, specificity 68.9%, accuracy 81.3%; nonexperts with AI: Sensitivity up from 69.8% to 78.0%, specificity up from 67.3% to 72.7%	AI significantly improved nonexperts’ diagnostic performance and confidence; comparable accuracy to experts; highlights human AI interaction dynamics
Ahmad et al[49], 2021		Colonoscopy		Identify top research priorities for AI implementation in colonoscopy	15 international experts from 9 countries; 3 Delphi rounds	Not performance focused; methodology scores used for consensus	Provides a structured framework to guide future AI implementation research in colonoscopy; emphasizes clinical trial design, data annotation, integration, and regulation
Lazaridis et al[50], 2021		CE		Assess adherence to ESGE guidelines and future perspectives on CE use	217 respondents from 47 countries via ESGE survey	Not model based; survey: 91% performed CE with appropriate indication; 84.1% classified findings as relevant/irrelevant	Highlights variation in guideline adherence; AI identified as top development priority (56.2%); suggests need for standardization and formal CE training
Tian et al[51], 2024		EUS	CNN with attention module	Automatic identification of 14 standard BPS anatomical sites on EUS	6230 training images (1812 patients), internal: 1569 images (47 patients), external: 85322 images (131 patients from 16 centers)	Sensitivity: 89.45%-99.92%, specificity: 93.35%-99.79%, accuracy (internal): 92.1%-100%, kappa: 0.84-0.98	Outperforms beginners, comparable to experts; enables efficient, high quality anatomical identification in EUS; potential for training and standardization
He et al[52], 2020		Upper GI endoscopy (EGD)	CNN models (DenseNet 121, ResNet 50, VGG, etc.)	Automated classification of 11 anatomical sites for quality control and reporting	3704 images from 211 routine EGD cases (Tianjin Medical University Hospital, Tianjin, China)	(DenseNet 121): Accuracy approximately 91.11%, F1 scores up to 94.92% for specific sites	Supports automated quality assurance in EGD via accurate site classification; aids report generation and completeness verification

AI: Artificial intelligence; BERN: Barrett’s esophagus related neoplasia; GI: Gastrointestinal; EGD: Esophagogastroduodenoscopy; CNN: Convolutional neural network; BBPS: Boston bowel preparation scale; CBD: Common bile duct; ERCP: Endoscopic retrograde cholangiopancreatography; EGC: Early gastric cancer; RCT: Randomized controlled trial; DRSPs: Diminutive rectosigmoid polyps; NPV: Negative predictive value; ASGE: American Society for Gastrointestinal Endoscopy; PIVI: Preservation and Incorporation of Valuable endoscopic Innovations; SSA: Sessile serrated adenomas; CADe: Computer-aided detection; ADR: Adverse drug reaction; AGC: Advanced gastric cancer; SMT: Submucosal tumor; PU: Peptic ulcer; DCNN: Deep convolutional neural network; CE: Capsule endoscopy; AUC: Area under the curve; PPV: Positive predictive value; NPV: Negative predictive value; Grad-CAM: Gradient-weighted class activation mapping; AUROC: Area under the receiver operating characteristic curve; LLM: Large language model; ESGE: European Society of Gastrointestinal Endoscopy; EUS: Endoscopic ultrasonography; BPS: Biliopancreatic system.

AI-assisted surgical education and skill assessment

As AI continues to evolve within surgical practice, its applications in education and skill evaluation are becoming increasingly relevant. Rather than replacing educators, AI can serve as a scalable adjunct for performance assessment, decision making training, and guideline adherence. Table 6 summarizes recent studies exploring AI’s potential in surgical education, including the use of large language models (LLMs) like ChatGPT, as well as structured surgical training environments in robotics.

Table 6 Artificial intelligence based and robotics enhanced surgical education and performance assessment.

Ref.	Surgical procedure/task	AI method/model	Assessment modality	Data type/source	Performance metrics	Educational outcome
Garfinkle et al[53], 2022	Gastrointestinal and endoscopic surgery (priority setting)		Survey/Delphi	Survey data from SAGES member surgeons		Identified core needs: Video training, tech adoption
Huo et al[54], 2024	Surgical decision making for GERD	ChatGPT 3.5/4, copilot, Google Bard, Perplexity AI	Prompt based guideline comparison	Standardized clinical vignettes based on SAGES guidelines	Surgeons (accuracy): Bard 6/7 (85.7%), ChatGPT 4 5/7 (71.4%); patients (accuracy): Bard 4/5 (80.0%), ChatGPT 4 3/5 (60.0%); children: Copilot & Bard 3/3 (100.0%)	Revealed inconsistencies in LLM advice; need for medical domain training
Huo et al[55], 2024	Surgical management of GERD	Generic ChatGPT 4 vs customized GPT (GTS)	Prompt based guideline comparison	60 surgeon cases and 40 patient cases based on SAGES and UEG EAES guidelines	GTS (custom GPT): 100% accuracy for both surgeons (60/60) and patients (40/40), generic GPT 4 66.7% (40/60) for surgeons, 47.5% (19/40) for patients	Demonstrated impact of domain customization in LLMs
Nasir et al[56], 2021	Robotic rectal cancer surgery	No AI model used (robotics only)		RCTs, observational studies, registry data	Reduced conversion to open surgery (especially in obese/male patients); improved urogenital function; no difference in long term oncologic outcomes	Reinforced need for structured robotic training (e.g., EARCS)

AI: Artificial intelligence; SAGES: Society of American Gastrointestinal and Endoscopic Surgeons; GERD: Gastroesophageal reflux disease; LLM: Large language model; GTS: Gastroesophageal reflux disease tool for surgery; UEG: United European Gastroenterology; EAES: European Association for Endoscopic Surgery; RCTs: Randomized controlled trials; EARCS: European Academy of Robotic Colorectal Surgery.

As illustrated in Table 6, LLMs holds promise in enhancing surgical education but also pose risks if implemented without domain adaptation. Customized models like the gastroesophageal reflux disease tool for surgery vastly outperformed generic ones, reinforcing the need for medical context-specific training data. Meanwhile, robotic surgery continues to drive a parallel need for structured training curricula, which may eventually be augmented by AI-powered assessment tools such as video analysis and simulation scoring. Ultimately, future directions should aim to integrate AI into both formative (training) and summative (certification) surgical education systems.

DISCUSSION

This systematic review aimed to assess the current role of AI in GI surgery and to investigate the future trajectory of the area. The evidence collected from the included studies unequivocally demonstrates that AI is no longer a mere futuristic concept; it is being employed to improve surgical planning, decision-making in the operating room, postoperative patient monitoring, and the training of future surgeons. During the preoperative phase, AI has demonstrated significant potential for enhancing risk assessment. Models like random forests and ensemble learning have consistently surpassed conventional clinical standards in predicting the probability of choledocholithiasis and lymph node metastases. These technologies possess significant potential for facilitating more individualized surgical planning and may assist in decreasin for superfluous diagnostic tests or treatments. Intraoperatively, AI is becoming a significant support tool, functioning as an additional set of eyes. Models utilizing convolutional neural networks, YOLO algorithms, and reinforcement learning are employed to recognize anatomical features in real time, thereby augmenting surgical precision and enhancing safety, especially in intricate or high-risk procedures. Although additional validation in real-world, multi-center environments is necessary, the results are encouraging. AI is assuming a progressively significant role in the postoperative phase. Multiple models have shown proficiency in properly forecasting problems, including hemorrhage or protracted wound healing, through the analysis of data derived from electronic health records, imaging, and physiological signals. These systems possess the capacity to facilitate earlier clinical intervention, expedite recovery, and reduce hospital stays. In endoscopy, AI has facilitated improved diagnosis of GI malignancies and greater procedural documentation. Automated technologies have resulted in increased adenoma detection rates and less variability among endoscopists. Capsule endoscopy has particularly benefited from AI’s capacity to swiftly and precisely analyze large quantities of image data. The application of AI in surgical education, however nascent, demonstrates potential. The advancement of LLMs, like ChatGPT, has created new opportunities for education and training. Although general-purpose models currently lack reliability for clinical decision-making, domain-specific LLMs trained on surgical guidelines exhibit superior performance and may ultimately offer customized feedback or explanations during surgical training. Significant obstacles persist. Numerous studies examined here lacked external validation or long-term follow-up data, and several were constrained to retrospective analyses or performed within a singular institution. Interpretability continues to be a significant issue; for AI technologies to achieve broad clinical acceptability, they must be transparent and comprehensible to physicians.

Clinical implications

Our comprehensive synthesis suggests that AI-driven algorithms can meaningfully enhance GI surgical workflows by improving lesion detection rates, optimizing patient selection through more accurate risk models, and supporting real-time anatomic guidance. To translate these benefits into everyday practice, institutions must invest in staff training for AI interpretation, develop robust data governance and integration pipelines, and establish multidisciplinary teams that include surgeons and data scientists to oversee model validation and deployment.

Limitations of this study

Several methodological factors may limit the scope of our findings. First, by focusing solely on the WoS Core Collection, our search may have overlooked pertinent studies indexed in other bibliographic databases or published as conference proceedings and technical reports. While restricting our search to the “Surgery” category of the WoS Core Collection was intended to capture the most impactful GI surgery and AI publications, this methodological choice may have inadvertently overlooked pertinent studies indexed in adjacent disciplines, thereby introducing a degree of residual selection bias. Second, we did not conduct a formal grey literature search or include non-English language articles, which could introduce publication and language biases. Finally, heterogeneity in reporting standards across primary studies, particularly regarding external validation and performance metrics, constrained our ability to perform meta-analysis. Future reviews should consider broader database coverage, inclusion of unpublished and non-English sources, and standardized data extraction protocols to enhance completeness and reduce bias.

CONCLUSION

AI is rapidly transforming the practice of GI surgery and increasingly serving as a valuable and adaptable instrument in enhancing patient care, facilitating risk prediction prior to surgery, providing assistance during procedures, and monitoring recovery thereafter. Numerous systems examined in this study currently demonstrate performance comparable to, or occasionally superior to, that of seasoned physicians for specific jobs. However, for AI to integrate into routine surgical practice, several critical measures must be undertaken. These technologies must be trained on high-quality, diversified datasets. They must also be comprehensible and reliable for the medical teams utilizing them. Equally significant, they must seamlessly integrate with existing hospital systems without complicating operations. The objective is not to supplant surgeons, but to assist them by enhancing decision-making and increasing the safety and efficacy of surgical procedures. Through continuous collaboration among physicians, researchers, and developers, AI can serve as a dependable ally in both standard and intricate GI surgical procedures.