• info@icbcongress.com
• English
• Persian

• Login

Create Free Account

INTERNATIONAL CONGRESS OF

CancerGenetics

• The Role of Artificial Intelligence in Tumor Surgery: A Systematic Review of Clinical Applications, Diagnostic Accuracy, and Postoperative Outcomes

• Mehdi Esmaeili,^1,* Shahrzad Fayyazfard,² Payam Pasandi ,³ Kosar Hassan Pour,⁴
  1. Faculty of Veterinary Medicine, Shahid Chamran University of Ahvaz, Ahvaz, Iran.
  2. Faculty of Veterinary Medicine, Shahid Chamran University of Ahvaz, Ahvaz, Iran.
  3. Faculty of Veterinary Medicine, Shahid Chamran University of Ahvaz, Ahvaz, Iran.
  4. Faculty of Veterinary Medicine, Shahid Chamran University of Ahvaz, Ahvaz, Iran.
  
  
  
• Introduction: Tumor surgery, as one of the main pillars of cancer treatment, has always faced challenges such as accurately determining tumor boundaries, preserving adjacent healthy tissue, reducing local recurrence, and managing intraoperative and postoperative complications. In recent years, artificial intelligence (AI), especially the subfields of machine learning (ML) and deep learning (DL) using convolutional neural networks (CNN), has been able to create a fundamental transformation in tumor surgery. AI helps surgeons make more accurate decisions by analyzing multimodal data, including preoperative images (MRI, CT, PET), intraoperative data (hyperspectral images, fluorescence, ultrasound), and genomic and pathology information. This systematic review aims to comprehensively evaluate the current applications of AI in brain, liver, lung, breast, prostate, and colorectal tumor surgery, with an emphasis on accuracy in tumor margin detection, intraoperative navigation, prediction of complications, and patient survival outcomes.
• Methods: This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. A systematic search was conducted in the PubMed/MEDLINE, Scopus, Web of Science, and IEEE Xplore databases with no time limit until January 2026. Keywords used included a combination of "Artificial Intelligence," "Machine Learning," "Deep Learning," "Surgical Oncology," "Tumor Resection," "Image-Guided Surgery," and "Clinical Outcome." Inclusion criteria included original English studies with prospective or retrospective designs, validation studies of AI models, phase I-III clinical trials, and meta-analyses that quantitatively assessed the accuracy, sensitivity, specificity, or clinical outcomes of AI-based methods in tumor surgery. Case studies, case series reports with a sample size of less than 10 patients, and animal studies were excluded. In total, from 1,287 initial records, after removing duplicates and screening the title/abstract and full text, 24 articles were included in the final synthesis.
• Results: The results of the systematic review showed that AI applications in tumor surgery can be categorized into four main areas. First, the preoperative planning and tumor delineation; CNN-based models were able to detect the border of glioma tumors with high accuracy, which was significantly different (p<0.001) compared to the surgeon's naked eye (81.3%). In liver tumors, deep learning algorithms had a sensitivity of 94.7% and a specificity of 89.2% in detecting metastatic lesions <5 mm. Second, intraoperative navigation and image-guided surgery; AI-based augmented reality (AR) systems and alignment of preoperative images with the surgical field reduced the average spatial error to less than 2.1 mm. Third, residual tumor detection and intraoperative decision-making. Fourth, prediction of complications and outcomes; Machine learning algorithms using preoperative data (age, body mass index, blood markers, ASA scores) were able to predict the occurrence of major complications after surgery for liver tumors (Clavien grade ≥III) with 86.3% accuracy (sensitivity 81%, specificity 89%). AI models based on digital pathology images also predicted 5-year disease survival in patients with colorectal adenocarcinoma with an accuracy of 81.7% (c-index=0.81).
• Conclusion: This systematic review showed that AI has evolved from a nascent technology to a reliable adjunct in tumor surgery in less than a decade. The most important clinical achievements include increased tumor delineation accuracy (an average of 11% better than traditional methods), a 25–40% reduction in surgical margin positivity, higher complete resection rates in brain and liver tumors, and the ability to reliably predict postoperative complications. However, significant challenges remain, such as a lack of integration and standardization across different algorithms and platforms, the need for large, high-quality labeled datasets, issues related to patient data privacy, a lack of sufficiently powered randomized phase III trials, and the gap between laboratory accuracy and performance in real-world clinical settings with high anatomical and pathological variability.
• Keywords: Artificial intelligence, Surgical oncology, Image-guided tumor resection, Deep learning