Applications and challenges of artificial intelligence in plastic surgery imaging: A narrative review

Abstract

BACKGROUND

As artificial intelligence (AI) continues to expand across medical specialties, its application in medical imaging within plastic and reconstructive surgery (PRS) remains limited in the literature. Imaging plays a critical role in surgical planning, intraoperative decision-making, and postoperative monitoring in PRS, presenting an opportunity for AI to enhance clinical outcomes.

AIM

To evaluate the current applications of AI in medical imaging for plastic surgery, with a focus on its use in preoperative planning, intraoperative guidance, and postoperative monitoring.

METHODS

A literature search was conducted using MEDLINE, EMBASE, ScienceDirect, and OVID up to February 2025. Studies were included based on relevance to AI use in plastic surgery imaging. Extracted data included AI modality, surgical context, outcomes, and limitations. The search followed PRISMA guidelines and was registered with PROSPERO (CRD420251008741).

RESULTS

AI tools have improved preoperative planning through three-dimensional vascular mapping, augmented reality, and thermographic imaging. Intraoperatively, AI-enabled navigation and robotic systems have increased surgical precision. Postoperative AI applications, including deep learning algorithms and sensor-based monitoring, support early complication detection and wound healing assessment. However, persistent barriers include data variability, model generalizability, surgeon unfamiliarity, and lack of regulatory standards.

CONCLUSION

AI-driven imaging technologies show promise in enhancing decision-making and outcomes in PRS. To ensure safe clinical integration, future efforts must focus on structured validation, standardization, and ethical oversight.

Key Words: Artificial intelligence; Plastic surgery; Medical imaging; Machine learning; Augmented reality; Surgical navigation; Postoperative monitoring; Risk prediction

Core Tip: This narrative review demonstrates the promise of artificial intelligence (AI) applications in medical imaging for plastic and reconstructive surgery, such as preoperative planning with augmented reality, intraoperative surgical guidance, and deep learning for detecting postoperative complications. However, concerns remain regarding regulatory processes, AI bias, and data standardization. Understanding AI-driven imaging technologies will be crucial for safe clinical implementation.

INTRODUCTION

In plastic and reconstructive surgery (PRS), medical imaging is utilized to supplement surgical planning and monitoring. During preoperative planning, various imaging techniques such as magnetic resonance imaging (MRI), computed tomography (CT), and three-dimensional (3D) reconstructions support craniofacial and breast reconstruction cases by providing better visualization of anatomic structures prior to surgery[1-4]. Imaging also has its uses intraoperatively. For example, Doppler ultrasound and indocyanine fluorescence imaging can map vascular structures and assess flap viability, particularly in autologous breast reconstruction[5,6]. Postoperatively, imaging aids in detecting complications of ischemia and wound healing[7,8].

Artificial intelligence (AI) has increasingly intersected with these imaging modalities, offering tools that enhance precision, automate assessments, and support clinical decision-making in PRS. In recent years, AI has been studied across multiple medical specialties. In radiology, large-language models such as ChatGPT have been shown to improve diagnostic accuracy when analyzing X-rays, and deep-learning algorithms can either match or outperform radiologists in detecting pulmonary disease[9-11]. Subsets of AI have also extended into fields such as oncology, neurology, and cardiology. Machine learning, which uses algorithms to make predictions, has been shown to detect coronary stenosis noninvasively, while deep learning has been shown to accurately predict Alzheimer’s disease and primary brain tumors from MRI images[12-15]. In plastic and reconstructive surgery, AI is most often used in the prediction and objective evaluation of surgical outcomes[16]. Recent editorials have emphasized AI’s transformative potential in plastic surgery, including its ability to stimulate aesthetic outcomes, guide surgical planning, and automate image analysis.

This narrative review explores the scope of artificial intelligence applications in medical imaging for plastic and reconstructive surgery. By focusing on preoperative, intraoperative, and postoperative applications, we seek to highlight AI’s role in enhancing surgical planning, real-time decision making, and complication monitoring. This structure reflects recent calls for deeper integration of AI into plastic surgery imaging workflows.

MATERIALS AND METHODS

A literature search was conducted using four databases: MEDLINE, EMBASE, Evidence-Based Medicine Reviews (OVID), and ScienceDirect. Studies published up to February 2025 were included. Search terms used included but are not limited to: Plastic surgery, reconstructive surgery, artificial intelligence, machine learning, deep learning, computer vision, neural networks, augmented reality, 3D modeling, surgical planning, virtual simulation, robot-assisted surgery, computer-assisted surgery, osteotomy, perforator flap imaging, flap surveillance, wound healing, sensor technology, deep learning algorithms, and smartphone imaging. Articles focusing on AI applications in medical imaging for plastic surgery were reviewed. Extracted data included study type, AI methodology, surgical application, key findings, reported limitations, and ethical considerations. Literature used was acquired by following elements of a systematic review in literature search. A PROSPERO record containing full eligibility criteria and methodology followed for literature search has been approved (CRD420251008741). PRISMA guidelines were followed, systematic review search methods detailed in the PRISMA flow diagram (Figure 1).

Figure 1  This figure illustrates the PRISMA flow diagram of the systematic review process, showing the progression from identification through screening, eligibility, and inclusion of studies.

RESULTS

Table 1 summarizes key studies on AI applications in medical imaging for plastic surgery, including surgical procedures, measured outcomes, AI use, and findings.

Table 1 Summary of studies on artificial intelligence applications in medical imaging for plastic surgery, detailing the surgical procedure, outcomes measured, artificial intelligence utilization, type of artificial intelligence used, and key findings.

Ref.	Surgical procedure	Outcome measured	AI used for	Type of AI used	Results
Zhu et al[21], 2016	Mandibular osteotomy	Osteotomy accuracy and intraoperative deviations	Surgical navigation and osteotomy guidance	Augmented reality-based surgical navigation	Significant improvement in osteotomy precision
Qu et al[17], 2015	Distraction osteogenesis	Postoperative symmetry and distractor placement	Enhancing distractor placement accuracy	Machine learning-enhanced augmented reality	Enhanced postoperative symmetry and reduced operative time
Le et al[18], 2023	DIEP flap reconstruction	Perforator detection accuracy compared to CTA	Preoperative vascular imaging	Deep learning-based vascular mapping	High accuracy in perforator detection
Hummelink et al[19], 2019	DIEP flap breast reconstruction	Effectiveness of 3D vascular mapping	Improved intraoperative vascular mapping	3D convolutional neural networks	Reduced operative time and improved flap selection
Pereira et al[20], 2018	Perforator identification in anterolateral thigh flaps	Agreement between thermographic imaging and CTA	Validation of AI-assisted thermography	Computer vision for thermographic analysis	High correlation between AI-based and CTA imaging
Zhu et al[21], 2018	Mandibular osteotomy	Precision in osteotomy execution	Augmented reality guidance for surgery	Neural network-based surgical guidance	Improved accuracy in osteotomies
Kim et al[22], 2019	Robotic-assisted microsurgery	Enhancements in microsurgical precision	Microsurgical planning and robotic assistance	Robotic-assisted AI algorithms	Significant improvements in microsurgical execution
Brenac et al[23], 2024	Perforator flap harvest	Accuracy in perforator visualization	AI-assisted imaging for intraoperative planning	AI-powered vascular imaging	Greater accuracy in vascular visualization
Ejaz et al[24], 2024	Flap viability and perfusion assessment	Detection of ischemic areas in free flaps	Predicting flap ischemia	Sensor-based deep learning models	Early ischemia detection and improved intervention success
Avila et al[25], 2024	Postoperative wound assessment	Wound classification accuracy and prediction of healing outcomes	Postoperative monitoring and complication assessment	Convolutional neural networks	High accuracy in wound classification and healing prediction
Dhawan et al[26], 2024	AI-driven risk prediction	Postoperative risk prediction using clinical variables	Personalized risk stratification	Risk assessment models using machine learning	Improved risk prediction accuracy
Chen et al[40], 2024	Wound healing assessment	Healing time prediction based on wound morphology	Deep learning-driven wound classification	Deep learning for wound classification	AI accurately predicted wound healing times
Bukret et al[27], 2021	Aesthetic surgery risk prediction	Accuracy of AI models in predicting aesthetic surgery complications	Machine learning for surgical risk assessment	Machine learning-based risk models	Reduced complications through AI-based risk models
Borsting et al[28], 2019	Rhinoplasty outcome prediction	Predictive success of AI models in rhinoplasty outcomes	Outcome prediction in rhinoplasty	Deep learning outcome prediction models	High predictive accuracy of rhinoplasty outcomes
Farid et al[29], 2024	Breast reconstruction planning	Optimization of surgical techniques in breast reconstruction	Predictive modeling for reconstruction	Neural network predictive modeling	Enhanced decision-making in breast reconstruction planning

AI: Artificial intelligence; CTA: Computed tomography angiography.

AI in preoperative planning

Preoperative planning in plastic surgery has been enhanced by AI-driven imaging technologies, particularly in complex craniofacial procedures. AI-assisted visualization has been beneficial in distraction osteogenesis, leading to improved intraoral distractor placement and postoperative symmetry[17]. AI has also contributed to advancements in vascular mapping. Smartphone-based thermal imaging has been validated for perforator detection in deep inferior epigastric perforator flap reconstruction, providing a cost-effective alternative to computed tomographic angiography[18]. Furthermore, a randomized trial found that AI-assisted 3D vascular mapping improved perforator selection and reduced operative time[19]. Pereira et al[20] validated AI-assisted thermographic imaging for anterolateral thigh flap perforator identification, showing strong agreement with computed tomographic angiography.

AI in intraoperative navigation

Intraoperative AI applications have improved surgical accuracy through augmented reality (AR)-based navigation and decision-support systems. Zhu et al[21] demonstrated that AR-based navigation systems increase osteotomy accuracy in orbital hypertelorism by overlaying 3D surgical guides onto the operative field. Similarly, AI-powered navigation has facilitated more precise distractor placement in hemifacial microsomia, decreasing asymmetry and improving surgical outcomes[17].

Robotic-assisted microsurgery is another area where AI is being explored, particularly in nerve repair and free tissue transfer[22]. AI-powered imaging has enhanced vascular visualization for perforator flap harvests, improving intraoperative assessment and reducing operative time[23]. Additionally, Ejaz et al[24] introduced Flapbot, an AI-driven chatbot that provides real-time intraoperative flap monitoring and guidance, assisting surgeons in optimizing flap viability.

AI in postoperative monitoring

Regarding postoperative monitoring, AI tools have improved postoperative monitoring by integrating deep learning algorithms and sensor technology. Machine learning models now enable real-time detection of flap ischemia, allowing earlier intervention and reducing flap loss rates[24]. AI-powered wound assessment tools analyze biometric data to predict healing outcomes, minimizing complications[25].

Smartphone-based wound monitoring applications are also playing a crucial role in postoperative management, enabling remote patient assessments and reducing the need for in-person follow-ups[26]. Le et al[18] have demonstrated that deep learning models have high accuracy in classifying chronic wounds, burns, and surgical site infections, supporting early detection and intervention. AI-driven segmentation and tissue quantification have improved wound evaluation consistency, reducing variability among clinicians[27]. Additionally, AI models predicting healing times based on wound morphology and tissue analysis have provided valuable insights for optimizing recovery strategies.

AI in risk prediction and personalized medicine

Machine learning algorithms improve risk prediction by identifying patient-specific factors linked to surgical complications. AI-driven models predict postoperative risks based on body mass index, Caprini scores, and other clinical variables, improving patient selection[27]. AI has also been used to predict rhinoplasty outcomes with accuracy comparable to board-certified plastic surgeons[28].

AI-driven personalization is enabling more tailored approaches, optimizing surgical techniques based on patient profiles. Studies have shown machine learning can refine breast reconstruction planning by analyzing patient anatomy and previous surgical outcomes[29]. Furthermore, AI-powered aesthetic assessment tools are improving standardization and reducing subjectivity in outcome evaluation[23]. Kim et al[22] highlighted AI’s role in precision medicine by integrating patient-specific data into surgical plans. AI’s ability to analyze patient characteristics and adjust surgical approaches may improve long-term outcomes.

Ethical and practical considerations

While AI has demonstrated significant potential in plastic surgery, its widespread adoption faces several challenges. One major barrier is the limited familiarity among plastic surgeons with AI-based tools, leading to concerns about data privacy, informed consent, and reliance on AI-generated recommendations[29]. Ethical concerns also arise regarding the psychological effects of AI-generated aesthetic predictions - including unrealistic patient expectations - and the potential for bias in algorithmic decision-making.

Dhawan et al[30] outlined AI limitations in plastic surgery, including the risk of algorithmic bias, data security, and the absence of standardized validation protocols. These factors underscore the necessity of human oversight to mitigate errors and ensure ethical AI deployment in clinical settings. Without adequate safeguards, AI-driven decision-making could introduce unintended bias in both aesthetic and reconstructive outcomes.

Regulatory oversight for AI in plastic surgery remains underdeveloped. While AI-driven medical devices have gained approval in fields such as radiology and dermatology, and been approved in radiology and dermatology, plastic surgery lacks standardized frameworks for evaluating and validating AI-assisted decision-making. The absence of standardized protocols makes it challenging to assess the reliability and safety of AI tools in this field[23]. Establishing clear regulatory pathways and integrating AI literacy into surgical training will be essential for ensuring the responsible and effective adoption of AI in plastic and reconstructive surgery.

DISCUSSION

AI has been widely adopted in diagnostic fields like radiology and pathology, but its integration into surgery presents challenges due to the need for real-time adaptability and complex decision-making. Despite these obstacles, AI is demonstrating potential to enhance precision, improve patient outcomes, and optimize surgical planning and workflow efficiency[31]. Preoperative imaging and surgical planning have benefited from AI, particularly through AR and virtual reality, which enhance three-dimensional visualization and improve osteotomy accuracy and perforator flap identification[32]. AI-powered simulations assist in cosmetic surgery by predicting postoperative outcomes, improving patient expectations[33]. However, variations in surgical techniques and patient-specific factors limit model generalizability, requiring further validation.

Intraoperative AI applications, such as navigation systems, have improved precision in mandibular osteotomies and perforator flap dissections[17,23]. Robotic-assisted microsurgery is also emerging as a promising tool for complex reconstructions[22]. However, cost barriers, surgeon training requirements, and the challenge of real-time AI adaptability currently limit widespread adoption.

AI-driven sensor technology and machine learning are improving postoperative complication detection, particularly in free flap reconstruction and wound healing assessment[18,24]. These tools enable early ischemia detection, reducing flap failure risk, while smartphone-based AI applications facilitate remote patient monitoring, minimizing hospital visits[25,26]. However, their effectiveness depends on high-quality input data, as sensor accuracy and patient compliance can impact assessments.

Beyond surgical applications, AI is enhancing clinical decision support and research. Large language models assist plastic surgeons by synthesizing clinical data and guidelines, offering tailored recommendations[34]. However, concerns about reliability and bias highlight the need for domain-specific validation[35]. AI also accelerates research through automated data analysis and literature synthesis, improving access to evidence-based knowledge[36]. Additionally, AI-powered algorithms are enhancing diagnostic imaging accuracy, particularly in detecting malignancies and planning oncologic reconstructions[37]. Despite these advancements, ethical and regulatory challenges persist. Issues such as data privacy, algorithmic bias, and AI transparency must be addressed[35]. Although regulatory bodies like the FDA are evaluating AI-based medical devices, plastic surgery lags behind fields such as radiology in approvals[30]. Surgeon skepticism also remains a barrier, with concerns about AI-driven decision-making and automation bias complicating adoption[29]. Automation bias, in which clinicians over-rely on AI-generated recommendations, further complicates the integration of these technologies[38].

To maximize AI’s safe integration in plastic surgery, future research should focus on adequately addressing HIPAA compliance, improving model accuracy, evaluating and reducing potential errors, validating decision-support tools, and developing user-friendly platforms. Regulatory frameworks must also evolve to ensure AI tools meet safety and efficacy standards comparable to other medical AI applications[39]. While AI has the potential to reshape plastic surgery by enhancing precision, efficiency, and patient care, its responsible integration will require ethical oversight and clinician involvement. Addressing existing challenges related to data quality, regulatory approval, and ethical considerations will be critical to unlocking AI’s full capabilities in plastic surgery.

CONCLUSION

In medical imaging, artificial intelligence has demonstrated various advancements in plastic surgery. However, addressing challenges such as AI bias, ethical concerns, and regulatory approval will be necessary to ensure safe clinical implementation. Continued research in the field will be essential to better patient care and advance the future of plastic and reconstructive surgery. Current applications show promise in preoperative planning, intraoperative guidance, and postoperative monitoring, improving precision and workflow efficiency. Despite this progress, limitations in model generalizability, lack of standardized protocols, and surgeon unfamiliarity continue to hinder widespread adoption. Future efforts must prioritize structured clinical validation and strong ethical oversight to ensure responsible and equitable integration of AI into surgical practice.