Updated: Aug 24
5 Applications of Machine Learning in Surgery
Preoperative planning
Anatomical classification of medical images or volumes of organs or lesions.
Detection and spatial localization of regions of interest
Segmentation of organs from CT scans and MRI
Segmentation and localization of surgical instruments
Registration: the spatial alignment between two medical images, volumes, or modalities.
Complications prediction
Intraoperative guidance
3D shape instantiation
Endoscopic navigation
Tissue feature tracking
Surgical Robotics
Perception: tracking instruments by detecting them, and optimizing the interaction between surgical tools and the environment
Autonomous navigation
Camera Guidance
Human-robot interaction: tracking 2D/3D eye-gaze points of surgeons (touchless manipulation) to assist surgical instrument control and navigation
Autonomous manipulations: autonomous suturing
Evaluation of Surgical Skills
The evaluation of surgical skills has traditionally been a subjective practice, often conducted by other trained surgeons. As robotic technology becomes more commonly used in surgeries, researchers are exploring automated methods of measuring surgical techniques.
A study presented at the 2016 World Congress on Engineering and Computer Science discussed using machine learning to evaluate surgeon performance in robot-assisted minimally invasive surgery.
The research team evaluated data collected from suturing performance and classified surgeons into two categories: novice and expert. The machine learning algorithm was developed to measure the following six features:
Completion time
Path length
Depth perception
Speed
Smoothness
Curvature
AR/VR Headsets and platforms
Platforms using augmented reality and/ or virtual reality headsets in surgery, use ML to analyze data and optimize user experience.
Updated: Sep 21
A 2015 study assessed the cognitive engagement, mental workload, and mental state between novice and expert surgeons during robotic surgery.
Surgeons were divided into three groups based on the Dreyfus model :
Beginners
Competent and proficients
Experts
The surgeons performed basic skills such as ring peg transfer and ball placement, intermediate skills such as suturing and knot tying, and advanced skills such as urethra-vesical anastomosis.
The subjects were analyzed using tool-based metrics as well as cognitive-based metrics:
Time to completion
Times the camera moved
Errors such as instrument collision
Number of times the ball was dropped
Significant differences were found between the beginners and experts when performing basic and intermediate skills, as well as a number of instrument collisions.
Competent, proficient surgeons and expert surgeons differed in terms of cognitive metrics, but not tool-based metrics.
Updated: Aug 24
Despite the numerous opportunities and applications, several challenges are associated with integrating AI into surgical education and training.
One of the most significant challenges is the lack of standardization in surgical procedures. Surgical procedures can vary significantly from one surgeon to another, making it challenging to develop standardized training programs.
5 Challenges of AI in Surgical Education
1. Data privacy and security
Using AI in surgical education and training requires collecting and storing large amounts of sensitive data.
There is a risk of this data being misused or stolen, which could have serious implications for patient privacy and security.
Patient data is highly sensitive, and it is crucial to protect patients’ privacy when using their data in AI algorithms. This requires developing appropriate security measures to ensure that patient data is not misused, hacked, or leaked.
Confidentiality of data is also important in protecting the patient’s rights, such that any sharing of patient data should be conducted in compliance with privacy and data protection regulations.
With the use of AI in surgical education and training, patients’ data may be used to develop AI algorithms. Therefore, it is essential to obtain informed consent from patients before their data is used in this way.
2. Bias and discrimination
AI algorithms can be biased, and this can lead to discrimination in surgical education and training.
It is essential to ensure that AI algorithms are developed and used in a way that is fair and unbiased.
AI algorithms are trained on large datasets of surgical procedures, and the quality of the data is essential in determining the effectiveness of the algorithm. However, there is a risk of bias in the data used to train AI algorithms.
This bias could come from the type of surgeries that are being analyzed, the demographic of the patients, or even the surgeon’s experience.
3. Lack of regulatory and standardization
There is currently a lack of regulatory frameworks around the use of AI in surgical education and training.
This can make it difficult to ensure that the technology is used ethically and responsibly.
The use of AI in surgical education and training requires standardized surgical procedures. Without standardization, it is difficult to develop AI algorithms that can accurately analyze surgical performance.
The lack of standardization could lead to AI algorithms that are not effective in identifying areas for improvement or that provide inaccurate feedback to trainees.
It is essential to develop standardized surgical procedures that are followed by all surgeons to ensure that the AI algorithms are accurate and effective.
There is a question of responsibility when using AI in surgical education and training. Who is responsible for the accuracy and safety of the AI algorithms? Who is responsible if something goes wrong during a simulated surgery?
These questions need to be addressed before AI can be fully integrated into surgical training.
4. Overreliance on technology
There is a risk that surgical trainees may become over-reliant on AI technology, and this could lead to a reduction in the development of their surgical skills.
5. Cost
The development and implementation of AI technology can be expensive.
This cost may be a barrier to the widespread adoption of AI in surgical research and education.
To overcome these challenges, it is essential to develop standardized surgical procedures, establish guidelines for patient privacy and consent, and develop AI algorithms that can adapt to the individual needs of each trainee.