Dr Afsana Elanko
Senior Educationalist and Healthcare Leader
Prof Jim Khan
Consultant Colorectal Surgeon, Clinical Director, Portsmouth Hospitals NHS Trust.
Program chair, Head of Colorectal & Robotic Program, QAH, Portsmouth.
Director of Research ASLGBI (Association of Laparoscopic Surgeons of Great Britain and Ireland)
Surgeons and clinicians are essential in leading advancements in artificial intelligence (AI) for patient benefit. However, more research into the applicability of AI and its safe use is required.
In recent years, AI and machine learning (ML) techniques have shown great promise in predicting the diagnosis and prognosis of various diseases and health conditions.
Increase in colorectal cancer
Colorectal cancer is the third most common cancer, with over 1.8 million new cases in 2018; and estimations indicate there will be around 2.4 million cases worldwide in 2035. We owe it to our patients to look at new technologies and how these can provide better knowledge for predicting outcomes and making informed decisions in their cancer care.
AI application in risk assessment
AI has been used in predicting length of stay, readmission and mortality in colorectal cancer patients during research models. In a recent study, data analytics and AI were used to predict patient outcomes after colorectal cancer surgery.
A prospectively maintained colorectal cancer database was used, looking at patients who underwent colorectal cancer surgery between 2003 and 2019. There were 47 patient parameters (demographics, peri/post-operative outcomes, surgical approaches, complications and mortality) reviewed.
In a recent study, data analytics and AI were used to predict patient outcomes after colorectal cancer surgery.
Data analytics compared the importance of each variable, and AI prediction models were built for length of stay (LOS), readmission and mortality. Using the system, accuracies of at least 80% have been achieved. The significant predictors of LOS were age, operative time, presence or absence of a stoma, robotic or laparoscopic approach to surgery and complications.
The significant predictors of readmission were age, laparoscopic procedure, stoma performed, preoperative nodal (N) stage, operation time, operation mode, previous surgery type, LOS and the specific procedure. The significant predictors of mortality were age, ASA grade, BMI, the formation of a stoma, preoperative TNM staging, neoadjuvant chemotherapy, curative resection and LOS. The patient’s quality of life during cancer care is paramount, and with AI models, it may be possible to make more informed decisions.
Potential of AI
AI has the potential to automate detection (eg. image recognition software for pattern recognition using artificial neural networks), improve efficiency (eg. act as a second reader of scans in radiology) and reduce costs. However, models need to have good data with patient safety mechanisms in place and regular quality assurance systems.
All advances must be done ethically, and responsibility for any harm due to ‘technology failure’ or privacy compromise needs to be built with the systems.
