AI-assisted colonoscopy enhances the detection of abnormal polyps
Gastrointestinal disorders range from mild, self-limited to more serious illnesses that may pose health risks and diminish quality of life.
Today, technological advancements have made available artificial intelligence (AI) powered by deep machine learning to augment colonoscopy. This cutting-edge innovation provides real-time alerts and aids in determining whether detected polyps are potentially harmful.
Dr. Santi Kulpatcharapong, a gastroenterologist and hepatologist at MedPark Hospital, discusses the advantages of harnessing DeepGI AI technology in colonoscopy. To what extent can these innovations enhance precision and accuracy? Let us find out!
A colonoscopy can swiftly visualize and make an accurate diagnosis.
“Gastrointestinal endoscopy allows physicians to visualize the internal linings of the digestive system, particularly in areas suspected to harbor abnormalities. The linings are not visible externally; initial assessments rely primarily on taking a thorough medical history. However, endoscopy is warranted if concerning symptoms or signs develop.”
A colonoscopy is necessary for patients experiencing bleeding, weight loss, or suspected intestinal inflammation symptoms. Colonoscopy enables direct visualization and biopsy for diagnosis and treatment of abnormalities. It is suitable for screening, even without symptoms.
For colon cancer screening, it can begin at age 45. However, those with a family history of colon cancer or early-onset colonic polyps should consider screening at an earlier age.
DeepGI improves GI abnormality detection accuracy.
“AI in gastrointestinal endoscopy serves a powerful assistant role, helping physicians detect abnormalities more effectively, particularly in the colon. It significantly increases the likelihood of identifying polyps and tumors that could develop into colorectal cancer.”
Did you know? While not all colonic polyps become cancerous, some can progress to malignancy over 10–20 years if left untreated. Detecting and assessing polyps for their cancerous potential enables timely removal, effectively preventing colorectal cancer from developing in the next 10-20 years.
Current AI technology can detect abnormalities in real time while the colonoscope navigates the colonic lumen. It can also classify the type of polyp detected and assess its potential for future malignant transformation.
Abnormal polyps occur in 3 out of 10 individuals
Gastroenterologists, relying on their expertise, knowledge, and experience, can identify abnormalities in the colon during examination. Typically, when performing a colonoscopy in patients over 45, doctors detect abnormal polyps in at least 3 out of 10 individuals.
“Before AI, detection relied solely on visual inspection expertise. The more experienced a physician is with colonoscopy, the faster they can identify abnormalities, whereas an untrained eye might miss them. The more abnormal polyps we detect, the greater the reduction in colorectal cancer risk for patients.”
In the past, colonoscopes lacked the resolution we have today. Over time, technological advancement leads to higher-resolution images, facilitating the detection of polyps or other irregularities. Despite the light-filtering technologies to aid polyp detection, they pale in comparison to the capabilities of AI.
“Experienced physicians highly skilled in detecting abnormalities may mistakenly think AI does not offer much value. However, data shows that even for highly skilled doctors, AI can increase the likelihood of detecting polyps by up to 10%.”
“Research indicates that increasing the detection rate of polyps by only 1% can reduce the risk of cancer by up to 3%.”
Dr. Santi explains that AI integration enhances the ability to detect polyps and other abnormalities. As noted earlier, a higher detection rate of polyps corresponds to a higher reduction in colorectal cancer risk.
AI can assist in classifying polyps. When polyps with malignant potential are detected, appropriate treatment can begin promptly, eliminating the need for a biopsy, treatment steps, and healthcare costs. However, a biopsy may still be necessary sometimes to confirm the diagnosis.
“Crucially, AI will evolve and become more intelligent as the AI training datasets expand. In the future, its accuracy may improve to the point where biopsies may no longer be necessary.”
The AI technology at MedPark Hospital is versatile, continuously improving detection and accuracy as its datasets grow.
MedPark Hospital places great emphasis on innovative medical technologies. The gastroenterology and hepatology clinic of MedPark Hospital has significantly contributed to research on AI-assisted colonoscopy. It leverages DeepGI, an advanced technology collaboratively developed between Chulalongkorn Hospital and the Faculty of Engineering at Chulalongkorn University, to enhance colonoscopy examinations. MedPark Hospital collaborates as a key data-gathering site for Chulalongkorn Hospital in this research.
“Numerous AI systems are in use for medical endoscopy, with many companies developing proprietary AI designed to work exclusively with their make of endoscopes. However, the advantage of DeepGI, the system employed at MedPark Hospital, is its compatibility with endoscopes from any manufacturer. This flexibility enables the AI to collect a broader range of datasets, facilitating its continuous development and refinement, ultimately enhancing its accuracy.”
Colon cancer. Do not wait for symptoms to appear. Early screening is key.
Colon cancer screening with colonoscopy is a highly efficacious preventive measure. Starting at age 45, individuals should undergo screening, even without symptoms. This point is vital because, by the time symptom manifests, cancer is often already advanced, making treatment more complex and potentially compromising your quality of life.
“If one detects the precancerous polyps, treatment merely involves their endoscopic removal, eliminating the need for surgery. It offers peace of mind. We no longer have to contend with the polyps. From a medical standpoint, screening is the most effective strategy for preventing and managing disease.”
Dr. Santi also recommends that those with a family history of colon cancer should note the age at which the diagnosis occurred and begin screening 10 years earlier. For example, if your father was diagnosed with colorectal cancer at age 50, you should start the screening at age 40.
“Colorectal cancer is increasingly being diagnosed in younger patients. Therefore, it is crucial to be attentive to unusual symptoms, assess your risk factors, and undergo screening before reaching advanced age or the symptoms manifest. This proactive approach significantly reduces the risk of developing colorectal cancer.”
