COLONOSCOPY
Artificial intelligence in bowel cancer screening
Dr. Bu’ Hayee, deputy clinical director for gastroenterology, King’s College Hospital NHS Foundation Trust, discusses the role of AI technology in reducing variation in detection rates between endoscopists.
Colorectal cancer (CRC) or bowel cancer is the fourth most common cancer, yet the second biggest cancer killer in the UK responsible for 16,000 deaths each year1-4 The mortality rate from CRC can be reduced by colorectal screening,3,5
not only by
detecting established cancers at an earlier stage, but also by detection and removal of polyps before they develop into cancers. Most commonly, CRC starts as a polyp (adenoma), a benign growth on the lining of the colon or the rectum.5
Colonoscopy is
considered the gold standard for colorectal screening, allowing physicians to detect and remove polyps,5
there is an appreciable miss-rate for polyps even in expert hands.6
can later on develop into cancers, known as post-colonoscopy colorectal cancers (PCCRCs).5,7
These missed polyps The occurrence rate of PCCRCs
is estimated to be between 1.8% - 9.0%,8,9 so while it must be that not all polyps turn into cancer, there is a biologically-plausible and statistical link between higher rates of polyp detection and removal and fewer PCCRCs. To add to the pre-existing challenges of early CRC diagnosis, the COVID-19 pandemic resulted in the suspension of all endoscopy in March-April 2020. This has resulted in a significant backlog.10,11
It is widely accepted
that early diagnosis and detection is crucial for improved prognosis and delays negatively affect survival. The delays in diagnosis due to the pandemic could have a significant impact on the survival rates.3,11-13
Furthermore, as colonoscopy services resume patients may
be hesitant to return to hospitals due to anxieties around transmissible infection.10,14 There is multicentre UK data to reassure patients that this risk is miniscule.15
Prior
to the COVID pandemic, the demand for screening colonoscopy increased year-on-year in the UK (as high as 15% in some reports).1 The provision of high-quality colonoscopy to reduce miss-rates and maintain patient experience, therefore, represents a truly significant burden to training programmes and healthcare services worldwide.
Colorectal cancer: pre-existing diagnosis challenge
The adenoma detection rate (ADR) is a key quality and performance indicator for colonoscopy worldwide,16
including
It is widely accepted that early diagnosis and detection is crucial for improved prognosis and delays negatively affect survival. The delays in diagnosis due to the pandemic could have a significant impact on the survival rates.
34 l
WWW.CLINICALSERVICESJOURNAL.COM
quality assurance in national bowel cancer screening programmes. It is defined as the proportion of colonoscopy procedures in which at least one adenoma is identified.16 Notional benchmarks ADR are set at 20% in the UK as an aspirational target, although some have suggested a higher target of 25%, specifically 30% for men and 20% in women.16
can reduce the risk of CRC by 3%.6
Increasing the ADR by 1% A
significant number of adenomas are missed or undetected during colonoscopies, roughly between 24%-35% and multiple factors can account for this.6
Training and experience
are an obvious factor, but the procedure is reliant on attentiveness and research has shown that there is a 12.4% difference between the number of adenomas detected in the morning compared to the afternoon/ evening.12
Artificial intelligence: improving colonoscopy quality The application of artificial intelligence (AI) into colonoscopy has developed computer-assisted detection (CADe), with the potential to improve detection rates.6 The first-to-market product from Medtronic,
APRIL 2021 but it is well-recognised that
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88
