Rates of colonoscopy in the 3 years prior to a diagnosis of colorectal cancer in Western Australia: assessment of risk factors for missed and interval cancers. Preliminary results from a population-based study

Abstract

Introduction and Aims: Colonoscopy is the gold standard for diagnosing colorectal cancer (CRC) however there is a miss rate for cancers of up to 7.9%, with significant variability in this figure from different populations (1–5). Patient, procedural, and biological risk factors are thought to have an impact on this miss rate (1, 2, 6–9). We are performing a population-based, state-wide study to determine the proportion of missed and interval CRC that occur 3–36 months after a colonoscopy in an Australian population, and assess the characteristics of these cancer. Method: All individuals with a diagnosis of CRC identified through the WA Cancer Registry between January 1st 2000 to December 31st 2011 in Western Australia (WA) were included in the study. The WA Data Linkage Unit (DLU) is a state-wide comprehensive health data linkage system that can connect health events in all WA individuals across all hospital types. Individuals with a diagnosis of inflammatory bowel disease (IBD), CRC before 2000 or with multiple CRC's between 2000 and 2011 were excluded. Individuals who had a colonoscopy 3–36 months prior to the cancer diagnosis were classed as having a possible missed or interval cancer. A review of all these colonoscopy records is being undertaken to determine if the cancer was missed at the index colonoscopy, or is an interval cancer. A cancer was classified as missed if there was a significant abnormality seen at colonoscopy or procedural issues documented on the colonoscopy report that may have resulted in the cancer or pre-cancerous lesion being missed at the colonoscopy. Interval cancers were cancers that developed after a colonoscopy that did not document an abnormality in the anatomical location where the cancer developed. Results: There were 13, 828 CRCs diagnosed in WA over the 12-year period. Of these, 660 individuals had a colonoscopy performed 3–36 months before the cancer diagnosis. Preliminary results, after review of colonoscopy records, are available on 203 of these. In this subset 26 patients were excluded due to a history of IBD, familial adenomatous polyposis (FAP) or Lynch Syndrome. A further 41 records were excluded due to a time delay of 3 or more months from the diagnostic colonoscopy to definitive management. Interval cancers occurred in 64 cases and missed cancers in 72. The cancer characteristics are outlined in Table 1. There remain 457 records to be reviewed for study completion. Table 1: Characteristics of colorectal cancers. Sporadic Missed Interval 1. Right: including and proximal to the splenic flexure, Left: distal to the splenic flexure, NOS: not otherwise specified n 13168 72 64 Male (%) 56.6 52.7 43.7 Age (min-max) 68 (12–102) 74 (36–94) 72 (37–91) Cancer location N (%) Right 4168 (31.7) Left 8499 (64.5) NOS 501 (3.8) Right 39 (54.2) Left 29 (40.3) NOS 4 (5.6) Right 47 (73.4) Left 15 (23.4) NOS 2 (3.1) Conclusions: Our preliminary findings suggest possible differences in gender and cancer location between sporadic, missed and interval CRCs which warrant further assessment to gain insights into mechanisms for these differences.