Background: While colonoscopy has become established as more accurate
than double contrast barium enema for detecting colonic polyps and can
cers, as well as offering the opportunity for therapy, there are occas
ional instances where colonoscopy is misleading. The present study is
to determine what problems occur, with a view to finding a solution. M
ethods: The records of the Colorectal Project at the Princess Alexandr
a Hospital indicate retrospectively that 346 patients have been correc
tly diagnosed with cancer of the colon and rectum by colonoscopy in th
e 5 years up to October 1996. During the same time eight patients (2.3
%) were recorded at the same hospital as being misdiagnosed by colonos
copy, the lesion being either missed completely or misplaced. Results:
In five of these patients there was failure to recognize that the who
le colon had not been examined endoscopically, thereby missing a more
proximal lesion. In two patients the lesion was missed although the en
tire colon was examined. In one patient the lesion was discovered but
inaccurately sited. Six of these mistakes would have been obviated by
the routine use of fluoroscopy to confirm the totality of the colonosc
opy and to site any lesions found. The other two cases occurred becaus
e of failure to remember that colonic examination during withdrawal sh
ould be performed meticulously back as far as the anal canal. Failure
to diagnose a colon cancer on the initial colonoscopy led to an averag
e delay of 6 months for definitive care. Conclusions: It is recommende
d that fluoroscopy be used routinely during colonoscopy to site accura
tely any lesions found, and to confirm the completeness of insertion i
f reliable landmarks, including terminal ileum, are not clearly identi
fied.