Who Misses the Cancers at Colonoscopy? Reply to Fracasso et al.

In our study of ”Reasons for failure to diagnose colorectal carcinoma at colonoscopy” [1], we found that there was a delayed diagnosis of colorectal cancer in 5.9 % of patients who underwent colonoscopy, potentially affecting their prognosis. Many of the delays were in fact system failures, in which further investigation was recommended but did not occur. We found that when a colonoscopy was not recognized as inadequate, a delay with a median duration of 10 months occurred.

The letter from Fracasso and colleagues raises the question of a possible operator effect and whether there was a difference between surgeons and physicians - i. e., whether surgeons missed more cancers than physicians. The answer is not as straightforward as suggested by the authors of the letter.

All of our endoscopists - whether gastroenterologists, general surgeons, or colorectal surgeons - have the same training requirements for accreditation to conduct colonoscopy by a conjoint committee of both surgeons and gastroenterologists, as well as meeting the individual hospital requirements for accreditation.

During the study period, a total of 5055 colonoscopies were conducted in 4598 patients entered onto the computerized reporting system and database. There were 2738 colonoscopies undertaken by gastroenterologists and 2317 by surgeons. Of the 2317 undertaken by surgeons, 867 were undertaken by the two colorectal surgeons, the rest by the seven general surgeons. The cancers missed were 10 by general surgeons, none by colorectal surgeons, five by gastroenterologists, and two by both gastroenterologists and general surgeons (i. e., at more than one colonoscopy). The best results were therefore obtained by the colorectal surgeons and the poorest by the general surgeons.

Haseman et al. [2] found 27 missed cancers, of which 20 were not reached on colonoscopy. They found that gastroenterologists had a lower miss rate than nongastroenterologists, although the ”miss” rates varied markedly between operators. This difference was not found in our own study.

With regard to the polyp question, it has been our practice to biopsy polyps before removal and to attempt to retrieve them all. The percentage of retrieved polyps reported by Fracasso and colleagues in their letter seems very high, and we are not certain whether our retrieval rate would be that good.

References

• 1 Leaper M, Johnston M J, Barclay M. et al . Reasons for failure to diagnose colorectal carcinoma at colonoscopy.  Endoscopy. 2004;  36 499-503
• 2 Haseman J H, Lemmel G T, Rahmani E Y, Rex D K. Failure of colonoscopy to detect colorectal carcinoma: evaluation of 47 cases in 20 hospitals.  Gastrointest Endosc. 1997;  45 451-455