Last week the NEJM published two major papers on screening for colon and rectal cancer. The most notable finding supports that colonoscopy – when done properly and not necessarily often – saves lives.

The NCI estimates that doctors will find over 103,000 colon and 40,000 rectal cancers, and the number of deaths will exceed 51,000 this year in the U.S. According to the ACS, colorectal cancer ranks third as a cause of cancer mortality in men and in women. In light of these numbers, the potential for screening to reduce deaths and costs of treating people with advanced disease is great.

Both analyses are unfortunately – almost dauntingly – complicated. An accompanying editorial, by Drs. M. Bretthauer and M. Kalager lends some perspective.

colon adenoma pathology, H&E stain, (Wiki Commons: "Nephron")

The first report comes from a group of researchers led by Ann Zauber, PhD, a biostatistician at MSKCC. This team examined long-term outcomes among 2602 adults who had adenomatous polyps removed between 1980 and 1990, followed by colonoscopy recommended at varying intervals in a trial. With a median follow-up of 15.8 years, there were only 12 deaths from colon cancer in the study population – essentially half the number expected by comparison with SEER data.

The main limitations I see in this report are two. First, what might be considered a good thing – the high compliance rate: 81% of those with adenomas underwent some follow-up colonoscopy. And second – along a similar vein – that the colonoscopies were performed by highly-trained physicians at academic centers in a trial that mandated a certain degree of thoroughness and quality. Some criticism of the work is that the findings won’t translate to the community at large, as mentioned in the editorial and in the paper itself. That’s because some “real world” gastroenterologists don’t perform the procedure so carefully. Apart from the trial, many people are genuinely hesitant about having colonoscopy out of concern about its unpleasantness and also costs. Compliance with colonoscopy recommendations runs low.

These are valid concerns. But they don’t abrogate the value of the procedure. Rather, they point to the need for rigorous training of doctors who do colonoscopy, for close monitoring of facilities where it’s done (and in path labs, where the specimens are evaluated), and for insurance or a national health plan to enable patients, if they choose, to have this potentially life-saving screening test covered.

The second study, from a group in Spain, examined the relative merits of checking stool samples for blood every two years vs. colonoscopy every ten years in over 50,000 people. The preliminary finding – after just one “round” of colonoscopy in those assigned to that trial arm, is that a higher proportion complied with fecal blood testing than with colonoscopy. Among those who underwent colonoscopy, cancers and adenomas were found in a greater fraction. But the absolute number of cancers detected was essentially the same in each group, because more people assigned to fecal screening completed the task.

My take from these reports, combined, is that periodic colonoscopy has the potential to halve the number of deaths from colon cancer in the general population. But it’s an unpleasant, invasive and expensive test that does carry some risks. The quality of the test – both in terms of its thoroughness and risk of complication – would depend, in part, on the training and experience of the doctor who performs the test. So, as with mammography, I favor heavy regulation and careful certification of physicians who perform these procedures.

As to how colonoscopy relates to fecal blood testing as a screening method at the population level, and the optimal start and frequency of either test, those remain uncertain. Dr. Zauber, it turns out, heads the NCI-funded National Colonoscopy Study. This ongoing work will, hopefully, shed light on how testing for blood in stool samples compares with colonoscopy in colon cancer screening and, ultimately, costs and mortality from late-stage disease.

Related Posts: