The risk of colorectal cancer occurring during surveillance colonoscopies in carriers of pathogenic variants in mismatch repair genes has attracted much attention. A lot is known, but the biology of tumour development, the degree to which colonoscopy prevents colorectal cancers (through polypectomies), gene and gender specific considerations and intrafamilial risk, afford complexities which frustrates successful application of guidelines to prevent cancer. Further, mismatch repair variant carriers, like the rest of us, are not mice, and real-life issues impacting on adherence to scheduling deserve psychosocial research and artful consideration.

In this issue of the Journal, Van Liere et al. [1] report that two-yearly colonoscopy surveillance in Lynch syndrome is burdensome and frequently delayed. The optimal colonoscopy interval in Lynch syndrome remains a matter of debate.

There are biological issues at play. The concept that risk should relate linearly to frequency of colonoscopy ignores the logical nexus which should relate frequency of colonoscopy to velocity of tumour growth. The risk maybe high, even very high, but if the tumours grow slowly, then the need for frequent colonoscopy is not rational. On the other hand, rapid tumorigenesis justifies frequent colonoscopies. Are we dealing with hares or tortoises [2].

Of course, if we do have data that allows a relaxation of frequency, then the burden of colonoscopies for the individual, and the healthcare system, is mitigated. Most patients would be happy with that, and adherence might follow.

So how does the data look? Influential cohort experiences have been published commencing with early Lynch Syndrome registry data from the Netherlands and Germany [3, 4]. Cancers were found in surveillance, escaping any prophylactic potential of colonoscopy. Some even occurred early during follow up in those programmes offering annual colonoscopy. However, cancer staging was notably favourable for screen-detected cancers. Recent analysed experience from Canada also supports the need for diligent, frequent (1–2 yearly) follow-up colonoscopies [5]. The Canadian experience is convincing in that the population of 429 Lynch Syndrome patients was dichotomized between those who developed adenomas and those who did not. In the adenoma group there was a clear signal of reduced colorectal cancer risk for those who had colonoscopy every 1–2 years, compared to those scoped at 2–3 yearly intervals in both males and females. Frequency mattered in this experience. Equally intriguing, those who had no adenomas during surveillance (half the cohort) had the highest colorectal cancer risk at 20 years. These observations support the need for frequent colonoscopies both for those destined to form adenomas (due to this increased risk) and for non-adenoma bearers (to detect early cancer developing independently of polyps) [6].

Evidenced-based modelling studies of gene and gender stratified populations suggest a start age and frequency later and less frequent, for MSH6 and PMS2 variant carriers (35 years and 2–3 yearly) [7, 8]. The penetrance of MLH1 and MSH2 variant carriers on the other hand does not allow the later start time at least as agreed in most guidelines. This differential surveillance by gene is now being implemented in various guidelines around the world. Such prescription however needs prudent implementation given variable intragenic and even intravariant penetrance estimates in large studies [9]. Note that the confidence intervals in these estimates is relatively large.

However, there are challenging lines of evidence to suggest that frequent colonoscopies, even in MLH1 and MSH2 variant carriers, may represent over-surveillance [10,11,12,13]. Although there has not been a randomized controlled trial of surveillance intervals in Lynch Syndrome, observational and ecological studies point to a lack of survival benefit, and even increased incidence, in colorectal cancer of programmes of frequent versus less frequent colonoscopies in Lynch Syndrome.

How could this be? The brilliant work from Heidelberg, Germany, identifying mismatch repair deficient crypts in Lynch Syndrome, and the immunogenic nature of unstable micro-satellites occurring characteristically in Lynch Syndrome, provides an explanation that perhaps the polyps and antecedent mismatch repair deficient crypts, and even macroscopic cancers, can be immunoedited from the colon [14]. Even many macroscopic MSI-H cancers diagnosed through annual colonoscopies may have been so deleted if a three-year window were implemented, meaning potentially it is not the careful colonoscopy with advanced detection techniques (high definition, narrow band imaging) but our clever immune systems that mostly have done the trick. No need for concern about compliance to frequent colonoscopic scheduling here! Other support for this hypothesis comes from the comparison of the International Mismatch Repair Consortium and the Prospective Lynch Syndrome Database [9, 10]. More, rather than less, cancers have been documented in the prospective compared with retrospective cohort. The IMMRC data is retrospective and although some have argued its colonoscopy scheduling and outcomes are less intense than the PLSD (which is the premise of the hypothesis supporting immunoediting, but has never been substantiated), retrospective series would not have been supported by the more advanced and effective colonoscopic facilities generally available contemporarily to the PSLD investigators. A priori, one would expect more cancers in the retrospective observations over time given the potential for polypectomy to reduce cancers in the more intensive colonoscopy practice in the prospective series.

So how do we resolve this? It is not true, as stated by some authors, that “colonoscopy fails to prevent colorectal cancer”, but it is likely that “not all colorectal cancers are prevented by colonoscopic surveillance”. A randomized controlled trial – multinational, InSiGHT sponsored, would be a nice way to address this. In the meantime, the art and science of clinical decision making prevails – factoring in what we know about gene, gender, intrafamilial risk overlaid by the important role of aspirin chemoprevention, dietary and lifestyle factors [15] and hopefully, vaccination against the common neo-antigens springing up from those mismatch repair deficient crypts [14].

There is still the “here and now” which needs to engage carriers and optimize adherence to whatever scheduling is prudent and advised. How do we do this? There are clear messages in reports of surveillance. First, registries [16,17,18,19]. Registries keep track of carriers, can provide automated time alerts for the need to attend scheduling (including sometimes non-colorectal cancer scheduling), and provide the basis for observation on interval adenomas and cancers. Registries save lives. Secondly, patient navigators [19]. Time and time again it has been shown that navigators personally chaperone carriers through the complexities of the healthcare system, to actually reach colonoscopy. They are worth their cost-benefit weight in gold at least in the syndromic scenarios we are discussing. Thirdly, a pleasant experience helps maintain adherence – the least offensive and effective bowel preparation, a comfortable colonoscopic experience (sedation is important for adherence), and compassionate staff. Efficient throughput (less time for patients) on the day is also valuable. Bowel preparation regime should be personalized against previous experience. This is a particular contribution of the current paper [1]. Fourthly, the governance of the risk management service needs to be comprehensive (not just gastrointestinal), patient-centred, interactive with the patient navigators, general practitioners. associated familial cancer clinics (not losing opportunities for time and age appropriate cascade testing across the family), and hospital electronic medical records. Of equal importance, the strategy itself. The best strategic advice is of no use if it is not one that is adhered to. What the field has missed is implementation science and an understanding of the factors that drive contemplation to action [1, 20,21,22,23]. The current study amongst others is important if we are to benefit from the wealth of science that has been uncovered in the understanding of familial cancer. Minimizing the burden of colonoscopies is an important variable, warranting that debate herein described.