Numerous studies demonstrated variations in CRC incidence among different LS-associated genes, yet limited knowledge exists regarding the customisation of surveillance and treatment strategies based on these variations. Therefore, our observational cohort study aimed to sort variations in occurrence and management of CRC and its precursor lesions during surveillance across different LS gPV groups. In line with current knowledge [16], we found that individuals with a gPV in MLH1 and MSH2 have higher life-time risks to develop CRC compared to those with MSH6. Specifically, the CRC risk is higher in men than women. Furthermore, we confirmed that CRC in individuals with MLH1 and MSH2 gPVs often develops without visible precursor lesions, whereas in individuals with MSH6 and PMS2 gPVs, significantly more visible precursor lesions are detected. Additionally, we observed considerable variation in performed treatments for T1 CRC.

This study possesses several strengths. To our knowledge, it represents one of the largest single-center LS surveillance cohorts with complete surveillance data. It reports CRC detection rates during surveillance across four different LS-associated gPVs over a span of four decades. Additionally, beyond CRC incidence, the study also analyses the incidence of precursor lesions, metachronous CRCs, TNM stages for all CRCs and treatment strategies. In contrast to prior research, this study exclusively included individuals with confirmed LS-associated gPVs. All participants were under surveillance at a specialised LS centre staffed with experienced endoscopists. Lastly, to guarantee optimal quality, all data were double-checked, minimizing the risk of information bias.

Regarding the big data databases, such as InSIGHT and PLSD; these are valuable for identifying patterns and are therefore possibly useful as predictive models. These databases include a tenfold larger number of LS carriers compared to our single-center cohort. However, there are several limitations to the PLSD database. First, the distribution of the four MMR genes is skewed with MSH6 and PMS2 gPVs in the minority. Because of this data for individuals with MSH6 and PMS2 gPVs should be interpreted with caution. Additionally, the data from these large databases come from multiple hospitals, each including a significantly lower number of LS carriers compared to our cohort. Furthermore, InSIGHT and PLSD data are reported by consulting specialists without double-checking, which may introduce selection- and reporting bias. Lastly, these databases do not include data on precursor lesions and treatment variations.

This study also exhibits several limitations. The retrospective design unavoidably resulted in missing data. However, despite the amount of missing data, significant differences or trends were still observed. Second, the timing of precursor and carcinoma detection relied on the timing of colonoscopy procedures, which may have introduced information bias. In addition, improvements in imaging quality over time may have influenced lesion detection, potentially resulting in a higher number of lesions being detected in patients under surveillance nowadays compared to those monitored in the past. Though, in our cohort, 69% of individuals with LS have been under surveillance during the last 15 years, indicating that the majority of individuals with LS have been monitored with the most modern endoscopic equipment. Both the timing of colonoscopy procedures and improvements in imaging quality may have influenced the overall detection rate, but these limitations are not expected to have directly impacted the proportion of detected lesions among mutations groups, as all patients underwent similar surveillance protocols. Lastly, a limitation of our single-centre design is that the generalizability of these findings to the broader global LS population may be questioned.

Performing biennial surveillance colonoscopies in individuals with LS has been shown to be essential, as it substantially improved the prognosis of individuals with LS caused by a gPV in one of the MMR genes, due to earlier detection of CRC and premalignant adenomas [8]. However, the effectiveness of the surveillance program heavily relies on adherence, and several studies already demonstrated adherence to timely colonoscopy surveillance in individuals with LS is not optimal [9, 10]. In a study of Van Liere et al. [17], it was shown that 57% of individuals with LS perceived the surveillance program as extremely burdensome, with 10% even considering quitting surveillance. Many individuals (60%) preferred a less invasive surveillance method. The current biennial surveillance program, therefore, appears to negatively influence quality of life, particularly if biennial checks may not even be necessary for certain LS associated gPV groups.

Earlier studies have already demonstrated variability in the lifetime risks of developing CRC among different individuals with LS, with carriers of a gPV in MLH1 and MSH2 being at the highest hereditary CRC risk [16]. Consistently, in this study, we observed that individuals with MLH1 gPV exhibited significantly higher CRC incidences at initial colonoscopy compared to other LS-associated gPV groups (Table 1). In addition, with regard to detected lesions during surveillance colonoscopies, higher CRC incidences were observed in the MLH1 and MSH2 compared to the MSH6 and PMS2 gPV groups. At the same time, the observed non-advanced adenoma incidence was lower in MLH1 and MSH2 gPV groups. This findings aligns with recent insights into CRC pathways in LS, which suggest three different pathways [15]. It is proposed that CRC development in individuals with MLH1 and MSH2 gPVs follows a more accelerated pathway with MMR deficient precursor foci in the colonic crypts [18]. In individuals with a MLH1 gPV, these foci may be located beneath the mucosal layer making them easily overlooked [18]. Thus, since the tumour biology of a MLH1 gPV differs from that of an MSH2 gPV, they cannot be considered as an equal entity.

The observed cumulative lifetime CRC incidence was higher in the MLH1 compared to MSH6 and PMS2 variant groups, as well as in MSH2 compared to the MSH6 variant group (Fig. 2). These findings suggest that, given the substantially lower lifetime risk of developing CRC and considering the negative impact of biennial surveillance on quality of life, surveillance intervals for individuals with MSH6 and PMS2 gPVs could potentially be safely extended.

We additionally analysed differences in TNM staging across the different gPV groups, as well as the corresponding treatments for early-stage (T1) CRC. As expected from the implementation of biennial surveillance, most CRCs detected during surveillance colonoscopies were T1 CRCs (52%) without the presence of lymph node- (85%) and distant (97%) metastasis (Table 3). When examining the relatively small number of patients with treated early-stage colorectal T1 CRCs (n = 25), we observed considerable variation in the performed surgical or endoscopic treatment (Table 4). In 68% of the cases, physicians deviated from the applicable guideline, which recommended subtotal colectomy for early stage (T1) CRC [12]. Less invasive local endoscopic treatments were performed more frequently than the resection advised by the current national guideline (36% vs. 32%), clearly demonstrating preference and confidence in more local methods of treatment. Given the significant advancement in endoscopic techniques [19, 20] and improved functional outcome after more local treatments [21], it is not surprising there has been a growing preference for minimally invasive interventions in recent years. The current guideline for non-LS related CRC even already includes more tailored treatment recommendations for T1 tumours (https://richtlijnendatabase.nl/richtlijn/colorectaal_carcinoom_crc/startpagina_-_colorectaal_carcinoom.html#:~:text=De%20richtlijn%20colorectaal%20carcinoom%20richt,bij%20(verdenking%20op)%20darmkanker), with endoscopic techniques considered superior to surgical resection in most cases. Specific histological risk factors for non-LS related colorectal T1 CRCs, including poor differentiation, lymph angioinvasion and high-grade tumour budding, have been identified as predictors of metastasis and recurrence [22]. After endoscopic removal of non-LS related colorectal T1 CRCs, the presence or absence of these histological risk factors can subsequently assist in determining the need for additional surgical resection and/or in formulating the appropriate surveillance strategy [22]. Interestingly, only one of the LS patients treated for T1-stage CRC experienced a second CRC twelve years after a CRC was detected during the index colonoscopy, suggesting less invasive organ preserving treatment for LS-related colorectal T1 CRCs might be a viable option. Nonetheless, evaluation of recurrence rates in the long term is necessary given that local treatments were conducted more recently, leaving the possibility for recurrences to emerge.

In our opinion, future studies should focus on distinguishing more personalised surveillance intervals for different LS associated gPV groups, possibly enabling extension of surveillance intervals for individuals with MSH6 and/or PMS2 gPVs given the smaller lifetime risks of developing CRC. Regarding treatment strategies for colorectal T1 CRCs, it is crucial to validate these findings in larger patient cohorts, with particular emphasis on monitoring the long-term outcomes of endoscopic treatments. In addition, it is interesting to investigate whether the current histological risk factors for non-LS related colorectal T1 CRCs also apply to LS-related colorectal T1 CRCs, and if so, whether these risk factors differ across the various LS associated gPV groups.

In conclusion, this large single centre analysis showed a higher precursor lesion detection rate as well as a lower CRC incidence for individuals with LS caused by a gPV in MSH6 or PMS2, suggesting the potential for extended surveillance intervals for these patients. Furthermore, consistent with biennial surveillance protocols, the majority of CRCs detected during surveillance were T1 CRCs without metastases. Regarding treatment of these T1 CRCs, considerable variation was observed with nearly equivalent recurrence rates, highlighting the feasibility of less invasive endoscopic interventions for individuals with LS. To further refine surveillance strategies, additional research into extending surveillance intervals for individuals with LS caused by a gPV in MSH6 and PMS2 is needed. Moreover, assessing the long-term outcomes of diverse treatment modalities and identifying histological risk factors specific to LS-related colorectal T1 CRCs are essential steps towards optimising future treatment for individuals with LS.