Background
In 1985 it was suggested that inherited colon cancer should be termed Lynch Syndrome I, and inherited colon with extracolonic cancers Lynch Syndrome II [
1]. (OMIM
# 120435). In 1989 an international network of researchers (ICG-HNPCC) set out to identify the genetic variants causing what they termed the Hereditary Non-Polyposis Colon Cancer (HNPCC) syndromes [
2]. It was discovered that a major fraction of HNPCC tumours were characterised by micro-satellite instability (MSI) and caused by inherited pathogenic variants (path_) affecting the mismatch repair (
MMR) genes. In 2009 the term Lynch Syndrome (LS) was redefined to denote this hereditary condition [
3]. That paper, however, erroneously stated that LS was identical to HNPCC, while in fact variants in several non-
MMR genes cause HNPCC without MSI tumours. In 2009 another group stated that Lynch syndrome includes both individuals with an existing cancer and those who have not yet developed cancer [
4]. These different definitions have created conceptual confusion, especially the latter because Mendelian inheritance by definition is describing inherited traits (phenotypes). How to explain the original nomenclature to integrate the concept of probability by age to demonstrate an inherited trait is challenging and may be why the discussions on inherited cancers have separated from the networks for inherited disorders diagnosable at birth or in infancy. Nomeclature for LS should comply with consented medical concepts delineating diseases from normal variation, and nomenclature should be applied as for the other inherited cancer and inherited disease syndromes. Using the same annotation for healthy carriers as for cancer cases is confusing and may be misunderstood and in conflict with both the scientific, ethical and legal platforms of medical genetics. Without defined and consented concepts and nomenclature communication to reach consensus is difficult.
ICG-HNPCC established the Amsterdam I clinical criteria to identify families with highly penetrant and dominantly inherited colon cancer.
Path_MLH1 and
path_MSH2 variants were identified as causative in some such families. Based on the logical circle that returned the selection criteria as results, it was concluded that LS was a dominantly inherited colorectal cancer (CRC) syndrome with high penetrance. It became clear that endometrial cancer was part of LS [
5] and the revised Amsterdam II clinical criteria were agreed, including endometrial cancer as an affected phenotype [
6] and consistent with
path_MSH6 being a cause of LS. It soon became evident, however, that the Amsterdam criteria were insensitive in identifying LS families caused by
path_MLH1 or
path_MSH2 variants, and even less sensitive in identifying LS caused by
path_MSH6 or
path_PMS2 variants [
7]. Despite these shortcomings, these clinical criteria are still in use as a clinical pre-test to select cases for genetic testing. The result has been that most LS families identified historically have fulfilled these criteria and have dominantly inherited CRC/endometrial cancer with high penetrance, while relatively few
path_MSH6 and very few
path_PMS2 families have been identified. It also became clear that while in former generations most patients died from their first cancers, a substantial number now survive their first cancer and live on to develop further cancers that are often in other organs. In summary, knowledge of LS a decade ago was by and large derived from retrospective family studies based on questionable concepts as were the clinical guidelines on how to manage both healthy
path_MMR carriers and affected LS patients [
8]. Because it was recognized that colonoscopy conducted every 3 years did not fully prevent CRC, guidelines were revised advocating a reduction of the interval between colonoscopies to 1–2 years, with no evidence that this would reduce CRC incidence.
Researchers from several collaborating European centres agreed to establish the PLSD during a meeting in Palma, Mallorca on May 4th 2012. The aims were to challenge and test assumptions based upon retrospective information, to determine empirical prospectively observed cancer incidences and survival in path_MMR carriers and to observe the effects of interventions and categorize these by age, gene and gender.
Methods
To validate the assumptions upon which clinical guidelines were based, the data entered into PLSD had to be assumption-free. The data recorded included gender, age of inclusion, age last observation, age at death, diagnosis of any cancer, age at diagnosis of cancer and the inherited
path_MMR variant that had been identified. The data had to be complete for these variables, and all carriers known at each reporting centre had to be contributed. Later, cancer stage at diagnosis and time since last colonoscopy at cancer diagnosis were requested for all prospectively detected CRCs and added to the information already filed. Reported pathogenic variants were assumed germline. The data were included in an Oracle relational database. Details relevant to an understanding of its capabilities and interpretation of outputs are discussed in our previous reports [
9,
10].
To control lead-time bias, all cancers diagnosed at the same age as inclusion were considered prevalent (first round cancers), and all cancers diagnosed later were scored as prospective. Some carriers had been followed for a long time, and there are time-trend biases in the technical development of the screening techniques that were applied, in understanding of what to look for during screening and in changing intervals between colonoscopies. There are length-time biases when no obligatory examinations were undertaken at right-censoring observation time. Length-time bias will most probably result in an artificially low incidence of CRC. The longer the observation time, the more impact time-trend biases will have, and the less impact lead- and length-time biases will have. Generally, in screening trials, there should be a randomized control group, but this approach is considered impossible for ethical reasons in LS carriers. Time-trend and length-time biases were accepted in order to maximize the number of observation years. Updated information on the carriers filed in the PLSD may be added to re-analyse the series, correcting for time-trends and length-time bias.
Survival was measured as overall/crude survival, because disease-specific survival includes assumptions.
Any study has a selection procedure to identify the cohort to be studied –
a selection bias. Results from any study should be interpreted based on the selection procedures, to avoid returning the selection criteria as the results of the studies. A selection artefact included in the PLSD dataset is that genetic testing was usually done in cancer families: there may be additional genetic and/or environmental factors causing disease in such families [
11] resulting in artificially high prospective average cancer incidences in carriers. A selection bias is the low number of low-penetrant variants. This bias may also be considered a result demonstrating the low penetrance of these variants.
Based on power calculations, the first PLSD dataset was censored when 25,000 observation years had been filed, and the first three descriptive papers were published: 1) incidence rates for cancers in carriers without prior or prevalent cancers [
12], 2) incidence rates for cancers in carriers who had prior and/or prevalent cancers [
13], and 3) - because papers 1 and 2 gave similar results – a combination of the first two papers into one study including all carriers with or without cancer prior to or at inclusion [
14]. With these three papers the original goal was reached. When an additional independent series of about 25,000 observation years were filed, we compared this independent replication cohort with the first series, reaching the conclusion that the results were similar. We then combined all cases in one large data set, refining our estimates of cancer risk and survival by age, gene and gender [
15]. At that time more contributors expressed their interest in participating, and the PLSD database is still growing.
In addition to the four descriptive reports described above, three hypothesis-testing papers have been published: CRC incidence related to the interval between colonoscopies [
16], clinic-pathological stage of colon cancer related to time since last colonoscopy [
17] and survival after colon cancer related to time since last colonoscopy [
18].
What is Lynch syndrome?
The definition of LS has changed repeatedly. Currently used definitions are contradictory, in conflict with ethical and scientific paradigms, and some results provided by PLSD are in conflict with all of them. The definition ‘
Lynch syndrome is a highly penetrant hereditary cancer syndrome caused by pathogenic germline variants in DNA mismatch repair (MMR) genes’ [
24] excludes male
path_MSH6 carriers and all
path_PMS2 carriers. Because of their biological similarities and responses to treatment, one may suggest to consider all MSI CRC cases as LS, if so most cases would not be inherited. If considering families with clinically dominantly inherited MSI tumours as LS, not all families have demonstrable pathogenic
MMR variants [
32]. Variants in additional DNA repair genes cause urothelial cancer [
31]. Also, it is increasingly evident that different classes of variants in the
MMR genes are associated with different penetrance – the emerging evidence for variants associated with differential splicing being one example [
33] and which may be more frequent than is currently recognized [
34]. Gene panel testing in both blood and tumours will identify many variants in these genes in incident cancer cases and there is a need to conceptualize and categorize interpretation of the results. The umbrella term ‘Lynch syndrome’ has been practically and scientifically useful but may longer be so. It appears timely to reconsider data from all sources in relation to LS and to be more precise in how we define it. For example, it may be clinically practical to group cancer cases who will benefit from similar treatment modalities. Better defined and individualized prospective probabilities of cancer may be needed for genetic counselling and planning of preventive interventions. Understanding associations between genetic variants and carcinogenetic and biological mechanisms may be objectives for further research. These topics are overlapping but not identical and will have different outputs relevant to decision-making in these different contexts.
Acknowledgments
The Prospective Lynch Syndrome Database would not have been possible without the contributors and without the initial support from the core members of The European Hereditary Tumour Group (former Mallorca group) contributing all their follow-up data for the first PLSD version. Mev Dominguez-Valentin (curator of the PLSD database), Julian Sampson and Toni Säppäla first-authored the last four PLSD papers. Gabriela Möslein and Gabriel Capella last-authored the first three papers. The close collaboration with John-Paul Plazzer and Finlay Macrae in the InSiGHT database has been essential, Finlay Macrae also for his continous support from his gastroenterological perspective. The enthusiastic support from (alphabetially mentioned) Aysel Ahadova, John Burn, Gareth Evans, Elke Holinski-Feder, Eivind Hovig, Mette Kalager, Matthias Kloor, Noralene Lindor, Jukka-Pekka Mecklin, Rodney Scott and Lone Sunde together providing a broad cross-professional platform for designing the studies and interpreting the results has been essential. The website
www.PLSD.eu is managed by Sigve Nakken. A special thank you to Julian Sampson for commenting the finial manuscript in details.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.