Introduction

Lynch syndrome (LS) is one of the most common genetic cancer syndromes and tends to occur in the form of various types of tumors at a young age, such as colorectal cancer (CRC) and endometrial cancer. Aldred S Warthin, M.D. first reported LS in 1913 [1], while Henry T Lynch, M.D. contributed to the concept of genetic cancer syndromes [2]. LS has a population prevalence of approximately 1 in 250–1,000 individuals [3,4,5,6] and accounts for 1–4% of all CRC cases [7,8,9]. The number of patients with LS in Japan is estimated to be more than 100,000. Recent progress in genomic medicine has led to new approaches that could be useful for the management of LS. This review describes the current clinical knowledge related to LS screening, differential diagnosis, surveillance, and therapy.

Clinical features

LS is an autosomal dominant disease caused by a pathogenic germline variant in a mismatch repair (MMR) gene. The causative genes are the four MMR genes MSH2, MLH1, MSH6, PMS2, and EPCAM [3, 4, 10]. EPCAM is upstream-adjacent to MSH2, and germline deletion of the 3′ end of EPCAM causes MSH2 silencing via methylation of the MSH2 promoter region [10]. The role of DNA MMR is to maintain genomic stability by correcting base mismatches and insertion–deletion mismatches that can arise during DNA replication. When the DNA MMR function is impaired, the sequence repeat number in simple repetitive sequences (microsatellites) is prone to changes. The altered number of repetitive sequences in microsatellites is termed microsatellite instability (MSI). Thus, LS-associated tumors, with variants in MMR genes, commonly show a high frequency of MSI (MSI-H) [3, 11].

Carriers of MMR gene variants are at high risk of early-onset LS-associated tumors, which include CRC, endometrial cancer (EC), gastric cancer, ovarian cancer, pancreas cancer, renal pelvic/ureteral cancer, biliary tract cancer, brain tumors, sebaceous tumors, keratoacanthoma, and small intestinal cancer [12]. In MSH2 and MLH1 gene variant carriers, the cumulative cancer risk up to age 70 years is 30–82% for CRC and 14–60% for EC [13, 14]. The risk of gastric cancer is 6–13% in Western countries [13] and 24% for Japan (up to age 60 years) [15]. The cumulative cancer risk for CRC in MSH6/PMS2 gene variant carriers is 10–22%, which is lower than that in MSH2/MLH1 variant carriers [13, 14]. Furthermore, the cumulative cancer risk for EC in MSH6 gene variant carriers is 16–71%, which is equivalent or higher than the risk for MSH2/MLH1 gene variant carriers (14–60%) [13, 14]. In addition, the cumulative risk of urinary tract cancer in MSH2/MLH1 gene variant carriers is high at 1–7% [13]. The tumor spectrum of LS has changed over time. The spectrum of LS in the family with LS first reported by Warthin included mainly gastric cancers and endometrial cancer [1]. However, follow-up reports of this family showed that in later generations, CRC was the most common tumor [2]. This change suggested that the cancer spectrum of LS is influenced by environmental factors, such as dietary and lifestyle habits, smoking, alcohol consumption, and obesity which interact with a host’s genetic factors, and may modulate the risk of developing cancer. Due to the high frequency of death due to gastric and biliary tract cancer among patients with LS in Japan, management of gastric cancer and biliary tract cancer is as important as that of CRC and EC [16] (Table 1).

Table 1 Amsterdam criteria I
Full size table

Diagnosis

Approach for diagnosis

Since CRC and EC are the most common LS-associated tumors and develop at earlier ages, their development often can be a clue for diagnoses of LS. Traditionally, as the first step for the diagnosis of LS, clinical criteria such as the Amsterdam criteria II (Table 2) [17] or revised Bethesda guidelines (Table 3) [12] had been used for selecting individuals for further testing. However, screening by using these criteria or guidelines could miss more than one-fourth of LS cases [7]. In 2009, therefore, the Evaluation of Genomic Applications in Practice and Prevention working group recommend screening all patients with CRC using either MSI testing or immunohistochemistry (IHC) [18]. IHC can be used, as a complement to MSI testing, to evaluate the expression of the MLH1, MSH2, MSH6, and PMS2 proteins in tumor tissues [3, 9]. In LS, 90% CRC show high-frequency MSI (MSI-H) or abnormality in IHC [3, 11].

Table 2 Amsterdam criteria II
Full size table
Table 3 The revised Bethesda guidelines for colorectal cancers for microsatellite instability testing
Full size table

Ten to fifteen percent of all CRCs show MSI-H in Western countries [3, 13, 14], whereas, in Japan, 6–7% show MSI-H [5, 19, 20]. Tumors that show MSI-H or abnormalities in IHC for MMR proteins are called deficient MMR (dMMR). Individuals with dMMR tumor are subjected to MMR gene testing. However, dMMR is not specific for LS. Presence of dMMR, in most cases, indicates epigenetic hypermethylation of the MLH1 promoter region. To rule out sporadic MSI-H CRC, BRAF V600E testing can be used [3, 13, 14]. BRAF V600E somatic variant is observed in approximately 40% of sporadic MSI-H CRC cases [21, 22] but rarely in LS. IHC for BRAF protein expression (clone VE1) can also be used to rule out sporadic MSI-H CRC [23]. It is important to remember that the BRAF testing cannot be used for EC [24].

Nowadays, dMMR CRC screening is thought to be useful not only as a diagnostic tool for LS, but also as a predictive, prognostic, and therapeutic marker. [11, 25]; currently, many guidelines recommend universal screening for patients with CRC and/or EC [13, 14].

Given that, even universal screening would miss a small but significant minority (5–10%) of individuals with LS, and the frequency of the de novo variant in LS is low at 2.3% [26], it should be noted that recording family history is not reliable, but still an important complement to tumor testing. Additionally, because of the increased cost and heavy workloads for universal screening, genetic testing of relatives (known as cascade testing) is important for implementing the universal screening effectively.

Founder variants

More than 50 types of founder variants associated with LS have been reported worldwide [27]. In Japan, MLH1 variant (exon5 c.381–431_c.453 + 717del1221) is considered a founder variant [28, 29]. Notably, a Japanese nationwide study conducted by the Japanese Society for Cancer of the Colon and Rectum reported that large deletion or duplication was common (26.6%) in Japanese patients with LS [30] including the MLH1 founder variant. Therefore, if no pathological variant is detected using Sanger sequencing, the MLPA or other methods should be used to rule out large deletion or duplication.

Multi-gene panel testing

Conventionally, genetic testing is carried out in a step-wise manner; beginning with the most suspected gene, and if no genetic variant is detected, it proceeds to the second suspected gene. However, there are many types of tumors associated with LS, and phenotypes of LS overlap with a substantial number of other conditions and differential diagnoses. Therefore, stepwise genetic testing is too expensive and time consuming. With recent advances in NGS technology, multi-gene panel testing has been introduced since 2012 [31].

The benefits of multi-gene panel testing include: (1) high detection rate of pathologic variants, (2) cost-effectiveness, (3) suitability for sequencing a wide area of tumor suppressor genes in which hotspots are rare, and (4) clarification of the genotype-phenotype relations. Interestingly, the phenotypes for LS and BRCA1/2-associated hereditary breast and ovarian cancer syndrome have been shown to overlap. A study investigated individuals with suspected LS by multigene panel testing and identified 114 individuals with LS and 71 with variants in other cancer predisposition genes including 15 with BRCA1 or BRCA2 [32].

In contrast, the disadvantages of multi-gene panel testing include: (1) high detection rate of variants of uncertain significance, (2) lack of clinical management guidelines for moderate and low-penetrance genes, and (3) few standard genetic counseling models for multi-gene panel testing [33].

Differential diagnosis

There are many tumors associated with LS, and phenotypes of LS commonly overlap with other conditions. Typical differential diagnoses which show nonpolyposis colon cancer are as follows.

Constitutional mismatch repair deficiency (CMMRD) syndrome

CMMRD, also called biallelic MMR deficiency, is caused by homozygous or biallelic germline variants in MMR genes. During childhood, most patients develop CMMRD related tumors. Brain tumors are the most common, followed by CRC and hematological tumors [34]. Most patients have café au lait spots on their skin resembling neurofibromatosis 1 [35]. Gastrointestinal polyposis is observed in half of the patients [36]. In LS, variants of MSH2 or MLH1 are more common than MSH6 or PMS2. In contrast, the most common MMR gene observed in CMMRD is PMS2 followed by MSH6, MLH1, and MSH2 [37]. Interestingly, a recent study reported that MSH3, which is one of MMR genes but has not been proven to cause LS, also causes CMMRD [38]. These findings suggested that biallelic variants of MSH2 or MLH1 might more likely to be embryonic lethal [39]. Currently, very intensive surveillance programs are proposed for CMMRD [39].

Sporadic MSI-H CRC

Epigenetic hypermethylation of the promoter region of the MLH1 gene is the main cause of sporadic MSI-H CRC. The clinical features of sporadic MSI-H CRC include elderly females, poorly differentiated adenocarcinoma, and right-sided colon preponderance. IHC shows absence of MLH1 protein expression. The somatic variant of BRAF V600E is observed in approximately 40% of sporadic MSI-H CRC [11], but rarely in LS.

Lynch-like syndrome

Patients with LS-associated tumors that show dMMR without hypermethylation of the MLH1 promoter in the absence of a germline variant of MMR gene or EPICAM are termed to have “Lynch-like syndrome”. Lynch-like syndrome is a heterogeneous condition and mainly caused by epigenetic bi-allelic variant of MMR genes [35, 40]. Lynch-like syndrome accounts for 18–71% of patients with dMMR LS-associated tumors without MLH1 hypermethylation. CRC or EC in Lynch-like syndrome develops at earlier ages similar to LS. [5, 40, 41].

Familial CRC type X (FCCTX)

Families that meets Amsterdam criteria I (Table 1) [42], but who lack a germline variant in an MMR gene and an MSI-H tumor, are termed to have “familial CRC type X” [3, 43]. FCCTX is a heterogeneous condition. Candidates of causative gene include BRCA2, SEMA4, NTS, RASSF9, GALNT12, KRAS, BRAF, APC, BMPR1A, and RPS20, among others [44]. FCCTX accounts for approximately 40% of cases meeting Amsterdam criteria I. The onset age of CRC in FCCTX is 7.5 years more than that in LS, with the left-sided colon accounting for 70% of CRC cases. Histologically, this type is similar to sporadic CRC. The adenoma/carcinoma ratio is high and, the rate of development of cancer from adenoma is slow compared to LS [45, 46]. The risk of developing extracolonic LS-associated cancer is not increased in patients with FCCTX [15]. For surveillance of FCCTX, colonoscopy is conducted at intervals of 3–5 years, starting 5–10 years earlier than the earliest onset age of CRC in the family [45].

Polymerase proofreading-associated polyposis (PPAP)

PPAP is an autosomal dominant disease caused by germline variants in the proofreading (exonuclease) domains of DNA polymerases (DNA polymerase ɛ, POLE and DNA polymerase δ, POLD1) [47, 48]. Two hot spots of the gene variant have been reported (POLE p.Leu424Val and POLD1 p.Ser478Asn). Patients with germline POLE variants present oligo-adenomatous colorectal polyposis and CRC at an early age. Patients with germline POLD1 variants present EC and brain tumor as well as oligo-adenomatous colorectal polyposis and CRC at an early age. Although PPAP associated tumors show stable microsatellites, these tumors are hypermutated or ultramutated due to loss of polymerase proofreading function [49]. Therefore, PPAP associated tumors might benefit from immune-checkpoint inhibitor therapy [50].

Surveillance

The efficacy of surveillance in LS has only been demonstrated for CRC; a 3-year endoscopic surveillance was reported to reduce mortality due to CRC by 65% [51]. However, because cancers can occur over the testing interval of 3 years, colon surveillance with endoscopy should be performed from 20 to 25 years of age at intervals of 1–2 years [3, 13, 14]. Because the progression from colon adenoma to cancer in patients with LS is faster than that in the general population, adenomas regardless of their size should be removed in patients with LS [11]. Gastric surveillance with endoscopy should be performed from 30 to 35 years of age at intervals of 1–2 years. The presence of Helicobacter pylori infection should be evaluated from approximately 25 years of age, and if detected, treatment should be performed [11, 52]. Small intestinal cancer is another LS-associated cancer, and when performing gastroscopy or colonoscopy, the duodenum and ileum should be observed whenever possible. Uterine/ovary surveillance with transvaginal ultrasound, endometrial biopsy/cytology, and serum CA125 should be conducted from an age of 30–35 years at an interval of 6 months to one year. Urinary tract surveillance with urine tests/cytology should be conducted from an age of 30–35 years at intervals of 1–2 years [3, 13, 14, 52].

Risk reduction surgery

Risk reduction surgery for CRC

An extended colectomy for CRC could reduce the risk of metachronous CRC. It also makes colonoscopic surveillance of the residual colon easier [53]. Therefore, extended surgery such as total colectomy or total proctocolectomy might be a good option. However, in clinical studies, extended surgery did not show better prognosis of CRC compared to partial colectomy, and might reduce bowel function. The US Multi-Society Task Force on colorectal cancer strongly recommended extended colectomy for patients with colon cancer in LS based on moderate evidence [14], however, a European group (Mallorca group) could only provide a weak recommendation, which was based on their expert opinion [52]. Thus, at present, no consensus has been reached on whether extended surgery is better management for patients with LS.

Risk reduction surgery for EC, and ovarian cancer

It has been reported that prophylactic hysterectomy and bilateral salpingo-oophorectomy can greatly reduce the risk of EC and ovarian cancer, respectively [54]. Especially, at the time of CRC surgery, these prophylactic surgeries have been considered to be a good option. Before making a decision, both patients and surgeons should take into account the stage of CRC, age of patients, their desire to have children, menopausal status, and gene variant type, among other factors [3, 13, 14]. Notably, a recent prospective study described a good prognosis for EC (98% 10-year crude survival) or ovarian cancer (89% 10-year crude survival) in patients with LS without prophylactic surgery [55].

Chemoprevention/drug therapy

Chemoprevention of CRC

The CAPP2 trial was a randomized controlled trial of LS variant carriers that evaluated the efficacy of aspirin (600 mg/day) on CRC, extra-colonic LS-associated tumors, and colon adenoma. Long-term aspirin use (at least 2 years) significantly lowered the risk of developing CRC and extra-colonic cancers [56]. Currently, CAPP3 trial is ongoing to study the optimal dose and administration period of aspirin.

Postoperative adjuvant chemotherapy for CRC

Postoperative adjuvant chemotherapy with 5-FU did not improve prognosis in Stage II colon cancer with MSI-H [57]. In contrast, postoperative adjuvant chemotherapy with oxaliplatin improved prognosis in Stage III colon cancer with MSI-H. However, only a few studies have been conducted with a focus on patients with LS.

Chemotherapy for advanced/metastatic CRC

MSI-H CRC is reported to be more common among stage II (∼20%) than III (∼12%), and less frequent in stage IV CRC (∼4%) [58]. Efficacy of chemotherapy against metastatic dMMR CRC in patients with or without LS has not yet been clarified. Therefore, regimens for sporadic colorectal cancers are generally used for CRC in patients with LS.

Immune therapy for advanced or metastatic CRC (immune-checkpoint inhibitors) (Table 4)

Table 4 Comparison of efficacy of immune checkpoint inhibitors against dMMR CRC with or without Lynch syndrome
Full size table

The efficacy of immune-checkpoint inhibitors has been demonstrated against dMMR solid tumors. The US Food and Drug Administration (FDA) approved the use of checkpoint inhibitors in the treatment of dMMR solid tumors.
PD-1 inhibitor monotherapies demonstrated no significant difference in the objective response rate for patients with or without LS: Pembrolizumab; 46 vs. 59%, and Nivolumab; 33% versus 29%, respectively [59, 60]. Combination immune therapy (Nivolumab plus Ipilimumab) demonstrated high response rates (55%) and disease control rate (80%) for ≥ 12 weeks [61].