SEOM clinical guideline on hereditary colorectal cancer (2019)

Lynch syndrome (LS) is an autosomal dominant inherited cancer predisposition disease. Its estimated general population prevalence is 1 in 279 [4]. LS accounts for about 3% of colorectal cancer (CRC) [5] and 2% of endometrial cancer (EC) cases [6]. CRC is the most common associated tumor, usually right-sided, poorly differentiated, with mucinous features or medullary growth pattern, abundant infiltrating lymphocytes, propensity for synchronous and/or metachronic tumors [7], and a better prognosis in the non-metastatic setting [8, 9]. LS also predisposes to EC, small intestinal, urinary tract, pancreaticobiliary, gastric and ovarian tumors, and slightly to breast and prostate cancers [7, 10‐12]. Muir–Torre syndrome is characterized by skin tumors (sebaceous neoplasms, keratoacanthomas) and Turcot’s syndrome includes glial brain tumors [7, 10]. The age of onset is younger compared with sporadic counterparts: 45–60 years old (years) for CRC and 50 years for EC [13].

Lynch syndrome is caused by pathogenic germline variants in the DNA mismatch repair (MMR) genes MLH1, MSH2, MSH6, or PMS2 (and in the non-MMR gene EpCAM, in which deletions induce epigenetic silencing of MSH2) [10, 14]. When a second “hit” of somatic mutation occurs, the MMR function fails leading to cancers with microsatellite instability (MSI) and hypermutation phenotype. Deficient MMR (dMMR) and MSI are not exclusive to the LS and are also likely in sporadic cancers caused by MLH1 promoter hypermethylation [15] or double somatic MMR mutations [16].

The age-specific cumulative risk could depend on genotype and sex [12, 13, 17‐19] (Table 2). MSH2 mutation carriers have a higher risk of extracolonic cancers [12]. EpCAM deletions close to the MSH2 promoter are associated with increased risk of EC [20]. MSH6 and PMS2 mutations have a lower penetrance [4], except for EC in MSH6 carriers, who also present with cancer at later ages [19, 21].

Universal strategy with molecular analyses in unselected CRC or EC adds diagnostic sensitivity for LS over clinical criteria, with a favorable cost-effectiveness profile [22‐25]. This information has prognostic and therapeutic value for common clinical practice. Multigene NGS in CRC, including MMR and BRAF genes, with computational tools for MSI testing, has a higher diagnostic sensitivity compared with a universal strategy based on immunohistochemistry (IHC) and BRAF analyses (100% sensitivity [95% CI 93.8–100] versus 89.7%, [95% CI 78.8–96.1], p = 0.04) [26]. In a prospective study of > 15.000 unselected cancer patients with 50 different histologies, a similar NGS device increased LS diagnostic sensitivity over revised Bethesda Criteria (rBC) plus universal strategy [27]. Somatic NGS panels are available in clinical practice for precision oncology due to their predictive value. Direct germline multigene NGS in unselected patients with CRC increases diagnostic sensitivity for less prevalent hereditary syndromes more than for LS [5, 28] (Fig. 1). When there are no tumor samples, fulfillment of rBC or a ≥ 2.5% likelihood of LS on the validated PREMM₅ prediction model [29] can be used for referral to a GCU. Although NGS is continuously less expensive, cost-effectiveness studies for LS diagnostic strategies that incorporate these platforms are lacking.

Recommendation Different screening strategies for LS of all newly diagnosed CRC and EC can be considered including tumor tests for defective MMR function and/or high-level MSI and/or NGS tumor sequencing including BRAF.

In case of lack of expression of MLH1 and PMS2 by immunohistochemistry, BRAFV600E mutation and/or MLH1 promoter hypermethylation should be carried out to rule out sporadic cases.

Patients with molecular profiles compatible with LS should be referred to GCU for appropriate counseling and NGS germline genetic testing.

In families with fulfillment of rBC or a ≥ 2.5% likelihood of LS on the PREMM₅ prediction model, prevalent and/or previous CRC and/or EC should follow the same screening procedure before considering referral to GCU (evidence level B, strength 1).

Adjuvant chemotherapy with 5-fluorouracil did not result in a survival benefit in subgroup analyses of patients with stage II colon cancer with dMMR [35].

In patients with dMMR metastatic CRC and previous cytotoxic agents failure, anti-programmed death 1 (PD-1) antibody nivolumab resulted in objective response rate (ORR) of 31.1% (95% CI 20.8–42.9) (median follow-up time, 12.0 months) [36]. In the same setting, the combination of nivolumab plus ipilimumab (anti-cytotoxic T-lymphocyte antigen 4 antibody) reached an ORR of 55% (95% CI 45.2–63.8) (median follow-up time, 13.4 months), with progression-free survival and overall survival rates at 12 months of 71% and 85%, respectively, and a manageable treatment-related grade 3–4 toxicity in 32% of patients [37]. In dMMR cancers across 12 different histologies, ORR of 53% (95% CI 42–64) and complete responses in 21% of patients (median follow-up time, 12.5 months) were observed with anti-PD-1 antibody pembrolizumab [38].

It is an autosomal dominant inherited disease caused by germline mutations (> 85% point mutations, 10–15% large rearrangements) in APC which encodes a protein with a significant role in the Wnt-β-catenin signaling. Up to 30% of carriers are due to de novo mutations or to somatic mosaicism. There are genotype–phenotype correlations. The clinical presentations are: (a) classic FAP, ≥ 100 polyps, appearing between 10 and 30 years, first located in rectum and sigma afterwards along the colon and development of CRC in almost 100% if untreated, at a mean age of 39 years. (b) Attenuated FAP (AFAP), < 100 polyps located in proximal colon and lesser risk of CRC, 70%, at a mean age of 50–55 years. (c) Gastric adenocarcinoma and proximal polyposis of the stomach (GAPP), none or few polyps in colon.

It is an autosomal recessive disease caused by bi-allelic, homozygous or compound heterozygous, germline mutations (> 99% point mutations, < 1% large rearrangement) in MUTYH which encodes a glycosylase of DNA base excision repair system. Somatic G:C to T:A transversions result in genes implicated in CRC carcinogenesis, such as APC or KRAS. The most frequent phenotype is AFAP, at a mean age of 45 years, but maybe classic FAP, serrated polyposis and few individuals develop CRC without polyposis. Genotype–phenotype correlation has been described. The risk of CRC is 43–100% at the age of 50 years.

For the study of FAP, single gene testing has been the traditional approach, but the use of a multigene panel should be specially considered in attenuated FAP.

Recommendation Criteria for referral to a GCU and APC/MUTYH or multigene panel testing (evidence level B, strength 1):

In classical FAP, flexible sigmoidoscopy or colonoscopy should be carried out every 1–2 years, starting at age 12–15 years. If adenomas are found at sigmoidoscopy, then it should be annual colonoscopy [43]. In AFAP, colonoscopy should be performed every 1–2 years starting at age 18–20 years and surgery is indicated if there is a high number of adenomas. In MAP, colonoscopy should be performed every 1–2 years, starting at 18–20 years and if polyps cannot be controlled endoscopically, colectomy should be considered [43, 44]. For MUTYH heterozygote, colonoscopy should be performed every 5 years, beginning at age 40 years or 10 years prior to age of first-degree relative’s age at CRC diagnosis [42, 44].

Surgical options in FAP patients are total abdominal colectomy with ileorectal anastomosis (TAC/IRA) or total proctocolectomy with pouch anal anastomosis (TPC/IPAA) [42‐44]. Surveillance of rectum depends on the type of surgery [42, 43].

For gastroduodenal adenomas, upper endoscopy (including complete visualization of the ampulla of Vater) should be performed starting at 25–30 years or at the time of diagnosis of colonic polyposis [43]. Surveillance intervals are based on the Spigelman’s stage of duodenal polyposis: 0 (no polyposis): 4 years; I (1–4 tubular adenomas, 1–4 mm): 2–3 years; II (5–19 tubular adenomas, 5–9 mm): 1–3 years; III (≥ 20, ≥ 1 cm): 6–12 months; and IV (dense polyposis or high grade): surgery [42, 44]. Duodenal adenomas are managed by endoscopic polypectomy, although duodenectomy or duodenal pancreatectomy may be necessary in advanced cases [43].

Patients with classical FAP have a lifetime thyroid cancer risk of 2–6% and annual surveillance is recommended [42, 44].

Treatments for desmoid tumors are surgery, non-steroid anti-inflammatory drugs (NSAIDs), anti-oestrogen agents, chemotherapy, imatinib, sorafenib, pazopanib and radiotherapy [44].

The absolute risk for hepatoblastoma in FAP is estimated at less than 2% and it occurs prior to the age of 3 years [42].

The use of NSAIDs (sulindac or celecoxib) has been shown to reduce the number and extent of CRC and duodenal adenomas, but without the clinical benefit of decrease in cancer risk [42, 43]. Due to the cardiovascular risk of NSAIDs, no drug has been approved.