Comparison of immediate germline sequencing and multi-step screening for Lynch syndrome detection in high-risk endometrial and colorectal cancer patients

Abstract

Objective

Lynch syndrome (LS) is a hereditary cancer predisposition syndrome with a significantly increased risk of colorectal and endometrial cancers. Current standard practice involves universal screening for LS in patients with newly diagnosed colorectal or endometrial cancer using a multi-step screening protocol (MSP). However, MSP may not always accurately identify LS cases. To address this limitation, we compared the diagnostic performance of immediate germline sequencing (IGS) with MSP in a high-risk group.

Methods

A total of 31 Taiwanese women with synchronous or metachronous endometrial and colorectal malignancies underwent MSP which included immunohistochemical staining of DNA mismatch repair (MMR) proteins, MLH1 promoter hypermethylation analysis, and germline sequencing to identify pathogenic variants. All patients who were excluded during MSP received germline sequencing for MMR genes to simulate IGS for the detection of LS.

Results

Our findings indicate that IGS surpassed MSP in terms of diagnostic yield (29.0% vs. 19.4%, respectively) and sensitivity (90% vs. 60%, respectively). Specifically, IGS successfully identified nine LS cases, which is 50% more than the number detected through MSP. Additionally, germline methylation analysis revealed one more LS case with constitutional MLH1 promoter hypermethylation, bringing the total LS cases to ten (32.3%). Intriguingly, we observed no significant differences in clinical characteristics or overall survival between patients with and without LS in our cohort.

Conclusion

Our study suggests that IGS may potentially offer a more effective approach compared to MSP in identifying LS among high-risk patients. This advantage is evident when patients have been pre-selected utilizing specific clinical criteria.

Synopsis

Lynch syndrome is prevalent among patients with synchronous/metachronous colorectal and endometrial cancers. In this high-risk patient group, immediate germline sequencing is more effective for detecting Lynch syndrome compared to multi-step screening approaches. No significant differences were observed in clinical parameters between patients diagnosed with Lynch syndrome and those without in this patient group.

INTRODUCTION

Lynch syndrome (LS) is an inherited cancer predisposition syndrome resulting from germline mutations in one of four mismatch repair (MMR) genes (MLH1, MSH2, MSH6, and PMS2) or due to constitutional epigenetic silencing of MLH1 or MSH2 [1,2,3]. Individuals with LS face a 50%–70% lifetime risk of colorectal cancer (CRC), a 40%–60% risk of endometrial cancer (EC), and increased risk of developing various other solid tumors [1]. The global prevalence of LS in CRC and EC is estimated to be near the 3% [4,5]. Malignancies associated with LS are typically characterized by the loss of MMR protein expression, as observed through immunohistochemistry (IHC). Furthermore, the disruption of the MMR system – a highly conserved DNA repair pathway – leads to a pronounced mutator phenotype accompanied by microsatellite instability (MSI) [6].

Numerous clinical criteria have been suggested for identifying high-risk individuals eligible for LS screening, with the Revised Bethesda Guidelines standing out as the most sensitive option [7,8]. In recent years, there has been a shift from selective screening of high-risk individuals based on these guidelines to a more comprehensive, universal LS screening approach that encompasses all newly diagnosed CRC patients [4,9]. This universal approach has since been expanded to incorporate patients diagnosed with EC as well [10,11,12]. Following the identification of germline mutation carriers, it is crucial to initiate genetic counseling and testing for blood relatives to ensure comprehensive care [13,14,15].

To pinpoint patients with LS, either within high-risk groups or universally, a standard screening strategy consists of three stages: 1) detecting MMR-deficient tumors by conducting immunohistochemical staining on tumor tissue slides for MMR proteins or performing MSI analysis, 2) excluding MMR-deficient tumors exhibiting MLH1 promoter hypermethylation through MLH1 promoter methylation or BRAF p.V600E mutation analysis, and 3) recognizing pathogenic or likely pathogenic variants via germline sequencing of MMR genes in patients who have MMR-deficient, MLH1 non-methylated (or BRAF wild-type) tumors [16]. This multi-step screening protocol (MSP) has gained recognition as a cost-effective and highly sensitive method for universal LS screening among patients with CRC or EC [4,17].

However, it is important to note that MSP has limitations in detecting LS in certain cases – as approximately 6.9% of LS-associated CRC may appear to have proficient MMR expression based on IHC or MSI analysis [18]. Moreover, MLH1 promoter hypermethylation or the BRAF p.V600E mutation, which typically exclude patients from germline sequencing in MSP, can arise as a second hit in LS patients carrying pathogenic MLH1 variants [19,20,21]. Immediate germline sequencing (IGS), which involves the analysis of MMR genes irrespective of MMR deficiency status, could theoretically offer greater sensitivity in comparison to MSP. Although the extensive deployment of IGS for universal LS screening is currently impeded by significant cost factors [17,18], it has the potential to emerge as a superior screening technique for high-risk subpopulations. In turn, this would significantly improve the detection rates of LS.

Here, we hypothesized that IGS has the potential to surpass MSP in identifying LS among high-risk patients who have been pre-selected based on clinical criteria. The Revised Bethesda Guidelines [8] define one of these high-risk groups as patients exhibiting synchronous or metachronous CRC or other LS-associated malignancies, such as EC [22] and, less frequently, ovarian, urothelial, gastric, pancreatic, and brain cancers [23,24]. In this study, our primary objective was to compare the diagnostic yield and sensitivity of IGS and MSP in a cohort of patients with synchronous/metachronous CRC and EC. Furthermore, we sought to analyze the clinical characteristics and overall survival (OS) among study participants, distinguishing between those with and without a definitive LS diagnosis.

MATERIALS AND METHODS

1. Study patients

Participants in this study were identified through a retrospective analysis of patient records. Our eligibility criteria included women diagnosed with synchronous/metachronous EC and CRC (n=31) who were consecutively treated in our hospital between November 2000 and May 2017 (Fig. 1). Synchronous tumors were defined as those diagnosed within a three-month duration from the primary cancer’s initial diagnosis, whereas metachronous cancers were recognized as those diagnosed more than three months apart. We excluded patients who 1) initially underwent non-surgical treatments, or 2) possessed tissue samples that were insufficient or inappropriate for high-quality DNA extraction, making them unsuitable for sequencing. The follow-up period spanned 10.6 years, ranging from 3.6 to 21.4 years. According to the International Classification of Diseases for Oncology, Third Edition (ICDO-3) codes, EC included the following histological types: endometrioid carcinoma (83803), serous carcinoma (84413, 84602, and 84613), mucinous carcinoma (84803), clear cell carcinoma (83103), undifferentiated carcinoma (80203), and adenocarcinoma, not otherwise specified (81403). CRC histological classification comprised adenocarcinoma, not otherwise specified (81403), mucinous carcinoma (84803), and signet-ring cell carcinoma (84903). The Institutional Review Board of the Chang Gung Medical Foundation approved the study (approval number: 201801202B0 and granted a waiver for patient consent due to its retrospective design. Nonetheless, written informed consent was obtained for germline mutation screening.

Fig. 1. Flowchart of the multi-step screening protocol employed for detecting Lynch syndrome in women presenting with synchronous/metachronous endometrial and colorectal cancers. Abbreviations: IHC, immunohistochemistry; MMR, mismatch repair.

2. IHC for MMR protein expression

IHC was utilized to assess the expression levels of MLH1, MSH2, MSH6, and PMS2 proteins in cancer tissue samples. The heat-induced epitope retrieval method was employed using an EDTA-based buffer with a pH of 9.0 (BOND Epitope Retrieval Solution 2; Leica Biosystems, Nußloch, Germany) at 100°C. Following this, sections were immunostained using antibodies against MLH1 (1:50 dilution, catalog number 60-0079; Genemed Biotechnologies, South San Francisco, CA, USA), MSH2 (1:50 dilution, catalog number G219-1129; Zeta Corporation, Arcadia, CA, USA), MSH6 (1:50 dilution, catalog number ab92471; Abcam, Cambridge, UK), and PMS2 (1:50 dilution; catalog number 556415; BD Biosciences, San Jose, CA, USA). The immunoreactivity of these proteins was then evaluated using a BOND Polymer Refine Detection system (Leica Biosystems). The results of IHC were meticulously reviewed in viable, well-preserved regions by an experienced pathologist (RCW), who adhered to following criteria. The MMR protein expression was considered as a loss or negative if the viable tumor cells either demonstrated an absolute lack of definite nuclear staining or displayed only a faint, indeterminate nuclear staining focally. Conversely, the expression was characterized as intact or positive when unequivocal nuclear staining was observable in viable tumor cells. These findings were contextualized with a positive internal control, described by nuclear staining in immune cells, stromal cells, or the adjoining non-tumor epithelium. Any punctate nuclear or cytoplasmic staining patterns were deemed as artifacts, and thus considered a loss if there was an absence of unambiguous nuclear staining.

3. DNA extraction

We used both plasma and tissue samples for germline mutation screening due to the occasional unavailability of plasma specimens. This was especially the case for some patients who had unfortunately succumbed to the disease. DNA was isolated from plasma samples and formalin-fixed paraffin-embedded (FFPE) tissue specimens by employing the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) and the QIAamp DNA FFPE Tissue Kit (Qiagen), respectively [25].

4. MLH1 promoter methylation analysis

The analysis of MLH1 promoter methylation (from position –209 to –181 bp) was conducted on DNA obtained from EC/CRC tissues and plasma samples. This process was carried out with a commercially available pyrosequencing assay (PyroMark Q24 CpG MLH1 kit; Qiagen). Following bisulphite conversion using the EpiTect Fast DNA Bisulfite Kit (Qiagen), PCR amplification for the pyrosequencing analysis was performed with the PyroMark PCR Kit (Qiagen).

5. Germline mutation screening

Screening for germline mutations in MMR genes was performed on DNA extracted from either plasma or normal tissues. The construction of AmpliSeq multiplexed libraries utilized the Ion AmpliSeq Library Kit 2.0 (Life Technologies, Carlsbad, CA, USA). As previously described [26], raw sequence data were analyzed systematically. A variant was considered pathogenic or likely pathogenic if at least one such submission was identified in the ClinVar database [27]. Out of the 31 patients, 14 had DNA samples extracted from plasma, which were screened for mutations using the Ion AmpliSeq Library Kit 2.0 and a customized panel targeting 30 cancer-related genes, including MLH1, MSH2, MSH6, PMS2, and EPCAM. The remaining 17 patients underwent screening of normal tissue samples employing the QIAseq Human Comprehensive Cancer Panel, covering 275 cancer-related genes, and encompassing MLH1, MSH2, MSH6, PMS2, and EPCAM.

6. MSI testing

MSI testing in tumor samples was performed utilizing the Promega MSI analysis system (MD1641; Promega, Madison, WI, USA). This assay incorporates five microsatellite markers, including NR-21, BAT-25, MONO-27, BAT-26, and NR-24.

7. Somatic mutation screening

The QIAseq Human Comprehensive Cancer Panel (DHS-3501Z, Qiagen) – a commercially available tool covering 275 cancer-related genes – was utilized to screen somatic mutations in EC/CRC tissues from a single patient. This panel includes frequently mutated oncogenes, tumor suppressor genes, and signaling pathway members that are potentially actionable by targeted therapies. Genomic DNA (40 ng) extracted from FFPE tissue specimens was used to generate massively parallel sequencing DNA libraries (Kapa Biosystems, Wilmington, MA, USA) containing barcoded universal primers. Each hybrid capture pool underwent deep coverage sequencing in a single paired-end lane of an Illumina flow cell (Illumina, San Diego, CA, USA). The libraries were sequenced on an Illumina NextSeq 500 system (paired-end, 2×150 bp) following the manufacturer’s protocol. Subsequently, raw FASTQ files were uploaded and trimmed, attaching an adapter coupling molecular tag sequence to the read identifier. The raw reads were aligned to the hg19 reference genome, and variant calling was executed using smCounter [28]. A 95% uniformity of coverage was maintained for all analyzed samples. Tissue specimen variants with frequencies below 5% were excluded. ANNOVAR was employed for variant annotation [29].

8. Multiplex ligation-dependent probe amplification

We utilized a commercially available multiplex ligation-dependent probe amplification (MLPA) kit (SALSA MLPA ME011-D1 Mismatch Repair Genes, MRC-Holland, Amsterdam, the Netherlands) to assess the methylation status of specific GCGC sites within the promoter regions of the MLH1, MSH2, MSH6, and PMS2 genes. Additionally, this kit enabled the detection of deletions in the 3’ region of the EPCAM gene and the promoter regions of the MLH1, MSH2, MSH6, and PMS2 genes. DNA samples from three apparently healthy individuals served as negative controls. The MLPA analysis was conducted using an Applied Biosystems 3500Dx Genetic Analyzer (Thermo Fisher Scientific, Waltham, MA, USA), while the Coffalyser.Net software was used for data analysis. Aberrant methylation was deemed present when the calculated methylation percentage exceeded 30%.

9. Statistical analysis

Patient characteristics are summarized using descriptive statistics. To examine differences between groups for continuous and categorical variables, the Mann-Whitney U test and Fisher’s exact test were employed, respectively. OS was calculated as the time elapsed from the initial cancer diagnosis to the date of death, regardless of the cause. The last follow-up date was used for censoring purposes (i.e., administrative censoring). Kaplan-Meier plots were generated to estimate OS curves, and differences between the groups were evaluated using a log-rank test. All analyses were undertaken in SPSS, version 26.0 (IBM, Armonk, NY, USA). Two-tailed p values <0.05 were considered statistically significant.

RESULTS

1. Patient characteristics

The study encompassed 31 women exhibiting synchronous or metachronous EC/CRC, as detailed in Table 1. Out of these patients, 21 (64.5%) had a positive family history of cancer. The median age for the initial malignancy diagnosis was 57 years (range: 39–89 years). Predominantly, patients were diagnosed with early-stage EC, displaying endometrioid histology. Among the 31 participants, five (16.1%) experienced synchronous EC/CRC, 16 (61.5%) encountered EC succeeded by CRC (median interval: 2.1 years; range: 0.9–9.6 years), and 10 (32.3%) had CRC followed by EC (median interval: 8.1 years; range: 1.1–15.6 years).

Table 1. Characteristics of the study patients (n=31).

Characteristics		Entire cohort (n=31)
Age at first cancer diagnosis (yr)		56.5 (38.7–88.9)
	<50	7 (22.6)
	≥50	24 (77.4)
Age at diagnosis of EC (yr)		57.6 (45.4–88.9)
	<50	5 (16.1)
	≥50	26 (83.9)
Age at diagnosis of CRC (yr)		58.5 (38.7–90.1)
	<50	4 (12.9)
	≥50	27 (87.1)
EC histology
	Endometrioid	28 (90.3)
	Non-endometrioid	3 (9.7)
CRC adenocarcinoma		31 (100)
EC FIGO stage
	Stage I–II	28 (90.3)
	Stage III–IV	3 (9.7)
CRC stage
	Stage I–II	20 (64.5)
	Stage III–IV	11 (35.5)
Temporal sequence of synchronous/metachronous tumors
	Synchronous EC/CRC	5 (16.1)
	Metachronous	26 (83.9)
	EC followed by CRC	16 (51.6)
	CRC followed by EC	10 (32.3)

Data are presented as counts, with percentages in parentheses or median (range).

CRC, colorectal cancer; EC, endometrial cancer; FIGO, International Federation of Gynecology and Obstetrics.

2. Multi-step screening of LS

Immunohistochemical staining for MMR proteins

The immunohistochemical staining results for MMR proteins among the 31 study patients are presented in Fig. 1. Out of these participants, 15 (48.4%) displayed intact expression of MLH1, MSH2, MSH6, and PMS2 in both EC and CRC tissues, suggesting sporadic malignancies. In contrast, the remaining 16 patients had a loss of one or more MMR proteins: 11 (35.5%) demonstrated concurrent MLH1 and PMS2 loss, four (12.9%) displayed concurrent MSH2 and MSH6 loss, and one (3.2%) showed isolated MSH6 loss. Upon further examination, three patients (#13, #23, and #27) presented inconsistent results between their endometrial and CRC tissues. In these cases, MLH1 and PMS2 were absent in the endometrial tissue but present in the colorectal tissue. All three patients exhibited MLH1 promoter hypermethylation in their EC tissues, which likely contributed to the MMR deficiency (discussed in the following section).

MLH1 promoter methylation testing and germline mutation screening

We subsequently focused on patients exhibiting MLH1 and PMS2 loss (n=11) and subjected them to MLH1 promoter methylation testing (Fig. 2). Out of these patients, four (#13, #15, #23, and #27) displayed MLH1 promoter hypermethylation and were classified as sporadic cancers, whereas the remaining seven did not. Patients with normal MLH1 promoter methylation status (n=7) and those with MSH2 and/or MSH6 loss (n=5) underwent germline mutation screening (details provided below). Among these 12 patients, six yielded negative results and were categorized as having sporadic cancer. The other six patients had pathogenic mutations and were diagnosed with LS (Fig. 1).

Fig. 2. Flowchart showcasing the simulation of immediate germline sequencing for MMR genes in women diagnosed with synchronous/metachronous endometrial and colorectal cancers. Lynch syndrome cases are highlighted using color. Abbreviations: IHC, immunohistochemistry; MMR, mismatch repair.

We have summarized the variants identified through germline sequencing in Table 2. Specifically, three out of seven patients with MLH1/PMS2 loss, two out of four patients with MSH2/MSH6 loss, and one patient with isolated MSH6 loss exhibited MMR gene mutations and were diagnosed with LS (#5, #9, #21, #22, #28, and #31). Among these patients, #9, #21, and #28, whose tumors lacked MLH1 and PMS2 expression, were carriers of pathogenic/likely pathogenic MLH1 mutations (p.Q346H, p.R497Pfs*6, and p.T117M, respectively). Patients #5 and #22, whose tumors lacked MSH2 and MSH6 expression, carried an MSH2 splicing variant (c.1076+1G>T) and a pathogenic MSH2 variant (p.R227Kfs*3), respectively. Patient #31 had tumors with isolated MSH6 loss and carried a likely pathogenic MSH6 variant (p.G685D). For the two patients with MSH2 and MSH6 loss but no evidence of MMR gene mutations (#19 and #25), we performed MLPA testing to analyze MSH2 promoter methylation. However, we did not detect any cases of promoter hypermethylation or deletion of the 3’ region of EPCAM.

Table 2. Detailed characteristics of the six patients with synchronous/metachronous endometrial and colorectal cancers accompanied by loss of MMR proteins and diagnosed with Lynch syndrome.

Patient identifier	Age at cancer diagnosis (yr)		Family history	Disease stage		Loss of MMR proteins		Germline pathogenic/likely pathogenic mutations or epimutations
	EC	CRC		EC	CRC	EC	CRC	Blood or non-tumor tissue specimens
#9	71	60	Yes	IIIC1	IIA	MLH1, PMS2	MLH1, PMS2	MLH1 (p.Q346H)
#21*	53	71	Yes	II	IIB	MLH1, PMS2	MLH1, PMS2	MLH1 (p.R497Pfs*6)
#28	59	52	Yes	IA	IIIA	MLH1, PMS2	MLH1, PMS2	MLH1 (p.T117M)
#5	55	57	Yes	IA	IIIC	MSH2, MSH6	MSH2, MSH6	MSH2 (splicing:c.1076+1G>T)
#22	55	55	Yes	IA	IIB	MSH2, MSH6	MSH2, MSH6	MSH2 (p.R227Kfs*3)
#31	55	56	Yes	IA	I	MSH6	MSH6	MSH6 (p.G685D)
#15†	60	65	Yes	IA	IIIB	MLH1, PMS2	MLH1, PMS2	MLH1 promoter hypermethylation

CRC, colorectal cancer; EC, endometrial cancer; MMR, mismatch repair.

^*Patient also diagnosed with gastric cancer.

^†Patient also diagnosed with liver cancer.

In conclusion, our combined results from IHC, MLH1 promoter methylation testing, and germline mutation screening revealed a 19.4% (6/31) prevalence of LS in our cohort.

3. Germline sequencing for patients excluded through the MSP

Out of the 31 study participants, 19 would have been disqualified from germline mutation analysis in a multi-step screening process. Of them, 15 displayed normal immunohistochemical staining for MMR proteins, while four exhibited MLH1 promoter hypermethylation. Nevertheless, germline sequencing uncovered five of these 19 participants (specifically, #2, #3, #14, #18, and #24) possessed germline genetic variants in MMR genes (Table 3). Among them, patient #3 and #14 were found to carry pathogenic/likely pathogenic variants (MSH6 p.A1320Efs*7 and MSH6 p.F432L, respectively) and was diagnosed as LS. In contrast, patient #24 carried a likely benign variant (MLH1 p.Q701K), while patients #2 and #18 had variants of uncertain significance (VUS) as per the ClinVar database [27]. It is important to highlight that the variant detected in patient #2 (MSH6 p.Y427D) only had a single submitted record in ClinVar.

Table 3. Detailed characteristics of the five patients with synchronous/metachronous endometrial and colorectal cancers without loss of MMR proteins who harbored germline genetic variants.

Patient identifier	Age at cancer diagnosis (yr)		Family history	Disease stage		MSI		Gene mutations
	EC	CRC		EC	CRC	EC	CRC	EC	CRC	Germline
#2*	45	41	Yes	IA	IIIB	MSS	MSI-H	MSH6 (p.Y427D)	MSH6 (p.Y427D)	MSH6 (p.Y427D)
								MSH6 (p.F1088Lfs*5)	MSH6 (p.R495X)
#3*	65	69	Yes	IB	IIIA	MSS	MSS	NP	NP	MSH6 (p.A1320Efs*7)
#14*	72	72	Yes	IA	IIB	MSS	MSI-H	NP	NP	MSH6 (p.F432L)
#18†	60	59	Yes	IB	IIC	MSS	MSS	NP	NP	MLH1 (p.L259S)
#24†	55	58	Yes	IA	IIB	MSS	MSS	NP	NP	MLH1 (p.Q701K)

CRC, colorectal cancer; EC, endometrial cancer; MMR, mismatch repair; MSI, microsatellite instability; MSI-H, microsatellite instability-high; MSS, microsatellite stable; NP, not performed.

^*Patient diagnosed with Lynch syndrome.

^†Presence of non-pathogenic variants.

To further examine the pathogenicity of the VUS variant (MSH6 p.Y427D) found in patient #2, we conducted a somatic mutation analysis on EC and CRC tumor tissues. Based on Knudson’s two-hit hypothesis [30], we anticipated discovering “second hits” in tumor tissues if the germline MMR variant was indeed pathogenic. Interestingly, tumor sequencing in patient #2 revealed a truncating pathogenic variant (MSH6 p.F1088Lfs*5) in the EC tissue and another variant (MSH6 p.R495X) in the CRC tissue, which likely represent the second hits subsequent to the germline missense mutation (MSH6 p.Y427D) (Table 3). Moreover, additional MSI analysis disclosed MSI-high in the CRC tissue of patient #2, indicating the functional impact of her germline variant despite the presence of intact MSH6 protein expression (Table 3). Based on these observations, we conclude that patient #2 likely had LS as well.

4. Analysis of germline MLH1 promoter methylation

Although constitutional MLH1 promoter hypermethylation is an infrequent cause of LS, it can still be present in a small percentage of cases [2,3,31]. In such instances, individuals with LS due to constitutional MLH1 hypermethylation should also display MLH1 promoter hypermethylation within their tumor tissue. This can be identified during the second phase of MSP, potentially leading to exclusion from further germline sequencing. To pinpoint patients with constitutional MLH1 hypermethylation, we carried out an in-depth analysis on four patients (#13, #15, #23, and #27) whose tumors showed MLH1 promoter hypermethylation. We evaluated the MLH1 promoter methylation status in DNA extracted from their plasma samples (Fig. 2). Our analysis identified one patient (#15) with evidence of germline hypermethylation in the MLH1 promoter (Table 2). This patient was diagnosed as LS due to the presence of constitutional MLH1 promoter hypermethylation [2,3,31].

5. Comparison between IGS and the MSP

Among the 31 patients with synchronous/metachronous EC and CRC, MSP identified six LS cases. In contrast, employing IGS would have detected nine LS cases, resulting in a 29.0% diagnostic yield compared to MSP’s 19.4%. Moreover, IGS demonstrated a 90% sensitivity rate, surpassing the 60% sensitivity rate of MSP. Further analysis of germline MLH1 promoter methylation in patients with MLH1 hypermethylated tumors led to the identification of an additional LS case, enhancing the diagnostic yield to 32.3%.

The IGS methodology relies entirely on the germline mutation screening process, which typically mandates a timeframe of between 10 to 30 days. Conversely, the MSP technique introduces a longer turnaround time, adding an additional two to six days. The latter requires a minimum of two extra working days for IHC staining. In cases where there is a loss of MLH1 expression, necessitating the inclusion of the MLH1 promoter methylation test, the procedure could extend by up to six additional working days. Financially, costs related to the IHC for MMR proteins, the MLH1 promoter methylation testing, and the germline mutation screening equated to NT$5,416, NT$1,781, and NT$25,000, respectively. The MSP process entailed 31 IHC tests, 11 MLH1 promoter methylation tests, and 12 germline mutation screenings. The total expense for this approach was NT$487,487, leading to the detection of six LS cases. This resulted in an average cost of NT$81,248 per identified case. As for the IGS approach, all 31 patients underwent germline mutation screening, resulting in a total cost of NT$775,000. This strategy successfully identified nine LS cases, equating to a cost of NT$86,111 per identified LS case, which is only six percent more than the MSP approach (NT$81,248).

6. Comparison of patients with and without LS

Among the 31 patients with synchronous/metachronous EC and CRC, ten (32.3%) were diagnosed with LS. Six of these LS cases (#5, #9, #21, #22, #28, and #31) were identified during MSP, three were detected through supplementary germline sequencing of MMR genes (#2, #3, and #14), and one was discovered via further germline MLH1 promoter methylation analysis (#15). We evaluated the characteristics of study participants based on the presence or absence of LS (Table 4). However, we observed no significant differences between the groups concerning age at initial diagnosis, age at EC or CRC diagnosis, EC or CRC stage, or the temporal sequence between the first and second cancer diagnoses. The median follow-up duration for censored patients was 12.5 years (range: 3.6−21.4 years). Additionally, no significant difference was found in OS between patients with and without LS (log-rank test, p=0.246; Fig. S1).

Table 4. Characteristics of the study patients with and without a conclusive diagnosis of Lynch syndrome.

Characteristics		LS (n=10)	Non-LS (n=21)	p
Age at first cancer diagnosis, years; median (range)		57.4 (41–72.4)	56.5 (38.7–88.9)	0.447
Age at first cancer diagnosis (yr)				0.379
	<50	1 (10)	6 (28.6)
	≥50	9 (90)	15 (71.4)
Age at diagnosis of EC, years; median (range)		59.3 (45.4–72.5)	56.5 (49.2–88.9)	0.916
Age at diagnosis of EC (yr)				1.000
	<50	1 (10)	4 (19)
	≥50	9 (90)	17 (81)
Age at diagnosis of CRC, years; median (range)		58.5 (41–72.4)	58.5 (38.7–90.1)	0.612
Age at diagnosis of CRC (yr)				1.000
	<50	1 (10)	3 (14.3)
	≥50	9 (90)	18 (85.7)
EC histology				1.000
	Endometrioid	9 (90)	19 (90.5)
	Non-endometrioid	1 (10)	2 (9.5)
CRC adenocarcinoma		10 (100)	21 (100)
EC FIGO stage				1.000
	Stage I−II	9 (90)	19 (90.5)
	Stage III−IV	1 (10)	2 (9.5)
CRC stage				0.423
	Stage I−II	5 (50)	15 (71.4)
	Stage III−IV	5 (50)	6 (28.6)
Temporal sequence of metachronous tumors				1.000
	EC followed by CRC	2 (20)	3 (14.3)
	CRC followed by EC	5 (50)	11 (52.4)
	Synchronous EC/CRC	3 (30)	7 (33.3)

Data are presented as counts, with percentages in parentheses, unless specified otherwise.

CRC, colorectal cancer; EC, endometrial cancer; FIGO, International Federation of Gynecology and Obstetrics; LS, Lynch syndrome.

DISCUSSION

In this study, we sought to compare the diagnostic performances of two LS screening approaches (IGS and MSP) in a high-risk group of 31 patients presenting with synchronous/metachronous colorectal and ECs. Our analysis revealed that IGS demonstrated superior performance compared to MSP, both in terms of diagnostic yield (29.0% vs. 19.4%, respectively) and sensitivity (90% vs. 60%, respectively). Specifically, IGS effectively pinpointed nine cases of LS, signifying a substantial 50% improvement in detection compared to the patients identified by MSP. Our findings lead us to conclude that, among high-risk groups, IGS surpasses MSP as a superior strategy for accurately identifying individuals with LS.

Despite MSP being regarded as the most cost-effective method for universal LS screening, it does come with certain drawbacks. A primary concern is its relatively low sensitivity, which falls within the range of 80% to 90%. Consequently, this could result in some patients with LS being overlooked during the initial screening stages, thereby excluding them from germline sequencing [17,32]. For example, as many as 6.9% of LS-associated CRCs may not display MMR deficiency when assessed using IHC or MSI analysis, resulting in potential underdiagnosis [18].

Furthermore, it is important to note that approximately 70% of methylated tumors exhibit BRAF p.V600E mutations, which serve as a surrogate marker for MLH1 promoter methylation. Patients with LS may have somatic BRAF p.V600E mutations present in their tumors [20] which can render them ineligible for germline sequencing. Similarly, the occurrence of hypermethylation in the MLH1 promoter could arise as a second hit in LS patients who possess pathogenic MLH1 variants. Alternatively, it could manifest as a germline epimutation in the MLH1 gene [19,21]. Besides the issue of suboptimal sensitivity, the intricate and multi-stage nature of the MSP procedure may contribute to misunderstandings among both patients and medical professionals. This complexity raises the possibility of inadequate follow-up care [33].

A comparative economic assessment of various universal LS screening methods determined that direct germline sequencing of MMR genes offered the highest sensitivity and specificity; however, it would only be as cost-effective as MSP if the germline sequencing costs were to decrease to $368 or lower [17]. A 2014 cost-effectiveness analysis indicated that direct germline MMR gene testing was a financially viable approach for LS screening solely for CRC patients diagnosed under the age of 60 [18,34]. As next-generation sequencing becomes more affordable and accessible, we contend that IGS has evolved into a superior cost-effective strategy for LS screening in highly susceptible populations, including those fulfilling the criteria outlined in the Revised Bethesda Guidelines.

Germline hypermethylation of the MLH1 promoter represents an alternate mechanism predisposing individuals to LS [31,35]. In a study examining 331 patients with suspected LS, Niessen et al. [31] discovered germline MLH1 promoter hypermethylation in 0.6% of those screened. In our study, we observed that patient #15 exhibited MLH1 protein loss on immunohistochemical staining, despite not having any pathogenic MLH1 germline mutations. Notably, both plasma and EC/CRC tissue samples from this woman displayed hypermethylation of the MLH1 promoter. This finding substantiates the notion that constitutional MLH1 promoter hypermethylation is present in a small fraction of patients experiencing MLH1 protein loss [2]. Consequently, it may be beneficial to consider germline MLH1 promoter methylation testing for this patient group, either universally or following IHC screening for MLH1 loss.

Significantly, all three LS cases detected by IGS but missed by MSP were found to have MSH6 germline mutations. MSP failed to identify these cases because MSH6 expression remained intact on IHC in both CRC and EC tissue samples, even though MSI-high was present in two CRC specimens but absent in the EC tissue (Table 3). In a prior investigation by Berends et al. [36], the authors discovered that one third of tumors in carriers of MSH6 germline truncating mutations still exhibited MSH6 expression on IHC. Furthermore, tumors associated with LS due to MSH6 mutations typically display lower levels of MSI, particularly in EC [36,37]. Considering that a substantial percentage of tumors with MSH6 mutations may appear normal on IHC or MSI tests, researchers have proposed that MSH6 sequencing should be contemplated for all patients suspected of having LS, irrespective of IHC or MSI findings [38]. When processing the MMR VUS in sequencing data, it is critical to revisit variant classifications through an in silico analysis that integrates clinical data, predicts alterations from RNA to protein sequence, and employs functional assays [19]. In line with this recommendation, our study also endorses the utilization of IGS, which proved more efficient than MSP in detecting LS cases caused by MSH6 mutations.

The prognostic implications of LS in patients with metachronous EC/CRC remain uncertain. Our investigation revealed no significant differences in clinical features or OS among patients with or without LS. In contrast, a propensity score-matched study by Xu et al. [39] demonstrated that LS patients experienced improved long-term survival outcomes compared to those with sporadic CRC, despite having comparable recurrence-free survival. The lack of significant differences in clinical characteristics between LS and non-LS patients in our cohort may be due to the limited sample size or the potential presence of undiagnosed pathogenic germline mutations/epimutations, which could be more prevalent in our high-risk patient population.

This study has several limitations that warrant consideration. First, obtaining consent for sequencing remains a significant barrier to implementing universal screening in routine practice [40]. Consequently, the methodology outlined in this research may not always be applicable in real-world settings, necessitating independent validation. Second, comprehensive family history information was not available for all study patients, potentially impacting the results. Finally, germline mutation screening was conducted using two distinct matrices, namely plasma and normal tissue samples, due to the unavailability of plasma specimens for all participants. This variability could also influence the findings.

In conclusion, our research uncovered a considerable prevalence of LS among patients presenting with synchronous/metachronous colorectal and ECs. The enhanced diagnostic yield and sensitivity of IGS advocate for its adoption as a superior alternative to MSP in detecting LS among high-risk patients who fulfill the clinical criteria outlined in the Revised Bethesda Guidelines. This advanced approach may provide these patients with more precise genetic counseling, particularly concerning the application of immunotherapy for tumors with MMR deficiency, the potential development of other malignancies, and the associated cancer risk for their family members. By adopting a streamlined one-step testing process, IGS eliminates the complex procedures involved with MSP, thereby delivering more efficient and effective results in a clinical setting.

SUPPLEMENTARY MATERIAL

Fig. S1

Kaplan-Meier curves comparing overall survival in patients with and without a conclusive Lynch syndrome diagnosis. Although patients with Lynch syndrome exhibited a trend towards improved overall survival, the observed difference did not achieve statistical significance.