Patients and samples

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Human Research Ethics Committee (HREC), Faculty of Medicine, of the Prince of Songkhla University (PSU), Thailand (REC.64-223-10-1; approval date: January 5, 2022) and the ethical committee of Ubon Ratchathani Cancer Hospital (URCH), Thailand (EC 013/2021; approval date: October 20, 2021). Participant recruitment took place between January 29, 2022, and December 29, 2023. Due to the necessity of accessing clinical data, family history, and histopathological information, the authors WW, NS, and PS could access information that could identify individual participants during and after the data collection process.

All patients and recruited family members were informed of their inclusion in the registries, and written informed consent was obtained during genetic counseling sessions. The individual in this manuscript has given written informed consent to publish these case details. For participants aged 18 years and above, written informed consent was obtained directly. For participants under 18 years of age, written informed consent was obtained from their parents or legal guardians, and written assent was obtained from the participants themselves. All minors were accompanied by their parents or guardians during the counseling sessions, and an information letter explaining the study’s purpose was provided. There were no deviations from the approved study protocol

Initially, a total of 132 patients with familial CRC and/or EC and with/without other cancer-related LS diagnosed and treated at the Songklanarind Hospital or Ubon Ratchathani Cancer Hospital between October 2001 and December 2023 were recruited for the study. All of them met at least one criteria of the RBG which were younger than 50 years when the cancer was detected, and/or had synchronous/metachronous CRC or other LS-associated tumors at any age, and/or having first- or second-degree relatives with CRC or LS-associated tumors diagnosed at an early age or in at least two family members [12]. Their histopathological demographics such as age at diagnosis, tumor site, staging, histologic grade, type, and metastasis, were collected from the Hospital Information System (HIS).

DNA extraction from blood and QC

Genomic DNA was isolated from peripheral blood lymphocytes using the GeneJET Genomic DNA Purification Kit (Thermo Scientific, Lithuania) following the manufacturer’s instructions. The extracted DNA was qualified and quantified using a Nanodrop One fluorescence spectrophotometer (Thermo Scientific, USA) and stored at 4°C until used.

DNA extraction from FFPE tissue

Tumor DNA was extracted from formalin-fixed paraffin-embedded (FFPE) tumor QIAamp Tissue Kit or the Reliaprep FFPE gDNA Miniprep System (Promega) according to the manufacturer’s recommendations. DNA concentration was measured by a Nanodrop One fluorescence spectrophotometer (Thermo Scientific, USA), and samples were stored at 4°C.

Immunohistochemical staining of MLH1, PMS2, MSH2, and MSH6 proteins

Immunohistochemical (IHC) staining of DNA mismatch repair (MMR) proteins including MLH1, PMS2, MSH2 and MSH6, was performed using mouse monoclonal MLH1, MSH2, MSH6 and PMS2 antibodies (Abcam, USA). Expression of MMR proteins in the nucleus of tumor cells was defined as positive staining of these proteins. Normal expression of protein was defined as the presence of nuclear staining in colon cancer cells. Loss of staining in carcinoma with concurrent positive staining in nuclei of normal colon epithelial cells indicated absent expression of protein.

Microsatellite analysis

Analysis for microsatellite instability (MSI) was performed with the automated Idylla™ MSI test prototype cartridges using seven microsatellite markers (ACVR2A, BTBD7, DIDO1, MRE11, RYR3, SEC31A, and SULF2). Following the manufacturer’s protocol, sections of 10-μm thickness were obtained from formalin-fixed paraffin-embedded (FFPE) tissue samples, covering an area of 100–200 mm². Each section contained at least 20% tumor cells. The prepared tissue sections were then placed into a sample cartridge, which was subsequently inserted into the instrument as per the previously established method [24]. The Idylla™ Console software analyzed results and generated a final report of microsatellite status as microsatellite stable (MSS) or microsatellite instable (MSI).

Whole genome sequencing and bioinformatic pipelines

Genomic DNA extracted from peripheral blood lymphocytes of 132 suspected LS patients underwent whole genome sequencing (WGS) on the DNBSEQ-T7 platform (MGI Tech Co.). The MGI MGIEasy FS PCR-Free DNA Library Prep Set was used to prepare DNA libraries. Reads (>100 bp) with Q20 > 90% and Q30 > 80% were aligned to GRCh38 with decoy genomes using the Burrows-Wheeler Aligner (BWA) [25]. Quality control with SAMtools and GATK CollectWGSMetrics required mapping rate ≥ 90%, unique mapping ≥ 85%, mean depth > 30x, > 90% of bases at ≥ 20x, and > 95% of bases at ≥ 10x. Across LS–associated genes; EPCAM (NM_002354.3), MSH2 (NM_000251.3), MSH6 (NM_000179.3), MLH1 (NM_000249.4), and PMS2 (NM_000535.7), the mean depth was 38x with 86–96% of target regions covered at ≥20x, supporting high-confidence variant calls (S1 Table). Short variants (SNPs and INDELs) were called following GATK4.0 best practices [26].

Duplicates were marked with Picard tools, followed by base quality score recalibration (BQSR) and local realignment around indels. Variants were called using GATK HaplotypeCaller, and filtered using VQSR/VariantFiltration. High-quality variants were annotated with Ensembl Variant Effect Predictor (VEP) build 113 [27] using dbNSFP 4.9a [28] and gnomAD v 4.1 [29]. TAPES [30], was used to classify variants as pathogenic (P), likely pathogenic (LP), variant of uncertain significance (VUS), likely benign (LB), or benign (B) according to the 2015 ACMG/AMP guidelines [31]. Sample-level annotated variants are available through the Thailand Variant Annotation and Prioritization Platform (v@pp) (https://vapp.genomicsthailand.com/), which provides direct link-outs to curated knowledge bases including GeneCards, OMIM, GeneReviews, gnomAD, and VarSome.

Structural variants (SVs) were called from BAM files using Manta [32] and annotated with AnnotSV [33]. We focused on EPCAM, MSH2, MSH6, MLH1, and PMS2, and classified variants by pathogenicity (P > LP > VUS), according to the 2015 ACMG/AMP guidelines. All variants were cross-referenced with InSiGHT, ClinVar, and gnomAD databases.

For promising VUSs, pathogenicity was assessed using in-silico tools, including PolyPhen-2, Mutation Taster, SpliceAI, and Priors. We further applied the preprinted ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis Variant Curation Expert Panel (VCEP) specifications for MMR genes v 1.0.0 to refine variant classification where possible.

MSH2 c.1237C > T genotyping study

The genomic DNA of family members was extracted from peripheral blood leukocytes using the GeneJET Genomic DNA Purification kit (Thermo Scientific, Lithuania) according to the manufacturer’s instructions. DNA samples were stored at −20°C until used. Primers were designed with Primer3web version 4.1.0 (https://primer3.ut.ee/); the sequences of forward and reverse primers were 5′- AAGATGCAGAATTGAGGCAGA-3′ and 5′- GGACAGCACATTGCCAAGTA-3′, respectively. The expected PCR product size was 242 bp. PCR was performed following standard protocol. Briefly, the PCR mixtures contained 10 μl of 2X PCR SuperMix (Bio-Helix Co. Ltd.), 1 μl of each primer at 10 μM, 8 μl of distilled water, and 1 μl of the DNA template, a total volume of 20 μl for each reaction. The reactions were conducted in the C1000 Thermal Cycler (BioRad Laboratories, USA) under the following conditions as initial denature of 95^oC for five minutes, followed by five cycles of denaturation at 95^oC for 30 sec, annealing at 60^oC for 30 sec, and extension at 72^oC for one minute, and then a final extension at 72^oC for five minutes. After amplification, 2 μl of each PCR product was visualized by gel electrophoresis in a 1.5% agarose gel mixed with RedSafe™ Nucleic Acid Staining Solution (iNtRON Biotechnology). PCR products were purified using NucleoSpin Gel and PCR Clean-up Kit (Macherey-Nagel GmbH and Co. KG, Düren, Germany). The purified PCR products were bidirectionally sequenced using the Sanger method at Macrogen Co. Ltd. (Seoul, Korea).

BRAF V600E and KRAS G12D analyses

To examine the BRAF V600E and KRAS G12D variants in tumor DNAs, we performed PCR and direct sequencing following standard protocol as described above. The 228 bp fragment of BRAF encompassing exon 15 and/or 225 bp fragment of KRAS encompassing exon 1 were amplified using 2X PCR SuperMix (Bio-Helix Co., Ltd.) and primers BRAF forward 5’-TGCTTGCTCTGATAGGAAAATG-3’ and reverse 5’-AGCATCTCAGGGCCAAAAAT-3’, and/or primers KRAS forward 5’-TTAACCTTATGTGTGACATGT TCTAA-3’ and reverse 5’-AGAATGGTCCTGCACCAGTAA-3’. PCR product was purified using Purelink™ Genomic DNA Mini kit (Invitrogen) and then submitted to Macrogen’s sequencing service (Macrogen Co. Ltd.) for the sequencing.

Screening of family members of probands with MSH2 PV and segregation analysis

The identity-by-descent (IBD) was performed by using KING software [34] to identify the relatedness among these patients. Genetic counselling was offered to family members of the probands who carried an MMR PV/LPV. Genetic testing was carried out in high-risk relatives with/without previous cancer diagnosis if they consented. Particularly, the family members of probands with MSH2 c.1237C > T, we contacted them by telephone and invited them to the hospital for interviewing in order to obtain their family histories. Their clinical data was retrieved from the Hospital Information System (HIS). A detailed family history and pedigree chart were created after interviewing to see whether the segregation of the PV corresponds with the LS-associated cancers.

Penetrance rate and cumulative risk calculations

Given the small number of cohort, a crude penetrance was calculated as a simple proportion of carriers who had any LS-associated cancer divided by the total number of carriers. To obtain more refined estimates, cumulative risk was calculated using AI-assisted computation (ChatGPT, OpenAI) in conjunction with actuarial life-table analysis with 10-year age intervals, following previously established approaches for hereditary cancer risk estimation [35–37]. Sex-stratified life tables were subsequently constructed by generating separate age-specific risk sets for males and females, with parallel calculation of interval-specific probabilities, survival functions, cumulative risks [35–37].

Cost–benefit analysis

Cost–benefit analysis is an approach used to evaluate whether a project, policy, or investment is worthwhile. It compares all expected costs (monetary and non-monetary) against all expected benefits, usually over time, to determine if the benefits outweigh the costs. Cost–Benefit Analysis begins by clearly defining the project or decision and setting its scope and objectives. Next, all potential costs and benefits are identified, including direct, indirect, tangible, and intangible factors. These are then quantified in monetary terms, often using financial proxies for non-market values. Since money has different values over time, future costs and benefits are discounted to present values using a discount rate. To evaluate economic viability, decision criteria like the Benefit–Cost Ratio (BCR) are computed. Finally, conclusions are drawn, and recommendations are made based on whether the benefits outweigh the costs. Key indicators used to consider whether the benefits outweigh the costs are as follows:

Benefit–Cost Ratio (BCR): Ratio of present value of benefits to present value of costs.

Decision rule: If BCR > 1, project is economically justified, whereas BCR < 1, costs outweigh benefits.

Clinicopathological data of 132 suspected Lynch syndrome patients

A total of 132 suspected Lynch syndrome (LS) patients (56 male and 76 female) who were diagnosed with colorectal cancer (CRC) and/or endometrial cancer (EC) and with other cancer-related LS at Prince of Songkhla University and Ubon Ratchathani Cancer Hospital between 2001 and 2023 were recruited in the study. The age of onset was 16–93, with the average age at 52. All patients met the RBG [12] but only 59 patients (45%) fulfilled the ACII [13]. While the majority of patients (92%) developed CRC with/without extracolonic cancer (average age of 52), the minority of them (8%) developed EC with/without OC (average age of 50). Sixty-one patients (46%) were younger than 50 years, thirty-one patients (24%) were 50–59 years, twenty-nine patients (22%) were diagnosed at 60–69 years, and eleven patients (8%) were older than 70 years at the age of diagnosis (Table 1).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Clinicopathological features of 132 suspected LS patients.

https://doi.org/10.1371/journal.pone.0348867.t001

Analysis of microsatellite status and immunohistochemical staining of DNA MMR proteins

The microsatellite instability (MSI) test was performed in 61 of 132 suspected LS patients. Eighteen of them exhibited MSI-High (MSI-H), while 43 were MSS. Of the 18 MSI-H cases, eight showed loss of MLH1 and PMS2 expression, and five showed loss of MSH2 and MSH6 expression. For the remaining five MSI-H patients, the MMR IHC staining result was unavailable. Of the 43 MSS cases, IHC staining of MMR proteins was performed in eight cases. All of them showed normal expression of all four MMR proteins, consistent with their MSS status (Table 2).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Number of patients with microsatellite status and expression analysis of MMR proteins.

https://doi.org/10.1371/journal.pone.0348867.t002

Of the 71 cases who did not undergo MSI testing, IHC staining was done in five cases. Two showed loss of MLH1 and PMS2 expression, one showed loss of MSH2 and MSH6 expression, and one exhibited normal expression of all four MMR proteins.

Identification of germline pathogenic and likely pathogenic variants in DNA MMR genes in 26 of 132 suspected LS patients

Analysis of MLH1, MSH2, MSH6, PMS2, and EPCAM identified 8 PVs/LPVs in 23 suspected LS patients; five variants in MLH1 (11 patients), two in MSH2 (11 patients), and one in PMS2 (1 patient). In addition, three promising VUSs were detected in MMR genes (Table 3). In total, 26 patients carried a PV/LPV/promising VUS. Among them, twelve with MLH1 variants were diagnosed between 23–65 years (mean ± SD: 49.0 ± 12.5), twelve with MSH2 variants between 23–71 years (mean ± SD: 44.5 ± 12.2), and two patients with PMS2 variants at 47 and 65 years (mean 56). No PV/LPV/promising VUS was detected in MSH6 or EPCAM.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Clinicopathological characteristics of 26 suspected LS patients carrying PVs/LPVs/promising VUSs in DNA MMR genes.

https://doi.org/10.1371/journal.pone.0348867.t003

Of the three promising VUSs, MLH1 c.1883T > C; p.Leu628Ser occurred in a patient (GN065) who developed duodenal cancer and CRC at the ages of 65 and 68 respectively. Pathological study showed loss of MLH1 & PMS2 expression and BRAF V600E negativity. Referring to his family history, his three brothers developed CRC at the ages of 30, 40, and 50. All of them are deceased. In addition, a brother of his mother had also developed CRC (unknown age) and was deceased. This VUS is absent from gnomAD. It is reported once on InSiGHT as it was found in one LS patient who had a positive family history and twice on ClinVar; one was found in an LS patient and the other was reported with an unknown phenotype. Since segregation of the variant with the disease could not be verified, the current evidence does not meet the criteria to classify it as PV/LPV although the pathological findings and family history support a likely association.

MSH2 c.923_937del; p.Arg308_Phe313delinsIle, detected in a female patient (GN048) with early-onset CRC and EC at 41 and 47 years with MSI-H and loss of MSH2 and MSH6 expression, was absent from InSiGHT, ClinVar, and gnomAD and classified as VUS under the ClinGen InSiGHT Hereditary Colorectal Cancer/Polyposis VCEP MMR specifications because in-frame length change alone is not considered sufficient for LPV classification without functional or segregation data.

PMS2 c.4dup; p.Glu2GlyfsTer4, a frameshift variant was found in a male patient (GN053) who developed rectal cancer and prostate cancer at the ages of 65 and 70, respectively. This germline variant is absent from gnomAD and ClinVar, but reported once on the InSiGHT database, found in a breast cancer patient [38]. The Franklin, an online variants classification tool (https://franklin.genoox.com/clinical-db/home), based on the 2015 ACMG/AMP guidelines [31], classified this variant as a LPV (PVS1, PM2). Consistent with the Franklin classification, we classified this variant as a PV based on the following criteria: 1) it is a frameshift, protein length changing variant (PM4); 2) loss of function is a well-established disease mechanism for this gene (PVS1); and 3) it is absent from gnomAD, the population database (PM2).

Among MLH1 PV/LPV carriers, most developed CRC (92%) with one extracolonic tumor (8%). In contrast, among MSH2 PV/LPV carriers, both CRC and extracolonic cancers occurred with comparable frequency (Table 3).

Structural variants (SVs) were detected with Manta and annotated using AnnotSV (S2 Table). A 79-bp deletion in MSH2 was observed in 111 patients (47 heterozygous, 64 homozygous), and a separate 80-bp deletion in MSH2 was found in one patient. In MLH1, a 56-bp deletion was identified in two patients. All deletions were intronic and classified as VUS.

Family histories and pedigree analysis of patients with MSH2 c.1237C > T, p. Gln413Ter reveal one big family tree in Phatthalung province

In our study, MSH2 c.1237C > T PV emerged as the most frequent finding. The IBD analysis demonstrated genetic relatedness among seven patients carrying this variant. Specifically, GN085 and GN007 were identified as a parent-offspring pair, GN012 and GN013 as full siblings, and GN097 and GN360 as full siblings as well. GN073 was found to be a second-degree relative of GN097 and GN360. (S3 Table). We contacted the probands and/or their family members. Based on the genetic counselling, we identified that the seven probands are originally from Phatthalung province and they share the ancestors (I.1 and I.2) (Fig 1). Patient GN027 was found to have the same last name as individual I.2. His maternal grandfather was born in Phatthalung, and he vaguely recalled being a cousin of the extended family, although he could not confirm whether the exact relationship is from the paternal or maternal side. Similarly, patient GN064 reported that her ancestors originally resided in Phatthalung but migrated to Songkhla during her great-grandparents’ or great-great-grandparents’ generation. In contrast, patient GN197 had no identifiable connection to this extended family. None of his family members has developed cancer. Additionally, their ancestors did not reside in Phatthalung but originated from Songkhla.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Pedigree of the nine probands with MSH2 c.1237C > T, p.Gln413Ter; seven probands (GN007, GN012, GN013, GN073, GN085, GN097, and GN360) are from the same family in Phatthalung Province, and two probands (GN027 and GN064) are in some way relatives.

The pedigree of the proband GN197 is not shown.

https://doi.org/10.1371/journal.pone.0348867.g001

Patient IDs written in red with the arrow represent the probands; Patient IDs written in black represent the family members who were tested but are not carriers; Patient IDs written in purple represent the family members who were tested and carry the MSH2 PV.

Histopathological studies of 10 suspected LS patients with MSH2 c.1237C > T PV

Of the ten patients, six (4 male, 2 female) developed CRC at the ages of 23, 30, 46, 47, 58, and 70, and four female patients developed EC at the ages of 35, 45, 47, and 47 (mean age ± SD: 44.8 ± 13.4). Four patients had MSI-H, whereas the microsatellite status was unknown to the other six patients (Table 3).

Pathogenicity classification of MSH2 c.1237C > T, p.Gln413Ter

The MSH2 c.1237C > T, p.Gln413Ter has already been reported as PV three times (1 LS, 2 unknown) on the InSiGHT database and reported six times (5 PVs, 1 LPV) on ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/ variation/218030/). Three of the five PVs were identified in LS while the other three reports did not provide the condition of the patients. The MSH2 c.1237C > T, p.Gln413Ter is not reported on gnomAD.

Germline variant screening for MSH2 c.1237C > T in family members

After interviewing and genetic counseling, we were able to collect venous blood from 56 family members for an analysis of the germline MSH2 c.1237C > T variant (Fig 1). By sequencing, we can identify MSH2 c.1237C > T in 15 members (27%) (5 females, 10 males). The ages of carriers were between 13 and 84 years (mean ± SD: 44 ± 14.6). A review of the medical history revealed that one male (GN109) developed ampullary cancer at the age of 73, while two members; one female (GN415) and one male (GN498) developed CRC at the age of 33 and 52 years, respectively. During a surveillance program, we detected an early-stage CRC in a 55-year-old male, and we found polyps in two members (ages 34 and 58 years). The other nine members; 5 males (GN014, GN478, GN504, GN530, GN637), and 4 females (GN028, GN029, GN037, GN505) with the age of 34–65 years (mean ± SD: 47 ± 12.5) were still healthy. Nonetheless, the 13-year-old boy (GN637) has not undergone colonoscopy.

Referring to patient GN027, who vaguely recalled being a cousin of the main family, the MSH2 sequencing showed that he inherits MSH2 c.1237C > T from his carrier mother GN028 (49 years old). The PV is likely from his maternal grandmother who developed CRC at the age of 59. She also passed this PV to his son GN498 who developed rectal cancer at the age of 52.

Penetrance rate analysis and cumulative risk assessment

Among 25 MSH2 carriers, 10 probands and 15 family members, 14 had been diagnosed with either colorectal cancer (CRC), endometrial cancer (EC), or Ampulla of Vater cancer. While two of them have had polyps, the other nine members were still healthy at the time of surveillance. Therefore, the penetrance rate was simply calculated to be 56%.

Cumulative risk was assessed by stratifying individuals into age groups and calculating the progressive probability of developing CRC or EC over time. The results show that the cumulative risk of CRC or EC increases steadily with age, reflecting the age-related nature of cancer incidence. The cumulative risk analysis revealed a progressive increase in the probability of developing CRC or EC with age, reaching 30% by age 20–29, 63% by age 40–49, and 100% by age 60–69. This suggests that nearly all individuals who developed CRC or EC in the dataset were diagnosed by their late 60s. Additionally, the analysis showed that females had a higher cumulative risk than males. The cumulative risk in females reached 100% by age 50–59 years while males reached 100% cumulative risk at 60–69 years, indicating a later onset compared to females.

Economic analysis

Following standard clinical guidelines, ten probands and their 56 family members would require annual endoscopic surveillance. Based on the Familial Adenomatous Polyposis (FAP) model (34 endoscopies for seven family members), we estimated an average of five endoscopies per person by age 25 [39]. Assuming each undergoes five endoscopies at 2,300 THB per session, the total endoscopic surveillance costs 759,000 THB (approximately 23,000 USD). The total surveillance expenditure would be substantial. Genetic testing costs include 490,000 THB (49,000 THB/proband) for ten probands and 11,200 THB (200 THB/person) for targeted mutation testing across 56 family members, resulting in a total cost of 501,200 THB. The analysis demonstrates that genetic testing achieves a 34% cost reduction (257,800 THB) compared to endoscopy-based surveillance. Furthermore, genetic testing generated indirect benefits through early detection and prevention. Avoided late-stage CRC treatment costs included early polyp removal (two cases), early-stage CRC treatment (one case) and late-stage CRC treatment (two cases). The polyp removal cost varies depending on the polyp size and treatment technique. The average cost of polyp removal is 20,000 THB. Early-stage treatment costs approximately 400,000 THB/person. Late-stage CRC treatment costs an average of 500,000 THB/person (minimum 1,000,000 THB for two cases, including surgery, chemotherapy, and hospitalization) [40]. Combined, these indirect benefits total 1,440,000 THB (43,600 USD).

It is unnecessary to discount all benefits and costs at the present value because they are assumed to occur in the same period. Therefore, the BCR equals 3.39, calculated by dividing the sum of direct (257,800 THB) and indirect benefits (1,440,000 THB) by genetic testing costs (501,200 THB). This ratio indicates that each 1 THB invested in genetic testing generates 3.39 THB in economic returns, thus demonstrating high cost-effectiveness. Additional indirect benefits that have not been monetized or incorporated into the BCR calculation, such as recovery from CRC, improving the quality of life of patients and their families.