Evaluation of the Accuracy of a Polymerase Chain Reaction–Based Assay for Polymerase Epsilon Mutation Detection in Endometrial Carcinoma

This study enrolled a total of 649 patients who received a diagnosis pathologically of endometrial carcinoma and underwent standard staging surgery at the Women’s Hospital Zhejiang University School of Medicine (Hangzhou, China) between January 2015 and December 2020. Clinicopathologic data were collected from all patients, except 9 (1.4%) who were excluded because of incomplete clinicopathologic data. Among the remaining 640 patients (98.6%), FFPE tissue specimens were resected, with 6 serial sections obtained from each specimen. One slice was used for hematoxylin-eosin staining and pathologic review to identify regions with greater than 30% tumor and less than 20% necrosis. The remaining 5 slices were used for DNA extraction (27 patients [4.2%] were further excluded because of insufficient DNA quality). Thus, 613 specimens (95.8%) were included in the final analysis (Figure 1).

All pathologic diagnoses were reviewed independently by 2 pathologists, who followed the World Health Organization classification of tumors for endometrial carcinoma.²⁰  The study was approved by the ethics committee of the hospital (reference No. IRB-20200363-R).

DNA was extracted from 5 serial slices 10 μm thick for each specimen using a NuClean FFPE DNA kit (CWbio, catalog No. CW 2646) according to the manufacturer’s instructions. The purity and concentration of the DNA extracts were determined using a NanoDrop 2000c spectrophotometer (ThermoFisher Scientific).

A total of 9 nucleotide mutation sites in exons 9 to 14 of the POLE gene were selected for detection based on a literature review and reference to The Cancer Genome Atlas (https://www.cancer.gov/ccg/research/genome-sequencing/tcga), the Clinvar database (http://www.ncbi.nlm.nih.gov/clinvar), and the Cosmic database (https://cancer.sanger.ac.uk/cosmic). The 9 mutation sites were c.857C>G, c.890C>T, c.1100T>C, c.1231G>T/C, c.1270C>G/A, c.1307C>G, c.1331T>A, c.1366G>C, and c.1376C>T, referring to 11 types of nucleotide mutations and corresponding to 10 amino acid variations, namely, P286R, S297F, F367S, V411L, L424V/I, P436R, M444K, A456P, and S459F, respectively (Supplemental Table 1, see Supplemental Digital Content file containing 1 figure and 5 tables, at https://meridian.allenpress.com/aplm in the August 2024 table of contents). Among these, c.1231G>T and c.1231G>C shared the same amino acid variation (V411L). Primers specific to 11 types of nucleotide substitutions were designed, synthesized, and assembled into the Dalton-MIT kit (Supplemental Table 2). The kit contained 6 PCR tubes with the corresponding primer probe combinations. DNA (1.5 µL) was added to each tube for PCR. The PCR procedure was as follows: 37°C for 5 minutes, 95°C for 3 minutes (1 cycle), 95°C for 10 seconds, 53°C for 10 seconds, and 72°C for 35 seconds (40 cycles). Signals of the channels FAM, VIC, and Cy5 for each tube, signals of FAM and VIC for tube C, and signals of VIC for the external control tube F (globin as internal reference gene) were collected in the second stage at 72°C. The ΔCT for each mutation site was calculated as the CT value of a specific site minus the CT value of the external control. Next-generation sequencing of the pathogenic mutations for each corresponding site was used as the standard for comparison. The threshold (ΔCT) value of the Dalton-MIT assay to identify a POLE mutation was determined for each mutation site, and an equal to or less than certain ΔCT value was defined as POLE mutation positive (POLE-mut). Greater than certain ΔCT values was defined as POLE mutation negative (non–POLE-mut) (Supplemental Table 3). All 613 samples were tested using the Dalton-MIT assay.

Among 613 specimens, 384 (62.6%) were selected for next-generation sequencing. Of these, 70 specimens (11.4%) that were defined as POLE-mut by the Dalton-MIT assay were included, and 314 specimens (51.2%) that were classified as non–POLE-mut were randomly assigned by computer at a ratio of approximately 2:1. A total of 19 specimens (4.9%) failed sequencing. Thus, 365 specimens (95.1%) yielded high-quality sequencing data and were included in the final analysis (Figure 1).

Multigene target sequencing was performed using VariantPro Capture Technology by LC Sciences (Hangzhou, China), as described previously.^21–23  The presence of the desired fragments and the purity of the indexed libraries were analyzed on an Agilent 2100 Bioanalyzer (Agilent Technologies) using a High-Sensitivity DNA Analysis Kit (Agilent Technologies). The concentrations of the libraries were measured using a Qubit Fluorometer (ThermoFisher Scientific). Sequencing was performed using the Illumina MiSeq Reagent v2 kit (500 cycles) on the Illumina MiSeq instrument following the manufacturer’s instructions (Illumina). Mutation analysis of exons 9 to 14 was performed using Mutation Surveyor software (SoftGenetics, State College, Pennsylvania) following manual inspection.

Fisher exact test and independent t tests were used to analyze the correlation between POLE mutation status and clinicopathologic data. Nonparametric testing was used for variables not meeting the assumptions of the t test equivalent. Survival analysis was performed by using Kaplan-Meier curves and log-rank tests. All analyses were performed using SPSS version 25.0 (IBM). All P value tests were 2 sided, and a P value <.05 was considered statistically significant.

Of 613 specimens, Dalton-MIT detected 70 POLE-mut specimens (11.4%), all of which were single-nucleotide mutations (Supplemental Table 4). Among the 9 detected mutation sites, the 3 most common were c.857C>G (5.38%), c.1231G>T/C (3.26%), and c.1366G>C (1.14%). The top 3 sites for mutation frequencies in all POLE-mut specimens were c.857C>G (47.14%), c.1231G>T/C (28.57%), and c.1366G>C (10.00%), and their corresponding amino acid changes were P286R, V411L, and A456P, respectively.

The clinicopathologic characteristics of 613 patients are shown in Table 1. Compared with POLE-mut patients, non–POLE-mut patients were at a significantly higher FIGO stage and had more lymph node metastasis (both P < .05). These results suggest that POLE-mut patients present less aggressive clinical and pathologic characteristics than non–POLE-mut patients.

All 613 patients were followed up for a median of 22 months (range, 5–75 months). In total, 26 patients (4.2%) were lost to follow-up. The overall survival (OS) and progression-free survival (PFS) of POLE-mut and non–POLE-mut patients were analyzed. Of the 587 patients (95.8%) who completed follow-up, 22 (3.7%) experienced recurrence, and 14 (2.4%) died during the follow-up period. None of these patients were in the POLE-mut group. POLE-mut patients had higher OS (mean ± SD, 25.88 ± 1.786 versus 23.89 ± 0.565 months) and PFS (mean ± SD, 25.88 ± 1.786 versus 23.37 ± 0.557 months) than non-POLE-mut patients, but the difference was not significant (Figure 2, A and B). In patients with high-grade (G3) endometrioid carcinoma or FIGO stages II to IV, differences were more obvious but remained nonsignificant (Figure 2, C through F). Our results indicate that patients with POLE-mut may have a better prognosis than those with non-POLE-mut.

POLE pathogenic mutations were detected in 64 of 365 specimens (17.5%; Supplemental Table 5). The detected mutation sites were c.857C>G, c.1231G>T/C, c.1366G>C, c.1376C>T, c.1100T>C, c.890C>T, c.1270C> G, c.1331T>A, and c.1307C>G. The 3 most common sites were c.857C>G (45.3%), c.1231G>T/C (29.7%), and c.1366G>C (9.4%), and the top 3 mutation frequencies were c.890C>T (33.4%), c.1331T>A (24.5%), and c.1307C>G (23.5%).

Furthermore, we performed a direct comparison between Dalton-MIT and next-generation sequencing in 365 specimens (59.5%). All 9 mutation sites were detected by both assays, but only 10 of 11 types of nucleotide mutations were detected by next-generation sequencing (Table 2). In 2 samples, POLE-mut was detected by Dalton-MIT but not next-generation sequencing. The POLE-mut was further verified by Sanger sequencing in these 2 samples (Supplemental Figure 1).

POLE mutation detection by Dalton-MIT was compared to next-generation sequencing as the standard. The sensitivity of the Dalton-MIT assay for detecting POLE pathogenic mutations was 100%, the specificity was 99.3%, the Youden index was .993, and the κ value was .981 (P < .001), as shown in Figure 3, A through C. Our data demonstrate good consistency between Dalton-MIT and next-generation sequencing for detecting POLE mutations.

Primers for the most common mutations, including P286R, V411L, S297F, A456P, and S459F, were assembled in the Dalton-MIT kit. POLE mutations were detected in 70 of 613 patients (11.4%), which is consistent with previous reports that POLE mutations are found in 7% to 12% of endometrial cancer cases, as detected by next-generation sequencing.¹⁹  The most common mutation was P286R (47.14%), followed by V411L (28.57%) and A456P (10.00%). Next-generation sequencing was performed on 365 of 613 specimens (59.5%) in order to compare the 2 technologies; only 2 specimens showed discrepant results. The sensitivity of Dalton-MIT for detecting POLE pathogenic mutations was 100%, the specificity was 99.3%, the Youden index was .993, and the κ value was .981 (P < .001).

POLE is a catalytic subunit of DNA POLE and is involved in nuclear DNA repair and replication. Pathogenic mutations in the POLE gene identified a subtype of endometrial cancer with a nonprogression prognosis. Multiple POLE mutation sites are detected by next-generation sequencing, but only those that correlate with favorable prognosis are defined as pathogenic mutations.¹³  The 5 most common mutations (P286R, V411L, S297F, A456P, and S459F) have been confirmed. The pivotal study by The Cancer Genome Atlas found substitutions primarily in P286R and V411L, although S297F, A456P, M444K, and L424I substitutions were also detected.^4,13  León-Castillo et al¹³  developed a scoring system to estimate the pathogenicity of novel POLE mutations based on the presence or absence of genomic alterations associated with known POLE pathogenic mutations. As reported, approximately 7% to 12% of endometrial carcinomas demonstrate pathogenic mutations in the POLE gene. In our study, POLE mutations were detected by the Dalton-MIT kit in 70 of 613 specimens (11.4%), which is consistent with previous reports.

The PCR technique is a fast, simple, and convenient technology with wide applications in disease diagnosis. However, it is limited by its ability to only detect genes with known mutated sites because it can only amplify preselected DNA segments.¹⁴  Kim and colleagues developed a droplet digital PCR assay for detecting 5 types of POLE mutations (P286R, V411L, S297F, A456P, and S459F), which only account for 84% of the 11 pathogenic mutations.²⁴  However, the Dalton-MIT kit, covering a wider range of POLE mutations, has a better accuracy. The Dalton-MIT kit is also more applicable in a clinical setting because of its low cost and convenient handling.

The results of the current study demonstrate that Dalton-MIT detection in patients with endometrial cancer with POLE-mut presented with a better prognosis than those with non–POLE-mut. All of the 22 patients who experienced recurrence (3.7%) and the 14 who died during the follow-up period (2.4%) were non–POLE-mut patients. Survival analysis showed that both the OS and PFS of POLE-mut patients were longer than those of non–POLE-mut patients, even though the difference was not significant. A possible reason for this could be that 90.4% of patients enrolled in this study were classified at an early stage (stage I or II), and therefore had a good prognosis. In the current study, the recurrence rates of these patients were 9.9% (8 of 81) and 7.6% (10 of 131), respectively. This study is limited by its small single-center retrospective design. Future studies should include more cases and expand to multicenter prospective designs in order to validate the POLE mutation detection by the Dalton-MIT kit.

POLE mutations are generally identified by next-generation sequencing or Sanger sequencing. Both methods have certain limitations. Next-generation sequencing is expensive and time-consuming, whereas Sanger sequencing has low sensitivity. PCR is a technology that is easy to operate and is inexpensive. The Dalton-MIT kit was developed to address these limitations and is a convenient and straightforward tool for POLE mutation detection. Using the Dalton-MIT kit to detect POLE mutations can be widely applied for molecular stratification in the clinic to accurately predict prognosis.

The Dalton-MIT assay targeted 9 POLE mutation sites, and POLE mutations were detected in 11.4% of samples from 613 patients with endometrial cancer. Patients with POLE-mut had better clinicopathologic characteristics and prognosis than those with non–POLE-mut. There was a good consistency between PCR and next-generation sequencing for detecting mutations. Our results suggest that Dalton-MIT is a promising alternative assay to detect POLE mutations. It will facilitate the wider application of molecular classification of endometrial cancer in the clinical setting.

We gratefully acknowledge the kind cooperation of all the members in our department.