Authors: Attila Patócs, Petra Nagy, János Papp, Anikó Bozsik, Bálint Antal, Vince Kornél Grolmusz, Tímea Pócza, Henriett Butz
Categories: Clinical Research Article, NGS, diagnostic, genetic testing, germline, hereditary endocrine tumor, multigene panel, AcademicSubjects/MED00250
Source: The Journal of Clinical Endocrinology and Metabolism
Heterogenous clinical manifestations, overlapping phenotypes, and complex genetic backgrounds are common in patients with endocrine tumors. There are no comprehensive recommendations for genetic testing and counseling of these patients compared to other hereditary cancer syndromes. The application of multigene panel testing is common in clinical genetic laboratories, but their performance for patients with endocrine tumors has not been assessed.
As a national reference center, we prospectively tested the diagnostic utility and cost-efficiency of a multigene panel covering 113 genes representing genetic susceptibility for solid tumors; 1279 patients (including 96 cases with endocrine tumors) were evaluated between October 2021 and December 2022 who were suspected to have hereditary tumor syndromes.
The analytical performance of the hereditary cancer panel was suitable for diagnostic testing. Clinical diagnosis was confirmed in 24% (23/96); incidental findings in genes not associated with the patient's phenotype were identified in 5% (5/96). A further 7% of pathogenic/likely pathogenic variants were detected in genes with potential genetic susceptibility roles but currently no clear clinical consequence. Cost-benefit analysis showed that the application of a more comprehensive gene panel in a diagnostic laboratory yielded a shorter turnaround time and provided additional genetic results with the same cost and workload.
Using comprehensive multigene panel results in faster turnaround time and cost-efficiently identifies genetic alterations in hereditary endocrine tumor syndromes. Incidentally identified variants in patients with poor prognoses may serve as a potential therapeutic target in tumors where therapeutic possibilities are limited.
Keywords: multigene panel, NGS, hereditary endocrine tumor, germline, diagnostic, genetic testing
Among all tumor types, the genetic predisposition is the highest in endocrine-related diseases; still, there are no recommendations for comprehensive and consolidated genetic testing and genetic counseling compared to other hereditary cancer syndromes, such as hereditary breast and ovarian cancer (1-3) Nevertheless, recently, Brock et al (4). summarized all the available management guidelines in hereditary endocrine neoplasia syndromes.
From both the clinical geneticist and clinical laboratory point of view, it is important that clinical manifestations can overlap in some tumor types (eg, parathyroid adenoma/hyperplasia, pituitary neuroendocrine tumors, and pheochromocytoma/paraganglioma), when the suspicion of more than 1 clinical condition can be raised, especially in young patients when the disease has not yet become full-blown. Hence, a multigene approach in molecular genetic testing is highly relevant for patients with endocrine-related tumors.
As a primary testing method, next-generation sequencing (NGS) based technologies are now routinely used in molecular genetic testing laboratories due to their increased cost-benefit value (4-6)
While the techniques are becoming more and more advanced, there are still technical standards that should be followed to provide reliable results (7, 8). This has a great impact as in many cases clinical laboratories use ‘research-use-only or “in-house” solutions for the lack of in vitro diagnostics qualified (IVD) test kits in germline genetics. Besides technical validation, there are ongoing challenges in variant interpretation with the continuously evolving information and knowledge regarding the classification of germline variants (9-12).
In high-throughput designs, only variants of genes with established associations with the disease are recommended to report due to ethical considerations (13). However, in multigene panel kits, manufacturers include other genes as well that lack robust data related to disease association or management guidelines (4). Diagnostic and research settings should be separated in the clinical diagnostic laboratory participating in both activities. Variants identified in numerous nondisease-associated genes present an additional reporting incidental findings. The American College of Medical Genetics and Genomics (ACMG) provides recommendations for reporting these secondary, unexpected findings, but, most ethically, the necessity of reporting should be clarified during pretest genetic counseling led by a clinical geneticist or human geneticist (14).
Recent recommendations stress the need for confirmation of results for all sequence variants classified as pathogenic or likely pathogenic using another method, eg, retesting the same or an independent sample and, if it is available, testing first-degree relatives (15). In this way, sample switch could also be avoided in addition to providing technical validation. Continued updating of variant interpretation is also encouraged by experts (9, 15), which allows variant reevaluation from time to time.
Before implementing NGS-based methods in a clinical laboratory, first, appropriate validation studies are needed to determine the performance characteristics of the test, and second, clear interpretation and reporting principles should be declared following the actual guidelines (9, 12, 16, 17).
While in many conditions, NGS-based panel performance has been tested and reported (6, 18-20), in endocrine-related tumor syndromes there is scarce information available (21).
The clinical genetic diagnosis of endocrine tumor syndromes is intricate due to several (1) the presence of overlapping clinical features that may suggest more than one genetic syndrome; (2) the recommendation to test multiple genes for certain tumor types (Fig. 1); (3) the rarity of these syndromes, necessitating time to collect sufficient samples; and (4) the common use of research-use-only tests in clinical molecular genetic testing laboratories, requiring evaluation and optimization for diagnostic purposes to ensure quality-assured results.
Therefore, as a national comprehensive cancer center for molecular genetic testing for patients suspected of hereditary cancer syndrome, we aimed to implement a comprehensive, 113-gene panel designed for hereditary tumor predisposition. The gene set was developed by the manufacturer together with experts on genetic hereditary tumor predisposition. We aimed to (1) assess a hereditary cancer gene panel's clinical (diagnostic) utility; (2) reduce the turnaround-time of the genetic diagnosis; and (3) validate the performance of a frequently used, research-use-only hereditary cancer panel in the clinical diagnostic setting. The primary genetic test result aims to confirm the genetic basis of the clinical diagnosis and facilitate the implementation of gene-specific surveillance protocols. Identifying secondary findings is crucial for preventing the development of additional tumors and may also reveal potential targets for therapeutic interventions.
In this study, we investigated 1279 consecutive Hungarian patients referred for genetic testing by clinical geneticists due to suspicion of hereditary predisposition to tumor development to the Department of Molecular Genetics of the National Institute of Oncology, Comprehensive Cancer Center, Budapest, Hungary between October 2021 and December 2022. Of these, 96 patients had endocrine-related diseases (Table 1). Considering all manifestations of syndromes associated with endocrine tumors/conditions, patients with some nonendocrine tumors, such as neurofibromas, hemangioblastomas, leiomyomas, or clear cell renal cell carcinomas, have been also included. Indeed, NF1 disease-causing variants can be related to pheochromocytoma-paraganglioma (PPGL) or adrenocortical cancer (ACC), FH to the development of PPGL, or VHL to clear cell renal cell carcinoma. In addition, leiomyomas can be developed in both multiple endocrine neoplasia type 1 (MEN1) and fumarate hydratase tumor predisposition syndrome. Furthermore, penetrance also has to be considered, as not all manifestations are present in the patient at the time of the genetic test.
We followed the national guideline published in 2020 by the Board of Clinical Geneticists regarding the criteria for germline testing of patients with cancer (http://www.hbcs.hu/uploads/jogszabaly/3278/fajlok/2020_EuK_20_szam_EMMI_szakmai_iranyelv_2.pdf; accessed on June 26, 2023) and international guidelines summarized by Brock et al 2020 (4). Briefly, patients with rare or specific tumor types (eg, medullary type of thyroid cancer), tumors where a high rate of positive genetic results can be expected (eg, pheochromocytoma), bilateral-multiplex appearance (occurrence of multiple tumors within the same individual), and young age of onset have been particularly referred for a molecular genetic test.
In all cases, clinical genetic counseling was done according to the Hungarian legal and ethical regulations before molecular genetic testing in our department. The study was approved by the Institutional Ethical Board and the Research and Ethics Committee of the Hungarian Health Science Council (ETT-TUKEB 53720-4/2019/EÜIG). After genetic counseling, all included patients gave written informed consent for the genetic testing. According to the national regulations, genetic results were returned to the patient in the frame of posttest counseling and to the referring clinician. All patients were queried regarding their preference for including not only variants of suspected disease-associated genes but also secondary findings in the genetic test report. This process is necessary to avoid unethical approaches such as undesired testing and placing an unsolicited burden of tumor risk on patients.
Germline genetic variants were analyzed using total DNA extracted from peripheral blood using a Gentra Puregene Blood Kit (cat no. 158389, Qiagen, Hilden, Germany) following the manufacturer's instructions.
Multigene panel testing was carried out using the TruSight Hereditary Cancer Panel (#20029551, Illumina, San Diego, CA, USA) library preparation kit, including 10,341 probes that target 113 genes related to cancer predisposition. The gene set was selected by Illumina following input and feedback from experts on genetic hereditary tumor predisposition. NGS was run on Illumina MiSeq or NextSeq 550Dx Instrument with MiSeq Reagent Kit v3 (600 cycles) (#MS-102-2002, Illumina) or NextSeq 500/550 High Output Kit v2.5 (300 cycles) (#20024908 #MS-102-2002, Illumina). Data analysis was performed using the Illumina Dragen Germline pipeline (Dragen version 4.0.3, Illumina) to uncover sequence variants, copy number alterations, and potential structural variants. GRCh37 genome build and NCBI MANE Select transcripts were used as reference sequences.
Variants were classified following the ACMG guideline for variant interpretation (9, 14). Variants were also cross-checked in the NCBI ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/; accessed between February 9, 2021, and February 28, 2023) and Franklin (https://franklin.genoox.com/clinical-db/home; accessed between February 9, 2021, and February 28, 2023) databases. Variant interpretation and cross-referencing in different databases were done between September 2, 2021, and December 31, 2022. Secondary findings were evaluated according to Miller et al (14). During the interpretation of test results, a multidisciplinary team including clinical and molecular geneticists and experts in laboratory diagnostics, oncology, and endocrinology was involved.
During the whole process, we followed the quality control and quality assurance recommendations by ACMG clinical laboratory standards for NGS (F.4.) (7). Based on this guideline in NGS methods, “similar to Sanger-based sequencing, positive controls do not need to be tested concurrently with routine clinical tests.” Instead, general quality requirements are formulated regarding DNA samples and processing, general quality, and bioinformatics. Accordingly, in our laboratory, we ensured the quality of the following yield and quality of DNA samples, yield and quality of cDNA library, initial DNA fragmentation, error rates during the sequencing run, and postrun/preanalysis assessment of the read quality. Including all these, the following parameters have been assessed during bioinformatics in our routine quality of fastq generation, mapping and variant calling, unique read/base enrichment, number of total and unique aligned reads, percent of duplicate aligned reads, percent Q30 bases, percent aligned bases, percent target coverage at 1X-10X-15X-20X-50X, percent callability, fragment length parameters (min-max-median-SD, number/Het-Hom ratio/Ts-Tv ratio of single nucleotide variants; these parameters are included in the multiqc reports generated by the software for each sample in Pdf and Html formats). The built-in Dragen workflow applies default quality cut-offs and includes all the mentioned parameters in multiqc reports generated for each sample in both Pdf and Html formats. In case of any suboptimal performance or reads/samples failing to meet the designated quality cut-offs, they were excluded and measurement was repeated to ensure high-quality results.
All germline pathogenic/likely pathogenic (P/LP) variants and variants of unknown significance (VUSs) in endocrine-related and clinically actionable genes were validated by the gold-standard bidirectional Sanger sequencing on an independent blood sample. PCR products were purified by ExoSAP-IT™ reagents (Thermo Fisher Scientific, Waltham, MA, USA), and then purified amplicons were sequenced bidirectionally on an ABI3500 Genetic Analyzer (Applied Biosystems, Thermo Fisher Scientific) using a BigDye™ Terminator v.1.1 kit (Thermo Fisher Scientific).
All copy number alterations were validated using the multiplex ligation-dependent probe amplification method using SDHA-P429 (research use only), SDHB/SDHC/SDHD/SDHAF1/SDHAF2-P226 (IVD in Europe), PTEN-P225 (IVD in Europe), VHL-P016 (IVD in Europe), PALB2-P260 (IVD in Europe), BRCA1-P002 (IVD in Europe), and BRCA2-P045 (IVD in Europe) probe sets (MRC-Holland, Amsterdam, the Netherlands).
Genotype and proband characterization proportions and 95% confidence intervals by a modified Wald method were calculated using GraphPad QuickCalcs (https://www.graphpad.com/quickcalcs/confInterval1/; accessed between March 1-28, 2023). For statistical analysis, 2 × 2 contingency tables were applied and P-values were calculated by Fisher's exact test. P-values were considered statistically significant at <.05. For comparison of the age of onsets, a T-test and Mann–Whitney U test were applied depending on data distribution evaluated by the Shapiro–Wilks test.
Sequence and copy number analysis of 113 tumor predisposition genes were performed on germline DNA samples of 1279 cancer patients including 96 cases with endocrine-related tumors. The age of onset of these cases was between 6 months and 83 years (Fig. 1A and 1B).
Phenotype-related genes were associated with the panel genes based on relevant recommendations (4, 22-30) Among the 113 tumor predisposition genes, 26 are related to tumors or conditions of endocrine syndromes (Fig. 1C).
We analyzed the tested genes in 3 groups based on their clinical (1) phenotype-related (endocrine condition-related) genes, (2) clinically actionable but not related to the actual clinical condition [so-called secondary (incidental) findings (defined by ACMG and National Comprehensive Cancer Network (NCCN) guidelines], and (3) genes with potential tumor predisposition role based on literature data but lacking clinical evidence regarding their actionability.
The mean coverage of phenotype-related (endocrine-related cancer) genes was 242 (min: 37 and 630). The less covered genes by this panel were PTEN and NF1, in which 0.6% and 0.5% of genes were not reliably detected (covered <10 reads). Using 10 reads as a cut-off, among all endocrine tumor-related genes the regions of interest (coding exons ± 30 bp) were covered in 99.79%.
Among cancer-related but nonphenotype-associated genes determined by ACMG, the average coverage was 225 (min: 28 and 586). Among ACMG cancer-related genes and NCCN-defined clinically actionable genes [(14), https://www.nccn.org/guidelines/category_2; accessed on June 26, 2023] WT1 and RB1 genes had the lowest coverage, affecting 0.7% and 0.6% of their region of interests, respectively.
Regarding the sequence variants, we validated 48 fragments (>14 000 bp) by Sanger sequencing. The technical sensitivity and specificity of the sequence variant detection was 100% compared to Sanger sequencing as each detected base was concordant between the 2 methods. However, due to a few not properly covered bases the diagnostic sensitivity for the endocrine-related gene panel was 99.8%, where the sensitivity corresponds to the mean coverage of depth for endocrine genes. Regarding copy number variations (CNVs), we performed a comparison between NGS and gold-standard CEIVD multiple ligation or probe amplification in 138 cases. Our NGS CNV method had 100% sensitivity, 41% specificity, 28% positive predictive value, and 100% negative predictive value.
Altogether 48 variants (P/LP) and VUSs were detected in phenotype-associated genes in patients with endocrine conditions (Table 2). In 23 patients (23/96; 24%) the disease-causing genetic alteration was identified and confirmed the clinical diagnosis, while in 7 patients, VUSs were detected (Fig. 2A). In clinically actionable genes recommended by ACMG and NCCN guidelines (14) (https://www.nccn.org/guidelines/category_2; accessed on June 26, 2023) P/LP variants and VUS were identified in 9 and 15 patients, respectively. Among genes that may also predispose individuals to hereditary tumor syndromes but have no clinical recommendations regarding their actionability, pathogenic/likely pathogenic (P/LP) variants and variants of uncertain significance (VUS) were detected in 7 and 52 cases, respectively (Fig. 2A).
Figure 2. (A**)** Diagnostic yield of phenotype-related, clinically actionable, and other tumor susceptibility gene sets. (B) Age of tumor onset according to identified genetic variant classes. (C) Diagnostic rate according to gene sets, variant class, and phenotype.Abbreviations: ACC, adrenocortical cancer; FHH, familial hypocalciuric hypercalcemia; FIPA, familial isolated pituitary adenoma; HLRCC, hereditary leiomyomas and renal cell cancer; MEN1, multiple endocrine neoplasia type 1; MEN2, multiple endocrine neoplasia type 2; NF1, neurofibromatosis type 1; PHTS, PTEN hamartoma tumor syndrome; P/LP, pathogenic/likely pathogenic variant; PPGL, pheochromocytoma-paraganglioma; VHL, von Hippel-Lindau syndrome; VUS, variant of uncertain significance.
We did not find any differences in age of first tumor onset between patients with P/LP variants compared to those where the genetic background was not identified (Fig. 2B). However, by analyzing syndrome-related manifestations in MEN1-suspected patients, expectedly, patients with more than 1 clinical manifestation the P/LP variant detection rate was significantly higher (P = .015; odds 20.25; 95% confidence 2.432 to 143.7) compared to patients having only 1 manifestation.
Among the investigated 3 gene groups, the detection rate was the highest in endocrine-related genes (Fig. 2C). The diagnostic yield according to the suspected diagnosis on referral was the highest in neurofibromatosis type 1, PPGL, multiple endocrine neoplasia type 2, and adrenocortical cancer (Fig. 2C).
As some manifestations can be present as part of more than 1 hereditary tumor syndrome (eg, pheochromocytoma, parathyroid adenoma/hyperplasia, or pituitary neuroendocrine tumors), the detection rate was investigated according to tumor types as well (Fig. 3). The highest diagnostic rates were observed in neurofibromas, multiplex uterine leiomyomata, neuroendocrine tumors, and PPGLs at 60%, 50%, 40%, and 34%, respectively.
Figure 3. The detection rate of variants with pathogenic/likely pathogenic and uncertain significance among different tumor types.Abbreviations: ACC, adrenocortical cancer; LM, leiomyomatosis; MTC, medullary thyroid carcinoma; NET, neuroendocrine tumor; NF, neurofibroma; PitNET, pituitary neuroendocrine tumor; P/LP, pathogenic/likely pathogenic variant; PPGL, pheochromocytoma-paraganglioma; PTH, parathyroid adenoma/hyperplasia; THY, thymoma; VUS, variant of uncertain significance.
In 3 cases heterozygote P variants in the MUTYH gene were identified; however, these variants in heterozygote form are not reportable according to ACMG recommendations due to their autosomal recessive inheritance. Either these patients or their families developed any MUTYH-related malignancies.
In 7 cases, additional P/LP variants were found in the third gene group (genes without clinical relevance) (7%, 7/96), which should be treated as research data instead of a clinical diagnostic test result.
While most somatic alterations give the basis of targeted therapies, there are some options regarding specific treatment among germline genetic variants as well (Tables 3 and 4).
Interestingly, among ACC patients, 1 P/LP variant in MSH6 and 1 P/LP in NF1 were identified. One other patient had a variant of uncertain significance in the NF1 gene. The patient carrying an MSH6 variant had no manifestation in association with Lynch syndrome; however, this variant may raise the potential sensitivity of this particular tumor for immunotherapy when other therapeutic options fail. Similarly, NF1-associated ACC with P/LP variants may represent a potential additional therapeutical option for cancer with such a dismal prognosis.
In 4 patients, ATM P/LP variants were detected. ATM is implicated in DNA damage response and may respond to PARP inhibition even if the tumors themselves are outside of the manifestations of classic ATM-related cancer types.
In clinically actionable genes, P/LP variants were found in 7 patients (7/96; 7%) (Fig. 2A, Table 3). However, VUSs were observed more frequently among these genes compared to endocrine tumor-associated genes; between 10% to 50% were related to the different tumor types (Fig. 3).
Among other potential tumor susceptibility genes but not included in the ACMG gene list, P/LP variants were observed in another 7 cases (Fig. 2A).
Only 2 patients harbored 2 pathogenic 1 with multiple endocrine neoplasia type 1 and 1 with hyperparathyroidism jaw tumor syndrome (Fig. 4). In these patients, the causal pathogenic variants were associated with a relatively common CHEK2 genetic variant (CHEK2: c.470T > C) as an incidental/secondary finding. Additionally, P/LP variants in clinically actionable genes were detected in 7 patients (7/96) in whom no endocrine-related genetic background was identified (Fig. 4). The most frequent P/LP variants were detected in ATM and CHEK2 genes; however, an incidental APC P variant was also revealed. Among these secondary findings in all CHEK2 cases the frequent (CHEK2: c.470T > C) variant was identified. Despite being frequently classified as P/LP, this variant is lately considered a low-penetrance variant, and it represents rather a genetic modifier instead of being a classic tumor susceptibility variant (32). If we do not consider this variant as P/LP, then no double P/LP double heterozygosity was identified in our patients with endocrine tumors, and the number of patients with clinically actionable genetic variants dropped to 5% (5/96). Among real incidental findings, disease-causing variants in APC and ATM genes were revealed in 1 and 4 patients, respectively (Fig. 4). In these patients, no colon polyposis, colon cancer, or breast or ovarian cancer were detected.
Figure 4. Clinically actionable incidental/secondary finding frequencies in association with phenotype-related genetic variants.Abbreviations: B, benign/likely benign variant; P/LP, pathogenic/likely pathogenic variant; VUS, variant of uncertain significance.
During our prospective study period between October 2021 and December 2022, sequencing kits were selected by sample availability to achieve coverage greater than 30x. During this period, we investigated 1279 patients. Pretest genetic counseling was given to all patients, and sampling for genetic testing was performed after obtaining their consent. Due to these restrictions, we analyzed 48 samples in 1 sequencing run every second week. Our calculation is based on this setup. The cost of reagent for 1 sample was ∼ 196 USD (180 euro) with a turnaround time of 8 weeks including the posttest genetic counseling. The posttest genetic counseling is challenging due to the availability of medical geneticists.
The cost of a custom endocrine tumor-related gene library preparation panel based on our previous experiences with custom-designed panels (6, 21) was around the same price, as only primers/probes are different and can be used with the same library preparation kits. Using a targeted endocrine panel would have necessitated 6 months (24 weeks) to collect 48 samples, given our study period and experience indicating an average of approximately 96 endocrine patients tested per year. This approach would have cost the same 196 USD (180 euros). However, maintaining the same turnaround time of 8 weeks would inflate the price by 12-fold, totaling 2350 USD.
The time required for bioinformatic analysis on the Dragen cloud server is independent of the sample number (for 1-96 samples running on parallel nodes); to it takes 45-47 minutes, and it does not significantly differ in the case of 26-gene or 113-gene panels. After filtering (by minor allele frequency, ClinVar significance, and ACMG classification), an average of 1 to 5 variants/sample is needed to be evaluated manually by our expert molecular biologist using specific databases (eg, VHL, SDHx UMD, or BRCA Exchange) followed by validation performed by a physician qualified in molecular genetic diagnostics.
Therefore, at the number of 113 genes, neither the time for variant interpretation nor the cost are significantly different compared to a targeted endocrine tumor gene panel (covering 26 genes).
During this 1-year prospective study, in our national reference center for germline genetic testing of hereditary cancer, we tested the laboratory performance and potential clinical diagnostic utility of the TruSight hereditary cancer panel on 1279 Hungarian patients referred for genetic testing by clinical geneticists due to a suspicion of hereditary predisposition to cancer. Among these patients, 96 harbored endocrine-related tumor types who underwent the same clinical genetic and molecular genetic workflow used for patients with potential hereditary tumor conditions. The 96 cases present the heterogenous disease spectrum; they belong to 10 disease groups (Table 1.).
The number of tested cases is in line with the estimated prevalence of these diseases. Some of them are rare/ultrarare (ie, von Hippel-Lindau syndrome, hyperparathyroidism-jaw tumor syndrome), while others [ie, pituitary tumors, neurofibromatosis type 1 (NF1) or MEN1] are more prevalent conditions. However, all of them are much more uncommon than hereditary breast and ovarian cancer or hereditary nonpolyposis colon cancer (22, 26, 31, 33-35)
The turnaround time for molecular genetic tests performed for endocrine-related cancers depends on multiple factors including the availability of clinical genetic consultation, the infrastructure and capacity of the molecular genetic laboratory, and the used workflow, but, in general, it is between 3 weeks and 6 months. Many laboratories perform only disease-specific testing with conventional methods. The workload and cost of these individual tests are comparable with the cost of NGS-based methods. We introduced a comprehensive hereditary cancer panel consisting of 113 genes, which includes testing for genes associated with both endocrine and non-endocrine solid tumors. This change in our workflow led to a shorter turnaround time and cost-effectively yielded more comprehensive results compared to single gene testing.
We found that the analytical performance of this hereditary cancer panel in our experimental settings was suitable for diagnostic testing. We set the CNV detection bioinformatic pipeline to be very sensitive, resulting in several false-positive cases; therefore, we performed 3 times more validation for these types of variants than the real positive cases. This ensured patient safety, allowing us not to lose any patients with P/LP CNV (0 false-negative cases) yielding 100% negative predictive value.
We performed cost-benefit and turnaround-time analysis. We found that at the number of 113 genes, neither the time for variant interpretation nor the cost are significantly different compared to a targeted endocrine tumor gene panel (covering 26 genes). However, on the other hand, it significantly decreased the turnaround time for test report issuing.
The diagnostic yield of this multigene panel was also assessed. We divided the analyzed genes into 3 (1) phenotype-related genes, (2) incidental/secondary findings, and (3) genes investigated only in research settings without clinical relevance.
The used hereditary cancer panel yielded high-diagnostic utility in patients with endocrine-related tumors (23/96; 24%) that exceeds the 10% set as a cut-off for the genetic testing indication (36). Naturally, this rate was different according to tumor types; however, this panel resulted in similar detection rates reported earlier [eg, 34% detection rate in PPGL in our cohort was concordant with Neumann et al (25, 26)]. ACC is rare, and therefore data regarding its hereditary predisposition are less reliable (23). Apart from childhood occurrence where the detection rate of germline TP53 mutations is high (∼80%), hereditary genetic background of adult-onset ACC is infrequent (∼5%) (23, 27, 37). The most common hereditary syndromes associated with ACC are Li-Fraumeni syndrome (2–4%), Lynch syndrome (3%), and MEN1 (1–2%). Occasionally familial adenomatous polyposis or Beckwith–Wiedemann syndrome are also described with ACC (23, 27). In our settings, a disease-causing variant was detected in a quarter of patients, which is higher compared to the literature data, and it is probably due to the small sample number. Indeed, the LP variant in the MSH6 gene was detected in 1 of our patients, and 1 NF1 P/LP variant was also identified in an ACC patient. To date, only a handful of ACC cases have been reported in patients with NF1 (38). However, in some cases, the causal effect was proved with loss of heterozygosity at NF1 locus and germline NF1 disease-causing variant (38), which may qualify NF1 as a novel susceptibility gene for ACC.
Our data regarding the identification of the causative genetic variants in these syndromes are in line with earlier reports and highlight that in some entities (ie, NF1) a strong genetic background (60% of cases with the clinical diagnosis of NF1) was confirmed, while in others (ie, MEN1) a smaller fraction (15% of cases) of clinical diagnosis were confirmed. Our earlier data demonstrated that in MEN1 the phenocopy is high (39), which may be the reason for this result.
Besides the disease-associated variants, a high rate (7/96; 7%) of incidental findings was observed in patients with endocrine-related tumors, mostly in genes associated with hereditary breast and ovarian cancer syndrome (ATM and CHEK2 genes). Although both of these are moderate penetrance genes, they still have clinical relevance as P/LP variant carriers benefit from recommended surveillance protocols (https://www.nccn.org/guidelines/category_2; accessed on June 26, 2023).
In addition, variant interpretation should be carefully performed as in many cases it may lead to uncertainties. For instance, the CHEK2: c.470T > C p.(Ile157Thr) variant was identified as the most frequent secondary finding in endocrine tumor patients (4/98; 4%). The interpretation of this particular variant is conflicting (ClinVar ID: 5591, Accession: VCV000005591.80, accessed on June 26, 2023). While this variant previously was classified as an LP genetic alteration, its high prevalence (2%) in the general population suggests that this should be treated rather as a risk allele for tumor predisposition. This was also strengthened by a retrospective cohort study investigating phenotypes of 3783 participants with CHEK2 pathogenic variants (32). While P/LP variants of CHEK2 mainly predispose to hereditary breast cancer and slightly increase the risk of kidney and thyroid cancer, the authors demonstrated that the CHEK2:c.470T > C p.(Ile157Thr) was not associated with non-breast cancers (32). Therefore, the interpretation of this variant and genetic counseling should consider these data to avoid unnecessary burdens on patients. Without this particularly common CHEK2 variant, only ATM and APC can be considered incidental, clinically actionable variants (7%).
While targeted therapy is usually based on somatic variant detection, in some cases germline genetics can also serve as the basis for these novel therapeutical options (40). Genes participating in DNA damage response or implicated in immunotherapy are frequently identified as secondary findings. While PARP inhibitors (targeting DNA damage response) and immune checkpoint inhibitors are not the first-line treatment options in endocrine-related cancer types, they may be effective, even as off-label applications, in cases when other therapeutical options run out in patients with P/LP variants in homologue recombination genes in the context of a clinical trial (40, 41).
Based on our findings, the role of other potential susceptibility genes can be raised in an additional 7% (7/96) of patients (in a research setting), where the clinical conclusions have not been revealed yet. However, these data may serve as additional data shortly in research settings regarding the investigations of their potential cancer-susceptibility genes, and these data should not be reported to avoid raising false hope or causing unnecessary stress. Thus, the distinction between clinical and research testing is essential, and counseling and consent have to be managed accordingly.
Clinical genetic counseling is important to determine the indication of genetic testing, estimate tumor risk, recommend surveillance and preventive measures, and offer family screening (4). This has an additional significance when overlapping phenotypes are seen or secondary/incidental findings are detected.
Multigene panels have many advantages in the clinical testing of patients with endocrine tumors. They yield a high diagnostic rate, are fast, and have a relatively reasonable cost. Due to the clinical significance of a hereditary genetic test result, the technical validation of the research-use-only assays routinely used in clinical laboratories should be precisely performed by molecular genetic laboratories.
Regarding high-risk genes, low penetrance cases can be identified only by high-throughput testing. In the case of moderate-risk genes, multigene panel testing also provides an effective tool for detection, and, due to the gene-specific management protocols (surveillance, family screening, genetic counseling), they also have an impact on clinical practice.
In addition to technical issues, variant interpretation and genetic counseling also represent a challenge. It is essential to follow the international recommendations in reporting principles (which genes, which variants) to gain the most benefit for patients and avoid unnecessary stress on them. Accurate variant interpretation in the right context is essential, and this can be obtained through a multidisciplinary molecular tumor board.
We thank all clinicians and patients and their relatives participating in our study.
Attila Patócs, HUN-REN Hereditary Tumors Research Group, Hungarian Research Network, H-1089 Budapest, Hungary; Department of Laboratory Medicine, Semmelweis University, H-1089 Budapest, Hungary; Department of Molecular Genetics and the National Tumor Biology Laboratory, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary.
Petra Nagy, Department of Molecular Genetics and the National Tumor Biology Laboratory, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary.
János Papp, HUN-REN Hereditary Tumors Research Group, Hungarian Research Network, H-1089 Budapest, Hungary; Department of Molecular Genetics and the National Tumor Biology Laboratory, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary.
Anikó Bozsik, HUN-REN Hereditary Tumors Research Group, Hungarian Research Network, H-1089 Budapest, Hungary; Department of Molecular Genetics and the National Tumor Biology Laboratory, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary.
Bálint Antal, Semmelweis University, National Academy of Scientist Education, H-1085 Budapest, Hungary.
Vince Kornél Grolmusz, HUN-REN Hereditary Tumors Research Group, Hungarian Research Network, H-1089 Budapest, Hungary; Department of Molecular Genetics and the National Tumor Biology Laboratory, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary.
Tímea Pócza, Department of Molecular Genetics and the National Tumor Biology Laboratory, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary.
Henriett Butz, HUN-REN Hereditary Tumors Research Group, Hungarian Research Network, H-1089 Budapest, Hungary; Department of Laboratory Medicine, Semmelweis University, H-1089 Budapest, Hungary; Department of Molecular Genetics and the National Tumor Biology Laboratory, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary; Department of Oncology Biobank, National Institute of Oncology, Comprehensive Cancer Center, H-1122 Budapest, Hungary.
This work was supported by the Hungarian Scientific Research Grant of the National Research, Development and Innovation (NRDI) Office Fund of the Ministry of Culture and Innovation under the National Laboratories Program [National Tumor Biology Laboratory (2022-2.1.1-NL-2022-00010)] and the Hungarian Thematic Excellence Program (TKP2021-EGA/TKP2021-NVA/TKP2021-NKTA, MOLORKIV to A.P.) grant agreements with the NRDI. H.B. is supported by NRDI NKFI-FK135065, the New National Excellence Program of the Ministry of Human Capacities (ÚNKP-23-5-SE-4), and the Bolyai Research Fellowship of the Hungarian Academy of Sciences.
A.P.: Conceptualization, Data curation, Writing—original draft, Writing—review & editing, Resources, Funding acquisition, Supervision; P.N., J.P., A.B., VK.G., T.P.: Methodology, Investigation; A.B.: Formal analysis, Investigation; H.B.: Conceptualization, Formal analysis, Data curation, Writing—original draft, Writing—review & editing, Conceptualization, Funding acquisition.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
All relevant data presented in the study is contained within the article. All genetic variants in this study were deposited in the publicly available database LOVD (Leiden Open Variation Database) under Owner 3628. Any further inquiries can be directed to the corresponding author.
All relevant data presented in the study is contained within the article. All genetic variants in this study were deposited in the publicly available database LOVD (Leiden Open Variation Database) under Owner 3628. Any further inquiries can be directed to the corresponding author.