Abstract
Background/Aim: Targeted next-generation sequencing (NGS) is a well-established technique to detect pathogenic alterations in tumors. Indeed, it is the cornerstone of targeted therapy in precision medicine. We investigated the clinical utility of next-generation sequencing in real-world cases. Patients and Methods: We retrospectively selected six representative cancer cases, wherein targeted NGS played a pivotal role in the diagnosis and treatment of patients. Additionally, we analyzed three cases with rare, unusual pathogenic alterations. Results: Our NGS analysis revealed that four patients had TPR-ROS1, EGFR-RAD51, and NCOA4-RET fusions and MET exon 14 skipping mutation, respectively, which can be treated with targeted therapy. Furthermore, we used NGS as a diagnostic tool to confirm the origin of unknown primary malignant tumors in two cases. Interestingly, NGS also helped us identify the following cases: patients exhibiting BRCA1 and TP53 mutations that exhibited histological and immunohistochemical characteristics consistent with endometrioid carcinoma, patients with high-grade serous carcinoma not possessing a TP53 mutation, and patients with small cell lung cancer with a ERBB2 mutation and displaying no loss of RB1. Conclusion: We recommend targeted NGS for the diagnoses and targeted therapy of cancer patients.
Cancer is a genetic disease that is caused by mutations in genes involved in the cell cycle and in cell signaling, cell growth, proliferation, survival, cancer invasion, and immunity (1, 2). In the era of precision oncology, molecular analysis of patient samples is essential for tumor classification, treatment planning, and prognostic stratification (3). Currently, several pathology labs follow the “one-gene one-test” approach, wherein they perform techniques such as polymerase chain reaction (PCR), Sanger sequencing, and pyrosequencing to detect targetable gene alterations, such as those in EGFR, BRAF, KRAS, ALK, ROS, MET, and PIK3CA. However, all these tests require a substantial amount of template DNA and are only applicable to the hotspot regions of targetable genes (4-7).
Remarkably, next-generation sequencing (NGS), massive parallel sequencing, simultaneously detect hundreds of genes (8, 9). Although whole exome sequencing and whole genome sequencing are widely used for research purposes, they have limited clinical applications owing to their low coverage depth and high cost (10). Thus, targeted NGS is the most widely employed sequencing approach for molecular analysis in clinical practice (11, 12).
Conventionally, cancer is diagnosed by histological evaluation. Indeed, a direct visual inspection of the patient’s sample by light microscopy provides considerable information regarding diagnosis, prognosis, and a treatment plan (13-15). Despite determining the overall pathological landscape of a patient sample, this approach does not reveal the underlying genetic alterations in cancer. Additionally, interobserver discrepancy and “gray zone” are some other limitations of histologically classifying cancers (16-19). Alternatively, immunohistochemistry and molecular techniques, such as PCR and fluorescence in situ hybridization, have also been used as ancillary tests for enhancing diagnostic efficiency and for establishing a treatment plan (20-28). However, targeted NGS can simultaneously identify multiple genetic events and transform patient care in this era of precision medicine (3, 12, 29, 30).
In this study, we retrospectively investigated the cases wherein NGS played a critical role in cancer diagnosis and treatment planning.
Patients and Methods
Study population. Among 125 cancer cases that had been subjected to targeted NGS from 2021 to 2022 in the Department of Pathology at Kyungpook National University Chilgok Hospital, we retrospectively studied nine cases. Four of these cases were selected because NGS played a critical role in finding rare targetable genetic alterations that had been undetected in other molecular methods. On the other hand, we chose two cases where NGS had been used as a diagnostic tool. The remaining three cases that were selected were dilemma cases, as they exhibited significant discrepancies between the NGS, histological, and immunohistochemical data. The clinicopathological data, including age, sex, histological diagnosis, and tumor stage, of the selected cases were retrieved from the electronic medical records of the hospital. The study was conducted in accordance with the guidelines of the Declaration of Helsinki and approved by the Daegu Joint Institutional Review Board (No. DGIRB 2021-12-004-001). The requirement for written informed consent from the patients was waived because of the retrospective nature of the study.
Pathological evaluation. Surgical specimens were fixed in 10% neutral-buffered formalin and embedded in paraffin blocks. The entire tumor specimen was microscopically examined if the tumors were smaller than 3 cm in diameter. For each selected case, at least three representative tumor sections (one section per 1 cm of the tumor) were submitted for pathologic evaluation. These sections were cut into 4-μm thick sections and stained with hematoxylin and eosin to obtain formalin-fixed, paraffin-embedded (FFPE) tissue blocks. Two pathologists experienced in surgical pathology (MK and JYJ) reviewed all these stained sections. Thereafter, the tumors were diagnosed and classified according to the fifth edition of the World Health Organization Classification of Tumors.
Immunohistochemistry. The FFPE sections were deparaffinized and rehydrated with xylene and alcohol. Thereafter, these sections were incubated with antibodies against ALK (rabbit monoclonal, clone D5F3, 1:50; Cell Signaling Technology, Danvers, MA, USA), CD56 (mouse monoclonal, clone CD564, 1:100; Novocastra, Newcastle, UK), chromogranin A (mouse monoclonal, clone DAK-A3, 1:500; Dako, Glostrup, Denmark), cytokeratin AE1/AE3 (mouse monoclonal, clone AE1/AE3, 1:600; Agilent, Santa Clara, CA, USA), ER (mouse monoclonal, clone GF11, 1:750; Novocastra), ki-67 (mouse monoclonal, clone MB-1, 1:150; Dako), p16 (mouse monoclonal, clone E6H4, prediluted; Ventana Medical Systems, Oro Valley, AZ, USA), p53 (mouse monoclonal, clone DO7, 1:300; Novocastra), PAX2 (rabbit monoclonal, clone EP3251, 1:300; Abcam, Cambridge, UK), PAX8 (rabbit polyclonal, 1:50; Cell Marque, Rocklin, CA, USA), PR (mouse monoclonal, clone PGR 312, 1:150; Novocastra), ROS1 (rabbit monoclonal, clone D4D6, prediluted; Cell Signaling Technology), synaptophysin (mouse monoclonal, clone DAK-SYNAP, 1:900; Dako), SMAD4 (mouse monoclonal, clone B-8, 1:50; Santa Cruz Biotechnology, Dallas, TX, USA), TTF-1 (mouse monoclonal, clone 8G7G3/1, 1:150; Dako), vimentin (mouse monoclonal, vim 384, 1:200; Dako), and WT 1 (mouse monoclonal; clone 6F-H2, 1:200; Cell Marque). The sections were chromogenically visualized using an ultraView Universal DAB Detection Kit (Ventana Medical Systems) or EnVision FLEX/HRP (Dako) and were subsequently counterstained with hematoxylin. These stained slides were analyzed by MK and JYJ.
PCR analysis for EGFR and ROS1 fusion. A representative section comprising the largest tumor volume (at least 30%) was subjected to molecular analysis in each case. To detect EGFR mutations, we performed peptide nucleic acid (PNA)-mediated real-time PCR using the PNAClamp EGFR Mutation Detection Kit (Panagene, Daejeon, Korea) or the PANAMutyper EGFR Kit (Panagene) according to the manufacturer’s instructions. Additionally, to detect ROS1 rearrangements, we performed real-time PCR using the ROS1 Gene Fusions Detection Kit (AmoyDx, Xiamen, PR China).
Targeted NGS. We detected cancer-related genetic alterations by performing targeted NGS using a customized cancer panel (ONCOaccuPanel, NGeneBio, Seoul, Republic of Korea) as per the manufacturer’s instructions. The ONCOaccuPanel was designed to identify potential single nucleotide variants (SNVs) in 323 cancer-related genes. This panel also identified insertions/deletions (indels) and copy number variations (CNVs) and potential fusion variants of six genes. To conduct targeted NGS, we first extracted DNA from the FFPE sections using the QIAGEN AllPrep DNA/RNA FFPE Kit (Qiagen, Hilden, Germany). Following this, 100 ng of the extracted DNA was subjected to hybridization capture-based target enrichment and paired-end sequencing (2×150 bp) performed using a MiSeq sequencer (Illumina, San Diego, CA, USA) according to the manufacturer’s instructions. Furthermore, Burrows-Wheeler Aligner and Genome Analysis Tool Kit were used to analyze the SNVs, indels, and CNVs in the samples. Variants with a total depth of at least 100× and allele frequencies of at least 3% were included in the variant analysis; STAR-Fusion and FusionCatcher were used to evaluate the fusion variants. Thereafter, we conducted variant interpretation and annotation based on the recommendations of the Association for Molecular Pathology, the American Society of Clinical Oncology, and the College of American Pathologists (31).
Results
Patient cohort and the genomic landscape of the selected cases. Table I presents the clinicopathological characteristics of nine cancer cases evaluated in this study, designated as cases 1-9. The median age of nine patients, four men and five women, was found to be 63 years (range=45-73 years). Furthermore, we identified four advanced lung adeno-carcinoma cases, two ovarian cancer cases and one anaplastic thyroid carcinoma case. However, the remaining two cases were malignancies of unknown origins. Additionally, all patients had stage IV cancer except for the case 7 patient who had stage 1A cancer.
Next, we performed targeted NGS at a mean depth of 373.3× (range=243.9×-444.7×) and assessed the mean number of pathogenic mutations to be 2.8 (range=1-6) (Table II). Mean number of variants harboring mutations of uncertain significance was 5.3 (range=3-9).
Cases 1-4. NGS played a critical role in finding rare targetable alterations. Case 1 was diagnosed as lung adenocarcinoma (Figure 1), and the patient had multiple lymph node and liver metastases. Furthermore, our immunohistochemical analysis revealed that the tumor was ALK-negative. Additionally, the PCR analysis determined that the tumor was negative for ROS1 fusion and wild-type EGFR. Notably, we detected TPR-ROS1 fusion by targeted NGS (Figure 2); this alteration has not been previously reported in lung cancer. Subsequent immuno-histochemical analysis demonstrated diffuse and strong cytoplasmic positivity for ROS1 in the tumor cells. Thereafter, the patient began crizotinib treatment and remained stable exhibiting no symptoms.
In case 2, the patient had lung adenocarcinoma with multiple lymph node, pleura, and bone metastases. Our PCR and immunohistochemical analyses revealed that the tumor was negative for EGFR and ROS1 and ALK, respectively (Figure 1). Subsequently, the NGS analysis detected an EGFR-RAD51 fusion (Figure 2) and RICTOR deletions.
Case 3 was an anaplastic thyroid carcinoma with multiple lymph node and lung metastases. Total thyroidectomy was performed in this case. On histological examination we observed that the entire tumor had a sarcomatoid pattern (Figure 1). Remarkably, we detected a NCOA4-RET fusion and pathogenic alterations in TP53 C238Y and TERT C228T by targeted NGS (Figure 2). Subsequently, the patient was treated with selpercatinib. Post one month of treatment, a follow-up chest computed tomography (CT) revealed a significant decrease in size of the metastasized tumor.
In case 4, the patient had a lung mass and multiple lymph node, peritoneum, bone, and small bowel metastases. Consequently, small bowel mass excision was performed. We diagnosed this patient with sarcomatoid carcinoma, as the tumor exhibited histologic characteristics of both carcinoma and sarcoma (Figure 1). Moreover, our immunohistochemical analysis revealed that the tumor was positive for TTF-1, cytokeratin (AE1/AE3), and vimentin. On the other hand, targeted NGS detected exon 14 skipping in MET (Figure 2) and a pathogenic mutation in KRAS Q61L, and an amplification of KRAS. Therefore, capatinib treatment was initiated for this patient. Of note, a follow-up chest CT post six months of treatment demonstrated a significant decrease in the size of the lung and metastasized masses.
Cases 5-6. NGS played a critical role in diagnosis. In case 5, the patient had a history of pancreatic cancer five years prior. A follow-up positron emission tomography-CT revealed a focal fluorodeoxyglucose uptake in the endometrium. The tumor was histologically diagnosed as an adenocarcinoma upon biopsy, and total abdominal hysterectomy was performed. Remarkably, the endometrial tumor was histologically indistinguishable from the metastatic pancreatic adenocarcinoma (Figure 3). Notably, our immuno-histochemical analysis determined the endometrial tumor to be negative for ER, PR, PAX2, and PAX8; SMAD4 loss was also detected. We also performed targeted NGS of both the pancreatic and endometrial specimens to identify the origin of the endometrial tumor. In fact, the NGS analysis revealed that both the specimens had identical driver mutations: KRAS, TP53, CDKN2A, and SMAD4. Both the tumors also had identical mutational variants of uncertain significance. Interestingly, they also had identical amino acid changes in sequences of proteins encoded by all genes except SMAD4 (SMAD4 p.H261Ifs*75 in the pancreatic specimen vs. SMAD4 p.Tyr133* in the endometrial specimen).
The case 6 patient was diagnosed with lung cancer and presented with clinical symptoms of multiple lymph node, adrenal gland, chest wall, and abdominal metastases. Of note, lung and lymph node biopsies had significantly different histological and immunohistochemical characteristics. For instance, the lymph node biopsy sample had a typical lung adenocarcinoma histology and exhibited TTF-1 nuclear positivity in immunohistochemistry. In contrast, the lung biopsy sample was TTF-1 negative and demonstrated a solid growth pattern with a few signet-ring cells (Figure 3). Thus, we speculated that this case was of a double primary cancer. Indeed, targeted NGS determined that both the specimens shared all pathogenic genetic alterations (KRAS G12V, STK11 D194Y, and SMARCA4 S1409*) and exhibited identical amino acid changes.
Cases 7-9. Dilemma cases demonstrating significant discrepancies between NGS, histological, and immuno-histochemical data. In case 7, the patient visited the hospital because of an incidentally detected 15.0 cm-sized left ovarian mass. Consequently, total abdominal hysterectomy, bilateral salpingo-oophorectomy, and lymph node dissection were performed. Histological examination demonstrated that the tumor had back-to-back arrangements of small glands lined by an endometrioid-like epithelium. The tumor cells had nuclei that were round or oval in shape and occasionally were observed to undergo mitosis. Remarkably, the tumor was positive for ER, PR, and PAX8, and focally positive for WT1, as determined by immunohistochemistry. The immunohistochemical analysis also detected weak and focal cytoplasmic expression of p53 (Figure 4). Thus, an endometrioid adenocarcinoma diagnosis of the ovarian tumor was made in line with the histological and immunohistochemical findings. However, subsequent NGS analysis identified BRCA1 W1815* and TP53 T256Hfs*89 mutations that are typical for ovarian high-grade serous carcinoma. We also detected a pathogenic mutation in ASC2 E1523Vfs*5.
Histological and immunohistochemical findings in case 8 were also consistent with ovarian high-grade serous carcinoma. The tumor demonstrated a solid, papillary, and glandular architecture with marked nuclear atypia. Moreover, the tumor cells underwent frequent mitosis. Immunohisto-chemical analysis of the tumor demonstrated WT1- and PAX8-positive nuclear expression, diffuse p16 expression, and complete negative p53 staining (mutant pattern; Figure 4). Furthermore, subsequent NGS analysis determined pathogenic mutations in the BRCA2 D2242Ffs*2, ARID1A Q557*, and PARP1 K15* genes. However, we did not detect genetic alterations in TP53.
Case 9 was diagnosed with small cell lung carcinoma and had consistent histological (Figure 4) and immuno-histochemical findings (positive expression for chromogranin A, synaptophysin, and cytokeratin AE1/AE3 positive and a >90% Ki-67 labeling index). Interestingly, our NGS analysis detected no RB1 loss, which is a characteristic of small cell lung carcinoma. Instead, we detected pathogenic alterations in ERBB2 R896G, TP53 R248L, CREBBP R1446L, RUNX1 R162K, MAP3K1 E1127*, and FANCL Q75*. We also identified deletions of TP53, and PTEN.
Discussion
In this study, we present real-world cases wherein NGS was beneficial in making an appropriate diagnosis and establishing an effective treatment plan for cancer patients.
First, targeted NGS identified targetable genetic alterations that were not detected by immunohistochemistry or PCR. For instance, we detected TPR-ROS1 fusion in case 1, which has not been previously reported in lung adenocarcinoma. Of note, ROS1 fusion occurs in 2-3% of invasive non-mucinous lung adenocarcinoma (32); such cases are treated with targeted tyrosine kinase inhibitors (TKIs) (33). ROS1 frequently fusses with the following genes: cluster of differentiation 74 (CD74), ezrin (EZR), syndecan 4 (SDC4), and tropomyosin 3 (TPM3) (34). Indeed, the AmoyDx ROS1 Gene Fusions Detection Kit did not include a probe for translocated promoted region (TPR), thereby failing to detect the TPR-ROS1 fusion. TPR encodes a large coiled coil protein that interacts with nuclear pore complexes (35); oncogenic fusion of TPR with several kinase genes has been reported in multiple types of neoplasms (36, 37). The breakpoint of the TPR-ROS1 fusion detected in case 1 had an intact kinase domain but lacked other driver mutations, indicating an oncogenic nature. This speculation is further supported by the diffuse and strong ROS1-positive expression we observed upon immunohistochemistry. Furthermore, EGFR-RAD51 fusion was detected in case 2. Although the majority of EGFR-activating mutations in lung adenocarcinomas are insertions and deletions (indels) or SNVs (38), EGFR fusion has been occasionally reported in previous studies. Although such cases of lung adenocarcinomas are limited in number, they have demonstrated a good response to EGFR TKIs (39-41). Since there is no commercially available kit specifically designed for the detection of EGFR fusion, NGS analysis is an essential alternative to identify such a rare fusion. In case 3, we detected RET fusion accompanied with TP53 and TERT promoter mutations. Although RET fusion has rarely been reported in anaplastic thyroid carcinoma (42), the patient in case 3 exhibited a remarkable response to capatinib treatment during follow-up. Adenocarcinoma with MET exon 14 skipping occurs in a subset of lung adenocarcinomas. In fact, sarcomatoids with MET exon 14 skipping in lung cancer have been previously described (43-45), consistent with our observations in case 4. In conclusion, targeted NGS should be the first choice for multiplex testing to detect these rare but targetable genetic alterations.
Second, we investigated two cases wherein targeted NGS helped identify the origin of malignant tumors. Pancreatic adenocarcinoma is one of the deadliest cancers that metastasizes to many organs (46); nonetheless, metastasis to the uterus has rarely been reported in pancreatic adenocarcinoma (47). Interestingly, the endometrial mass in case 5 was histologically and immunohistochemically indistinguishable from the primary pancreatic adenocarcinoma. Since the primary pancreatic adenocarcinoma and endometrial mass shared most of the genetic alterations, we concluded that the endometrial mass was a metastatic pancreatic cancer. However, the two tumors had different amino acid changes in protein encoded by SMAD4, which is attributable to intratumor heterogeneity and branched clonal evolution (48-51).
Lung and lymph node biopsy samples in case 6 had different histological and immunohistochemical profiles. Nevertheless, we performed targeted NGS to confirm that both tumors shared identical genetic alterations and were of the same lineage. Although histological assessment and immunohistochemical evaluation are the cornerstone of pathological diagnosis, studies have reported inconsistent histological and immunohistochemical data between primary and metastatic tumors. For example, Bruehl et al. (2021) suggested that histologic assessment alone can mislead judgement regarding the possibility of a double primary cancer; instead, next-generation sequencing can be used to distinguish primary tumors from metastatic tumors (52).
Lastly, we assessed interesting cases that demonstrated substantial discrepancies between the histological, immunohistochemical, and NGS findings. It is challenging to distinguish ovarian high-grade serous carcinoma from endometrioid carcinoma, particularly poorly differentiated carcinoma, because of a morphological overlap (53). Furthermore, high-grade serous carcinoma has solid, pseudo-endometroid, and transitional cell patterns (54) designated as “SET” variant. While aberrant p53 staining pattern, diffuse WT1- and p16-positive staining, and negative ER and PR staining are immunohistochemical characteristics of high-grade serous carcinoma, endometroid adenocarcinoma exhibits contrasting characteristics (53). Histologic findings in case 7 were characteristic of an endometrioid carcinoma demonstrating round, punched-out glandular spaces. While we did not detect any psammoma bodies, our immunohistochemical data revealed that the tumors were ER- and PR-positive and were only focally WT1-positive. Additionally, the tumor demonstrated weak and focal cytoplasmic p53 expression. On the other hand, histological and immunohistochemical evaluations in case 8 implicated it to be high-grade serous carcinoma. However, while the genomic profile of case 7 indicated at high-grade serous carcinoma with concurrent mutations of BRCA1 and TP53, that of case 8 suggested endometrioid carcinoma possessing an ARID1A mutations and no TP53 mutation. This might be of therapeutic significance because ovarian cancers with BRCA mutations or homologous recombination deficiency are treatable with PARP inhibitors (55). Taken together, ovarian cancer subtypes that are not high-grade serious carcinoma histology can also benefit from targeted NGS, although the number might be small. Furthermore, case 9 was a small-cell lung carcinoma. Bi-allelic inactivation of RB1 and TP53 is essential for small cell lung carcinoma (56). Our NGS analysis of case 9 identified pathogenic mutations in TP53, ERBB2, CREBBP, RUNX1, MAP3K1, and FANCL; however, an RB1 mutation was not detected. Since inactivation of RB1 often occurs through complex genomic rearrangements (57), it is possible that RB1 loss was not detected by targeted NGS. Notably, we detected an ERBB2 mutation in case 9. ERBB2 mutations are frequently found in non-small cell lung cancer; in contrast, ERBB2 is rarely mutated in small cell lung cancer (58). Indeed, we did not identify any pathogenic alterations in ERBB2 upon screening the cBio Cancer Genomics Portal (http://cbioportal.org). Importantly, oncogenic mutation of ERBB2 in non-small cell lung cancer is an indication for targeted therapy, such as ado-trastuzumab emtansine (59). However, there is a lack of evidence supporting targeted therapy of ERBB2 mutations in small cell lung carcinoma.
This study had certain limitations. Firstly, owing to our use of a selected gene panel for targeted NGS, rare but targetable genetic alterations or mutations of diagnostic significance may have been missed. Secondly, only selected representative cases have been included in the study. Thus, this study does not provide the overall mutational landscape of tumors in real-world samples.
Despite these limitations, we demonstrated the usefulness of applying targeted NGS in clinical practice. Additionally, targetable genetic alterations that were not detected in other molecular methods were successfully detected by massive parallel sequencing. Targeted NGS was also proven to be efficient in confirming the origin of a malignant tumor. Therefore, NGS can be widely used for the management of cancer patients, as it can detect unexpected targetable alterations that go undetected during histological and immunohistochemical examinations.
Acknowledgements
The Authors are grateful for the support provided by Kyungpook National University Chilgok Hospital Molecular Pathology Laboratory.
Footnotes
Authors’ Contributions
JYP conceived and designed the manuscript. JYJ and KMS drafted the manuscript. MSK, NJP analyzed previous articles on aggressive thyroid cancer. MSK, JYJ, NJP, and JYP reviewed and revised the manuscript carefully. All Authors have read and approved the final manuscript.
Conflicts of Interest
The Authors declare no conflicts of interest.
Funding
This work was supported by a Biomedical Research Institute grant, Kyungpook National University Hospital (2021).
- Received February 27, 2022.
- Revision received March 25, 2022.
- Accepted April 4, 2022.
- Copyright © 2022, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).