Comprehensive genetic profiling using tissue and blood in locally advanced tumors

Introduction

Comprehensive genomic profiling (CGP) using next-generation sequencing (NGS) has become a clinical practice for the effective selection of treatments based on druggable genomic variants. Identifying somatic variants of tumors using tissue is widely implemented, whereas circulating tumor DNA (ctDNA) detection is a noninvasive method for assessing genomic profiles using peripheral blood (1,2). ctDNA constitutes only a minor fraction of the cell-free DNA (cfDNA) circulating in cancer patients, complicating ctDNA detection (3,4). Genomic profiling using ctDNA has the potential to predict recurrence, survival and response to therapy (5,6). However, advanced tools are required to accurately personalize treatment decisions following the detection of somatic variants and monitoring of tumor clone and subclone evolution (7).

Tumor heterogeneity is associated with the treatment outcomes of metastatic cancers (8); however, only a small number of studies have examined whether tumor heterogeneity in ctDNA can be used to predict treatment outcomes (9). Recent studies have demonstrated that the number of somatic variants in ctDNA can be used to noninvasively assess tumor heterogeneity (10–13). Therefore, multiple ctDNA analyses following tumor resection may capture molecular and genetic heterogeneity by identifying somatic variants for targeted therapy and monitoring tumor evolution and resistance in real time (14).

Ion Torrent™ Genexus™ Integrated Sequencer (Genexus, Thermo Fisher Scientific, Inc.) was developed to automate all targeted NGS workflows from library construction using either nucleic acids of formalin-fixed paraffin-embedded (FFPE) tissues or cell-free total nucleic acids (cfTNA) of plasma to data analyses and delivers results within 24 h. Therefore, Genexus enables clinicians to provide patients with somatic variant information leading to more efficient cancer treatment (15,16).

This study, thus, aimed to evaluate the concordance rate of somatic variants between resected tumor tissue and ctDNAs at post-resection in patients with various types of solid tumor by using Genexus. Furthermore, the presence of recurrence was evaluated at one year post-resection.

Patients and methods

Patients and samples

The study was approved by the Institutional Review Board of Kurume University Hospital (approval no. 2022008) and was registered in the Japan Registry of Clinical Trials (April 10, 2023; no. 072230003). All methods were performed in accordance with relevant guidelines and regulations. Among the patients scheduled for curative surgery at Kurume University Hospital (Kurume, Japan) in April and May 2023, patients with breast, lung, pancreas, and head and neck cancer were selected in the order of the earliest scheduled surgery, aiming to enroll 2 patients per cancer type. All of the candidates were asked to participate in this study and the patients who agreed to this study were enrolled. A total of eight patients with resectable solid tumors who agreed to participate in this study were enrolled: Two patients with breast cancer, two with lung cancer, two with pancreatic cancer and two with head and neck cancer. Patients were excluded if they had synchronous and metachronous advanced cancers or severe complications. Participants were fully informed of the purpose and procedures of the study and had adequate time to ask questions and contemplate their voluntary participation. Written informed consent was obtained from all patients prior to enrollment. Detailed patient information is presented in Table I and Fig. 1A. Blood samples were obtained from the enrolled patients before surgery. Radical resection was performed and the resected tumor tissues were diagnosed by pathologists at Kurume University Hospital (Kurume, Japan). Following pathological diagnosis, DNA and RNA were extracted from FFPE tumor blocks. Blood samples were also collected in the first (between fourth to fifth week) and second (between eighth to nineth week) postoperative months, and ctDNA was promptly extracted. Tissue samples were analyzed using the Genexus Oncomine Comprehensive Assay v3 (OCAv3), while blood samples were analyzed using the Genexus Oncomine Precision Assay (OPA). The gene list for each test is provided in Table SI.

Figure 1. Study design and clinical course of the cases. (A) Blood samples were obtained before surgery and 1 and 2 months after surgery. A tissue sample was obtained during surgery. All samples were analyzed using Genexus-NGS (OPA for blood and OCAv3 for tissue samples). (B-D) Clinical information and somatic variants of (B) patient 5, (C) patient 6 and (D) patient 8. The CT image on the upper right in B shows the recurrence (yellow arrows) in the abdomen at nine months postoperatively. M, month(s); y.o., years old; NGS, next-generation sequencing; ctDNA, circulating tumor DNA; RT, radiation therapy; Pre ope, preoperative; OCAv3, Genexus Oncomine Comprehensive Assay v3; OPA, Genexus Oncomine Precision Assay; TS-1, tegafur/gimeracil/oteracil.

Table I. Patients' characteristics.

Table I.

Patients' characteristics.

Patient no.	Age, years	Sex	Primary tumor	Histological findings	Clinical staging	Preoperative therapy	Surgical procedure	Disease status at testing	Postoperative therapy
1	57	Female	Breast	Invasive ductal carcinoma	cT1cN0M0 Stage I	Anastrozole (2M)	Skin-sparing mastectomy	pTisN0M0 Stage 0	None
2	72	Female	Breast	Invasive ductal carcinoma	cT2N1M0 Stage IIB	Dose-dense PTX (4 cycles)	Partial mastectomy	pT1N0M0 Stage 0	None
3	71	Female	Lung	Papillary adenocarcinoma	cT2aN0M0 Stage IB	None	Lower lobectomy	pT1cN0M0 Stage IA3	None
4	76	Male	Lung	Squamous cell carcinoma,	cT2N0M0 Stage IB	None	Upper lobectomy	pT1bN0M0 Stage IA2	None
5	79	Male	Pancreas	Invasive ductal carcinoma	cT3N0M0 Stage IIA	None	Pancreatoduodenectomy	pT3N1aM0 StageIIB	TS-1 (6 cycles)
6	73	Male	Pancreas	Invasive ductal carcinoma	cT3N0M0 Stage IIA	None	Pancreatoduodenectomy	pT3N1M0 Stage IIB	TS-1 (6 cycles)
7	29	Female	Parotid gland	Secretory carcinoma	cT2aN0M0 Stage I	None	Partial parotidectomy	pT3N0M0 Stage III	None
8	63	Male	Hypopharynx	Squamous cell carcinoma	cT4N3bM0 Stage IVB	TPF (2 cycles)	Laryngopharyngectomy	pT3N3bM0 Stage IVB	Cisplatin (2 weeks) plus radiation (20 Gy/10 Fr)

[i] M, month(s); PTX, paclitaxel; TPF, docetaxel, cisplatin and 5-fluorouracil; TS-1, tegafur/gimeracil/oteracil.

Samples, DNA extraction and quantification

DNA and RNA from FFPE tissue specimens were extracted using a Maxwell RSC Instrument (cat. no. AS4500; Promega, Corp.), Maxwell RSC FFPE Plus DNA kit (cat. no. AS1720; Promega, Corp.) and Maxwell RSC RNA FFPE kit (cat. no. AS1440; Promega, Corp.). Whole blood for plasma analysis was processed within 15 min of collection. To obtain blood plasma, 14 ml of whole blood with EDTA-2Na was cooled and centrifuged (4°C, 2,000 × g, 10 min) twice. cfTNA, which included both DNA and RNA, was extracted using a Maxwell RSC Instrument with a Maxwell RSC miRNA Plasma and Serum Kit (cat. no. AS1680; Promega, Corp.). DNA and RNA concentrations were measured using the QuantiFluor ONE dsDNA System (cat. no. E4871; Promega, Corp.) and the QuantiFluor RNA System (cat. no. E3310; Promega, Corp.), respectively. Nucleic acid >1.1 ng/µl for DNA and >0.95 ng/µl for RNA extracted from tissues, and >1.33 ng/µl from blood were recommended as the cutoffs for further processes. The DNA and RNA fragment lengths were evaluated using an Agilent 4200 TapeStation system (Agilent Technologies, Inc.).

Genexus sequencing

The Ion Torrent Genexus Integrated Sequencer (Thermo Fisher Scientific, Inc.) is a fully automated NGS system that integrates library preparation, including cDNA synthesis, template preparation, sequencing and data analysis of purified and quantified nucleic acids (DNA, RNA and cfTNA). In this study, DNA and RNA derived from FFPE tissues were used as samples for OCAv3, and cfTNA for OPA. A total of 25 µl of DNA (>1.1 ng/µl) and RNA (>0.95 ng/µl) for OCAv3 and 20 µl of cfTNA (>1.33 ng/µl) for OPA were recommended for library preparation. After assigning the assay to the Genexus Integrative System (https://assets.thermofisher.cn/TFS-Assets/LSG/manuals/MAN0017910_GenexusIntegratedSequencer_UG.pdf), the sample type and sequence run settings were specified and they were loaded onto sample plates. The Multiplex I cfDNA Reference Standard Set (HD780; Horizon Diagnostics) was used to evaluate the performance of OCAv3 and OPA (Table SI) (17).

Genexus variant analyses

Ion Torrent™ Genexus™ Software ver. 6.6.2.1 (Thermo Fisher Scientific, Inc.) supports the Ion Torrent™ Genexus™ Integrated Sequencer (Thermo Fisher Scientific, Inc.) workflow for research use purposes from sample preparation through library preparation, template preparation, and sequencing. During and after sequencing, the software generates base calls, trims reads and determines quality values (primary analysis), then aligns reads, calls variants, and generates reports (secondary analysis) (https://assets.thermofisher.com/TFS-Assets/LSG/manuals/MAN0024953_IonTorrentGenexusSoftware6.6RUO_UG.pdf). The sequencing data were mapped to the standard reference genome, the human genome assembly 19, using the software and aligned using the torrent mapping alignment program. After the initial mapping, somatic variants were identified using the Torrent Variant Caller. The variants were selected using a built-in variant filter. The four major classes of variants evaluated were single nucleotide variants, insertions and deletions, copy number amplifications and gene fusions. Variants with a low frequency of <0.3%, an allele count of ≤5, and a C-T or G-A change were excluded from the results as potential deaminations for the OPA of the liquid samples (17).

Statistical analysis

The concordance rate was calculated for the 45 genes in common between the genes in OCA-V3 and OPA (Table SI). The concordance rate for each case was calculated as the number of concordant genes regardless of the presence of somatic variants/45 genes. In addition, the number of genes with somatic variants/45 genes was also calculated using JMP 16 software (SAS Institute).

Results

Patients' treatment course

A total of eight patients were enrolled in this study: Two patients with breast cancer, two with lung cancer, two with pancreatic cancer and two with head and neck cancer (Table I). Patient 1 was a 57-year-old female with cT1cN0M0 stage I breast cancer who received anastrozole as preoperative therapy (daily oral anastrozole 1mg for 2 months). Following mastectomy, the final stage was pTisN0M0, stage 0. No postoperative therapy was administered. Patient 2 was a 72-year-old female with cT2N1M0, stage IIB breast cancer who received paclitaxel as preoperative therapy (dose-dense paclitaxel, paclitaxel 175 mg/m2 every 2 weeks for 4 cycles). A mastectomy was performed and the final stage was pT1N0M0, stage 0. No postoperative therapy was administered. Patient 3 was a 71-year-old female with cT2N1M0, stage IIB lung adenocarcinoma. Following pulmonary lobectomy, the final stage was pT1cN0M0, stage IA3. No pre- and postoperative therapy was administered. Patient 4, a 76-year-old male, had cT2N1M0, stage IIB lung squamous cell carcinoma. Pulmonary lobectomy was performed and the final stage was pT1bN0M0, stage IA2. No pre- and postoperative therapy was administered. Patient 5 (Fig. 1B) was a 79-year-old male with cT3N0M0, stage IIA pancreatic cancer. Following pancreatoduodenectomy, the final stage was pT3N1aM0 stage IIB. TS-1 (tegafur/gimeracil/oteracil) was administered as adjuvant chemotherapy (6 cycles of daily oral TS-1 for 14 days followed by a 14-day break). Patient 6 (Fig. 1C), a 73-year-old male with cT2N1M0 stage IIB lung squamous cell carcinoma underwent pulmonary lobectomy, and the final stage was pT3N1M0, stage IIB. TS-1 was administered as adjuvant chemotherapy (6 cycles of daily oral TS-1 for 14 days followed by a 14-day break). Patient 7 was a 29-year-old female with cT2aN1M0 stage I secretory carcinoma. Partial parotidectomy was performed and the final stage was pT3N0M0, stage III. No pre- and postoperative therapy was administered. Patient 8 (Fig. 1D), a 63-year-old female, had cT2aN1M0 stage I squamous cell hypopharyngeal cancer and received preoperative therapy with docetaxel, cisplatin and 5-fluorouracil (2 cycles of TPF; cisplatin, fluorouracil, and docetaxel). Laryngopharyngectomy was performed and the final stage was pT3N3bM0, stage IVB. Cisplatin plus radiation (20 Gy/10Fr) was administered postoperatively [weekly cisplatin with daily radiation (66 Gy/33Fr) was planned, but only 2 weeks had implemented due to patient refusal of treatment].

Somatic variants in tissue and blood

All eight paired tissue samples and three blood samples (prior to resection and 1 and 2 months after resection) were analyzed using Genexus. A total of 26 somatic variants were detected in the resected tumor tissues (Table II). The overall concordance rate in all genes between tissue and postoperative blood was 94.2% (91.1–97.8%), whereas the overall concordance rate in genes with somatic variants was 4.76% (0.0–33.3%), but the concordance rate in genes with somatic variants was low (4.76%). Actionable mutations, ETV6-NTRK3 fusion and EML4-ALK fusion, were found in two patients (Patients #3 and #7). The two somatic variants were concordant between the two patients (TP53 variant in patient 6 and MAP2K1 variant in patient 8). In patient 6 (Fig. 1C), the TP53 p.P152L variant was concordant in the blood samples obtained before surgery and that obtained 1 month after surgery. The TP53 p.P152L variant was detected in ctDNA before surgery. The same variant (TP53 p.P152L) was found 1 month postresection but was not detected 2 months after resection and in the resected tissue samples, and the patient did not experience recurrence at 1 year after surgery. In patient 8 (Fig. 1D), the MAP2K1 p.L215P variant was concordant in tissues obtained during surgery and in the blood sample obtained 1 month after surgery. The MAP2K1 p.L215P variant was identified in a tissue sample. The same variant (MAP2K1 p.L215P) was identified 1 month after resection but was not detected in ctDNA prior to surgery and at two months post-resection. The patient was diagnosed with lung recurrence [a 10-mm tumor identified via chest computed tomography (CT)] 9 months post-operatively. A retrospective review of the chest CT image obtained 6 months after surgery revealed the presence of a small pulmonary nodule (measuring 3 mm on the chest CT scan) in the same region. In patient 5 (Fig. 1B), the somatic variants KRAS G12V and TP53 R175H were identified in the tissue samples. Somatic variants were not identified in the ctDNA at 1 and 2 months postresection. However, CA19-9 levels were elevated 9 months post-surgery and recurrence in the abdomen was observed on the CT scan.

Table II. Somatic variants from surgical tissue and pre- and post-operative blood.

Table II.

Somatic variants from surgical tissue and pre- and post-operative blood.

					Ope (tissue)
		Pre-ope (blood)						Post-ope 1M (blood)			Post-ope 2M (blood)			Recurrence at one year after resection
					Gene	Variant	Allele freq, # of leads, # of copies
Patient	Tumor	Gene	Variant	Allele freq, %				Gene	Variant	Allele freq	Gene	Variant	Allele freq, %
1	Breast		ND		PIK3CA	H1047R	5.8%		ND			ND		ND
					ESR1-CCDC170	Fusion	88 reads
					KRAS	CNV	64 copies
2	Breast		ND			ND			ND			ND		ND
3	Lung		ND		EML4-ALKa	Fusion	189058 reads		ND			ND		ND
					EML4-ALKa	Fusion	39 reads
					MDM2	CNV	23 copies
						ND		TRIM24-BRAF	Fusion	243 reads
4	Lung		ND		PMS2	K541EQfs*3	16.7%		ND			ND		ND
					EGFR	CNV	16 copies
					FGF19	CNV	11 copies
					KRAS	CNV	17 copies
					PIK3CA	CNV	10 copies
									ND		PIK3CA	R108C	0.39
											PIK3CA	G106V	0.30
5	Pancreas		ND		KRAS	G12V	17.0%		ND			ND		Recurrence
					TP53	R175H	11.9%							(intra-
														abdomen)
6	Pancreas				TP53	P250L	4.1%							ND
					KRAS	G12V	3.7%
					FGFR1	CNV	16 copies
					KRAS	CNV	16 copies
		TP53	P152L	0.57		ND		TP53	P152L	0.17%
			ND			ND			ND		FGFR3	R399C	0.31
7	Parotid		ND		ETV6-NTRK3a	Fusion	347801 copies		ND			ND		ND
	gland				EML4-ALKa	Fusion	5403 copies
						ND		UBN2-BRAF	Fusion	135 copies
8	Hypopharynx		ND		PMS2	K541Efs*3	3.5%		ND			ND		Recurrence
					TP53	H179L	54.2%							(lung)
					CCND1	CNV	9.5 copies
					EGFR	CNV	41 copies
					FGF19	CNV	12 copies
					FGF3	CNV	8.8 copies
					MAP2K1	L215P	0.30%	MAP2K1	L215P	0.08%

{ label (or @symbol) needed for fn[@id='tfn2-or-54-4-08967'] } In patient 8, blood from 2M after surgery was not tested due to a personal issue.

a Actionable mutation. ND, somatic variant was not detected; M, month(s); freq, frequency; CNV, copy number variation; Pre ope, preoperative.

Discussion

The present findings demonstrate that the overall concordance rate in all genes between tissue and postoperative blood was high (94.2%), but the concordance rate in genes with somatic variants was low (4.76%). However, in one patient (Patient #8) with somatic variants, which corresponded to the tissue and postoperative blood samples, distant recurrence was identified during subsequent surveillance. This finding suggests that detecting a somatic variant in postoperative ctDNA matching the same variant in the tissue may predict the occurrence of metastatic relapse in localized cancer. However, sensitivity must be validated, as recurrences were also noted in cases in which no somatic variants were identified in the postoperative blood. This pilot study, with its limited sample size, serves as a preliminary investigation into the potential benefits of multiple CGP testing. Subsequent studies with larger sample sizes and more extensive research designs are necessary.

Regarding the concordance of somatic variants between tissue and blood, high concordance has been reported between tumor tissue NGS and cfDNA in various studies: EGFR alterations in non-small cell lung cancer; multiple genes (KRAS, TP53, APC, FBXW7, SMAD4) in pancreaticobiliary cancers; KRAS variants in exons 12–13 in colorectal cancer; BRAF V600E and KIT variants in melanoma; and BRAF, EGFR, KRAS and PIK3CA across a variety of advanced cancers (7,18–21).

In a study investigating the clinical utility of ctDNA in plasma in patients with locally advanced esophageal cancer (n=11) who received neoadjuvant chemotherapy followed by esophagectomy, somatic variants from the primary tumor and ctDNA after resection were analyzed. Patients with the TP53 somatic variant in both ctDNA and resected tissues had a 1-year recurrence-free survival rate of 90% compared to 0% in patients with ctDNA negativity (22,23). This suggests that identification of the same variant in the resected tissue and ctDNA following resection is a valuable indicator for predicting recurrence. In the present study, a case of recurrence showing somatic variant matching in the tissue with the postoperative blood was encountered, indicating that detecting a somatic variant in postoperative ctDNA and identifying the same variant in the tissue may predict recurrence. However, the present study also included a case of recurrence with no somatic variants in the postoperative blood, suggesting that the specificity of predicting recurrence based on variants in postoperative blood may be high, but without high sensitivity. However, the present study showed that the concordance rate in the genes with somatic variants was low, but the overall concordance rate in all genes was high. Potential explanations for the lack of concordance include spatial and temporal tumor heterogeneity, differences in sampling intervals and potential germline DNA contamination.

In the present study, a major cause of discordance could be heterogeneity. A study comparing tumor tissue NGS (FoundationOne) and ctDNA (Guardant360) showed significant discordance between the same NGS panels in a limited number of patients with diverse solid tumors, including those of the breast, lung, pancreas and salivary glands (24). Additionally, advanced-stage urothelial cancer studies have shown significant discordance between clinically available NGS panels, even when collected at approximately the same time (25). A major cause of discordance could, in part, be differences in tumor type, timing of specimen collection, intratumoral heterogeneity, clonal evolution, discrete gene alteration types and assays performed (24–26). Furthermore, in a study with multiple CGP, the concordance rate was lower in cases with longer intervals between CGPs. EGFR mutation was analyzed in patients with lung cancer and it was found that increasing timing intervals between tumor and cfDNA sampling from <2 weeks to >6 months led to significantly lower concordance (P=0.038) (27). In a similar study on lung cancer, the overall concordance of EGFR mutations between tissue and blood varied depending on sampling time; concordance for time intervals of 0.8 and >0.8 months were 88.2 and 64.7%, respectively (28). The tissue and blood intervals in this study were 1 and 2 months. Considering the previous studies, the relatively long intervals in the present study may be associated with the low concordance rate. Additionally, the use of fresh frozen samples for tumor tissue instead of FFPE marginally increased the concordance rate from 57.1 to 66.7%, suggesting fragmentation of DNA in FFPE processing may be significant, particularly when the detection assay relies on amplicon-based amplification (29). The use of FFPE for the analysis of tissues in this study may also be associated with the low concordance rate. To conclude, the reasons for the low concordance rate between tissue and blood in this study were considered to be the presence of heterogeneity, the different collection time between tissue and blood, and nucleic acid extraction using FFPE.

The major limitation of the present study is that the number of samples and types of tumors were limited. The type of cancer and biopsy site can influence the results of the tests to identify variants. Pancreatic, ovarian, colorectal, breast, bladder, gastroesophageal, melanoma and hepatocellular carcinomas are more likely to possess detectable cfDNA than are primary brain, renal, prostate and thyroid cancers (30). Additionally, the potential limitations of ctDNA may be related to either quantity, such as the amount of ctDNA accessible in the peripheral blood, or quality, such as the purity of noncancerous cells in the tumor microenvironment, which may complicate cfDNA assays (31). Moreover, selection bias might be included in this study, because we have selected the enrolled patients. As this is a preliminary study with a limited sample size, subsequent studies with larger sample sizes and more extensive research designs are warranted.

In conclusion, the present study demonstrated that the overall concordance rate in all genes between tissue and postoperative blood was high, but the concordance rate in genes with somatic variants was low. Subsequent surveillance revealed that a single case in which matched somatic variants identified in tissue samples and postoperative blood samples exhibited distant recurrence. These findings indicate the potential of using postoperative ctDNA for recurrence detection. However, further studies on sensitivity assessment are warranted with a large sample size because the present study also included a case of recurrence despite the absence of somatic variants in the postoperative blood. The early detection of recurrence and the initiation of treatment can be facilitated by the detection of somatic variants that correspond to resected tissue.

Supplementary Material

Supporting Data

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The data derived from the Ion Torrent™ Genexus™ Sequencer are available in the DNA Data Bank of Japan (DDBJ; http://www.ddbj.nig.ac.jp) as Bioproject (no. PRJDB20670; http://ddbj.nig.ac.jp/search/entry/bioproject/PRJDB20670) and DRA (no. DRA021249; http://ddbj.nig.ac.jp/public/ddbj_database/dra/fastq/DRA021/DRA021249/).

Authors' contributions

YN and KF designed the study and wrote the protocol. KF, YW, TS, SN, KY, HU, MN and YN contributed to the design. NM, JA, HA, YI, KI and FY extracted DNA and implemented NGS. KF and YN managed the literature search and analyses. RS, SO, YO, TO, KO, MK, RK, HK, AK, RT, RU, UT, KH, TH, SH, MM and DM included the patients and performed the clinical assessments. KF, NM and YN checked and confirmed the authenticity of the raw data. KF and NM wrote the first draft of the manuscript. YN and TS supervised the whole process and critically reviewed the article. All authors read, reviewed and approved the final version of the manuscript.

Ethics approval and consent to participate

The study was approved by Kurume University Hospital's Institutional Review Board (Kurume, Japan; approval no. 2022008) and was registered in the Japan Registry of Clinical Trials (April 10, 2023; no. 072230003). All methods were carried out in accordance with relevant guidelines and regulations. Participants of this study were fully informed of the purpose and procedures of the study and had adequate time to ask questions and contemplate their voluntary participation. Written informed consent was obtained from all patients before enrollment.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Glossary

Abbreviations

Abbreviations:

References