Liquid Biopsy Using Ascites and Pleural Effusion Supernatant for Genomic Proling in Gastrointestinal and Lung Cancers

Background: Precision medicine highlights the importance of incorporating molecular genetic testing into standard clinical care. Next-generation sequencing (NGS) can detect cancer-specic gene mutations, and molecular targeted drugs can be designed to be effective for one or more specic gene mutations. For patients with special site metastasis, it is particularly important to use appropriate samples for genetic proling test. Methods: Tissues, plasma, ascites (ASC) supernatants, pleural effusion (PE) samples from gastrointestinal (GI) patients with peritoneal metastasis and lung cancer patients with pleural metastasis were collected for comprehensive genomic proling. The sample were performed on next-generation sequencing (NGS) using 59 or 1021 cancer-relevant genes panel. Results: The study enrolled 156 tissues, 188 plasma, 45 ascites supernatants, and 1 pleural effusion in 304 GI group and 446 pleural effusion supernatants, 122 tissues, 389 plasma, 45 pleural effusion sediments in 407 lung cancer group. The MSAF were signicantly higher in ASC and PE supernatant than plasma ctDNA (50.18% ± 32.03% vs 12.31% ± 19.90%, p < 0.0001 and 33.74% ± 28.34% vs 6.28% ± 12.17%, p < 0.0001, respectively). ASC supernatant had a higher actionable mutation rate than plasma. ASC supernatant accounts for more actionable alterations than plasma ctDNA in 26 paired samples. PE supernatant had higher total actionable mutation rate than plasma (80.3% vs 48.4%, p < 0.05). PE supernatant had a higher frequency of uncommon variations than plasma no matter had distant organ metastasis. Conclusion: ASC and PE supernatants better

PE contain oating malignant cells as well as tumor cfDNA in the supernatant, and the cytological and supernatants were used to detect genetic mutations using NGS in multiple types of tumors [8][9][10]. However, the performance of genomic pro ling using ASC and PE in the real-world settings has not been fully investigated.
In this study, we collected ASC, plasma, tissue, and PE samples from GI cancer who had peritoneal metastasis and plasma, tissue, PE samples from lung cancer who had pleura metastasis to verify the e cacy of ASC and PE in detecting genetic mutations. Paired samples from a subset of patients were compared to systematically evaluate the concordance of genomic pro les from different sample types.

Patients and samples
In the GI cohort, 304 patients with peritoneal metastasis were prospectively enrolled. A total of 390 samples using next-generation sequencing (NGS), included 156 tissues, 188 plasma, 45 ascites supernatants, and 1 pleural effusion, were used to analyze the e cacy of ascites in detecting genetic mutations.
In the lung cancer cohort, 407 patients with pleural metastasis were prospectively enrolled. A total of 1002 samples using NGS, included 446 pleural effusion supernatants, 122 tissues, 389 plasma, 45 pleural effusion sediments, were used to analyze the e cacy of pleural effusion in detecting genetic mutations.
Next-generation sequencing ASC and PE were isolated for the extraction of cfDNA and sezsyykydiments used for genomic DNA extraction. Circulating cell-free DNA (cfDNA) was isolated from plasma using the QIAamp Circulating Sequencing data were analyzed using default parameters [14]. The reads which removed adaptor sequences and low-quality reads were aligned to the reference human genome (hg19) using Burrows-Wheeler Aligner (BWA; version 0.7.12-r1039). Realignment and recalibration were performed by using GATK (version 3.4-46-gbc02625) [15]. Single nucleotide variants (SNV) were called using MuTect (version 1.1.4) and NChot, a software developed in-house to review hotspot variants. Small insertions and deletions (InDel) were determined by GATK. Somatic copy number variations (CNV) were identi ed with CONTRA (v2.0.8) [16]. The nal candidate variants were all manually veri ed using Integrative Genomics Viewer. Targetable genomic alterations simultaneously detected by this assay included SNV, Indel, CNV and fusions.

Statistical Analysis
Fisher exact test was used to test the difference between any groups. P-value of <0.05 was de ned as statistical signi cance. Statistical analyses were performed using GraphPad Prism 8.0.

Study design and GI cancer patient characteristics
The study enrolled 304 GI cancer patients with peritoneal metastasis at diagnosis or during disease progression ( Table 1). The median diagnosis age was 57 (range 20-93), and 145 (47.7%) were female.
The major cancer type was colorectal cancer (CRC), and gastric cancer (GC), followed with appendix cancer, pancreatic cancer, small intestinal and ampullary carcinoma, and esophageal cancer. One hundred and sixty-nine (55.6%) patients had distant organ metastasis except peritoneal. Two hundred and ve (67.4%) patients had previous system treatment. A total of 390 specimens were collected to analyze the e cacy of mutation detected ability among different samples in the real world, included 188 plasma, 156 tissue, 45 ascites supernatant, and 1 pleural effusion supernatant.
ASC supernatant in paired samples among 26 patients Twenty-six patients had more than one sample, 26 paired ASC supernatant and plasma ctDNA, 7 tissues, and 1 PE supernatant. All samples had detectable somatic alterations. In the different subtype groups, 69.2% (18/26) of ASC supernatant samples, 50% (13/26) of plasma ctDNA samples, 57.1% (4/7) of tissues, and 100% (1/1) PE supernatant samples had detectable actionable alterations (Table 2). ASC supernatant had a higher detectable rate in actionable alterations although there was no statistically signi cant difference. In the paired ASC supernatant and plasma ctDNA samples, 7 patients had actionable alterations detected only from ASC supernatant, 2 patients had actionable alterations detected only from plasma ctDNA. ASC supernatant accounts for more actionable alterations than plasma ctDNA. One CRC patient (P23) had paired ASC supernatant and PE supernatant when disease recurrence, and the actionable alterations in two samples had high consistency. Six patients had no distant organ metastasis, three patients (P01, P09, P10) had the same discovery in actionable alterations. However, actionable alterations were more detected from ASC supernatant compared with plasma ctDNA in the other three patients (P14, P15, P20). We then compared the MSAF of ASC supernatant and plasma ctDNA in the 26 patients. The MSAF in the ASC supernatant was signi cantly higher than plasma ctDNA (p=0.003) ( Figure 2). We speculated that the higher MSAF may explicate the superior detecting e cacy of ASC supernatant compared to plasma ctDNA, especially in the patient who had no distant organ metastasis.

Study design and lung cancer patient characteristics
The study enrolled 407 lung patients with pleural metastasis at diagnosis or during disease progression ( Table 3). The median diagnosis age was 60 (range 28-94), and 202 (49.6%) were female. The major histology subtype was adenocarcinoma (84.8%). One hundred and twenty-four (30.5%) patients had only pleural metastasis with stage M1a and 227 patients had the other organ metastasis with stage M1b/c. Three hundred and twenty-one (78.9%) patients had previous system treatment. A total of 1002 specimen were collected to analyze the e cacy of mutation detected ability among different samples in the real world, included 389 plasma, 122 tissue, 446 pleural effusion (PE) supernatant, and 45 pleural effusion sediment.

Discussion
Approximate 10%-40% of patients have peritoneal metastasis in gastric cancer and colorectal cancer [6,17]. An approximate 15% of patients with lung cancer have pleura metastasis at the rst presentation and 50% of patients develop a pleural effusion in the duration of disease.
[18] ASC and PE are the common manifestations of peritoneal metastasis of digestive system tumors and pleural metastasis of lung cancer, respectively. Abdominal puncture and thoracentesis are widely used in patients with these body uids. Content of ASC and PE are generally relatively large, which is average in a few thousand milliliters. Target therapy has become an important strategy in the clinic and matched therapies provided longer survival [20,21]. ASC and PE supernatant had a higher actionable mutation rate and more mutations detected than plasma in GI and lung cancers. We observed that ASC supernatant had more actionable mutations detected than plasma in 26 GI cancer patients who had paired samples. We also observed this phenomenon in limited paired samples of ASC supernatant and tissues. In a CRC patient (P23) who had simultaneous ASC and PE, the actionable variations were highly consistent; however, plasma and tissue had no actionable mutation detected.
With the development of drugs and the advancement of detection technology, patients have more opportunities to receive target therapy. Oncogenic fusions, such as ALK, NTRK1/3, and RET, were reported in gastrointestinal cancer [22,23]. These variations are of great signi cance in-clinic treatment. CRC patients who had oncogenic fusion can bene t from target therapy based on previous case series [24,25]. We observed FGFR2 fusion from ASC supernatant in gastric adenocarcinoma and RET fusion from tissue in colon adenocarcinoma. FGFR inhibitors, erda tinib and pemigatinib, were approved to treat FGFR fusion urothelial carcinoma and cholangiocarcinoma, respectively [26,27]. RET inhibitors, selpercatinib and pralsetinib, were approved to treat RET fusion non-small cell lung cancer etc [28,29]. We also observed a higher uncommon actionable mutation detect rate in PE supernatant than plasma in lung cancer.
Plasma had lower MSAF and actionable mutation rate in the patient who had only peritoneal or pleura metastasis than widespread. The mutation detection ability of ASC or PE supernatant were comparable regardless of stage.

Conclusion
In a conclusion, cfDNA from ASC and PE supernatants provide more information for tumor genomic pro ling than the plasma and PE sediment. ASC Tables   Due to technical limitations, table 1 to 3 pptx are only available    The mutation detect ability of PE supernatant and plasma in 139 lung cancer patients with paired samples. PE supernatant had higher MSAF (A) and more mutations (B) than plasma. PE supernatant had higher driver mutation detected rate than plasma (C).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.