Prognostic value of molecular cytology by one-step nucleic acid amplification (OSNA) assay of peritoneal washings in advanced gastric cancer patients

Peritoneal dissemination is a common form of gastric cancer (GC) recurrence, despite surgery with curative intent. This study aimed to evaluate the prognostic value of intraperitoneal lavage One-Step Nucleic Acid Amplification (OSNA) assay in advanced GC patients. OSNA assay targeting CK-19 mRNA was applied to detect free cancer cells (FCC) in intraperitoneal lavage samples obtained during gastrectomy. A total of 82 GC patients were enrolled to investigate the correlation between OSNA assay and patient’s prognosis. Of the 82 patients, OSNA assay was positive in 25 (30.5%) patients. The median OS in OSNA positive patients was significantly lower than in OSNA negative patients (19 vs 45 months). Positive OSNA assay result was a significant unfavourable prognostic factor in both, univariable (HR 3.45, 95% CI 0.95–12.48; p = 0.0030) and multivariable analysis (HR 3.10, 95% CI 1.22–8.54; p = 0.0298). Positive OSNA assay in intraperitoneal lavage is a valuable indicator of poor survival in advanced GC patients after multimodal treatment. After further confirmation on larger sample size, OSNA assay of peritoneal washings could be considered an adjunct tool to conventional cytology, the current gold standard, to provide precise intraoperative staging and additional prognostic information.

Although the incidence of gastric cancer (GC) has steadily declined over the past 50 years, it remains one of the most common and deadly cancers 1 . Worldwide, more than 1 million new cases and an estimated 783 000 deaths caused by GC were recorded in 2018, making it the fifth most frequently diagnosed cancer and the third-leading cause of cancer death 2 . Despite surgery with curative intent, peritoneal dissemination is a common form of GC recurrence 3 . It is believed that the cause of spread may be due to the free cancer cells (FCC) present in the peritoneal cavity during surgery 4 . Although the presence of FCC is associated with advanced GC stage and poor prognosis, there is no standard treatment for patients with microscopic peritoneal dissemination 3 . The REGATTA trial demonstrated that limited gastrectomy followed by systemic chemotherapy (CTH) did not affect survival compared with CTH alone in patients with peritoneal metastases from GC 5 . Japanese GC treatment guidelines 2018 (5th edition) recommend CTH for peritoneal washings cytology-positive (CY1) patients who underwent gastrectomy 6 . Furthermore, there is no agreement on the routine testing for FCC in peritoneal fluid, the methods of detection, and the relationship of FCC to the clinical and pathological variables 3 .
Determination of the clinical stage of GC at baseline provides the necessary information to develop an appropriate treatment strategy 7 . Computed tomography (CT) has good accuracy in GC staging, but sensitivity and specificity for detection of peritoneal metastases remain poor 8 . The utility of staging laparoscopy (SL) for preoperative staging of GC has been proposed in many studies 9,10 . It has been suggested that SL has better sensitivity, specificity and accuracy than other imaging modalities, especially in detecting peritoneal metastases (87%, 100% and 91%, respectively) 9,11 . SL with peritoneal washings enables detection of FCC by cytology 12  www.nature.com/scientificreports/ M1, a stage IV of GC 13 . The European Society for Medical Oncology (ESMO) recommendation is to perform the peritoneal washings analysis in all patients with a potentially resectable GC (stages IB-III) 14 .
Out of many peritoneal washings examination methods, conventional cytology with Papanicolau or Giemsa staining is most common 14,15 . Immunohistochemical (IHC) methods based on the reaction of antibodies against antigens presented on tumour cells improve cytological sensitivity 10 . Molecular diagnostics using RT-PCR has also been used to detect FCC in peritoneal fluid due to its high sensitivity 16 . The sensitivity and specificity of all these methods for predicting peritoneal recurrence vary widely 14 . The conventional cytology remains the standard for peritoneal washings examination due to its simplicity 17 . The cytology's sensitivity, specificity and overall accuracy in predicting peritoneal recurrence is 11-80%, 86-100% and 73-92%, respectively 15 . Due to the relatively low sensitivity of conventional cytology, many patients with negative cytology (CY0) may experience unexpected and rapid peritoneal recurrence after surgery performed with curative intent 12 .
It has been reported that molecular detection of tumour markers such as carcinoembryonic antigen (CEA), cytokeratin , cytokeratin 20 (CK-20) is better correlated with peritoneal recurrence and associated with adverse outcomes in patients with GC 10,[17][18][19] . It has been proven that CK-19 may serve as a suitable marker of metastases in GC patients. CK-19 is commonly expressed by cancer cells of epithelial origin but not by lymphoid or hematopoietic cells 20 . It has also been reported that CK-19 may exceed CK-20 in detecting circulating cancer cells in peripheral blood from GC patients by means of RT-PCR 21 . Results of the research demonstrate that CK19 mRNA detected by reverse transcription-loop-mediated isothermal amplification (RT-LAMP) may be suitable as an intraoperative diagnostic modality for detecting GC patients with high risk of recurrence even after clinically curative surgery who require appropriate adjuvant treatment 22 . A meta-analysis demonstrated that molecular analysis of peritoneal fluid could be a prognostic factor for patients with negative cytological findings and/or receiving curative treatment 23 .
As we previously described, One-Step Nucleic acid Amplification (OSNA) can be applied for fast detection of CK-19 mRNA in peritoneal washings, reflecting the amount of FCC in advanced GC patients 24 . The sensitivity and specificity of the peritoneal washings samples using OSNA were 83.3% and 87.8%, respectively. These results indicated that intraoperative OSNA assay may serve as a valuable alternative to conventional cytology. Due to its objectiveness and reproducibility, OSNA seems to be a reliable quantitative method of FCCs assessment in the peritoneal fluid 24 .
This study aimed to evaluate the prognostic value of intraoperative peritoneal lavage OSNA assay in advanced GC patients.

Results
Among the 82 patients included in the study, 44 (53.7%) were male, and 38 (46.3%) were female with a median age of 65 years (range 40-84). According to Lauren classification, there were 35 (42.7%) intestinal, 16 (19.5%) mixed, 22 (26.8%) diffuse type tumours, whereas 9 (11.0%) cases were not classified. Most of the patients included in the study (62.2%) had pT3/4 tumours. Additionally, 45 (54.9%) patients had lymph node metastases (pN+) and 11 (13.4%) patients had limited peritoneal metastases (pM1). The majority (75.6%) of the patients received neoadjuvant CTH. Proximal, distal, and total gastrectomy with a corresponding extent of D1 + /2 lymphadenectomy has been performed in 17.1%, 43.9%, and 39.0%, respectively. Twenty-five (30.5%) patients obtained positive results by OSNA assay, whereas 6 (7.3%) patients were positive by conventional cytology. There were two cases in which the cytology was positive, and OSNA assay was negative. One patient died as a result of postoperative complications, and the other was alive at the time of the last follow up. For OSNA positive patients, the values range from 24.6 to 71,000 cCP/µl. The characteristics of patients included in the study are shown in Table 1. Comparison of OSNA assessment results depending on demographic and clinical variables distribution was presented in Supplementary Table 1. In patients with lower (y)pT stages significantly more often OSNA negative results were observed ((y)pT0; (y)pT1; (y)pT2: 100%; 75%; 72.7%, respectively), while more frequent positive results were noted in patients with the (y)pT4b stage (71.4%) (p = 0.0292). In (y)pM1 patients compared to those classified as (y)pM0, positive OSNA assessment results were observed significantly more often (72.7% vs 23.9%; p = 0.0035). In patients with lower (y)pTNM stages significantly more often OSNA negative results were observed (0; IA; IB: 100%; 75%; 66.7%, respectively), while more frequent positive results were noted in patients in the IV stage (72.7%) (p = 0.0186).
We found statistically significant survival differences after combined cytology assessment and OSNA assay (p = 0.0051) (Fig. 2). Interestingly, a significantly higher risk of death was noted in patients with negative cytology and positive OSNA assay than double-negative patients (

Discussion
Our study focused on the prognostic value of OSNA assay for detecting single tumour marker (CK19) mRNA in intraoperative peritoneal lavage in advanced GC patients. Multiple studies on genetics biomarkers in peritoneal washings have demonstrated that PCR positive patients have worse OS and earlier peritoneal recurrence than PCR negative patients [25][26][27][28] . In the present study, the mOS in OSNA positive patients was significantly lower than in OSNA negative patients. Additionally, OSNA assay on peritoneal washings showed an additional prognostic value to conventional cytology, as reflected in both, univariate and multivariate analysis. Moreover, mixed type GC and nodal involvement were associated with a significant increase in the risk of death. It is known that mixed type carcinomas have more aggressive clinical behaviour correlated with greater tumour size, submucosal and lymphovascular invasion 29 . According to the Japanese classification, mixed histological types exhibit higher lymph node metastasis rates than pure differentiated or undifferentiated types 30 . However, for higher stage cancers, diffuse type may have worse prognosis 31 .
Although the first reports on FCC detection by cytology in GC patients were published nearly 70 years ago 32 , there is no consensus on the management of patients with positive peritoneal cytology. The treatment recommendations range from palliative CTH to neoadjuvant CTH followed by conversion surgery if regression is achieved 6,33 . In the REGATTA trial, gastrectomy followed by CTH did not demonstrate any OS benefit compared with CTH alone in advanced GC patients with a single non-curative factor 5 . However, in patients with GC and CY1 or low-volume peritoneal dissemination, perioperative S-1 based CTH 34 or hyperthermic intraperitoneal www.nature.com/scientificreports/ chemotherapy (HIPEC) 35 or intraperitoneal and intravenous paclitaxel plus S-1 36 may be considered as an addition to gastrectomy with curative intent. Patients with negative cytology findings may experience unexpected peritoneal recurrence after surgery 12 . These results indicate the lack of sensitivity on conventional cytology for detecting FCC and predicting peritoneal recurrence. Our findings indicate the additional value of the OSNA assay in detecting FCC in peritoneal washings when conventional cytology is negative. Studies showed that molecular biology techniques increase the sensitivity of FCC detection 37,38 . A positive result of CEA mRNA's cytology and transcription-reverse transcription concentrated reaction (TRC) was observed in 20% and 54% of patients, respectively 38 . Similarly, our study showed that OSNA assay increases the detection of FCC. We found 7.3% patients positive by cytology and 30.5% positive by OSNA examination. The positive result of RT-PCR assay has correlated with poor prognosis and short survival 39 . A recent meta-analysis on the prognostic significance of molecular analysis of peritoneal fluid in GC patients showed that even in CY0 patients, positive molecular status increases the risk of a poor prognosis by over twofold 23 . In our study, twenty-one patients were OSNA positive out of seventy-six CY0 patients. Therefore, nearly one-third of the study group gained additional prognostic information with the OSNA assay for intraperitoneal lavage assessment. Applying SL with intraoperative lavage as part of the diagnostic process in patients with GC may help identify microscopic peritoneal dissemination and choose an appropriate treatment modality 39 . A meta-analysis showed that CY0 before treatment is associated with more prolonged survival than CY1 40 .
Furthermore, it is postulated that cytology should be considered a modifiable factor as a conversion from CY1 into CY0 after systemic CTH is associated with OS improvement 40 . Although CY1 is considered stage IV disease, the prognosis and OS of patients with positive cytology or macroscopic peritoneal metastases are not equivalent 39 . The OS was significantly improved in the solitary CY1 group 40 . Neoadjuvant CTH resulted in the downstaging of the disease in 61% of CY1 patients 39 . On the other hand, nearly 25% of CY0 patients converted into CY1 after receiving neoadjuvant CTH 41 . Prediction of advanced GC patients response to neoadjuvant CTH remains a challenge.
In contrast to conventional cytology, which is based on the assessment of cell morphology, quantitative molecular methods have greater reproducibility and objectivity, favouring their application in laboratory practice 23 . However, several factors can impair the standardisation of a molecular method for routine diagnostic workup. The sensitivity of molecular methods is mostly determined by the marker gene expression level and the relevant background level in peritoneal lavage. According to the human protein atlas 42 , amplified CK-19 mRNA may originate from various cells. Expression at the background level demands a cut-off approach, especially for samples such as intraperitoneal lavage, in debris and various other cells 43 . Additionally, the expression of markers in FCC might differ from the tissue-resident cells or marker concentration may vary within the tissue due to intratumoural heterogeneity 44 . We could not sufficiently establish the specificity of the assay yet. CK-19 mRNA might be produced by other cells in the abdomen and therefore further studies and respective validation are needed. It is postulated that using a panel of various tumour markers may further increase the sensitivity and specificity of molecular detection of FCC 15 . Future studies to determine whether analysis of multiple tumour markers rather than a single gene may improve the diagnostic utility of OSNA examination.
Limitations of our study include the small sample size and relatively short follow-up time. Extended follow-up with larger sample size is essential to assess if OSNA examination of peritoneal washings can be utilised instead of conventional cytology. Another limitation of our study is the heterogeneity of the patient population, which

Conclusions
Positive OSNA assay in intraperitoneal lavage is a valuable indicator of poor survival in GC patients after multimodal treatment. After further confirmation on a larger sample size, OSNA assay of peritoneal washings could be considered as an adjunct tool to conventional cytology that is the current gold standard to provide precise intraoperative staging and additional prognostic information.

Methods
This observational study was conducted after obtaining institutional review board approval (Bioethical Committee of Medical University of Lublin, Ethic Code: KE-0254/297/2018). Written informed consent was obtained from the patients in line with the principles outlined in the Declaration of Helsinki. We collected data from database of patients operated on GC between July 2017 and Jun 2021 in the Department of Surgical Oncology, Medical University of Lublin, Poland. Inclusion criteria were: histologically confirmed gastric adenocarcinoma scheduled for gastrectomy either following neoadjuvant CTH or upfront surgery with adjuvant radiochemotherapy. Type of gastrectomy and extent of lymphadenectomy (D1+, D2) was performed at the surgeon's discretion. The exclusion criteria were no resection: explorative laparotomy (laparoscopy) or any palliative non-resection procedure (bypass, jejunostomy). The study included 82 consecutive patients in whom intraperitoneal lavage was performed after surgical exploration. Patients were staged following the 8th AJCC/TNM classification. Neoadjuvant treatment was received by 75% of patients, therefore the abbreviation (y)pTNM was introduced to emphasize that preoperative therapy did not apply to the entire study group. Modified Becker's system was used for the pathological evaluation of tumour regression grade (TRG) 45 .
Intraoperative lavage examination. After laparotomy and thorough exploration of the abdominal cavity, intraoperative lavage with 100 ml of saline solution was performed in all included patients. The washings obtained from the tumour area (50 ml) were divided into two equal parts (25 ml). One was intended for cytological examination, and the other one was centrifuged for 10 min at 1500×g in order to obtain cellular sediment. The cell pellet was subjected to OSNA examination.
OSNA examination. Peritoneal lavage samples for OSNA assay were centrifuged for 10 min at 1500×g to obtain a cell pellet. The cellular sediment was stored at − 80 °C until OSNA examination. As we previously described, peritoneal lavage was assessed according to the protocol for OSNA performance 24 . The first step of sample preparation was homogenisation of the cell sediment using LYNORHAG homogenising buffer, pH 3.5 (Sysmex, Kobe, Japan). The process was carried out with a RP-10 homogenizer (Sysmex, Kobe, Japan). One ml of LYNORHAG homogenising buffer, pH 3.5 (Sysmex, Kobe, Japan) was added to cellular pellet. The cells with homogenising buffer were transferred to LYNOPREP tube (Sysmex, Kobe, Japan), and homogenized for 90 s. at 10,000 rpm. In this process, CK-19 mRNA was released from the tumour cells.  49 . The RT-LAMP method measured the time required to exceed a defined threshold of turbidity caused by magnesium pyrophosphate, a by-product of the reaction. The change in turbidity correlates with the amount of CK19 mRNA calculated from the standard curve value. All results below 250 cCP/ µL (a cut-off set in RD-210 analyzer to distinguish positive and negative lymph node) were calculated based on the current standard curve. As we published before, the cut-off value for distinguishing positive and negative cases for peritoneal washings samples was set at 24.6 cCP/µL 25 . At this cut-off value for peritoneal washings samples, the sensitivity and specificity were 83.3% and 87.8%, respectively 24  www.nature.com/scientificreports/ data. Categorized or dichotomized variables were expressed as numbers and percentages. Overall survival (OS) was defined as the time from the date of surgery to the date of the patients death or the date of the last follow-up. None of the patients was lost from follow-up (median was 12 months). Survival data were updated in May 2021.
In univariable OS analysis, we used the log-rank test to calculate the proportional hazard ratio (survival curves were generated with the use of Kaplan-Meier estimation method). In multivariable analysis, Cox logistic regression models were used. All statistically significant results of the univariable analysis were considered potentially valuable for multivariable analysis. However, since not all patients underwent neoadjuvant chemotherapy, this variable was not included in the multivariable model. Moreover, the backward elimination method revealed that only (y)pTNM and OSNA assay were significantly related to survival and thus, only those variables were finally used for adjustment in multivariable analysis. The only exceptions were models for ypT and ypN in which results were adjusted by (y)pN and (y)pT, respectively (instead of (y)pTNM) as they are elements of this composite measure. In all analyses, we used two-tailed tests. All results with a p-value below 0.05 were considered to be statistically significant.

Data availability
All data generated or analysed during this study are included in this published article and its supplementary information files. www.nature.com/scientificreports/ Reprints and permissions information is available at www.nature.com/reprints.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.