Peritoneal tumour <scp>DNA</scp> —A promising peritoneal‐specific biomarker for diagnosis, prognosis, and monitoring treatment response

Gastric cancer carries a high tendency of peritoneal metastasis and unless it is at an advanced stage, routine imaging modalities such as CT and/or PET scans regularly fail to identify it, especially peritoneal micrometastasis. It is crucial to correctly stage peritoneal disease, as its presence alters disease prognosis and management plans. In Western countries, routine methods of peritoneal staging require laparoscopy and cytology, the sensitivity of which varies significantly with 20%–40% results returning as indeterminant.1 Therefore, a more accurate peritoneal-specific biomarker is required to better stage patients and direct their cancer management. As demonstrated by Yue et al.2 in their study, published in this issue, peritoneal lavage fluid cancer cell fraction (PLF CCF) may be such a prognostic biomarker in gastric cancer. PLF CCF refers to cell-free tumour-derived DNA found in the peritoneal lavage fluid collected immediately before and after surgical resection of the gastric cancer. In this derivation and validation analysis,2 Yue et al. described a tumour-informed mutational assay that achieved an impressive 100% sensitivity and 80% specificity in predicting peritoneal recurrence—a far superior predictor than peritoneal cytology status alone. Additionally, in patients who received intraperitoneal chemotherapy (IPC), dynamic changes in PLF CCF were observed corresponding to overall survival, potentially indicating that PLF CCF could be used to monitor treatment response. This proof-of-concept study greatly strengthens the potential clinical utility of peritoneal-based biomarkers in gastric cancer reported by other studies.3, 4 While the study findings are exciting and promising, we would like to highlight several areas for consideration. Firstly, in the West, peritoneal cytology is typically performed pre-treatment as part of clinical staging. Patients without macroscopic peritoneal disease and a cytology negative result are considered for curative resection. Patients with locally-advanced disease (cT3-4 or cN1+) who are at high risk of peritoneal recurrence typically receive perioperative chemotherapy in addition to surgery.5, 6 Contrastingly, in the current study, none of the patients in the exploratory cohort and only 21.6% of the validation cohort received neoadjuvant systemic treatment.2 This discrepancy represents differences in clinical practice between Eastern and Western countries, which limits the application of this study's findings when considering Western populations. Secondly, it remains unclear whether systemic therapies affect the level of PLF CCF, and whether changes in PLF CCF post-treatment carry prognostic significance. A deeper understanding of these effects will likely facilitate the generalisability of peritoneal-based biomarkers and its role in personalising anti-cancer treatments. Thirdly, performing PLF CCF at the time of surgical resection limits the interpretation of results to the post-operative setting which removes the opportunity to personalise pre-operative management. In this study, one patient received adjuvant IPC based on PLF CCF, highlighting the potential for this biomarker to inform treatment decision making.2 However, adjuvant treatment is typically limited by a reduction in patients' performance status, surgical complications, and drug toxicity. Indeed, in a recent international analysis of perioperative FLOT chemotherapy utilisation, only 50% of patients were able to complete adjuvant treatment.6 Therefore, consideration should be given to identifying patients at high risk of peritoneal recurrence at diagnosis and then implementing IPC in the neoadjuvant period to reduce peritoneal-specific relapse, as previously reported.7 Lastly, considerations need to be given to the practical implementation of PLF CCF into clinical practice. Here, factors such as assay cost and turn-around time need to be considered. A tumour-informed mutational assay whilst reportedly sensitive and specific requires biopsies of the primary tumour to identify tumour-specific DNA variants via sequence comparison with the patient's germline DNA, typically sourced from white blood cells. This informs assay design to detect the corresponding mutant DNA in the peritoneal fluid. In our experience,8 this is a costly and slow process, typically requiring a 6–8 week turn-around time. In contrast, other studies that have evaluated peritoneal tumour-derived DNA have employed methylomic-based tumour-agnostic assays,3, 4 where a pre-defined panel of characteristic DNA methylomic markers have been applied to analyse just the peritoneal fluid. The presence of these methylomic markers indicates the presence of peritoneal disease. Accordingly, a methylomic approach may trade higher assay sensitivity and specificity for lower cost and timeliness, which ultimately may impact the clinical translatability of these tests. Therefore, future research in this area should compare head-to-head a genomic versus methylomic approach for the detection of peritoneal tumour-derived DNA. In summary, the study by Yue et al.2 provides further demonstration for the use of peritoneal-based biomarkers in gastric cancer, with promising results demonstrating that PLF CCF is both sensitive and specific in predicting, detecting, and monitoring peritoneal metastasis. To implement such a biomarker into clinical use, one should consider the following: (1) practice variations between East and West, (2) the integration of such assays into pre-existing management pathways, and (3) practical factors such as cost and turn-around time. Peritoneal-based biomarker testing is a fast-evolving field. We look forward to future studies addressing those considerations and implementing peritoneal-based biomarkers into routine clinical use. Zexi Allan and Anthony Hamilton wrote the original draft. All authors were responsible for conceptualization and review and editing of the final version. Nicholas J. Clemons and David S. Liu provided supervision and were responsible for funding acquisition. The authors declare no conflicts of interest.