Considerations on the rationale and real-world applicability of conversion therapy in gastric cancer with peritoneal metastasis correspondence

Dear Editor,

We read with great interest the study by Yu et al in the International Journal of Surgery, which evaluated chemoimmunotherapy as a conversion strategy in gastric cancer with peritoneal metastasis (GC-PM)[1]. The authors demonstrated that selected patients could achieve R0 resection and improved survival. They also proposed a predictive model based on genomic and immune features, offering a novel precision approach for this challenging population. In compliance with the TITAN guideline for transparent reporting[2], we offer the following commentary.

Gastric cancer is a highly aggressive malignancy and a leading cause of cancer-related mortality worldwide[3]. Peritoneal metastasis (PM), a hallmark of stage IV disease, is particularly challenging due to its diffuse spread and resistance to conventional chemotherapy, resulting in a median overall survival of less than 20 months[4]. Recent advances in immunotherapy have rekindled interest in conversion therapy, which aims to downstage tumors and make curative resection feasible in select cases.

The study by Yu et al employed a retrospective, single-center design, which inherently limits generalizability and introduces potential risks of selection and intervention bias[1]. Although patients were stratified into surgical and non-surgical groups for outcome analysis, the decision to proceed with conversion surgery was not determined by predefined, standardized criteria. Instead, it was based on multidisciplinary team (MDT) reassessments conducted after every two cycles of chemoimmunotherapy, incorporating imaging findings, symptom improvement, and intraoperative judgment of resectability. While this approach aligns with real-world clinical practice, it introduces heterogeneity in treatment allocation and increases the likelihood of selecting patients with more favorable disease biology and therapeutic response.

HIGHLIGHTS

• The study employed a retrospective, single-center design without standardized surgical criteria, introducing potential selection bias and limiting generalizability.

• Treatment allocation was based on multidisciplinary reassessments rather than predefined criteria, raising concerns about heterogeneity and the influence of subjective judgment.

• The predictive model lacked regularization or internal validation, increasing the risk of overfitting and compromising applicability to broader clinical populations.

• Key clinical variables such as performance status, PCI, and tumor burden were omitted from model construction and validation, reducing its capacity to reflect real-world patient diversity.

• The conversion surgery approach, while innovative, deviates from established NCCN, ESMO, and JGCA guidelines, which caution against curative surgery in stage IV gastric cancer with peritoneal metastasis outside of clinical trials.

In constructing the model, the authors selected biologically informed variables based on information criteria (AIC and BIC), rather than applying regularization techniques such as LASSO regression or internal resampling procedures like cross-validation. While this approach was justified by the limited sample size and the need to preserve interpretability, the absence of statistical regularization or internal validation nonetheless raises concerns regarding potential overfitting and may constrain the model’s generalizability in broader clinical settings. According to current ESMO guidelines, surgical intervention for patients with peritoneal metastasis should be considered only in highly selected cases and preferably within the framework of prospective clinical trials. Given these methodological limitations, the observed survival advantage associated with conversion surgery in this study, while promising, should be interpreted with appropriate caution.

Although the model was externally validated in an independent cohort, several clinically significant confounding variables—such as performance status (PS), peritoneal cancer index (PCI), and baseline tumor burden—were not incorporated into the model nor adjusted for during validation. The omission of these factors, which are known to substantially influence immunotherapy response in patients with peritoneal metastasis, may limit the model’s ability to capture real-world clinical heterogeneity and reduce its reproducibility across diverse patient populations.

According to the consensus established by three major international guidelines—NCCN (2025), ESMO (2022), and JGCA (2021)—patients with peritoneal metastasis (PM) are uniformly classified as having stage IV (cM1) gastric cancer, for which the primary treatment objective is palliative systemic therapy aimed at prolonging survival and improving quality of life, rather than achieving curative resection. The NCCN explicitly states: “Palliative systemic therapy is the mainstay of treatment for patients with metastatic (cM1) gastric cancer. The goal is to improve quality of life and prolong survival, not cure” (GAST-9). The JGCA, citing data from the REGATTA trial, advises against reduction surgery in PM patients, concluding that it provides no survival benefit and may delay effective systemic therapy. Similarly, the ESMO guidelines recommend that surgical intervention in patients with PM or positive peritoneal cytology (CY⁺) be considered only in highly selected cases and preferably within the context of a clinical trial. In contrast, Yu et al. systematically enrolled PM patients into a conversion therapy pathway, with curative R0 resection as the central endpoint. While this approach offers exploratory potential, it deviates substantially from current guideline recommendations and lacks the level of evidence necessary to support its routine clinical adoption.

We sincerely appreciate the authors’ thoughtful contribution to the evolving field of gastric cancer with peritoneal metastasis. Their study offers valuable insights into the potential role of conversion surgery following chemoimmunotherapy and lays important groundwork for future investigations. While certain methodological limitations warrant careful consideration, the exploratory nature and clinical relevance of their findings are commendable and merit continued discussion and validation in prospective research settings.

Ethical approval

Not applicable.

Consent

Not applicable.

Source of funding

None.

Author contributions

Y.G. conceived and drafted the correspondence, and was responsible for organizing the structure and writing the manuscript. S.C. co-conceived the correspondence and contributed to manuscript writing, logical refinement, and language polishing. F.S. supervised the overall writing process, critically reviewed the content, provided key revisions, and approved the final version of the manuscript.

Conflicts of interest disclosure

The authors declare that they have no conflicts of interest.

Research registration unique identifying number (UIN)

Not applicable.

Guarantor

Feng Sun.

Provenance and peer review

Not applicable.

Data availability statement

Data sharing is not applicable to this article.

Presentation

None.