Response to Comments on “Prognostic Significance of Pretreatment [¹⁸F]ALF-NOTA-FAPI-04 PET/CT in Patients With Recurrent Gastric Cancer Undergoing Combined PD-1 Inhibitor and Chemotherapy”

We greatly appreciate the thoughtful and constructive comments of Drs. Liu and Liu [1] on our recently published article [2] regarding the prognostic value of pretreatment [¹⁸F]ALF-NOTA-FAPI-04 PET/CT in patients with recurrent gastric cancer receiving programmed cell death protein 1 inhibitor plus chemotherapy. These valuable insights have helped us identify critical issues to address and they have provided a clear direction for our follow-up research. We have responded to each concern in detail below.

REGARDING THE OPTIMAL CUT-OFF VALUES AND EXTERNAL VALIDATION

We agree that the cutoff values derived from our single-center cohort of 51 patients require external validation before widespread clinical application, and we have acknowledged this as a core limitation of the original manuscript.

In our study, we used median values of PET parameters (total lesion fibroblast activation protein expression [TLF] = 188.88 SUVbw·cm³, FAPI-avid tumor volum [FTV] = 44.17 cm³) as stratification thresholds. This is a conservative and widely adopted approach in small-sample retrospective prognostic studies designed to minimize the risk of overfitting from data-driven optimal cut-off values. In order to further verify the robustness of the prognostic associations, we performed supplementary analyses, treating TLF and FTV as continuous variables in Cox proportional hazards models. The results have confirmed that per-unit increases in TLF were independently associated with worse progression-free survival (hazard ratio [HR] = 1.01, 95% confidence interval [CI]: 1.01–1.01, P = 0.025), and per-unit increases in FTV were independently associated with worse overall survival (HR = 1.01, 95% CI: 1.01–1.01, P = 0.028). These findings are consistent with the dichotomized analyses, demonstrating that the prognostic value of these parameters is not dependent on the specific cut-off values.

Regarding the suggestion of restricted cubic spline (RCS) analysis, we recognize its advantage in characterizing nonlinear dose-response relationships between biomarkers and outcomes. However, given the relatively small sample size of our cohort, RCS modeling lacks sufficient statistical power and may yield unstable and unreliable estimates. We concur that external multicenter validation is essential for establishing generalizable thresholds. At present, we are initiating a multicenter validation study with a larger sample size in which we will apply the RCS and other robust statistical methods to refine the prognostic thresholds.

REGARDING MOLECULAR BIOMARKERS (PD-L1, MSI/MMR, EBV, TMB)

We acknowledge that molecular biomarkers, including programmed death-ligand 1 expression, microsatellite instability/mismatch repair status, Epstein-Barr virus status, and tumor mutational burden, are established predictors of immunotherapy efficacy in gastric cancer, and their absence represents an important limitation of our study.

This limitation is primarily attributable to the retrospective design and the clinical context of recurrent gastric cancer. Among the 51 patients, only 35 had histologically confirmed recurrence, and re-biopsy specimens with sufficient tissue for molecular profiling were available in <40% of cases. Most patients were diagnosed with recurrence based on a multidisciplinary radiological consensus and archived primary tumor specimens cannot reliably reflect the molecular profile of recurrent lesions. Therefore, complete molecular biomarker data were unavailable for the full cohort, preventing their formal inclusion in multivariable models.

The primary objective of our study was to evaluate FAPI PET/CT as a non-invasive imaging biomarker rather than to claim its superiority over molecular biomarkers. We fully agree that clarifying whether the FAPI PET parameters provide incremental prognostic value beyond established molecular markers is critical for establishing its clinical utility. Our ongoing prospective study will systematically collect paired tissue samples and molecular profiling data in order to address this important question.

HEAD-TO-HEAD COMPARISON OF PROGNOSTIC PERFORMANCE BETWEEN FAPI AND FDG PET/CT

We thank Drs. Liu and Liu [1] for this perceptive comment. Our subgroup analysis of 16 patients only compared semi-quantitative uptake parameters between [¹⁸F]ALF-NOTA-FAPI-04 and [¹⁸F]FDG PET/CT, and did not perform formal statistical comparisons of their prognostic predictive performance.

The main reason for this design was the very small sample size of the FDG subgroup (n = 16), which was underpowered for reliable comparisons of the receiver operating characteristic curves or survival endpoints. A formal head-to-head comparison of the prognostic accuracy would require a much larger matched cohort to draw statistically robust conclusions.

We agree that directly comparing the predictive values of the two tracers for durable clinical benefits and survival is a clinically high-impact question. We are currently expanding our cohort of patients who underwent both PET examinations, and our next study will systematically compare the diagnostic and prognostic performance of [¹⁸F]ALF-NOTA-FAPI-04 and [¹⁸F]FDG PET/CT, including ROC curve comparisons and incremental value analyses, to verify whether FAPI PET offers superior prognostic stratification.

REGARDING DISCORDANT CASES

We appreciate this in-depth observation regarding discordant cases and have performed supplementary analyses to explore the underlying clinical and pathological characteristics of the patients.

Among patients with low baseline FAPI uptake but early disease progression (nondurable clinical benefit), the majority presented with widespread peritoneal carcinomatosis or diffuse bone metastases at recurrence, and a high proportion had signet-ring cell carcinoma (SRCC) histology. These cases indicate that highly aggressive tumor cell biology may drive rapid progression even with relatively low stromal FAP expression. Conversely, among patients with high baseline FAPI uptake who achieved durable clinical benefits, most had only locoregional lymph node recurrence and subsequently received consolidative local therapy (radiotherapy or metastasectomy) after systemic treatment, which contributed substantially to favorable long-term outcomes.

We conducted exploratory stratified analyses to determine the impact of histological subtypes. In both SRCC/mucinous adenocarcinoma and conventional adenocarcinoma subgroups, elevated FTV and TLF showed trends toward worse survival, although the differences did not reach statistical significance owing to the small sample size in each stratum. We agree that the histological subtype may influence FAPI uptake patterns and prognostic associations, and this hypothesis warrants validation in larger histologically stratified cohorts.

CONCLUSION

In summary, we sincerely thank the correspondents for their insightful comments, which deepened our interpretation of the study’s findings and helped guide future research directions. We will continue to validate and expand our findings in larger, multicenter, prospective studies with integrated molecular profiling, to further establish the clinical value of [¹⁸F]ALF-NOTA-FAPI-04 PET/CT in precision oncology for gastric cancer.