Blood lncRNA RIME test refines who benefits from first-line ESCC immunochemotherapy

A simple blood test measuring a long non-coding RNA called RIME may help identify which patients with advanced esophageal squamous cell carcinoma stand to gain the most from first-line immunochemotherapy, according to an exploratory biomarker analysis of 447 patients from a multicenter cohort. The study found that patients with low pretreatment plasma RIME expression had markedly higher response rates and better survival when treated with a PD-1 inhibitor plus chemotherapy, while those with high RIME expression fared substantially worse. When researchers combined plasma RIME with the tumor PD-L1 score already used in the clinic, the separation between patient groups became even sharper: the RIME-low/PD-L1-high group had a median overall survival of 18.56 months, whereas the RIME-high/PD-L1-high group had a median overall survival of just 10.20 months, an 8.36-month gap between two groups that standard testing alone could not tell apart.

Esophageal cancer is one of the leading causes of cancer-related death worldwide and carries a particularly heavy burden in East Asia. For patients with advanced or metastatic esophageal squamous cell carcinoma, the prognosis has long been poor, and the benefit of systemic therapy remains limited. Multiple phase III randomized controlled trials have shown that adding a PD-1 inhibitor to platinum-based chemotherapy as first-line treatment improves objective response rates and survival outcomes, and this combination has become an important standard-of-care strategy. Yet even under combined immunochemotherapy, clinical outcomes remain strikingly heterogeneous. Some patients achieve durable responses that reshape the course of their disease, while others progress early or derive limited benefit from the same regimen. That heterogeneity has pushed researchers to look for stratification markers that can explain why the same treatment produces such different outcomes in different people.

Among currently available immunotherapy-related markers, PD-L1 expression remains the most commonly used stratification factor and has been prespecified in multiple clinical trials. Its predictive value in esophageal squamous cell carcinoma, however, remains controversial. Some studies suggest that higher PD-L1 expression is associated with more evident survival benefit, while results in lower-threshold-defined strata have been inconsistent across studies. Differences in antibodies, scoring systems such as CPS, TPS, or TAP, and cutoff definitions further limit cross-study comparison and clinical interpretation. Notably, a recent meta-analysis incorporating data presented at the September 2024 FDA Oncologic Drugs Advisory Committee meeting reported no significant overall survival benefit from PD-1 or PD-L1 inhibitors in PD-L1-negative ESCC, whereas a post hoc TPS-stratified analysis of the JUPITER-06 trial suggested that clinical benefit from a PD-1 inhibitor plus chemotherapy could be observed regardless of TPS level. Current evidence therefore supports viewing PD-L1 as a clinically meaningful but insufficient baseline stratification marker rather than a decisive biomarker.

The new analysis builds on the researchers’ previous work showing that RIME promotes immune escape in ESCC through a specific epigenetic mechanism. RIME increases H3K4me3 enrichment at the promoters of the PD-L1 and IDO-1 genes, thereby upregulating the expression of both immune-suppressive molecules. The result is reduced intratumoral T cell infiltration and activity, which contributes to immunotherapy resistance. Because RIME sits upstream of PD-L1 in this regulatory circuit, the team hypothesized that measuring circulating RIME in plasma before treatment could capture a layer of immune-regulatory information that PD-L1 staining alone misses. Unlike genomic or transcriptomic profiling strategies that rely on costly, technically demanding high-throughput platforms, plasma lncRNA measurement is comparatively accessible, making it attractive for routine clinical application.

To test that hypothesis, the researchers measured pretreatment plasma RIME expression in 447 patients with available baseline values, using RNA expression data normalized to a polyA spike-in control to correct for differences in RNA input and assay efficiency across samples. Technical replicate values were averaged, log-transformed, and used as the continuous expression variable. To derive a clinically interpretable binary grouping, the team used maximally selected rank statistics based on progression-free survival time and event status, classifying patients into RIME-high and RIME-low groups. The resulting threshold of 2.23 placed 79 of 218 patients (36.2%) in the chemotherapy group and 81 of 229 patients (35.4%) in the chemotherapy plus PD-1 inhibitor group into the RIME-high category. Baseline characteristics, including age, sex, ECOG performance status, disease status, number of metastatic sites, liver metastasis, prior chemoradiotherapy, and PD-L1 subgroup distribution, were generally comparable between groups.

The response findings were among the most striking in the study. In the chemotherapy-only group, objective response rates were similar between RIME-low and RIME-high patients, at 70.59% (96/136) and 65.79% (50/76), respectively, a difference that was not statistically significant. By contrast, in the overall chemotherapy plus PD-1 inhibitor group, the response rate was markedly higher in the RIME-low group than in the RIME-high group, at 84.25% (123/146) versus 61.54% (48/78). When the researchers stratified further by PD-L1 thresholds, the differences became even more pronounced. In the PD-L1 ≥1% stratum, response rates were 91.14% for RIME-low versus 60.47% for RIME-high; in the PD-L1 ≥5% stratum, 91.30% versus 63.89%; and in the PD-L1 ≥10% stratum, 88.68% versus 54.17%. In the PD-L1 <1% stratum, the same directional trend appeared but without statistical significance.

Survival analyses told a consistent story. When PD-L1 was used alone in the immunochemotherapy group, its discrimination was limited: median overall survival values were 16.59 months for PD-L1 <1%, 16.30 months for PD-L1 ≥1%, 17.28 months for PD-L1 ≥5%, and 15.64 months for PD-L1 ≥10%, with hazard ratios generally close to 1 and no significant differences in separate Cox models. Median progression-free survival was similarly flat across thresholds, at 7.72 months for PD-L1 <1% and 7.00 months for each positive-threshold population. RIME, in contrast, separated outcomes clearly. In the PD-L1 ≥1% stratum, RIME-high was associated with worse progression-free survival (HR 2.01) and worse overall survival (HR 1.74); in the PD-L1 ≥5% stratum, the corresponding hazard ratios were 1.83 and 1.73; and in the PD-L1 ≥10% stratum, the overall survival hazard ratio reached 2.01. Median progression-free survival in the PD-L1 ≥1% stratum was 8.74 months for RIME-low versus 5.49 months for RIME-high, and in the PD-L1 ≥10% stratum it was 9.89 versus 5.49 months.

Because multiple PD-L1 thresholds and clinical endpoints were examined, the team performed an exploratory Benjamini-Hochberg false discovery rate adjustment across 12 RIME-high versus RIME-low comparisons within the immunochemotherapy group, defined as four PD-L1 strata across response, progression-free survival, and overall survival. All nominally significant subgroup associations remained statistically significant after adjustment. Nested-model comparisons then addressed the critical question of whether RIME merely repartitions PD-L1-defined groups or provides genuinely new information. In the PD-L1 ≥1% subgroup, adding RIME to a model containing residual PD-L1 categories improved discrimination for response (AUC from 0.611 to 0.773), progression-free survival (Harrell’s C index from 0.523 to 0.604), and overall survival (0.511 to 0.583), with all improvements statistically significant. Treatment-by-RIME interaction analyses further showed that the relative benefit of adding a PD-1 inhibitor to chemotherapy differed by baseline RIME status, with the most consistent exploratory evidence observed for response and progression-free survival, providing an exploratory predictive signal for the biomarker.

Safety outcomes, notably, showed no meaningful differences by RIME status. In the immunochemotherapy group, any-grade treatment-emergent adverse events occurred in 99.3% of RIME-low and 100% of RIME-high patients, and the incidence of grade ≥3 events was similar between groups at 75.7% versus 74.1%. Serious adverse events occurred in 39.9% and 37.0% of patients, respectively, and immune-related adverse events in 87.8% and 82.7%. Treatment discontinuation due to adverse events was only slightly more frequent in the RIME-high group at 11.1% versus 8.1%. These findings indicate that the observed associations of plasma RIME are more apparent for efficacy and survival outcomes than for safety outcomes. Exploratory analyses of dynamic RIME changes during treatment also showed a consistent direction, with patients whose RIME declined tending to have better efficacy and prognosis than those whose RIME increased, though these differences did not reach statistical significance because of limited longitudinal samples.

The authors emphasize that the findings require careful interpretation and external validation. The biomarker analysis was exploratory and involved repeated comparisons across multiple endpoints and PD-L1 strata without a prespecified confirmatory testing hierarchy. The dichotomous RIME cutoff was derived from the same cohort used for subsequent outcome analyses, which may introduce overfitting; sensitivity analyses using median cutoffs, tertile groupings, and continuous modeling showed generally consistent directions but attenuated statistical strength, and bootstrap resampling indicated that the cutoff varied across samples. All analyses were conducted in a single biomarker-evaluable cohort without independent external validation, and because CPS was unavailable in the analyzed dataset, the study could not directly compare TPS- and CPS-based stratification. The researchers conclude that integrating plasma lncRNA RIME with PD-L1 may improve discrimination of clinical benefit in patients receiving first-line immunochemotherapy for advanced ESCC, but that the cutoff of 2.23 should be interpreted as an exploratory stratification threshold requiring validation in independent prospective multicenter cohorts rather than a definitive clinical decision threshold.

Subject of Research: A biomarker analysis of plasma RIME and PD-L1 for stratifying first-line immunochemotherapy outcomes in advanced ESCC

Article Title: A biomarker analysis of plasma RIME and PD-L1 for stratifying first-line immunochemotherapy outcomes in advanced ESCC

Article References: Liu, J., Ren, R.-Q., He, Y., Zhang, Q.-H., Zhang, X.-M., Luo, X.-J., Cao, Y.-Z., Wu, Q.-N., & Luo, H.-Y. (2026). A biomarker analysis of plasma RIME and PD-L1 for stratifying first-line immunochemotherapy outcomes in advanced ESCC. iScience, 29(11), Article 117995. https://doi.org/10.1016/j.isci.2026.117995

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117995

Keywords: esophageal cancer, biomarker, immunotherapy, RIME, PD-L1, lncRNA, esophageal squamous cell carcinoma, immunochemotherapy, precision oncology, analysis, plasma, stratifying