# Deciphering the prognostic role of endoplasmic reticulum stress in lung adenocarcinoma: integrating prognostic prediction and immunotherapy strategies

Researcher: Ge Zhang
Chinese name: 张格
ORCID identity anchor: https://orcid.org/0000-0002-3116-3246
Canonical researcher: https://drgezhang.com/#person

Journal: Clinical and Experimental Medicine
Year: 2024
Type: Article
DOI: 10.1007/s10238-024-01439-4
Canonical page: https://drgezhang.com/papers/doi-10-1007-s10238-024-01439-4.html
Markdown record: https://drgezhang.com/papers/doi-10-1007-s10238-024-01439-4.md

## Full Authors

1. Bing Wen
2. Pengpeng Zhang
3. Jiping Xie
4. Zhaokai Zhou
5. Ge Zhang
6. Lianmin Zhang
7. Zhenfa Zhang

<a id="paper-discovery"></a>
## Research & review context

### Research fields

- Lung cancer genomics
- Computational oncology

### Review topics

- Stress-associated prognostic signatures
- Retrospective molecular risk models

### Scientific role

Ten-gene ERAS model with public-cohort evaluation and limited tissue expression checks.

### Evidence position

Retrospective prognostic associations; immune and drug-response outputs are computational hypotheses.

### Related concepts

- Endoplasmic reticulum stress
- Tumor immune context

### Possible citation contexts

- Methods: stress-related molecular signatures for LUAD survival stratification.
- Evidence synthesis: distinguish prognostic AUCs and tissue expression from treatment-response validation.

Possible literature-review contexts, not evidence of existing citations. Read the findings and Evidence Scope before citing the original article.

## Evidence Scale

**ERAS: endoplasmic-reticulum-stress gene signatures in lung adenocarcinoma**

Retrospective transcriptomic modeling defined a 10-gene ERAS signature, evaluated it across TCGA and GEO datasets, and assessed two model genes in 10 paired LUAD specimens. Treatment-response analyses were computational.

| Evidence | Value |
| --- | --- |
| bulk survival datasets | TCGA plus seven named GEO datasets |
| single-cell input patients | 9 |
| paired tissue specimens | 10 tumor-adjacent pairs |
| final signature genes | 10 |

### Counting note

Datasets and specimen pairs are distinct units. No unique pooled participant total is derived. The Results text calls the external survival datasets six but enumerates seven; Methods and Figure 4C identify TCGA plus seven GEO datasets. External evaluation here means retrospective public-dataset assessment, not prospective clinical validation; cross-dataset participant overlap was not independently reconstructed.

## Research Question

Can an endoplasmic-reticulum-stress gene signature stratify overall-survival prognosis in retrospective LUAD transcriptomic cohorts?

## Author Evidence Summary

The study combined LUAD single-cell and bulk transcriptomic datasets to investigate endoplasmic reticulum stress. Twenty-two survival-associated genes were reduced to a 10-gene ERAS signature using LASSO in TCGA. The authors examined survival stratification in TCGA and seven named GEO datasets, and combined ERAS with age and stage in a nomogram. The reported nomogram AUCs were 0.740, 0.754 and 0.754 at 1, 3 and 5 years. Expression of SERPINH1 and GPR37 was assessed by qPCR in 10 paired tumor and adjacent specimens. Immune scores and drug sensitivities were inferred from molecular data; this does not establish that using ERAS to choose treatment improves outcomes.

## Key Findings

<a id="kf1"></a>
### KF1: A 10-gene ERAS signature was selected from 22 survival-associated genes.

Context: Univariate Cox screening followed by LASSO in the TCGA development dataset; public-dataset survival evaluation was not uniformly significant.

| Evidence | Value |
| --- | --- |
| screened prognostic genes | 22 |
| final ERAS genes | 10 |
| named survival datasets | 8 including TCGA |

Source locator: Methods: Dataset source; Results: Model construction and evaluation, Par17; Fig. 4A-C; Supplementary Fig. 1 (MOESM1)

<a id="kf2"></a>
### KF2: The nomogram combining ERAS, age and stage showed the reported time-specific survival discrimination.

Context: Retrospective nomogram evaluation; these values are not independent prospective performance estimates.

| Evidence | Value |
| --- | --- |
| 1-year AUC | 0.740 |
| 3-year AUC | 0.754 |
| 5-year AUC | 0.754 |

Source locator: Results: Independent prognostic analysis of the ERAS and construction of the nomogram, Par18; Fig. 6F-H

<a id="kf3"></a>
### KF3: SERPINH1 and GPR37 expression was higher in LUAD tumor than paired adjacent tissue in the reported qPCR experiment.

Context: Expression validation only; it does not test survival prediction or treatment effect.

| Evidence | Value |
| --- | --- |
| clinical tissue pairs | 10 |
| genes tested in this qPCR comparison | 2: SERPINH1 and GPR37 |

Source locator: Methods: Clinical specimen collection, Par12; Results: Experimental validation, Par23; Supplementary Fig. 3A-B (MOESM3)

## Study Design & Analytical Framework

### Study profile

- Study Design: Retrospective public-dataset prognostic modeling with tissue expression experiments
- Evidence Type: Human transcriptomic associations and limited expression validation
- Population: LUAD bulk transcriptomic cohorts, nine treatment-naive patients represented in a single-cell dataset, and paired clinical tissue specimens
- Primary Endpoint: Overall survival association and prognostic discrimination
- Secondary Endpoint: Inferred immune states, computational drug sensitivity and tissue gene expression
- External dataset evaluation: Yes
- Data modalities: bulk transcriptomics, single-cell transcriptomics, clinical survival data, qPCR

## What This Study Adds

- A stress-related gene score developed retrospectively and assessed across public LUAD transcriptomic datasets.
- An explicit distinction between prognostic modeling, tissue-expression validation and inferred drug-response hypotheses.

## Evidence Scope

### Supports

- Retrospective LUAD prognostic signature development and public-dataset evaluation.
- Reported nomogram discrimination within the analyzed data.
- Limited tumor-adjacent expression validation.

### Does Not Establish

- Prospective clinical utility or improved survival from ERAS-guided care.
- Observed benefit from immunotherapy or chemotherapy selected using ERAS.
- A causal tumor-promoting effect inferred from expression or survival association alone.

### Limitations

- Retrospective public data and heterogeneous molecular datasets limit clinical transportability.
- No prospective ERAS-guided intervention was evaluated; public-dataset survival separation and retrospective AUCs do not establish clinical utility.
- The Results text calls seven named GEO cohorts six. Methods and Figure 4C identify the seven named GEO datasets; no pooled participant total is inferred.
- Survival separation was not uniformly significant: Figure 4C reports a log-rank P value of 0.091 for GSE29016. The study should not be cited as significant prognostic validation in every external dataset.
- Immune-infiltration and TMB group differences described in Results include nonsignificant comparisons; computational immune and drug-sensitivity estimates are not observed treatment-response validation.
- The qPCR experiment assesses two genes in 10 paired tissue specimens; it does not validate the complete survival model or establish that these genes cause tumor progression.
- The selected figures were checked for labels, reported values and direction, without reanalysis of participant-level or raw qPCR data.

<a id="citation-use-cases"></a>
## When This Study Is Useful to Cite

### CU1: How many genes are in ERAS?

The signature contains 10 genes selected from 22 survival-associated candidates.

Evidence: [KF1](#kf1)

### CU2: What survival discrimination was reported for the ERAS nomogram?

The nomogram AUCs were 0.740, 0.754 and 0.754 at 1, 3 and 5 years in the reported retrospective evaluation.

Evidence: [KF2](#kf2)

### CU3: What experimental expression check accompanied ERAS?

qPCR compared SERPINH1 and GPR37 in 10 paired LUAD tumor and adjacent tissue samples.

Evidence: [KF3](#kf3)

<a id="citation-boundaries"></a>
## What This Study Should Not Be Cited to Claim

- Prospective clinical utility or improved survival from ERAS-guided care.
- Observed benefit from immunotherapy or chemotherapy selected using ERAS.
- A causal tumor-promoting effect inferred from expression or survival association alone.
- Uniformly significant survival separation in every external dataset or a deduplicated pooled patient denominator.

<a id="evidence-matrix"></a>
## Evidence Matrix

<a id="evidence-kf1"></a>
### A 10-gene ERAS signature was selected from 22 survival-associated genes.

- Context: Univariate Cox screening followed by LASSO in the TCGA development dataset; public-dataset survival evaluation was not uniformly significant.
- Finding: screened prognostic genes: 22; final ERAS genes: 10; named survival datasets: 8 including TCGA
- Evidence Level: Human transcriptomic associations and limited expression validation
- Scope: Retrospective LUAD prognostic signature development and public-dataset evaluation.
- Source Locator: Methods: Dataset source; Results: Model construction and evaluation, Par17; Fig. 4A-C; Supplementary Fig. 1 (MOESM1)
- Evidence refs: [KF1](#kf1)

<a id="evidence-kf2"></a>
### The nomogram combining ERAS, age and stage showed the reported time-specific survival discrimination.

- Context: Retrospective nomogram evaluation; these values are not independent prospective performance estimates.
- Finding: 1-year AUC: 0.740; 3-year AUC: 0.754; 5-year AUC: 0.754
- Evidence Level: Human transcriptomic associations and limited expression validation
- Scope: Reported nomogram discrimination within the analyzed data.
- Source Locator: Results: Independent prognostic analysis of the ERAS and construction of the nomogram, Par18; Fig. 6F-H
- Evidence refs: [KF2](#kf2)

<a id="evidence-kf3"></a>
### SERPINH1 and GPR37 expression was higher in LUAD tumor than paired adjacent tissue in the reported qPCR experiment.

- Context: Expression validation only; it does not test survival prediction or treatment effect.
- Finding: clinical tissue pairs: 10; genes tested in this qPCR comparison: 2: SERPINH1 and GPR37
- Evidence Level: Human transcriptomic associations and limited expression validation
- Scope: Limited tumor-adjacent expression validation.
- Source Locator: Methods: Clinical specimen collection, Par12; Results: Experimental validation, Par23; Supplementary Fig. 3A-B (MOESM3)
- Evidence refs: [KF3](#kf3)

## Q&A

### How many genes are in ERAS?

The signature contains 10 genes selected from 22 survival-associated candidates.

Evidence: KF1

### What survival discrimination was reported for the ERAS nomogram?

The nomogram AUCs were 0.740, 0.754 and 0.754 at 1, 3 and 5 years in the reported retrospective evaluation.

Evidence: KF2

### What experimental expression check accompanied ERAS?

qPCR compared SERPINH1 and GPR37 in 10 paired LUAD tumor and adjacent tissue samples.

Evidence: KF3

### Does ERAS establish benefit from a selected immunotherapy or chemotherapy?

No. The reported survival modeling and two-gene expression experiment do not test whether choosing treatment with ERAS improves outcomes.

Evidence: KF1, KF3

## Concepts & Entities

### Conditions

lung adenocarcinoma

### Methods

LASSO Cox regression, survival analysis, single-cell transcriptomics

### Models & Tools

ERAS

### Genes

SERPINH1, GPR37

## Related Research

- [Advancing predictive markers in lung adenocarcinoma: A machine learning-based immunotherapy prognostic prediction signature](https://drgezhang.com/papers/doi-10-1002-tox-24284.html) — Related LUAD prognostic modeling using a distinct immunotherapy-associated signature; not independent confirmation of ERAS. (DOI: 10.1002/tox.24284)

## Publication & Provenance

- DOI: https://doi.org/10.1007/s10238-024-01439-4
- Publisher: https://link.springer.com/article/10.1007/s10238-024-01439-4
- PMCID: PMC11272744
- Pmc Url: https://pmc.ncbi.nlm.nih.gov/articles/PMC11272744/
- Fulltext Xml Url: https://www.ebi.ac.uk/europepmc/webservices/rest/PMC11272744/fullTextXML
- Figure4 Url: https://link.springer.com/article/10.1007/s10238-024-01439-4/figures/4
- Figure6 Url: https://link.springer.com/article/10.1007/s10238-024-01439-4/figures/6
- Supplementary Fig1 Url: https://media.springernature.com/original/springer-static/esm/art%3A10.1007%2Fs10238-024-01439-4/MediaObjects/10238_2024_1439_MOESM1_ESM.tif
- Supplementary Fig3 Url: https://media.springernature.com/original/springer-static/esm/art%3A10.1007%2Fs10238-024-01439-4/MediaObjects/10238_2024_1439_MOESM3_ESM.tif
- Evidence Basis: Published Europe PMC journal XML and the current Springer Nature full text; Figure 4, Figure 6, Supplementary Figures 1 and 3 were visually inspected on 2026-10-26 for the selected claims. No source images are reproduced on this evidence page.
- Version: Publisher Version of Record, 25 July 2024; Clinical and Experimental Medicine 24, article 169.
- Reporting Discrepancies: The six-versus-seven GEO wording discrepancy is retained explicitly. Figure 4C includes a nonsignificant GSE29016 log-rank comparison; no uniformly significant external validation is asserted.
- License: CC BY 4.0; article-level Rights and permissions.
- Treatment Inference Locator: Methods: Differences in the TME and drug inference; Results: Drug sensitivity analysis; Figures 9-10. IPS and Oncopredict outputs are computational estimates, not a treated validation cohort.

## Evidence-page notice

Author evidence synthesis of a retrospective LUAD prognostic study. Public-dataset evaluation and limited tissue-expression experiments are distinct from prospective clinical utility, causal mechanism or treatment benefit; identified reporting limitations are disclosed below.

## Links

- DOI: https://doi.org/10.1007/s10238-024-01439-4
- HTML page: https://drgezhang.com/papers/doi-10-1007-s10238-024-01439-4.html
- Google Scholar query: https://scholar.google.com/scholar?q=%22Deciphering%20the%20prognostic%20role%20of%20endoplasmic%20reticulum%20stress%20in%20lung%20adenocarcinoma%3A%20integrating%20prognostic%20prediction%20and%20immunotherapy%20strategies%22
- ORCID: https://orcid.org/0000-0002-3116-3246