Abstract
One important aim of precision oncology is a personalized treatment of patients. This can be achieved by various biomarkers, especially imaging parameters and gene expression signatures are commonly used. So far, combination approaches are sparse. The aim of the study was to independently validate the prognostic value of the novel positron emission tomography (PET) parameter tumor asphericity (ASP) in non small cell lung cancer (NSCLC) patients and to investigate associations between published gene expression profiles and ASP. This was a retrospective evaluation of PET imaging and gene expression data from three public databases and two institutional datasets. The whole cohort comprised 253 NSCLC patients, all treated with curative intent surgery. Clinical parameters, standard PET parameters and ASP were evaluated in all patients. Additional gene expression data were available for 120 patients. Univariate Cox regression and Kaplan–Meier analysis was performed for the primary endpoint progression-free survival (PFS) and additional endpoints. Furthermore, multivariate cox regression testing was performed including clinically significant parameters, ASP, and the extracellular matrix-related prognostic gene signature (EPPI). In the whole cohort, a significant association with PFS was observed for ASP (p < 0.001) and EPPI (p = 0.012). Upon multivariate testing, EPPI remained significantly associated with PFS (p = 0.018) in the subgroup of patients with additional gene expression data, while ASP was significantly associated with PFS in the whole cohort (p = 0.012). In stage II patients, ASP was significantly associated with PFS (p = 0.009), and a previously published cutoff value for ASP (19.5%) was successfully validated (p = 0.008). In patients with additional gene expression data, EPPI showed a significant association with PFS, too (p = 0.033). The exploratory combination of ASP and EPPI showed that the combinatory approach has potential to further improve patient stratification compared to the use of only one parameter. We report the first successful validation of EPPI and ASP in stage II NSCLC patients. The combination of both parameters seems to be a very promising approach for improvement of risk stratification in a group of patients with urgent need for a more personalized treatment approach.
Similar content being viewed by others
Introduction
Treatment of non-small cell lung cancer (NSCLC) has rapidly changed during the last decade. Targeted therapies and immunotherapy have shown considerable benefit in metastatic stage IV patients1,2. The encouraging results of the PACIFIC trial have established consolidation immunotherapy for stage III patients who received definitive chemoradiation3,4. Due to the aggressive course of NSCLC, several trials are investigating additional targeted or immunotherapeutic approaches even in stage I and II disease. However, patient selection in these earlier stages is pivotal, since a large number of patients do not need further therapeutic escalation or do not benefit from potentially toxic adjuvant therapies. The biggest unmet clinical need for patient stratification concerns patients with stage II disease, where general recommendations reach from observation to platinum based adjuvant treatment, but may also include targeted therapy or immunotherapy.
Tumor asphericity is a measure of spatial irregularity, the asphericity values represents the fractional increase of the considered tumor’s surface area relative to that of a sphere exhibiting the same volume. Various publications have been able to show that the asphericity (ASP) of [18F-]fluorodeoxyglucose (FDG) uptake within primary tumors in staging positron emission tomography (PET) scans is associated with patient outcome5,6,7. ASP cut-off values successfully stratified NSCLC patients at high or low risk for tumor progression, with a large effect size and highest clinical relevance for stage II disease8. Only sparse data is available for biological explanation of the observed association of ASP and patient outcome. A significant correlation with the proliferation marker Ki-67 and a trend for a correlation with the expression of the VEGF receptor have been reported in a small cohort of NSCLC patients9. Another study was able to show a significant association of tumor ASP and EGFR mutations. EGFR mutated tumors exhibited lower ASP values than EGFR wildtype tumors10. However, further insights into the relationship between molecular alterations and ASP are missing so far.
Gene expression is the cell’s central mechanism to control its cellular response and identity. Therefore signatures can be extracted from the transcriptome that encode information about the properties of the cell(s) such as pathway activity, phenotype, cell/tissue type and disease state11,45,46,14. Next to the present state the transcriptome also carries information about the potential of the cells to respond and thus might contain information to predict the outcome of a disease or treatment. In the past 20 years, many disease-specific gene signatures have been identified, with varying prognostic or predictive value15.
The aim of our study was to independently validate the prognostic value of ASP in NSCLC patients and to investigate associations between published pan-cancer or NSCLC gene expression profiles and ASP.
Patients and methods
Data source
Original imaging data, patient characteristics, and follow up was analyzed from two European centers and three public repositories available in The Cancer Imaging archive: NSCLC Radiogenomics, TCGA LUSC/ LUAD and CPTAC LSCC/LUAD16,12,13,14,15,21. Additional gene expression data was available for patients from the Radiogenomics cohort (GEO accession number GSE103584) and from the TCGA cohort (NCI Genomic Data Commons (GDC)).
All patients were treated by primary surgery.
Image acquisition
PET CT imaging for staging was performed prior to surgery. Details of image acquisition for patients from The Cancer Imaging Archive can be found in the original publications17,22. Patients treated at university hospital Magdeburg, Germany were imaged on a Biograph mCT ~ 64 PET/CT (Siemens Medical Solutions Inc., Knoxville, TN, USA). Data acquisition started 65.9 ± 5.7 min (range 57.4–78.0) after injection of (234 ± 11.9) MBq 18F-FDG. Tomographic images were reconstructed using PSF + TOF reconstruction (2 iterations, 21 subsets). Patients treated at Greater Poland Cancer Centre, Poznan, Poland were scanned on a Gemini TF TOF PET/CT (Philips Healthcare, Best, The Netherlands). Data acquisition started 68.5 ± 11.2 min (range 51.3–85.8) after injection of 18F-FDG with mean activity of 373 ± 67.2 MBq. Tomographic images were reconstructed using the BLOB-OS-TF reconstruction (3 iterations, 33 subsets).
Image analysis
The metabolically active part of the primary tumor was delineated in the PET data by an automatic algorithm based on adaptive thresholding considering the local background23,24. The resulting delineation was inspected visually by an experienced observer and corrected manually where this was deemed necessary. This happened in 16 of 253 cases exhibiting only low diffuse tracer accumulation in the respective lesions. For the delineated ROIs the metabolic active tumor volume (MTV), maximum standardized uptake value (SUVmax), total lesion glycolysis (MTV x SUVmean, TLG), and ASP were computed, where ASP was determined as described previously6,25. ROI definition and analysis was performed using the ROVER software, version 3.0.51 (ABX, Radeberg, Germany).
Gene expression analysis
Various previously published gene expression signatures that have been proposed as pan-cancer signatures or in NSCLC were investigated. These signatures included hypoxia, extracellular matrix-related genes, immunotherapy related signatures and others. Details on the respective signatures can be found in the original publications26,22,23,24,25,26,27,28,29,35.
Raw reads for the radiogenomics data set were downloaded from the Sequence Read Archive (SRA, project number PRJNA401995). Reads were mapped to human reference genome GRCh38.p13 using STAR aligner v2.7.9c with GENCODE v38 annotation36. Aligned reads were quantified using featureCounts from subread v2.0.1.
Count data for the TCGA data set were downloaded using the R package TCGAbiolinks v2.25.337. Both data sets were filtered for genes with insufficient counts (> 1 count in whole data set) and subsequently normalized using the R package DESeq2 v1.37.638.
Gene signatures
Genes of the following signatures were used in this study (Supplemental Fig. 1):
Signature | Abbreviation | # Genes | Source |
---|---|---|---|
Tumor inflammation signature | TIS | 18 | |
T-effector and interferon-γ gene signature | TEffInfG | 8 | |
Radiosensitivity index genes | RadSensInd | 10 | |
T cell-inflamed Gene Expression Profile signature | TCellInfGEP | 18 | |
Immune signature for LUAD prognosis | ImmLUADProgn | 18 | |
Response to MAGE-A3 immunotherapeutic | MAGA3ImmTher | 91 | |
Common hypoxia metagene | HypoxMeta | 15 | |
ECM-related prognostic and predictive indicator | EPPI | 29 | |
Hypoxia-induced EMT | HypoxEMT | 17 | |
Small extracellular vesicle-associated signature | ExtraCellVes | 15 |
EPPI risk score (in this paper abbreviated as EPPI) was calculated using the original formular and coefficients published previously26.
To investigate if PET parameters and information of prognostic or predictive gene signatures are correlated with each other, several gene sets that have been investigated in NSCLC patients or as pan-cancer signatures were calculated in a sub-group of patients (radiogenomics and TCGA cohort). To see if the observed correlations are reproducible, both cohorts were investigated separately. The only gene set that showed a reproducible weak negative correlation with ASP in both cohorts was the extracellular matrix-related prognostic and predictive indicator (EPPI), published by Lim and colleagues (Supplementary Fig. 1).
Statistical analysis
Primary clinical endpoint was progression free survival (PFS), defined as absence of occurrence of any disease recurrence (loco-regional or distant) or death. In addition, loco-regional tumor control (LRC), freedom from distant metastases (FFDM), and overall survival (OS) measured from the date of surgery to death and/or event, were analyzed. Patients who did not keep follow-up appointments and for whom information on survival or tumor status therefore was unavailable were censored at the date of last follow-up.
The association of OS, LRC, FFDM, and PFS with clinically relevant parameters (sex, age, histology, T-stage, N-stage, and UICC-stage) as well as quantitative PET parameters and gene signatures was analyzed using univariate Cox proportional hazard regression in which the PET parameters were included as metric and as binarized parameters, respectively. The cutoff values used for binarization were calculated by performing a univariate Cox regression for each measured value. The values leading to the hazard ratio (HR) with the highest significance were used as cutoff. To avoid too small group sizes, only values within the interquartile range were considered as potential cutoff. Cutoff values were separately computed for OS, LRC, FFDM, and PFS. For validation of ASP, a previously published cutoff was applied without using the cutoff optimization method described here.
The probability of survival was computed and rendered as Kaplan–Meier curves. Independence of parameters was analyzed by multivariate Cox regression.
Statistical significance was assumed at a P-value of less than 0.05. Statistical analysis was performed with the R language and environment for statistical computing version 4.2.139.
Results
253 patients with NSCLC and curative intended surgery were analyzed. Most patients were male and UICC stage I. Univariate cox regression revealed a significant association of ASP with PFS and a trend for OS (Table 1). With binarized cut-off values, ASP significantly discriminated between high and low risk patients for the investigated endpoints PFS and LRC and showed a trend for FFDM (see Kaplan Meier plots Fig. 1 and Supplementary Table 1). Other PET parameters with significant association with outcome were, both, SUVmax and MTV, showed an association with LRC. SUVmax also significant a association with PFS.
Since the EPPI risk score was originally established using the TCGA database, we used patients of the radiogenomics cohort only to see if this signature can be validated independently. In this subgroup of 120 patients it was possible to validate the prognostic impact of the signature by univariate cox regression analyses (Table 1). Furthermore, the combined information of EPPI and PET measured tumor asphericity seems to provide additional prognostic information as shown by the Kaplan Meier plots in Fig. 2. Multivariate testing of ASP, EPPI risk score, and clinical parameters revealed a significant association of ASP in the whole cohort and a significant association of EPPI in the sub-group of patients with gene expression data. The results of multivariate testing are shown in Table 2. As a side note, SUVmax showed similar association with outcome upon multivariate testing (Supplementary Table 2).
Since both EPPI and ASP were initially evaluated in stage II disease, further analysis was restricted to this tumor stage. In stage II patients, ASP was the only PET parameter that was significantly associated with PFS (Table 3). Furthermore, after binarization using previously published cut-off values, ASP significantly discriminated low and high risk stage II patients, while other PET parameters did not show a significant discrimination between risk groups, despite cut-off optimization (Supplementary Table 3). Figure 3 shows Kaplan Meier plots for the previously published ASP cutoff and routinely used PET parameters.
In the radiogenomics sub-group of patients with PET and genomic data, the incremental value of additional gene expression information was investigated. Patients stratified according to ASP were further stratified based on their EPPI. This analysis revealed that EPPI seems to have an additional prognostic value, especially in low risk patients according to ASP, where EPPI classification significantly improved risk stratification as shown in Fig. 4. Multivariate testing of clinical parameters, ASP, and EPPI in stage II patients revealed EPPI risk score as the only significant parameter as depicted in Table 4. However, this finding has to be interpreted cautiously due to the low number of only 27 patients in this sub-group. When analyzing all patients with imaging data, ASP remained significantly associated with PFS, as shown in Table 4.
Discussion
We independently validated the novel PET parameter tumor asphericity and the extracellular matrix-related prognostic gene expression signature in NSCLC patients. Both parameters were significantly associated with progression-free survival of surgically treated patients and independent from established clinical parameters. ASP has previously shown most promising stratification potential in stage II patients and we were able to independently validate the published cut-off value in this group. Even EPPI risk score was initially investigated for stage I and II disease and the strong prognostic value of EPPI risk score could successfully be validated in the whole group, but also in stage II patients alone. Most importantly the combination of EPPI risk score and ASP seems to improve patient stratification compared to the use of only one biomarker. This result has to be interpreted cautiously due to the relatively low number of patients with stage II disease and gene expression data. Nonetheless, the observed clinical effect is strong and affects a group of NSCLC patients with an urgent need for more personalized treatment approaches. Recommendations for resected stage II patients include no further treatment or adjuvant platinum based chemotherapy. Additionally, a recently published phase III trial compared the checkpoint-inhibitor pembrolizumab with placebo in resected stage IB-IIIA NSCLC patients40. Additionally, the IMpower010 trial investigated the use of the checkpoint inhibitor atezolizumab in a similar setting and observed a disease-free survival benefit with atezolizumab versus best supportive care after adjuvant chemotherapy in patients with resected stage II–IIIA NSCLC41. Nevertheless, one has to bear in mind that median progression free survival in the trial cohort was about three years in the control arm and the effect of adjuvant pembrolizumab was moderate in the whole study population. Given the relevant toxicities of chemotherapy and immunotherapy, an improvement of patient risk stratification is urgently needed. The combination of gene and imaging data seems a very promising approach. Conventional gene expression data, especially if performed on biopsy specimens, is not able to reflect intratumor heterogeneity and characteristics of the tumor micromilieu. Single-cell profiling experiments have shown a high genetic heterogeneity in NSCLC42. Due to the complexity of single-cell analyses and high-costs, this approach is not affordable on a large scale, yet. PET imaging might therefore be a well-suited modality to assess individual tumor heterogeneity as a complement to conventional gene expression analyses, as shown by our study.
Data on the combined use of gene expression and PET imaging data is sparse and most of the published data has only exploratory character. One publication evaluated the combined use of PET parameters and Thymidylate Synthase Expression in stage IV NSCLC patients. In this group of patients, the authors found a significant correlation of Thymidylate Synthase Expression and TLG. TLG showed the most promising prognostic value regarding several important outcome measures including PFS. However, there did not seem to be added value by the combined use of both parameters43. In patients treated with curative intent, a recent analyses evaluated PET, CT, and genome data of surgically resected NSCLC patients44. In this study PET and CT information failed to predict disease recurrence but some gene expression profiles seemed deliver helpful information regarding disease recurrence. Nonetheless, the study was a retrospective single center analyses that would need further validation. Another interesting study used TCGA and institutional data to develop a PET CT radiomics signature that can predict tumor immune profiles in NSCLC. The resulting model could be successfully validated in an independent dataset and is therefore a promising predictive biomarker for immunotherapies45. Two other publications have investigated the combined use of tumor or circulating immune parameters and PET parameters of resected NSCLC patients46,47. Both publications reported some correlation with immune parameters and an association of PET parameters with clinical outcome. These data are interesting because adjuvant checkpoint inhibition would be a reasonable therapy for selective patients in case of a good predictive marker. Nonetheless, both publications only investigated very basic PET parameters like SUVmax and TLG. Additionally, the observed effect size was not that large and these parameters do not seem to be optimal for individual treatment personalization.
A big strength of our study is the independent validation of both, the ASP PET parameter and the EPPI gene expression signature. The additional finding that these biomarkers are not strongly correlated with each other and potentially provide independent prognostic value indicate that these biomarkers merit further prospective validation, which is a prerequisite for future clinical use. The retrospective nature of the current analyses is a major limitation due to the well-known shortcomings of retrospective data. Additionally, while the multicentric nature of our analyses is principally advantageous in terms of future application in a multicenter setting, it also can be viewed as a limitation: while all patients received surgical treatment, surgical techniques changed over time and might potentially affect outcome of patients. Furthermore, reconstruction algorithms have been shown to affect ASP48. Although in most cases discrepancies were only moderate and can potentially be addressed by smoothing/ re-scaling the data, this has not been investigated in the current analysis.
Overall, our analysis independently validates the strong prognostic value of ASP and shows great potential for treatment personalization when combined with EPPI risk score.
Data availability
The datasets generated and analysed during the current study are available in the Synapse repository, http://www.synapse.org with the synID syn52673953.
References
Herbst, R. S., Morgensztern, D. & Boshoff, C. The biology and management of non-small cell lung cancer. Nature 553, 446–454 (2018).
Tan, P. S., Bilger, M., de Lima, L. G., Acharyya, S. & Haaland, B. Meta-analysis of first-line therapies with maintenance regimens for advanced non-small-cell lung cancer (NSCLC) in molecularly and clinically selected populations. Cancer Med. 6, 1847–1860 (2017).
Antonia, S. J. et al. Durvalumab after chemoradiotherapy in stage III non-small-cell lung cancer. N. Engl. J. Med. 377, 1919–1929 (2017).
Antonia, S. J. et al. Overall survival with durvalumab after chemoradiotherapy in stage III NSCLC. N. Engl. J. Med. 379, 2342–2350 (2018).
Apostolova, I. et al. Quantitative assessment of the asphericity of pretherapeutic FDG uptake as an independent predictor of outcome in NSCLC. BMC Cancer 14, 896 (2014).
Apostolova, I. et al. Asphericity of pretherapeutic tumour FDG uptake provides independent prognostic value in head-and-neck cancer. Eur. Radiol. 24, 2077–2087 (2014).
Zschaeck, S. et al. Prognostic value of baseline [18F]-fluorodeoxyglucose positron emission tomography parameters MTV, TLG and asphericity in an international multicenter cohort of nasopharyngeal carcinoma patients. PLoS One 15, e0236841 (2020).
Rogasch, J. M. M. et al. Validation of independent prognostic value of asphericity of 18F-fluorodeoxyglucose uptake in non-small-cell lung cancer patients undergoing treatment with curative intent. Clin Lung Cancer 21, 264-272.e6 (2020).
Apostolova, I. et al. The asphericity of the metabolic tumour volume in NSCLC: Correlation with histopathology and molecular markers. Eur. J. Nucl. Med. Mol. Imaging 43, 2360–2373 (2016).
Whi, W. et al. Relationship of EGFR mutation to glucose metabolic activity and asphericity of metabolic tumor volume in lung adenocarcinoma. Nucl. Med. Mol. Imaging 54, 175–182 (2020).
Schubert, M. et al. Perturbation-response genes reveal signaling footprints in cancer gene expression. Nat. Commun. 9(1), 20. https://doi.org/10.1038/s41467-017-02391-6 (2018).
Rydenfelt, M., Klinger, B., Klünemann, M. & Blüthgen, N. SPEED2: inferring upstream pathway activity from differential gene expression. Nucleic Acids Res. 48, 307–312 (2020).
Ianevski, A., Giri, A. K. & Aittokallio, T. Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data. Nat. Commun. 13(1), 1246. https://doi.org/10.1038/s41467-022-28803-w (2022).
Subramanian, A. et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. 102(43), 15545–15550. https://doi.org/10.1073/pnas.0506580102 (2005).
Qian, Y. et al. Prognostic cancer gene expression signatures: Current status and challenges. Cells 10, 648 (2021).
Clark, K. et al. The Cancer Imaging Archive (TCIA): Maintaining and operating a public information repository. J. Digit. Imaging 26, 1045–1057 (2013).
Bakr, S., Gevaert, O., Echegaray, S., Ayers, K., Zhou, M., Shafiq, M. et al. Data for NSCLC Radiogenomics Collection [Internet] (2017). The Cancer Imaging Archive. https://wiki.cancerimagingarchive.net/x/W4G1AQ [cited 29 Jun 2022].
Albertina, B., Watson, M., Holback, C., Jarosz, R., Kirk, S., Lee, Y. et al. Radiology Data from The Cancer Genome Atlas Lung Adenocarcinoma [TCGA-LUAD] collection [Internet]. The Cancer Imaging Archive (2016). https://wiki.cancerimagingarchive.net/x/wgBp [cited 29 Jun 2022].
Kirk, S., Lee, Y., Kumar, P., Filippini, J., Albertina, B., Watson, M. et al. Radiology Data from The Cancer Genome Atlas Lung Squamous Cell Carcinoma [TCGA-LUSC] collection [Internet]. The Cancer Imaging Archive (2016). https://wiki.cancerimagingarchive.net/x/pAD1 [cited 29 Jun 2022].
National Cancer Institute Clinical Proteomic Tumor Analysis Consortium (CPTAC). Radiology Data from the Clinical Proteomic Tumor Analysis Consortium Lung Squamous Cell Carcinoma [CPTAC-LSCC] Collection [Internet]. The Cancer Imaging Archive (2018). https://wiki.cancerimagingarchive.net/x/WAIGAg [cited 29 Jun 2022].
National Cancer Institute Clinical Proteomic Tumor Analysis Consortium (CPTAC). Radiology Data from the Clinical Proteomic Tumor Analysis Consortium Lung Adenocarcinoma [CPTAC-LUAD] collection [Internet]. The Cancer Imaging Archive (2018). https://wiki.cancerimagingarchive.net/x/XQIGAg [cited 29 Jun 2022].
Cancer Genome Atlas Research Network et al. Distinct patterns of somatic genome alterations in lung adenocarcinomas and squamous cell carcinomas. Nat. Genet. 48, 607–616 (2016).
Hofheinz, F. et al. Automatic volume delineation in oncological PET. Evaluation of a dedicated software tool and comparison with manual delineation in clinical data sets. Nuklearmedizin 51, 9–16 (2012).
Hofheinz, F. et al. An automatic method for accurate volume delineation of heterogeneous tumors in PET. Med. Phys. 40, 082503 (2013).
Hofheinz, F. et al. Increased evidence for the prognostic value of primary tumor asphericity in pretherapeutic FDG PET for risk stratification in patients with head and neck cancer. Eur. J. Nucl. Med. Mol. Imaging 42, 429–437 (2015).
Lim, S. B., Tan, S. J., Lim, W.-T. & Lim, C. T. An extracellular matrix-related prognostic and predictive indicator for early-stage non-small cell lung cancer. Nat. Commun. 8, 1734 (2017).
Scott, J. G. et al. A genome-based model for adjusting radiotherapy dose (GARD): A retrospective, cohort-based study. Lancet Oncol. 18, 202–211 (2017).
Danaher, P. et al. Pan-cancer adaptive immune resistance as defined by the Tumor Inflammation Signature (TIS): Results from The Cancer Genome Atlas (TCGA). J. Immunother. Cancer 6, 63 (2018).
Buffa, F. M., Harris, A. L., West, C. M. & Miller, C. J. Large meta-analysis of multiple cancers reveals a common, compact and highly prognostic hypoxia metagene. Br. J. Cancer 102, 428–435 (2010).
Cristescu, R. et al. Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy. Science 362, eaar3593 (2018).
Wang, L. et al. A gene expression-based immune signature for lung adenocarcinoma prognosis. Cancer Immunol. Immunother. 69, 1881–1890 (2020).
Fehrenbacher, L. et al. Atezolizumab versus docetaxel for patients with previously treated non-small-cell lung cancer (POPLAR): A multicentre, open-label, phase 2 randomised controlled trial. Lancet 387, 1837–1846 (2016).
Ulloa-Montoya, F. et al. Predictive gene signature in MAGE-A3 antigen-specific cancer immunotherapy. J. Clin. Oncol. 31, 2388–2395 (2013).
Chen, Y.-L. et al. A 17 gene panel for non-small-cell lung cancer prognosis identified through integrative epigenomic-transcriptomic analyses of hypoxia-induced epithelial-mesenchymal transition. Mol. Oncol. 13, 1490–1502 (2019).
Cao, B. et al. An EV-associated gene signature correlates with hypoxic microenvironment and predicts recurrence in lung adenocarcinoma. Mol. Ther. Nucleic Acids 17, 879–890 (2019).
Dobin, A. et al. STAR: Ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Colaprico, A. et al. TCGAbiolinks: An R/Bioconductor package for integrative analysis of TCGA data. Nucleic Acids Res. 44, e71 (2016).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing.
O’Brien, M. et al. Pembrolizumab versus placebo as adjuvant therapy for completely resected stage IB–IIIA non-small-cell lung cancer (PEARLS/KEYNOTE-091): An interim analysis of a randomised, triple-blind, phase 3 trial. Lancet Oncol. 23(10), 1274–1286 (2022).
Felip, E. et al. Adjuvant atezolizumab after adjuvant chemotherapy in resected stage IB-IIIA non-small-cell lung cancer (IMpower010): A randomised, multicentre, open-label, phase 3 trial. Lancet 398, 1344–1357 (2021).
Wu, F. et al. Single-cell profiling of tumor heterogeneity and the microenvironment in advanced non-small cell lung cancer. Nat. Commun. 12, 2540 (2021).
Moon, S. H. et al. Predictive and prognostic value of 18F-fluorodeoxyglucose uptake combined with thymidylate synthase expression in patients with advanced non-small cell lung cancer. Sci. Rep. 9, 12215 (2019).
Kirienko, M. et al. Radiomics and gene expression profile to characterise the disease and predict outcome in patients with lung cancer. Eur. J. Nucl. Med. Mol. Imaging 48, 3643–3655 (2021).
Tong, H. et al. A Machine learning model based on PET/CT radiomics and clinical characteristics predicts tumor immune profiles in non-small cell lung cancer: A retrospective multicohort study. Front. Immunol. 13, 859323 (2022).
Mitchell, K. G. et al. 18F-fluorodeoxyglucose positron emission tomography correlates with tumor immunometabolic phenotypes in resected lung cancer. Cancer Immunol. Immunother. 69, 1519–1534 (2020).
Grizzi, F. et al. Independent expression of circulating and tissue levels of PD-L1: Correlation of clusters with tumor metabolism and outcome in patients with non-small cell lung cancer. Cancer Immunol. Immunother. 68, 1537–1545 (2019).
Rogasch, J. M. M. et al. Asphericity of tumor FDG uptake in non-small cell lung cancer: Reproducibility and implications for harmonization in multicenter studies. EJNMMI Res. 10, 134 (2020).
Funding
Open Access funding enabled and organized by Projekt DEAL.
Author information
Authors and Affiliations
Contributions
Study conception and design: S.Z., B.K. and F.H.; Drafting of manuscript: S.Z., B.K. and F.H.; Image processing and analysis: S.Z. ,F.H.; Gene expression data processing and analysis: B.K.; Interpretation of data: all authors; Final approval of manuscript: all authors.
Corresponding author
Ethics declarations
Competing interests
Holger Amthauer declares research grants, travel grants, and lecture fees from Sirtex Medical Europe; Dr. Amthauer confirms that none of the above funding sources were involved in the preparation of this paper. All other authors have nothing to disclose.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Zschaeck, S., Klinger, B., van den Hoff, J. et al. Combination of tumor asphericity and an extracellular matrix-related prognostic gene signature in non-small cell lung cancer patients. Sci Rep 13, 20840 (2023). https://doi.org/10.1038/s41598-023-46405-4
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-023-46405-4
- Springer Nature Limited