Matematik tabanlı veri analizi ile özellik seçimi ve sınıflandırma: Akciğer kanseri evrelerinin belirlenmesi

Abstract

Arka Plan—Akciğer kanseri, dünya genelinde kansere bağlı ölümlerin önde gelen nedenlerinden biri olup, hastalığın prognozu büyük ölçüde tanı anındaki evreye bağlıdır. Erken evrede teşhis edilen vakalarda sağkalım oranları önemli ölçüde artarken, mevcut tanı yöntemleri akciğer adenokarsinomunun moleküler heterojenliğini yeterince yansıtamamaktadır. Yüksek kapasiteli dizileme teknolojilerindeki gelişmeler, genomik ve klinik verilerin entegrasyonunu mümkün kılarak kanser evreleme ve prognoz tahmini için yeni fırsatlar sunmaktadır. Bununla birlikte, çoklu veri kaynaklarını etkin biçimde birleştiren ve yüksek doğruluk ile yorumlanabilirlik sağlayan hesaplamalı yaklaşımlara olan ihtiyaç devam etmektedir.

Amaç—Bu çalışmanın amacı, gen ekspresyon verileri ve klinik bilgileri kullanarak erken evre (Evre I–II) ve ileri evre (Evre III–IV) akciğer adenokarsinomu ayrımını gerçekleştiren bütünleşik bir hesaplamalı çerçeve geliştirmektir. Çalışma ile hem sınıflandırma performansının artırılması hem de hastalık progresyonuna ilişkin biyolojik olarak anlamlı çıkarımlar elde edilmesi hedeflenmektedir. Ayrıca, önerilen modelin klinik karar destek sistemlerine entegre edilerek erken teşhis ve tedavi planlamasına katkı sağlaması amaçlanmaktadır.

Yöntem—Çalışmada, TCGA-LUAD veri setinden elde edilen 193 hastaya ait RNA-seq gen ekspresyon verileri ve klinik bilgiler kullanılmıştır. Önerilen yaklaşım; ANOVA tabanlı özellik seçimi ile ayırt edici genlerin belirlenmesini, protein-protein etkileşim (PPI) ağı analizi ile biyolojik yapıların incelenmesini ve Cox orantılı Hazard modeli ile sağkalım analizinin gerçekleştirilmesini içermektedir. Ardından lojistik regresyon, destek vektör makineleri (SVM), rastgele orman ve gradyan artırma algoritmaları kullanılarak sınıflandırma işlemi yapılmıştır. Model performansı, tabakalı 5 katlı çapraz doğrulama yöntemi ile doğruluk, AUC-ROC, duyarlılık ve özgüllük metrikleri kullanılarak değerlendirilmiştir.

Bulgular—Lojistik regresyon modeli %94.9 doğruluk ve %98.6 AUC-ROC değeri ile en yüksek performansı göstermiş, SVM modeli ise benzer sonuçlar elde etmiştir. Sağkalım analizi, ileri evre hastalığın anlamlı derecede daha yüksek mortalite riski ile ilişkili olduğunu ortaya koymuştur (HR = 6.50; p < 0.001). PPI ağı analizi sonucunda TP53, EGFR, KRAS, AKT1 ve PIK3CA genlerinin ağda merkezi rol oynadığı belirlenmiştir. Elde edilen sonuçlar, genomik ve klinik verilerin entegrasyonunun tek modaliteli yaklaşımlara kıyasla daha yüksek sınıflandırma başarısı sağladığını göstermektedir.

Sonuç—Bu çalışma, akciğer adenokarsinomu evreleme ve prognoz tahmini için etkili ve yorumlanabilir birçok modlu yaklaşım sunmaktadır. Gen ekspresyon verileri ile klinik değişkenlerin birlikte kullanılması hem tahmin doğruluğunu hem de biyolojik anlamlılığı artırmaktadır. Önerilen çerçeve, klinik karar destek sistemlerine katkı sağlayabilecek potansiyele sahiptir. Gelecek çalışmalarda, bağımsız veri setleri ile doğrulama yapılması ve farklı veri türlerinin entegrasyonu önerilmektedir.

Keywords

• Akciğer adenokarsinomu
• TCGA-LUAD
• makine öğrenmesi
• gen ekspresyon analizi
• sağkalım analizi
• erken teşhis

Project Number

0

Feature selection and classification with mathematical data analysis: Determining lung cancer stages

Abstract

Context—Lung cancer is one of the leading causes of cancer-related mortality worldwide, with prognosis highly dependent on the stage at diagnosis. Early-stage detection significantly improves survival rates; however, conventional diagnostic methods often fail to capture the molecular heterogeneity of lung adenocarcinoma. Recent advances in high-throughput sequencing technologies have enabled the integration of genomic and clinical data, offering new opportunities for improving cancer staging and prognosis prediction. Despite these advancements, there remains a need for robust computational frameworks that effectively combine multi-source data to enhance classification accuracy and interpretability.

Objective—This study aims to develop an integrated computational framework for distinguishing early-stage (Stage I–II) from advanced-stage (Stage III–IV) lung adenocarcinoma using gene expression and clinical data. The primary objective is to improve classification performance while providing biologically meaningful insights into disease progression. Furthermore, it is expected that the developed model will be incorporated into clinical decision support systems.

Method—The proposed approach utilizes RNA-seq gene expression data and clinical information from 193 patients in the TCGA-LUAD dataset. A multi-step methodology is employed, including ANOVA-based feature selection to identify discriminative genes, followed by protein–protein interaction (PPI) network analysis to reveal key biological structures. Survival analysis is conducted using the Cox proportional Hazards model to evaluate prognostic factors. Subsequently, machine learning models, including logistic regression, support vector machines (SVM), random forest, and gradient boosting, are applied for classification. Model performance is evaluated using stratified 5-fold cross-validation and metrics such as accuracy, AUC-ROC, sensitivity, and specificity.

Results—The logistic regression model achieved the highest performance, with an accuracy of 94.9% and an AUC-ROC of 98.6%, followed closely by the SVM model. Survival analysis confirmed the strong prognostic significance of disease stage, with advanced-stage patients showing significantly higher mortality risk (HR = 6.50, p < 0.001). PPI network analysis identified TP53, EGFR, KRAS, AKT1, and PIK3CA as central hub genes, indicating their critical roles in disease progression. The integration of genomic and clinical data significantly improved classification performance compared to single-modality approaches.

Conclusion—The findings demonstrate that the proposed multimodal framework provides an effective and interpretable approach for lung adenocarcinoma staging and prognosis prediction. The integration of gene expression profiles with clinical variables enhances both predictive accuracy and biological relevance. This approach has strong potential for supporting clinical decision-making and early diagnosis. Future work should focus on external validation using independent datasets and the incorporation of additional data modalities to further improve model robustness.

Keywords

• Lung adenocarcinoma
• TCGA-LUAD
• machine learning
• gene expression analysis
• survival analysis
• early diagnosis

Supporting Institution

Istanbul Esenyurt University

Project Number

0

References

1. H. Sung, J. Ferlay, R. L. Siegel, M. Laversanne, I. Soerjomataram, A. Jemal, F. Bray, “Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries”, CA-A Cancer Journal for Clinicians, 71(3), 209–249, 2021. https://doi.org/10.3322/caac.21660.
2. R. L. Siegel, K. D. Miller, N. S. Wagle, A. Jemal, “Cancer statistics, 2023”, CA-A Cancer Journal for Clinicians, 73(1), 17–48, 2023, https://doi.org/10.3322/caac.21763.
3. World Health Organization, “Global Cancer Observatory: Cancer Today”, Lyon, France: International Agency for Research on Cancer, 2022, https://gco.iarc.who.int/en.
4. W. D. Travis, et al., “The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification”, Journal of Thoracic Oncology, 10(9), 1243–1260, 2015. https://doi.org/10.1097/JTO.0000000000000630.
5. P. Goldstraw, K. Chansky, J. Crowley, R. Rami-Porta, H. Asamura, W. E. E. Eberhardt, A. G. Nicholson, P. Groome, A. Mitchell, V. Bolejack, “The IASLC lung cancer staging project: Proposals for revision of the TNM stage groupings in the forthcoming (eighth) edition of the TNM classification for lung cancer”, Journal of Thoracic Oncology, 11(1), 39–51, 2016. https://doi.org/10.1016/j.jtho.2015.09.009.
6. N. Howlader, et al., “SEER Cancer Statistics Review, 1975–2018”, Bethesda, MD, USA, National Cancer Institute, 2021, https://seer.cancer.gov/csr/1975_2018/.
7. B. C. Bade, C. S. Dela Cruz, “Lung cancer 2020: Epidemiology, etiology, and prevention”, Clinics in Chest Medicine, 41(1), 1–24, 2020. https://doi.org/10.1016/j.ccm.2019.10.001.
8. P. J. Mazzone, G. A. Silvestri, L. H. Souter, T. J. Caverly, J. P. Kanne, H. A. Katki, R. S. Wiener, F. C. Detterbeck, C. Frank, “Screening for lung cancer: CHEST guideline and expert panel report”, Chest, 160(5), e427–e494, 2021. https://doi.org/10.1016/j.chest.2021.06.063.

9. D. R. Aberle, et al., “Reduced lung-cancer mortality with low-dose computed tomographic screening”, The New England Journal of Medicine, 365(5), 395–409, 2011. https://doi.org/10.1056/NEJMoa1102873.
10. H. J. de Koning, et al., “Reduced lung-cancer mortality with volume CT screening in a randomized trial,” The New England Journal of Medicine, 382(6), 503–513, 2020. https://doi.org/10.1056/NEJMoa1911793.
11. E. J. Jordan, et al., “Prospective comprehensive molecular characterization of lung adenocarcinomas for efficient patient matching to approved and emerging therapies”, Cancer Discovery, 7(6), 596–609, 2017. https://doi.org/10.1158/2159-8290.CD-16-1337.
12. E. A. Collisson, et al., “Comprehensive molecular profiling of lung adenocarcinoma”, Nature, 511(7511), 543–550, 2014. https://doi.org/10.1038/nature13385.
13. F. Skoulidis, J. V. Heymach, “Co-occurring genomic alterations in non-small-cell lung cancer biology and therapy”, Nature Reviews Cancer, 19(9), 495–509, 2019. https://doi.org/10.1038/s41568-019-0179-8.
14. N. Liu, G. Han, Q. Gu, Y. Zhang, M. Chen, “A new era of precision diagnosis and treatment for lung cancer: Artificial intelligence-driven multimodal data integration and clinical applications”, Cell Death & Disease, 17, 534, 2026. https://doi.org/10.1038/s41419-026-08769-z.
15. X. Huang, X. Huang, K. Wang, H. Bai, X. Lu, G. Jin, “2.5D deep learning radiomics and clinical data for predicting occult lymph node metastasis in lung adenocarcinoma”, BMC Medical Imaging, 25, 225, 2025. https://doi.org/10.1186/s12880-025-01759-1.
16. Y. Qin, et al., “A review of deep learning-based multimodal data integration of lung cancer”, Clinical and Experimental Medicine, 26, 275, 2026, https://doi.org/10.1007/s10238-026-02184-6.
17. P. S. Hammerman, et al., “Comprehensive genomic characterization of squamous cell lung cancers”, Nature, 489(7417), 519–525, 2012, https://doi.org/10.1038/nature11404.
18. Y. Farhang, “Hybrid optimization for K-means clustering learning enhancement”, Ph.D. dissertation, Universiti Teknologi Malaysia, Johor, Malaysia, 2016.
19. H. Fattahi, Y. Farhang, “A novel deep learning model for diabetes diagnosis”, International Symposium on AI-Driven Engineering Systems (ISADES2025), Tokat, Türkiye, 19 June 2025, 124-127. https://doi.org/10.36287/setsci.22.1.001.
20. A. Soleimany, Y. Farhang, A. Babazadeh Sangar, “Hierarchical federated learning model for traffic light management in future smart cities”, International Journal of Nonlinear Analysis and Applications, 14(12), 175–186, 2023. https://doi.org/10.22075/ijnaa.2022.29398.4150.
21. A. Cheerla, O. Gevaert, “Deep learning with multimodal representation for pancancer prognosis prediction”, Bioinformatics, 35(14), i446–i454, 2019. https://doi.org/10.1093/bioinformatics/btz342.
22. A.-L. Barabási, N. Gulbahce, J. Loscalzo, “Network medicine: A network-based approach to human disease”, Nature Reviews Genetics, 12(1), 56–68, 2011. https://doi.org/10.1038/nrg2918.
23. J. Menche, A. Sharma, M. Kitsak, S. D. Ghiassian, M. Vidal, J. Loscalzo, A. L. Barabási, “Uncovering disease-disease relationships through the incomplete interactome”, Science, 347(6224), 1257601, 2015, https://doi.org/10.1126/science.1257601.
24. M. Brede, “Networks—An introduction. Mark E. J. Newman (2010, Oxford University Press)”, Artificial Life, 18(2), 241–242, 2012, https://doi.org/10.1162/artl_r_00062.
25. D. R. Cox, “Regression models and life-tables”, Journal of the Royal Statistical Society: Series B, 34(2), 187–220, 1972, https://www.jstor.org/stable/2985181.
26. F. E. Harrell, K. L. Lee, D. B. Mark, “Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors”, Statistics in Medicine, 15, 361–387, 1996. https://doi.org/10.1002/(SICI)1097-0258(19960229)15:4%3C361::AID-SIM168%3E3.0.CO;2-4.
27. A. R. A. Anderson, V. Quaranta, “Integrative mathematical oncology”, Nature Reviews Cancer, 8, 227–234, 2008. https://doi.org/10.1038/nrc2329.
28. G. Carlsson, “Topology and data”, Bulletin of the American Mathematical Society, 46(2), 255–308, 2009. https://doi.org/10.1090/S0273-0979-09-01249-X.
29. H. Edelsbrunner, J. Harer, Computational Topology: An Introduction, Providence, RI, USA, American Mathematical Society, 2010, https://doi.org/10.1090/mbk/069.
30. E. Somasundaram, A. Litzler, R. Wadhwa, R. Barker-Clarke, J. Scott, “Persistent homology of tumor CT scans is associated with survival in lung cancer”, Medical Physics, 48(11), 7043-7051, 2021. https://doi.org/10.1002/mp.15255.
31. A. Takiyama, T. Teramoto, H. Suzuki, K. Yamashiro, S. Tanaka, “Persistent homology index as a robust quantitative measure of immunohistochemical scoring,” Scientific Reports, 7, 14002, 2017. https://doi.org/10.1038/s41598-017-14392-y.
32. D. Szklarczyk, et al., “STRING v11: Protein–protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets”, Nucleic Acids Research, 47(D1), D607–D613, 2019. https://doi.org/10.1093/nar/gky1131.
33. U. Brandes, “A faster algorithm for betweenness centrality,” The Journal of Mathematical Sociology, 25(2), 163–177, 2001. https://doi.org/10.1080/0022250X.2001.9990249.
34. P. Bonacich, “Power and centrality: A family of measures”, American Journal of Sociology, 92(5), 1170–1182, 1987. https://doi.org/10.1086/228631.
35. J. Pelletier, G. Thomas, S. Volarević, “Ribosome biogenesis in cancer: New players and therapeutic avenues”, Nature Reviews Cancer, 18, 51–63, 2018. https://doi.org/10.1038/nrc.2017.104.
36. W. H. Fridman, F. Pagès, C. Sautès-Fridman, J. Galon, “The immune contexture in human tumours: Impact on clinical outcome”, Nature Reviews Cancer, 12, 298–306, 2012. https://doi.org/10.1038/nrc3245.
37. M. Huarte, “The emerging role of lncRNAs in cancer”, Nature Medicine, 21, 1253–1261, 2015. https://doi.org/10.1038/nm.3981.
38. M. Olivier, M. Hollstein, P. Hainaut, “TP53 mutations in human cancers: Origins, consequences, and clinical use”, Cold Spring Harbor Perspectives in Biology, 2, a001008, 2010. https://doi.org/10.1101/cshperspect.a001008.
39. R. S. Herbst, D. Morgensztern, C. Boshoff, “The biology and management of non-small cell lung cancer”, Nature, 553, 446–454, 2018. https://doi.org/10.1038/nature25183.
40. F. Sanchez-Vega, et al., “Oncogenic signaling pathways in The Cancer Genome Atlas”, Cell, 173(2), 321–337, 2018. https://doi.org/10.1016/j.cell.2018.03.035.
41. F. Skoulidis, et al., “Sotorasib for lung cancers with KRAS p.G12C mutation”, New England Journal of Medicine, 384(25), 2371–2381, 2021. https://doi.org/10.1056/NEJMoa2103695.
42. E. J. Topol, “High-performance medicine: The convergence of human and artificial intelligence”, Nature Medicine, 25, 44–56, 2019. https://doi.org/10.1038/s41591-018-0300-7.
43. J. C. M. Wan, C. Massie, J. Garcia-Corbacho, F. Mouliere, J. D. Brenton, C. Caldas, S. Pacey, R. Baird, N. Rosenfeld, “Liquid biopsies come of age: Towards implementation of circulating tumour DNA”, Nature Reviews Cancer, 17, 223–238, 2017. https://doi.org/10.1038/nrc.2017.7.
44. M. Roa, L. Daza, M. Escobar, A. Castillo, P. Arbelaez, “SAMA: Spatially-aware multimodal network with attention for early lung cancer diagnosis”, Proceedings of the 11th International Workshop, ML-CDS 2021, Strasbourg, France, 01 October 2021, 48–58. https://doi.org/10.1007/978-3-030-89847-2_5.
45. D. Cohen-Steiner, H. Edelsbrunner, J. Harer, “Stability of persistence diagrams”, Proceedings of the Twenty-First Annual Symposium on Computational Geometry, Pisa, Italy, 06-08 June 2005, 263–271. https://doi.org/10.1145/1064092.1064133.
46. B. J. Stolz, J. Dhesi, J. A. Bull, H. A. Harrington, H. M. Byrne, I. H. R. Yoon, “Relational persistent homology for multispecies data with application to the tumor microenvironment”, Bulletin of Mathematical Biology, 86, 128, 2024. https://doi.org/10.1007/s11538-024-01353-6.
47. D. Brito-Pacheco, P. Giannopoulos, C. C. Reyes-Aldasoro, “Persistent homology in medical image processing: A literature review”, ACM Transactions on Computing for Healthcare, to be published, https://doi.org/10.1145/3820762.
48. Y. Cai, L. Ke, A. Du, J. Dong, Z. Gai, L. Gao, X. Yang, H. Han, M. Du, G. Qiang, L. Wang, B. Wei, Y. Fan, Y. Wang, “Machine learning-assisted multimodal early screening of lung cancer based on a multiplexed laser-induced graphene immunosensor”, ACS Nano, 19(28), 25697–25709, 2025. https://doi.org/10.1021/acsnano.5c02822.
49. A. W. F. Edwards, “R.A. Fischer, statistical methods for research workers, first edition (1925)”, Landmark Writings in Western Mathematics 1640–1940, I. Grattan-Guinness, R. Cooke, L. Corry, P. Crépel, and N. Guicciardini, Eds., Amsterdam, Elsevier, 2005, 856–870. https://doi.org/10.1016/B978-044450871-3/50148-0.
50. Y. Benjamini, Y. Hochberg, “Controlling the false discovery rate: A practical and powerful approach to multiple testing”, Journal of the Royal Statistical Society: Series B, 57(1), 289–300, 1995. https://doi.org/10.1111/j.2517-6161.1995.tb02031.x.
51. F. Pedregosa, et al., “Scikit-learn: Machine learning in Python”, Journal of Machine Learning Research, 12, 2825–2830, 2011.