Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Nov 18;24(1):1417.
doi: 10.1186/s12885-024-13190-w.

Artificial intelligence-based personalized survival prediction using clinical and radiomics features in patients with advanced non-small cell lung cancer

Junji Koyama  1 , Masahiro Morise  2 , Taiki Furukawa  3 , Shintaro Oyama  4   5 , Reiko Matsuzawa  1 , Ichidai Tanaka  1 , Keiko Wakahara  1 , Hideo Yokota  5   6 , Tomoki Kimura  7 , Yoshimune Shiratori  3 , Yasuhiro Kondoh  7 , Naozumi Hashimoto  8 , Makoto Ishii  1
Affiliations

Artificial intelligence-based personalized survival prediction using clinical and radiomics features in patients with advanced non-small cell lung cancer

Junji Koyama et al. BMC Cancer. .

Abstract

Background: Multiple first-line treatment options have been developed for advanced non-small cell lung cancer (NSCLC) in each subgroup determined by predictive biomarkers, specifically driver oncogene and programmed cell death ligand-1 (PD-L1) status. However, the methodology for optimal treatment selection in individual patients is not established. This study aimed to develop artificial intelligence (AI)-based personalized survival prediction model according to treatment selection.

Methods: The prediction model was built based on random survival forest (RSF) algorithm using patient characteristics, anticancer treatment histories, and radiomics features of the primary tumor. The predictive accuracy was validated with external test data and compared with that of cox proportional hazard (CPH) model.

Results: A total of 459 patients (training, n = 299; test, n = 160) with advanced NSCLC were enrolled. The algorithm identified following features as significant factors associated with survival: age, sex, performance status, Brinkman index, comorbidity of chronic obstructive pulmonary disease, histology, stage, driver oncogene status, tumor PD-L1 expression, administered anticancer agent, six markers of blood test (sodium, lactate dehydrogenase, etc.), and three radiomics features associated with tumor texture, volume, and shape. The C-index of RSF model for test data was 0.841, which was higher than that of CPH model (0.775, P < 0.001). Furthermore, the RSF model enabled to identify poor survivor treated with pembrolizumab because of tumor PD-L1 high expression and those treated with driver oncogene targeted therapy according to driver oncogene status.

Conclusions: The proposed AI-based algorithm accurately predicted the survival of each patient with advanced NSCLC. The AI-based methodology will contribute to personalized medicine.

Trial registration: The trial design was retrospectively registered study performed in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Nagoya University Graduate School of Medicine (approval: 2020 - 0287).

Keywords: Artificial intelligence; Machine learning; Non-small cell lung cancer; Precision medicine; Random survival forest.

PubMed Disclaimer

Conflict of interest statement

Declarations Ethics approval and consent to participate The trial design was retrospectively registered study performed in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Nagoya University Graduate School of Medicine (approval: 2020 − 0287). Because of retrospective design, the requirement to obtain informed consent was waived according to the ethical guidelines for life sciences and medical research involving human subjects of Japan, which was approved by institutional review boards of each participating institutions. Consent for publication Not applicable. Competing interests J Koyama has a patent JP2022-043291 pending. M Morise receives lecture fees from Taiho, Boehringer Ingelheim, Daiichi Sankyo, Eli Lilly, Chugai, Astra Zeneca, Ono, Pfizer, and MSD; is the principal investigator of clinical trials initiated by Chugai, Astra Zeneca, Ono, Pfizer, Merck Serono, Kissei, Taiho, and Novartis; and receives research grants as an investigator to initiate clinical trials for Boehringer Ingelheim and Eli Lilly outside of the submitted work. In addition, M Morise has a patent JP2022-043291 pending; T Furukawa has a patent JP2022-043291 pending; and H Yokota has a patent JP2022-043291 pending. Y Kondoh serves as a consultant to Asahi Kasei, Healios, Shionogi, Boehringer Ingelheim, Janssen, Chugai, and Taiho; received lecture fees from Asahi Kasei, Shionogi, Boehringer Ingelheim, Astra Zeneca, Eisai, Janssen, Bristol Myers Squibb, KYORIN, Mitsubishi Tanabe, Novartis, and Teijin outside of the submitted work. N Hashimoto reports research grants from Boehringer Ingelheim and non-financial support from Astra Zeneca, Glaxo Smith Kline, Novartis, and Boehringer Ingelheim outside of the submitted work. All remaining authors have nothing to disclose.

Figures

Fig. 1
Fig. 1
Flowchart for the training and external tests of survival prediction model. (A) training tests. (B) external tests. Abbreviations: PS, performance status; PD-L1, programmed cell death protein-1; TPS, tumor proportion score; ICI, immune checkpoint inhibitor; TKI, tyrosine kinase inhibitor; CT, computed tomography; PCA, principal component analysis; CPH, cox proportional hazard; RSF, random survival forest
Fig. 2
Fig. 2
External test of the RSF model. (A) Survival function predicted for each patient. (B) Kaplan–Meier analysis according to ERS. (C) Kaplan–Meier analysis of subgroup classified by initial treatment category: cytotoxic chemotherapy, (D) immune checkpoint inhibitor ± cytotoxic chemotherapy, (E) molecular targeted therapy, and (F) best supportive care. The cutoff value for the ERS was 154.6, except for the BSC subgroup. For the BSC subgroup, it was 244.9. Abbreviations: OS, overall survival; RSF, random survival forest; ERS, ensemble risk score
Fig. 3
Fig. 3
Feature importance generated in the RSF model. Abbreviations: LDH, lactate dehydrogenase; WBC, white blood cell; NLR, neutrophil–lymphocyte ratio; CRP, C-reactive protein; PS, performance status; Ccr, creatinine clearance; PC, principal component
Fig. 4
Fig. 4
Individual survival simulation for two patients in the test cohort who received with BSC. (A) A 73-year-old male patient who was expected to have a survival benefit from anticancer therapy. (B) An 81-year-old male patient who was suggested to have little survival benefit from anticancer therapy. Abbreviations: PS, performance status; PD-L1, programmed cell death ligand-1: TPS, tumor proportion score; BSC, best supportive care; ICI, immune checkpoint inhibitor
See this image and copyright information in PMC

References

    1. Soria JC, Ohe Y, Vansteenkiste J, et al. Osimertinib in untreated EGFR-Mutated Advanced Non-small-cell Lung Cancer. N Engl J Med. 2018;378:113–25. - PubMed
    1. Hida T, Nokihara H, Kondo M, et al. Alectinib versus Crizotinib in patients with ALK-positive non-small-cell lung cancer (J-ALEX): an open-label, randomised phase 3 trial. Lancet. 2017;390:29–39. - PubMed
    1. Peters S, Camidge DR, Shaw AT, et al. Alectinib versus Crizotinib in untreated ALK-Positive non-small-cell Lung Cancer. N Engl J Med. 2017;377:829–38. - PubMed
    1. Reck M, Rodríguez-Abreu D, Robinson AG, et al. Pembrolizumab versus Chemotherapy for PD-L1-Positive non-small-cell Lung Cancer. N Engl J Med. 2016;375:1823–33. - PubMed
    1. Herbst RS, Giaccone G, de Marinis F, et al. Atezolizumab for First-Line treatment of PD-L1-Selected patients with NSCLC. N Engl J Med. 2020;383:1328–39. - PubMed

MeSH terms