• Machine learning-based predict...

*Result*: Machine learning-based predictive model for high- grade cytokine release syndrome in chimeric antigen receptor T-cell therapy.

Title:
Machine learning-based predictive model for high- grade cytokine release syndrome in chimeric antigen receptor T-cell therapy.
Authors:
Yu X; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Wang Q; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Halimulati T; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Lv J; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Zhou K; Department of General Surgery, The First Affiliated Hospital, Army Medical University, Chongqing, China., Chen G; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Yin L; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Liu Y; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Bi J; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Xiang Z; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China., Wang Q; Clinical Medicine Research Center, Shandong Second Provincial General Hospital, Jinan, China.
Source:
Frontiers in immunology [Front Immunol] 2025 Nov 20; Vol. 16, pp. 1692892. Date of Electronic Publication: 2025 Nov 20 (Print Publication: 2025).
Publication Type:
Comparative Study; Journal Article; Observational Study; Validation Study
Language:
English
Journal Info:
Publisher: Frontiers Research Foundation] Country of Publication: Switzerland NLM ID: 101560960 Publication Model: eCollection Cited Medium: Internet ISSN: 1664-3224 (Electronic) Linking ISSN: 16643224 NLM ISO Abbreviation: Front Immunol Subsets: MEDLINE
Imprint Name(s):
Original Publication: [Lausanne : Frontiers Research Foundation]
MeSH Terms:
Immunotherapy, Adoptive*/adverse effects , Cytokine Release Syndrome*/etiology , COVID-19*/complications , Machine Learning* , Predictive Learning Models*, Retrospective Studies ; Humans ; Male ; Female ; Aged
References:
J Oral Pathol Med. 2020 Nov;49(10):977-985. (PMID: 32740951)
Front Immunol. 2020 Jun 26;11:1626. (PMID: 32714336)
World J Clin Cases. 2020 Oct 6;8(19):4370-4379. (PMID: 33083396)
Sci Rep. 2019 Aug 14;9(1):11862. (PMID: 31413285)
Front Oncol. 2022 Dec 19;12:938292. (PMID: 36601485)
Trials. 2020 Sep 03;21(1):758. (PMID: 32883328)
Nature. 2020 Aug;584(7821):463-469. (PMID: 32717743)
Front Med (Lausanne). 2023 Jan 09;9:930541. (PMID: 36698845)
Cell. 2021 Jan 7;184(1):149-168.e17. (PMID: 33278357)
Anaesth Rep. 2021 Jan 13;9(1):16-19. (PMID: 33490954)
Case Rep Hematol. 2015;2015:640528. (PMID: 26124967)
Biomed Pharmacother. 2022 Jan;145:112243. (PMID: 34840031)
J Infect Dis. 2020 Oct 1;222(9):1444-1451. (PMID: 32601708)
Crit Rev Oncol Hematol. 2025 Dec;216:104969. (PMID: 41072894)
Sci Rep. 2022 Dec 2;12(1):20805. (PMID: 36460741)
Front Pharmacol. 2022 Oct 14;13:950923. (PMID: 36313336)
Blood Adv. 2021 Jan 26;5(2):593-601. (PMID: 33496754)
Annu Int Conf IEEE Eng Med Biol Soc. 2022 Jul;2022:966-970. (PMID: 36086220)
Diabetes Metab Syndr. 2020 Nov-Dec;14(6):2245-2249. (PMID: 33395786)
Nat Mach Intell. 2020 Jan;2(1):56-67. (PMID: 32607472)
Blood Cancer J. 2021 Apr 6;11(4):69. (PMID: 33824268)
J Cell Mol Med. 2021 Dec;25(24):11063-11074. (PMID: 34734474)
Cancer. 2023 Jul 1;129(13):2035-2046. (PMID: 36991547)
Eur J Radiol. 2020 Nov;132:109336. (PMID: 33069986)
Cold Spring Harb Perspect Biol. 2014 Sep 04;6(10):a016295. (PMID: 25190079)
Front Immunol. 2022 Jan 28;13:807050. (PMID: 35154124)
Int J Infect Dis. 2022 Jul;120:88-95. (PMID: 35487339)
Trials. 2007 Jun 07;8:16. (PMID: 17555582)
J Med Internet Res. 2021 Feb 10;23(2):e24246. (PMID: 33476281)
J Thromb Haemost. 2020 Apr;18(4):844-847. (PMID: 32073213)
Int J Med Sci. 2021 Jan 14;18(5):1207-1215. (PMID: 33526982)
N Engl J Med. 2020 Dec 3;383(23):2255-2273. (PMID: 33264547)
Blood Coagul Fibrinolysis. 2023 Apr 1;34(3):179-183. (PMID: 36966807)
Bioinformatics. 2012 Jan 1;28(1):112-8. (PMID: 22039212)
Mediators Inflamm. 2022 Apr 6;2022:7137900. (PMID: 35431655)
J Inflamm Res. 2023 Dec 11;16:6055-6070. (PMID: 38107380)
F1000Res. 2021 Mar 22;10:228. (PMID: 35350705)
Bone Marrow Transplant. 2021 Jul;56(7):1642-1650. (PMID: 33608658)
Blood Rev. 2021 Jan;45:100707. (PMID: 32425294)
Br J Haematol. 2023 Apr;201(1):86-94. (PMID: 36503182)
Trop Med Infect Dis. 2020 Jul 03;5(3):. (PMID: 32635353)
Exp Biol Med (Maywood). 2022 May;247(9):734-742. (PMID: 35492022)
Lancet Respir Med. 2020 Jun;8(6):e46-e47. (PMID: 32353251)
Bone Marrow Transplant. 2021 Mar;56(3):570-580. (PMID: 32943758)
Sci Rep. 2023 Mar 9;13(1):3940. (PMID: 36894561)
Signal Transduct Target Ther. 2021 Oct 20;6(1):367. (PMID: 34667157)
Signal Transduct Target Ther. 2021 Jul 12;6(1):263. (PMID: 34248142)
PLoS One. 2022 May 19;17(5):e0268296. (PMID: 35588115)
Nat Med. 2020 Oct;26(10):1636-1643. (PMID: 32839624)
J Hematol Oncol. 2014 Jun 12;7:44. (PMID: 24919462)
Contributed Indexing:
Keywords: CAR-T therapy; COVID-19; XGBoost model; cytokine release syndrome; machine learning technique
Entry Date(s):
Date Created: 20251208 Date Completed: 20251215 Latest Revision: 20251215
Update Code:
20260130
PubMed Central ID:
PMC12675352
DOI:
10.3389/fimmu.2025.1692892
PMID:
41357188
Database:
MEDLINE

*Further Information*

*Introduction: The development of robust predictive models for high-grade cytokine release syndrome (CRS) in CAR-T recipients remains limited by sparse clinical trial data.
Methods: We analyzed of 496 COVID-19 patients revealed that CRS plays a pivotal role in disease progression and serves as a valuable data source for understanding CRS progression. Building on this insight, we evaluated and compared the predictive performance of three machine learning models, with the ultimate goal of developing a predictive model for high-grade CRS in patients receiving CAR-T therapy.
Results: Among evaluated algorithms (XGBoost, Random Forest, Logistic Regression), XGBoost demonstrated superior performance in high-grade CRS prediction. Feature importance analysis identified SpO2, D-dimer, diastolic blood pressure, and INR as key predictors, enabling development of a validated riskassessment algorithm. In an independent CAR-T cohort (n=45), the algorithm achieved impressive predictive performance for high-grade CRS prediction.
Discussion: Using machine learning, we identified key clinical biomarkers strongly associated with high-grade CRS. This tool efficiently predicts progression to high-grade CRS post-onset and shows significant potential for clinical deployment in CAR-T therapy.
(Copyright © 2025 Yu, Wang, Halimulati, Lv, Zhou, Chen, Yin, Liu, Bi, Xiang and Wang.)*

*The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.*