ATR-FTIR Spectral Fingerprinting of Serum with Machine Learning Enables Single-Measurement Diagnosis and Prognostic Risk Assessment for Sepsis

DOI: https://doi.org/10.62517/jmhs.202605201

Author(s)

Xuejie Wang1,2, Guoqiang Bao2,*

Affiliation(s)

1Department of General Surgery, Shaanxi University of Chinese Medicine, Xianyang, China 2Department of General Surgery, Tangdu Hospital, The Air Force Military Medical University, Xi’an, China

Abstract

Sepsis necessitates rapid, accurate diagnostic tools, as current biomarkers often lack sufficient speed or specificity. This study explores the novel integration of Attenuated Total Reflection Fourier Transform Infrared (ATR-FTIR) spectroscopy with machine learning as a promising strategy to address this gap. We analyzed serum samples from 146 subjects (sepsis patients and controls) using ATR-FTIR and compared 16 ML algorithms. The Extra Trees classifier demonstrated superior performance on an independent test set, achieving an area under the curve (AUC) of 0.899, an F1-score of 0.885, and a notably high recall of 0.958, indicating strong potential for minimizing missed diagnoses. For prognostic stratification, a simple clinical score based on NT-PROBNP, Troponin I, and platelet count effectively identified high-risk patients (28-day mortality of 24% vs. 0% in the low-risk group). Furthermore, a prognostic model based solely on spectral features achieved an AUC of 0.716 using a Naive Bayes classifier. Importantly, significant correlations were established between specific spectral features and established clinical biomarkers of infection (PCT, CRP), organ dysfunction (Scr), and age, providing a plausible biological basis for the spectral findings. This proof-of-concept study demonstrates that serum infrared spectral fingerprinting, enhanced by machine learning, can serve as a rapid, adjunctive tool for both diagnosing sepsis and stratifying patient risk. This approach directly responds to the critical care imperative for technologies that are faster, more informative, and capable of improving early clinical decision-making. Further validation in larger, multicenter cohorts is warranted to translate this potential into clinical practice.

Keywords

Sepsis; ATR-FTIR Serum Fingerprinting; Machine Learning; Diagnosis; Prognostic

References

[1] Fleischmann C, Scherag A, Adhikari N K J, et al. Assessment of Global Incidence and Mortality of Hospital-treated Sepsis. AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE, 2016, 193(3): 259-272. [2] Rudd K E, Johnson S C, Agesa K M, et al. Global, regional, and national sepsis incidence and mortality, 1990-2017: analysis for the Global Burden of Disease Study. LANCET, 2020, 395(10219): 200-211. [3] Fleischmann-Struzek C, Mellhammar L, Rose N, et al. Incidence and mortality of hospital- and ICU-treated sepsis: results from an updated and expanded systematic review and meta-analysis. Intensive Care Medicine, 2020, 46(8): 1552-1562. [4] Weng L, Zeng X Y, Yin P, et al. Sepsis-related mortality in China: a descriptive analysis. Intensive Care Medicine, 2018, 44(7): 1071-1080. [5] Evans L, Rhodes A, Alhazzani W, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Medicine, 2021, 47(11): 1181-1247. [6] Cecconi M, Evans L, Levy M, et al. Sepsis and septic shock. LANCET, 2018, 392(10141): 75-87. [7] Seymour C W, Gesten F, Prescott H C, et al. Time to Treatment and Mortality during Mandated Emergency Care for Sepsis. NEW ENGLAND JOURNAL OF MEDICINE, 2017, 376(23): 2235-2244. [8] Weng L, Xu Y, Yin P, et al. National incidence and mortality of hospitalized sepsis in China. Critical Care (London, England), 2023, 27(1): 84. [9] Bolanaki M, Winning J, Slagman A, et al. Biomarkers Improve Diagnostics of Sepsis in Adult Patients With Suspected Organ Dysfunction Based on the Quick Sepsis-Related Organ Failure Assessment (qSOFA) Score in the Emergency Department. Critical Care Medicine, 2024, 52(6): 887-899. [10] Wang Q, Peng G, Gan L, et al. The Value of Interleukin-10 in the Early Diagnosis of Neonatal Sepsis: A Meta-Analysis. Pediatric Critical Care Medicine: A Journal of the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies, 2021, 22(9): e492-e501. [11] Zaki H A, Bensliman S, Bashir K, et al. Accuracy of procalcitonin for diagnosing sepsis in adult patients admitted to the emergency department: a systematic review and meta-analysis. Systematic Reviews, 2024, 13(1): 37. [12] Dhudasia M B, Benitz W E, Flannery D D, et al. Diagnostic Performance and Patient Outcomes With C-Reactive Protein Use in Early-Onset Sepsis Evaluations. The Journal of Pediatrics, 2023, 256: 98-104.e6. [13] Zhang W, Wang W, Hou W, et al. The diagnostic utility of IL-10, IL-17, and PCT in patients with sepsis infection. Frontiers in Public Health, 2022, 10: 923457. [14] Friedman N, Yochpaz S, Zirkin S, et al. C-reactive protein and the neonatal early-onset sepsis calculator for the diagnosis of neonatal sepsis. European Journal of Clinical Microbiology & Infectious Diseases: Official Publication of the European Society of Clinical Microbiology, 2021, 40(6): 1227-1234. [15] Cong S, Ma T, Di X, et al. Diagnostic value of neutrophil CD64, procalcitonin, and interleukin-6 in sepsis: a meta-analysis. BMC infectious diseases, 2021, 21(1): 384. [16] Baker M J, Faulds K. Fundamental developments in clinical infrared and Raman spectroscopy. Chemical Society Reviews, 2016, 45(7): 1792-1793. [17] Kumar S, Verma T, Mukherjee R, et al. Raman and infra-red microspectroscopy: towards quantitative evaluation for clinical research by ratiometric analysis. Chemical Society Reviews, 2016, 45(7): 1879-1900. [18] Budidha K, Mamouei M, Baishya N, et al. Identification and Quantitative Determination of Lactate Using Optical Spectroscopy-Towards a Noninvasive Tool for Early Recognition of Sepsis. Sensors (Basel, Switzerland), 2020, 20(18): 5402. [19] Montes-Narváez O, García-Juárez M, Beltrán-Pérez G, et al. ATR-FTIR spectroscopy to evaluate serum protein expression in a murine cerebral ischemia model. Spectrochimica Acta. Part A, Molecular and Biomolecular Spectroscopy, 2025, 326: 125261. [20] Ainsworth C. Infrared spectrometry guides cancer treatment. Nature, 2024. [21] Lilo T, Morais C L M, Shenton C, et al. Revising Fourier-transform infrared (FT-IR) and Raman spectroscopy towards brain cancer detection. Photodiagnosis and Photodynamic Therapy, 2022, 38: 102785. [22] Su K Y, Lee W L. Fourier Transform Infrared Spectroscopy as a Cancer Screening and Diagnostic Tool: A Review and Prospects. Cancers, 2020, 12(1): 115. [23] Wang X, Shen X, Sheng D, et al. FTIR spectroscopic comparison of serum from lung cancer patients and healthy persons. Spectrochimica Acta. Part A, Molecular and Biomolecular Spectroscopy, 2014, 122: 193-197. [24] Tian Y, Fan X, Chen K, et al. Optical biomarker analysis for renal cell carcinoma obtained from preoperative and postoperative patients using ATR-FTIR spectroscopy. Spectrochimica Acta. Part A, Molecular and Biomolecular Spectroscopy, 2024, 318: 124426. [25] Eissa T, Leonardo C, Kepesidis K V, et al. Plasma infrared fingerprinting with machine learning enables single-measurement multi-phenotype health screening. Cell Reports. Medicine, 2024, 5(7): 101625. [26] Chechekina O G, Tropina E V, Fatkhutdinova L I, et al. Machine learning assisted rapid approach for quantitative prediction of biochemical parameters of blood serum with FTIR spectroscopy. Spectrochimica Acta. Part A, Molecular and Biomolecular Spectroscopy, 2025, 326: 125283. [27] Eshel Y D, Sharaha U, Beck G, et al. Monitoring the efficacy of antibiotic therapy in febrile pediatric oncology patients with bacteremia using infrared spectroscopy of white blood cells-based machine learning. Talanta, 2024, 270: 125619. [28] Korb E, Bağcıoğlu M, Garner-Spitzer E, et al. Machine Learning-Empowered FTIR Spectroscopy Serum Analysis Stratifies Healthy, Allergic, and SIT-Treated Mice and Humans. Biomolecules, 2020, 10(7): 1058. [29] Ghimire H, Venkataramani M, Bian Z, et al. ATR-FTIR spectral discrimination between normal and tumorous mouse models of lymphoma and melanoma from serum samples. Scientific Reports, 2017, 7(1): 16993. [30] Ghimire H, Garlapati C, Janssen E A M, et al. Protein Conformational Changes in Breast Cancer Sera Using Infrared Spectroscopic Analysis. Cancers, 2020, 12(7): 1708. [31] Baker M J, Hussain S R, Lovergne L, et al. Developing and understanding biofluid vibrational spectroscopy: a critical review. Chemical Society Reviews, 2016, 45(7): 1803-1818. [32] Wu M, Pu K, Jiang T, et al. Early label-free analysis of mitochondrial redox states by Raman spectroscopy predicts septic outcomes. Journal of Advanced Research, 2021, 28: 209-219.