An interpretable radiomics-machine learning model for diagnosing invasive fungal infections in community-acquired pneumonia: multicenter study

An interpretable radiomics-machine learning model for diagnosing invasive fungal infections in community-acquired pneumonia: multicenter study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article An interpretable radiomics-machine learning model for diagnosing invasive fungal infections in community-acquired pneumonia: multicenter study Wenzhang He, Yulin Xiong, Xuan Huang, Haoran Luo, Shaoquan Zhou, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8145288/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract Objectives To develop a high-resolution computed tomography (HRCT) radiomics-based interpretable machine learning model for diagnosing invasive fungal infection (IFI) in community-acquired pneumonia (CAP) patient. Methods A total of 570 CAP patients who underwent HRCT from July 2022 to August 2024 in Center 1 and Center 2 were recruited. A vb-net pneumonia automatic segmentation algorithm was employed. Three models, a radiomics model (HRCT-derived radiomics features), a clinical model (clinical variables), and a combined model (integrating both), were developed. The performance of these models was evaluated through receiver operating characteristic analysis with respect to the area under the curve (AUC). Clinical utility was evaluated by using decision curve analysis. The Shapley Additive Explanation tool was employed. Results 239 (mean age: 62.1 ± 19.3 years; 134 male), 101 (mean age: 57.5 ± 17.3 years; 44 male), and 230 (mean age: 68.4 ± 15.3 years; 153 male) patients were included in the training, internal validation, and external validation datasets. Based on linear discriminant analysis classifier, the AUCs of the clinical, radiomics, and combined models were 0.719, 0.724, and 0.808, respectively, in the internal validation dataset; and 0.707, 0.709, and 0.786, respectively, in the external validation dataset. The combined model yielded a superior net benefit relative to both the clinical and radiomics models alone. Age exerted the greatest influence on the predictions of the combined model, while the three most important radiomics features were all higher-order texture features. Conclusions A radiomics-based machine learning model can effectively diagnose IFI in CAP patients, demonstrating favorable interpretability. Clinical relevance statement: The radiomics-based interpretable model developed in this study has significant clinical relevance as it offers a non-invasive method for diagnosing invasive fungal infection in community-acquired pneumonia patients and holds promise as an early diagnostic tool. Biological sciences/Computational biology and bioinformatics Health sciences/Diseases Health sciences/Medical research Machine learning Invasive fungal infections Community-acquired pneumonia Radiomics Interpretability Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Key Points 1. The study addresses the challenge of diagnosing invasive fungal infections in patients with community-acquired pneumonia. 2. The developed radiomics-machine learning model, combining clinical variables and HRCT-derived radiomics features. 3. The model offers a non-invasive, interpretable tool for early invasive fungal infections diagnosis in community-acquired pneumonia patients, potentially reducing diagnostic delays. 1. Introduction Community-acquired pneumonia (CAP) represents a global health priority, ranking among the top five causes of mortality worldwide and imposing substantial socioeconomic burdens ( 1 ). CAP exhibits considerable etiological diversity, with potential involvement of either singular or polymicrobial pathogens, resulting in marked heterogeneity in both clinical course and patient outcomes ( 2 , 3 ). A subset of patients, especially immunodeficient individuals, frequently present with co-occurring invasive fungal infection (IFI) which is associated with substantial morbidity, having a mortality rate of 40% to 50% ( 4 ). The current diagnostic workflow, dependent on culture methods with suboptimal sensitivity (30–50%) and serological tests requiring 6–8 hours processing time, demonstrates significant limitations ( 5 , 6 ). This diagnostic inadequacy creates a critical 72–96 hours window of uncertainty, during which mortality risk escalates ( 5 , 7 ). This underscores the urgent need for advanced diagnostic frameworks capable of rapid, precise pathogen differentiation. High-resolution computed tomography (HRCT) plays a pivotal role in evaluating pulmonary infections due to its ability to provide detailed morphological information ( 8 , 9 ). Characteristic CT manifestations of invasive fungal pneumonia typically include multiple nodules accompanied by peripheral ground-glass opacities and/or wedge-shaped consolidations ( 10 ). Nevertheless, these radiographic features lack specificity and often demonstrate significant overlap with other pulmonary infectious processes. In polymicrobial infections, the radiological manifestations of co-existing pathogens can mask distinctive fungal signatures, resulting in reduced diagnostic recognition. Radiomics, an emerging field in medical imaging, offers a promising solution by extracting high-dimensional quantitative features from medical images and analyzing them by using machine learning algorithms ( 11 ). Radiomic features invisible to the naked eye can reveal differences at the protein, cellular, and tissue levels ( 12 , 13 ). Yan et al. established CT radiomics as a viable predictor of pulmonary invasive fungal infections in immunocompromised hosts ( 14 ). In another research, Yang et al. demonstrated that machine learning-enhanced radiomics improves severe CAP identification ( 15 ). Subsequent multicenter studies have validated radiomics' diagnostic utility in distinguishing pathogens causing pulmonary infections ( 16 – 18 ). However, no machine learning-based system integrating imaging biomarkers with clinical parameters has yet been developed for invasive fungal infection identification and clinical decision support in large, multicenter CAP cohort. By leveraging advanced computational techniques, we seek to identify robust imaging biomarkers that can identify IFI with high precision. This study aims to develop and validate a radiomics-based model for diagnosing IFI in CAP patients based on multicenter datasets. In addition, the study aimed to explore the interpretability of the radiomics model by SHapley Additive exPlanations analysis. 2. Materials and Methods The study was in line with the Declaration of Helsinki and was approved by the Institutional Review Board of Chongqing general hospital, Chongqing University [No. KYS2022-009-01]. 2.1. Patients The inclusion criteria were as follows: (I) age 18 years and over; (II) meeting the diagnostic criteria for CAP; (III) CT scan with thin section (slice thickness less than 1.5 mm); (IV) definitive evidence of pulmonary infiltration/consolidation on HRCT; and (V) pathogen detection tests on the day of hospital admission (including identifying pathogens by deep sputum culture, blood culture, bronchoalveolar lavage culture, histopathology, and macrogenomic sequencing). A total of 1212 CAP patients who underwent HRCT examination from July 2022 to August 2024 in Chongqing General Hospital (Center 1) and Chongqing University Central Hospital (Center 2) were initially recruited. The exclusion criteria were as follows: (I) the interval between CT examination and pathogen specimen testing exceeds 1 days (n = 276); (II) patient data missing exceeds 30% (n = 27); (III) patients who received antibiotic therapy (n = 265); (IV) CT image quality is insufficient for clinical diagnosis (n = 74). The patient selection process is depicted in Fig. 1 . Ultimately, 340 patients from Center 1 comprised the development cohort. These patients were randomly divided into the training dataset (n = 239) and the internal validation dataset (n = 101) in a 7:3 ratio. An independent cohort of 230 patients from Center 2 served as the external validation dataset. 2.2. Volume segmentation of the region of interest (ROI) The detailed HRCT acquisition information is provided in S1 . For lesion segmentation related to pneumonia, a vb-net pneumonia automatic segmentation algorithm was employed, which was developed by United Imaging Intelligence's one-stop research platform (uAI Research Portal, V20230515, https://urp.united-imaging.com/ ) ( 19 ). Previously, in a pneumonia infection segmentation task, it was proven to highly overlap with the manual sketch, and the similarity coefficient of dice was 91.6%. To secure segmentation accuracy, two chest radiologists manually reviewed and adjusted algorithm-generated segmentations (with 9 and 25 years of experience in chest imaging diagnosis, respectively). The experts based their annotations of the boundaries and regions of disease lesions on their clinical skills and medical knowledge. Figure S1 shows the ROI segmentation example. 2.3. Radiomics features extraction and selection Radiomics feature extraction method and feature details are presented in S2 . A total of 2,264 radiomic features were extracted. To mitigate overfitting and reduce model complexity, initial feature selection was performed using the Mann–Whitney U test, retaining only radiomic features that demonstrated a significant association with the endpoint outcome ( P 0.95), the feature with the higher P value from the Mann–Whitney U test was excluded. Finally, the most informative radiomic features for predicting the endpoint outcome were identified using the least absolute shrinkage and selection operator regression. The optimal penalty parameter (lambda) was determined through 10-fold cross-validation. 2.4. Selection of the clinical variables Clinical variables for each patient were extracted from the electronic medical record system, including radiological features (ROI Volume and ROI Average HU Value), epidemiologic factors (sex, age, non-productive cough, hemoptysis, dyspnea, chest pain, confusion, chest tightness, myalgia, duration of symptoms, and fever), and laboratory findings (white blood cell count, neutrophil count, lymphocyte count, neutrophil/lymphocyte ratio, monocyte count, eosinophil count, basophil count, red blood cell count, hemoglobin level, and C-reactive protein level). Univariate logistic regression analysis was performed to evaluate the association between each clinical variable and fungal infection status. Variables demonstrating a significant association ( P < 0.05) were subsequently included in the multivariate regression analysis, and variables with P < 0.05 in the multivariate model were selected for further analysis. 2.5. Model development and evaluation Prior to model development, multiple well-established and widely applied machine learning algorithms were evaluated, including linear discriminant analysis (LDA), k-nearest neighbor, logistic regression, naïve Bayes, random forest, and support vector machine classifiers. These models were compared within the development cohort using 4-fold cross-validation to ensure robustness and generalizability. Selection of the optimal classifier was based on the highest binary classification accuracy observed in the validation folds, while also considering the extent of overfitting, which was assessed by comparing model performance between the training and validation datasets. Based on the selected machine learning classifier, three distinct predictive models were constructed in the training dataset: a clinical model utilizing only the selected clinical variables, a radiomics model incorporating solely the selected radiomic features, and a combined model that integrated both clinical and radiomic features. The performance of these models was evaluated and compared in both the internal and external validation datasets through receiver operating characteristic analysis with respect to the area under the curve (AUC). Additionally, the detailed sensitivity, specificity, positive predictive value, and negative predictive value for each model were computed at the optimal cutoff point, which was determined using the maximum Youden index criterion. Model calibration was assessed using the Brier score, which quantifies the mean squared difference between the predicted probabilities and the actual endpoints (i.e., fungal infection status), with lower values indicating better calibration. Clinical utility was evaluated using decision curve analysis by comparing the net benefits of the models across a range of threshold probabilities. 2.6. SHAP analysis In order to assess the contribution of each feature in a predictive model, the SHapley Additive exPlanations (SHAP) analysis was performed to enhance model interpretability. Based on the concept of Shapley values from cooperative game theory, SHAP quantifies the impact of each feature on model predictions by decomposing the output into additive contributions of individual variables. This approach provides both global insights into overall feature importance and local explanations of specific predictions, offering a transparent, visual interpretation of model behavior. 2.7. Statistical analysis All statistical analyses were performed using SPSS software (version 21.0). One-way analysis of variance was used to compare quantitative variables across groups, and the chi-square test was used for categorical variables. Differences between two AUCs were assessed using the DeLong test. All statistical tests were two-sided, and a P value less than 0.05 was considered indicative of statistical significance. Model development and validation were conducted on the InferScholar platform (version 3.5, Infervision). SHAP analysis and decision curve analysis were conducted using R software (version 3.6.3) with the shapviz and rmda packages, respectively. 3. Results 3.1. Patient characteristics The prevalence of invasive fungal infection was 18.8% (45 of 239) in the training dataset, 18.8% (19 of 101) in the internal validation dataset, and 20.0% (46 of 230) in the external validation dataset, with no significant differences in fungal infection prevalence observed across the three datasets. As summarized in Table 1 , the external validation dataset demonstrated significantly higher values for ROI volume, male prevalence, age, prevalence of productive cough, white blood cell count, and neutrophil-to-lymphocyte ratio compared with the training and internal validation datasets. In contrast, the prevalence of nonproductive cough, chest tightness, myalgia, fever, duration of symptoms, hemoglobin level, and lymphocyte count were significantly lower in the external validation dataset than in the training and internal validation datasets. No significant differences were observed in the remaining clinical variables across the datasets. 3.2. Selection of radiomics features Following significance testing and pearson correlation analysis, 129 radiomics features were retained for least absolute shrinkage and selection operator regression analysis. As shown in Figure 2 , a total of six radiomics features associated with fungal infection, each with a nonzero coefficient, were selected for model development using the optimal tuning parameter (lambda = 0.0335). Heatmap of the selected radiomics features, based on standardized feature values, was presented in Figure S2 . 3.3. Selection of clinical variables A total of 24 clinical variables were collected and analyzed using univariate logistic regression. As shown in Table 2 , ROI volume, dyspnea, age, and sex were significantly associated with invasive fungal infection in the multivariate regression analysis and were subsequently incorporated into the development of the clinical and combined models. 3.4. Machine learning classifier selection The receiver operating characteristic (ROC) analysis of the radiomics models based on LDA, k-nearest neighbor, logistic regression, naïve Bayes, random forest, and support vector machine classifiers in four-fold cross-validation was presented in Figure 3 . The corresponding AUCs for each model were summarized in Table 3 . Model robustness was assessed by comparing AUCs between the training and validation datasets, and only the LDA- and LR-based models demonstrated good robustness without significant overfitting. Notably, the LDA-based model achieved the highest AUC in the validation dataset. Accordingly, the LDA classifier was selected as the optimal machine learning algorithm for constructing the clinical, radiomics, and combined models in this study. 3.5. Model performance evaluation As shown in Figure 4 , receiver operating characteristic (ROC) analysis demonstrated that the areas under the curve (AUCs) for the clinical, radiomics, and combined models were 0.777 (95% CI: 0.719–0.828), 0.765 (95% CI: 0.706–0.817), and 0.836 (95% CI: 0.783–0.881), respectively, in the training dataset; 0.719 (95% CI: 0.621–0.804), 0.724 (95% CI: 0.626–0.808), and 0.808 (95% CI: 0.718–0.880), respectively, in the internal validation dataset; and 0.707 (95% CI: 0.644–0.765), 0.709 (95% CI: 0.646–0.767), and 0.786 (95% CI: 0.728–0.837), respectively, in the external validation dataset. Across all datasets, the combined model demonstrated a significantly higher AUC than both the clinical and radiomics models; however, no significant difference was observed between the clinical and radiomics models. Sensitivity, specificity, positive predictive value, and negative predictive value for the models in the training and validation datasets are summarized in Table 4 . 3.6. Calibration and decision curve analysis Good calibration was observed for all models. The Brier scores for the clinical, radiomics, and combined models were 0.168, 0.166, and 0.152, respectively, in the internal validation dataset, and 0.162, 0.154, and 0.134, respectively, in the external validation dataset. Decision curve analysis in the validation datasets demonstrated that the combined model provided higher net benefit across nearly the entire range of threshold probabilities compared with both the clinical and radiomics models ( Figure 5 ). 3.7. SHAP analysis The feature importance and individual contributions to positive and negative invasive fungal infection predictions in the combined model are illustrated in the summary plot, with the mean SHAP value for each feature also displayed ( Figure 6 ). Age exerted the greatest influence on the predictions of the combined model, while the three most important radiomics features were all higher-order texture features, which are not readily appreciable by visual assessment. 4. Discussion This study developed an integrated machine learning model combining HRCT radiomics features and clinical variables to enable early diagnosis of invasive fungal infection (IFI) in community-acquired pneumonia (CAP). Comparative evaluation of six machine learning classifiers identified the Linear Discriminant Analysis algorithm as demonstrating optimal diagnosing performance and robustness, leading to its selection as our primary modeling framework. Through SHapley Additive exPlanations (SHAP) interpretability analysis, we determined that the model's decision-making was predominantly driven by four key predictors: patient age along with three high-order radiomics features. Our combined model, including clinical variables and radiomics features, showed robust and consistent performance in both internal and external validation datasets. Current gold-standard diagnostic methods for IFI, including histopathological examination and microbial cultures, typically require 3–7 days to yield results ( 7 ). Such diagnostic delays carry grave consequences, as mortality risk increases by 33% for every 24-hour delay in initiating appropriate antifungal therapy ( 20 ). Notably, invasive aspergillosis complicates 19% of severe influenza cases, with associated mortality rates doubling compared to controls ( 21 ). Current guidelines emphasize the importance of diagnosis-driven therapy for high-risk patients with evidence of IFI ( 22 ). However, conventional CT interpretation faces limitations, radiologist agreement for characteristic signs like the halo sign remains suboptimal, and most early IFI cases lack pathognomonic imaging features ( 23 , 24 ). Our model overcomes these challenges by extracting and quantifying subvisual biological information from HRCT images that reflect lesion heterogeneity. Importantly, while polymicrobial infections are common, HRCT provides direct visualization of pulmonary pathology. Our approach leverages existing HRCT data without requiring additional examinations, thereby enabling early IFI diagnosis without imposing additional economic or physiological burdens on patients, while simultaneously avoiding diagnostic delays inherent to conventional methods. Recent retrospective studies have highlighted the diagnostic potential of radiomics in detecting IFI in pulmonary diseases. Most research supports integrated models that combine clinical, radiological, and radiomic features, demonstrating their incremental diagnostic value. For instance, Yan et al. showed that a combined deep learning and radiomics approach effectively differentiated IFIs from bacterial infections in patients with hematologic malignancies, achieving an AUC of 0.844 in internal validation which significantly outperforming radiomics (AUC of 0.767) and clinical models (AUC of 0.696) ( 14 ). Similarly, Gong et al. reported that radiomics could discriminate IFI from bacterial pneumonia with high accuracy (AUC of 0.911 and accuracy of 0.837) in an internal validation cohort ( 16 ). Zhang et al. further validated the efficacy of a combined model integrating clinical, CT radiomic, and deep learning features for predicting invasive pulmonary aspergillosis ( 25 ). In our study, a combined clinical-radiological model exhibited robust diagnostic performance for IFI in CAP. Furthermore, validation in a large external cohort confirmed the model’s strong generalizability and reliability. Advanced age constitutes a major independent risk factor for IFI. Elderly patients typically present with a greater comorbidity burden, including chronic obstructive pulmonary disease, diabetes mellitus, and prolonged glucocorticoid/immunosuppressant use, all of which collectively compromise host defenses against fungal pathogens ( 26 , 27 ). At the immunological level, aging is characterized by progressive T-cell dysfunction (evidenced by CD4+/CD8 + ratio disturbances) and diminished neutrophil capacity ( 28 , 29 ). Clinically, these vulnerabilities translate to more severe disease presentations, as reflected by larger pulmonary infiltrates and increased dyspnea frequency. An intriguing gender disparity exists in IFI epidemiology, the incidence of IFI is higher in male patients ( 30 ). Current evidence suggests estrogen-mediated protection in females through augmentation of innate immune cell ( 31 ). SHAP analysis is an interpretable machine learning method based on game theory ( 32 , 33 ). To address the "black box" nature of machine learning models, we employed SHAP interpretability analysis to evaluate features in our model and quantify their respective SHAP values. This approach offers two key clinical advantages: ( 1 ) it enhances physicians' understanding of the model's decision-making logic, thereby mitigating concerns about algorithmic opacity; ( 2 ) for individual patients, it provides intuitive visualization of personalized risk factor combinations. This study had some limitations. Firstly, there may be data selection bias in this retrospective study. Although it was a multicenter study and the model was externally validated, data collection was conducted by professional radiologists to minimize bias. However, in order to provide more reliable models for clinical use, further prospective analysis is necessary in the future. Second, our study may be constrained by sample size limitations. While the inclusion of 579 patients provides meaningful data, the heterogeneous etiological spectrum of community-acquired pneumonia (CAP) warrants validation in more extensive multicenter cohorts to ensure generalizability. Third, microbiological confirmation posed challenges in our cohort. Although we rigorously excluded invasive fungal infections (IFIs) through standardized diagnostic criteria, approximately [X]% of cases lacked definitive pathogen identification. This diagnostic gap may reflect either technical limitations in conventional microbiological methods or infections caused by fastidious or atypical pathogens. A significant methodological consideration involves the exclusion of COPD as a covariate in our predictive model. This decision was based on two factors: ( 1 ) the predominance of first-episode pneumonia cases in our cohort with incomplete documentation of pre-existing COPD diagnoses in medical records. 5. Conclusions In conclusion, a radiomics-based machine learning model can effectively diagnose IFI in CAP patients, demonstrating favorable interpretability. The performance of diagnosing IFI is further improved compared to the radiomics and the clinical models. This model is further analyzed by SHAP analysis, providing an interpretable and reliable guidance tool for a personalized clinical assessment of IFI risk in CAP patients early. Abbreviations Community-acquired pneumonia (CAP); invasive fungal infection (IFI); High-resolution computed tomography (HRCT); region of interest (ROI); linear discriminant analysis (LDA); SHapley Additive exPlanations (SHAP); area under the curve (AUC). Declarations Ethics approval and consent to participate: The study was in line with the Declaration of Helsinki and was approved by the Institutional Review Board of Chongqing general hospital, Chongqing University [No. KYS2022-009-01]. In view of the fact that the study was retrospective, the Medical Ethical Committee waived the requirement for written informed consent from the patients. Consent for publication: Not applicable. Availability of data and material: The data that support the findings of this study are available from the corresponding author, upon reasonable request. Competing interests: No conflict of interest exists in the submission of this manuscript, and the manuscript is approved by all authors for publication and not under consideration for publication elsewhere, in whole or in part. Funding: This study was supported by grants from the Chongqing Science and Health Joint Medical Research Project [No. 2022MSXM037], the Basic Research and Frontier Exploration Project of Science and Technology Commission of Yuzhong District, Chongqing [No. 20190126]. Authors' contribution: Wenzhang He, Tie Deng, and Min Gu designed the study. Wenzhang He, Yulin Xiong, and Xuan Huang integrated and analyzed the data. Wenzhang He and Yulin Xiong collected the clinical and CT data and analysis. Min Gu obtained funding. Wenzhang He, Yulin Xiong, Xuan Huang, Haoran Luo, Shaoquan Zhou, and YongboTu prepared the figures. Wenzhang He, Yulin Xiong, Xuan Huang, Haoran Luo, Shaoquan Zhou, YongboTu, Tie Deng, and Min Gu revised the manuscript. Tie Deng and Min Gu supervised the study. All authors read and approved the final manuscript Acknowledgements: Not applicable. Study subjects or cohorts overlap: None. Methodology Single-center study Retrospective Diagnostic or prognostic study References Salluh JIF, Póvoa P, Beane A, Kalil A, Sendagire C, Sweeney DA, et al. Challenges for a broad international implementation of the current severe community-acquired pneumonia guidelines. Intensive Care Med. 2024;50(4):526-38. Aliberti S, Dela Cruz CS, Amati F, Sotgiu G, Restrepo MI. Community-acquired pneumonia. Lancet. 2021;398(10303):906-19. Niederman MS, Torres A. Severe community-acquired pneumonia. Eur Respir Rev. 2022;31(166). Chean D, Windsor C, Lafarge A, Dupont T, Nakaa S, Whiting L, et al. Severe Community-Acquired Pneumonia in Immunocompromised Patients. Semin Respir Crit Care Med. 2024;45(2):255-65. Arvanitis M, Anagnostou T, Fuchs BB, Caliendo AM, Mylonakis E. Molecular and nonmolecular diagnostic methods for invasive fungal infections. Clin Microbiol Rev. 2014;27(3):490-526. Richardson M, Page I. Role of Serological Tests in the Diagnosis of Mold Infections. Curr Fungal Infect Rep. 2018;12(3):127-36. Guarner J, Brandt ME. Histopathologic diagnosis of fungal infections in the 21st century. Clin Microbiol Rev. 2011;24(2):247-80. Chen Z, Fan H, Cai J, Li Y, Wu B, Hou Y, et al. High-resolution computed tomography manifestations of COVID-19 infections in patients of different ages. Eur J Radiol. 2020;126:108972. du Plessis AM, Andronikou S, Machemedze T, Griffith-Richards S, Myer L, Mahtab S, et al. High-resolution computed tomography features of lung disease in perinatally HIV-infected adolescents on combined antiretroviral therapy. Pediatr Pulmonol. 2019;54(11):1765-73. Chen W, Xiong X, Xie B, Ou Y, Hou W, Du M, et al. Pulmonary invasive fungal disease and bacterial pneumonia: a comparative study with high-resolution CT. Am J Transl Res. 2019;11(7):4542-51. Lambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, et al. Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749-62. Sun R, Limkin EJ, Vakalopoulou M, Dercle L, Champiat S, Han SR, et al. A radiomics approach to assess tumour-infiltrating CD8 cells and response to anti-PD-1 or anti-PD-L1 immunotherapy: an imaging biomarker, retrospective multicohort study. Lancet Oncol. 2018;19(9):1180-91. Bitencourt AGV, Gibbs P, Rossi Saccarelli C, Daimiel I, Lo Gullo R, Fox MJ, et al. MRI-based machine learning radiomics can predict HER2 expression level and pathologic response after neoadjuvant therapy in HER2 overexpressing breast cancer. EBioMedicine. 2020;61:103042. Yan C, Hao P, Wu G, Lin J, Xu J, Zhang T, et al. Machine learning-based combined nomogram for predicting the risk of pulmonary invasive fungal infection in severely immunocompromised patients. Ann Transl Med. 2022;10(9):514. Yang T, Zhang L, Sun S, Yao X, Wang L, Ge Y. Identifying severe community-acquired pneumonia using radiomics and clinical data: a machine learning approach. Sci Rep. 2024;14(1):21884. Gong M, Xu J, Li K, Li K, Xia Y, Jing Y, et al. A CT-based nomogram for differentiating invasive fungal disease of the lung from bacterial pneumonia. BMC Med Imaging. 2022;22(1):172. Hu Y, Zhong L, Liu H, Ding W, Wang L, Xing Z, et al. Lung CT-based multi-lesion radiomic model to differentiate between nontuberculous mycobacteria and Mycobacterium tuberculosis. Med Phys. 2025;52(2):1086-95. Li P, Wang J, Tang M, Li M, Han R, Zhou S, et al. A CT-based radiomics predictive nomogram to identify pulmonary tuberculosis from community-acquired pneumonia: a multicenter cohort study. Front Cell Infect Microbiol. 2024;14:1388991. Wu J, Xia Y, Wang X, Wei Y, Liu A, Innanje A, et al. uRP: An integrated research platform for one-stop analysis of medical images. Front Radiol. 2023;3:1153784. Albrich WC, Lamoth F. Viral-associated Pulmonary Aspergillosis: Have We Finally Overcome the Debate of Colonization versus Infection? Am J Respir Crit Care Med. 2023;208(3):230-1. Heylen J, Vanbiervliet Y, Maertens J, Rijnders B, Wauters J. Acute Invasive Pulmonary Aspergillosis: Clinical Presentation and Treatment. Semin Respir Crit Care Med. 2024;45(1):69-87. Ko BS, Chen WT, Kung HC, Wu UI, Tang JL, Yao M, et al. 2016 guideline strategies for the use of antifungal agents in patients with hematological malignancies or hematopoietic stem cell transplantation recipients in Taiwan. J Microbiol Immunol Infect. 2018;51(3):287-301. Wang JW, Yang FF, Zhang CY, Lin JZ, Wang HX, Xu WJ. Imaging Characteristics of Invasive Pulmonary Fungal Infection Secondary to Hematological Diseases and Comparison before and after Treatment. J Healthc Eng. 2021;2021:3736108. Obmann VC, Bickel F, Hosek N, Ebner L, Huber AT, Damonti L, et al. Radiological CT Patterns and Distribution of Invasive Pulmonary Aspergillus, Non-Aspergillus, Cryptococcus and Pneumocystis Jirovecii Mold Infections - A Multicenter Study. Rofo. 2021;193(11):1304-14. Zhang K, Zhao G, Liu Y, Huang Y, Long J, Li N, et al. Clinic, CT radiomics, and deep learning combined model for the prediction of invasive pulmonary aspergillosis. BMC Med Imaging. 2024;24(1):264. Christenson SA, Smith BM, Bafadhel M, Putcha N. Chronic obstructive pulmonary disease. Lancet. 2022;399(10342):2227-42. Martinez-Gomez D, Ortega FB, Hamer M, Lopez-Garcia E, Struijk E, Sadarangani KP, et al. Physical Activity and Risk of Metabolic Phenotypes of Obesity: A Prospective Taiwanese Cohort Study in More Than 200,000 Adults. Mayo Clin Proc. 2019;94(11):2209-19. Garrido-Rodríguez V, Herrero-Fernández I, Castro MJ, Castillo A, Rosado-Sánchez I, Galvá MI, et al. Immunological features beyond CD4/CD8 ratio values in older individuals. Aging (Albany NY). 2021;13(10):13443-59. Goyani P, Christodoulou R, Vassiliou E. Immunosenescence: Aging and Immune System Decline. Vaccines (Basel). 2024;12(12). Egger M, Hoenigl M, Thompson GR, 3rd, Carvalho A, Jenks JD. Let's talk about sex characteristics-As a risk factor for invasive fungal diseases. Mycoses. 2022;65(6):599-612. Zhao L, Huang S, Mei S, Yang Z, Xu L, Zhou N, et al. Pharmacological activation of estrogen receptor beta augments innate immunity to suppress cancer metastasis. Proc Natl Acad Sci U S A. 2018;115(16):E3673-e81. Ponce-Bobadilla AV, Schmitt V, Maier CS, Mensing S, Stodtmann S. Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development. Clin Transl Sci. 2024;17(11):e70056. Ghasemi A, Hashtarkhani S, Schwartz DL, Shaban-Nejad A. Explainable artificial intelligence in breast cancer detection and risk prediction: A systematic scoping review. Cancer Innov. 2024;3(5):e136. Tables Table 1. Comparison of patient characteristics across the training, internal validation and external validation datasets. Clinical variables Training dataset (n=239) Internal validation dataset (n=101) External validation dataset (n=230) P-value Radiological feature ROI Volume (Mean ± SD, cm 3 ) 218.3 ± 282.8 157.5 ± 191.9 340.8 ± 341.3 <0.001 ROI Average HU Value (Mean ± SD, HU) -508.1 ± 115.0 -497.2 ± 131.1 -499.3 ± 149.5 0.722 Epidemiologic factors Gender, male, n (%) 134 (56.1) 44 (43.6) 153 (66.5) <0.001 Age (Mean ± SD, years) 62.1 ± 19.3 57.5 ± 17.3 68.4 ± 15.3 <0.001 non-productive cough, yes, n (%) 30 (12.6) 12 (11.9) 7 (3.0) <0.001 productive cough, yes, n (%) 193 (80.8) 84 (83.2) 223 (97.0) <0.001 Hemoptysis, yes, n (%) 22 (9.2) 9 (8.9) 17 (7.4) 0.764 Dyspnea, yes, n (%) 115 (48.1) 46 (39.6) 74 (32.2) 0.001 Chest pain, yes, n (%) 35 (16.3) 15 (14.9) 26 (11.3) 0.503 Confusion, yes, n (%) 2 (0.8) 0 (0) 1 (0.4) 0.603 Chest tightness, yes, n (%) 203/36 79/22 220/10 <0.001 Myalgia, yes, n (%) 190/49 79/22 227/3 <0.001 Duration of Symptoms (Median, Q1, Q3, days) 7.0 (4.0, 15.0) 7.0 (4.0, 15.0) 5.0 (3.0, 10.0) <0.001 Fever, yes, n (%) 139 (58.2) 61 (60.4) 102 (44.3) 0.003 Laboratory findings White Blood Cell (Median, Q1, Q3, 10^9) 7.50 (5.59, 9.98) 7.44 (5.51, 10.04) 8.36 (6.49, 11.24) 0.022 Neutrophil (Median, Q1, Q3, 10^9) 5.57 (4.00, 8.05) 5.49 (3.76, 7.77) 6.48 (4.65, 9.27) 0.003 Lymphocyte (Median, Q1, Q3, 10^9) 1.16 (0.85, 1.55) 1.19 (0.89, 1.63) 0.99 (0.59, 1.42) 0.011 Monocyte (Median, Q1, Q3, 10^9) 0.53 (0.38, 0.69) 0.50 (0.35, 0.70) 0.52 (0.35, 0.73) 0.452 Eosinophil (Median, Q1, Q3, 10^9) 0.05 (0.01, 0.12) 0.06 (0.01, 0.15) 0.04 (0.01, 0.15) 0.627 Basophil (Median, Q1, Q3, 10^9) 0.02 (0.01, 0.03) 0.02 (0.01, 0.03) 0.02 (0.01, 0.03) 0.767 Red Blood Cell (Median, Q1, Q3, 10^12) 4.38 (3.95, 4.80) 4.44 (4.18, 4.77) 4.18 (3.70, 4.62) 0.917 Hemoglobin (Median, Q1, Q3, g/L) 131.0 (118.0, 144.0) 132.0 (120.0, 141.0) 125.0 (109.0, 139.0) <0.001 C-reactive Protein (Median, Q1, Q3, mg/L) 31.9 (10.8, 62.9) 31.9 (9.5, 73.9) 34.7 (8.8, 83.7) 0.665 Neutrophil/Lymphocyte Ratio (Median, Q1, Q3) 4.82 (3.16, 7.60) 4.75 (2.90, 6.98) 6.49 (3.57, 13.26) <0.001 Abbreviations: ROI, Region of interest; HU, Hounsfield Unit. Table 2. Univariate and multivariate analysis of the clinical variables. Variable Univariate Multivariate Odds ratio 95% CI P Odds ratio 95% CI P ROI Average HU Value 1.000 0.998-1.001 0.746 ROI Volume 3.137 1.829-5.382 <0.001 2.337 1.316-4.148 0.006 Chest pain 0.687 0.367-1.291 0.244 Chest tightness 0.955 0.522-1.748 0.882 Confusion 0.000 0.000-0.000 0.998 Duration of Symptoms 1.012 1.001-1.023 0.028 1.011 0.999-1.022 0.062 Dyspnea 1.867 1.270-2.745 0.002 1.775 1.178-2.675 0.006 Fever 0.826 0.587-1.163 0.274 1.039 0.995-1.085 0.081 Age 1.026 1.014-1.039 <0.001 1.016 1.001-1.030 0.032 Hemoptysis 0.787 0.372-1.663 0.530 Myalgia 0.290 0.123-0.684 0.005 0.474 0.194-1.157 0.101 Non-productive cough 0.382 0.149-0.983 0.046 0.604 0.226-1.617 0.316 Gender 1.887 1.248-2.853 0.003 1.588 1.011-2.494 0.045 Productive cough 1.594 0.813-3.123 0.175 Basophil Count 1.831 0.165-20.347 0.623 C reactive Protein 1.165 0.739-1.836 0.510 Eosinophil Count 0.805 0.553-1.173 0.259 Hemoglobin 0.574 0.391-0.842 0.005 Lymphocyte Count 0.778 0.531-1.141 0.199 Monocyte Count 1.026 0.683-1.539 0.903 Neutrophil Count 1.318 0.918-1.892 0.134 Red Blood Cell Count 0.556 0.385-0.801 0.002 0.839 0.619-1.138 0.259 White Blood Cell Count 1.176 0.831-1.664 0.360 NLR 1.007 0.995-1.020 0.260 Abbreviations: ROI, Region of interest; HU, Hounsfield Unit; CI, confidence interval; NLR, neutrophil-to-lymphocyte ratio. Table 3. Detailed performance of the machine learning classifiers for invasive fungal infection identification in the development cohort through cross-validation. Classifier Training Validation Significant Overfitting AUC 95% CI P-value AUC 95% CI P-value LDA 0.758 0.709-0.803 reference 0.741 0.691-0.787 reference No ( P = 0.066) KNN 0.906 0.870-0.935 <0.001 0.656 0.603-0.707 0.018 Yes ( P < 0.001) LR 0.723 0.672-0.770 0.055 0.645 0.591-0.696 0.004 No ( P = 0.100) NB 0.748 0.698-0.793 0.639 0.711 0.660-0.759 0.220 Yes ( P < 0.001) RF 1.000 0.989-1.000 <0.001 0.707 0.655-0.755 0.194 Yes ( P < 0.001) SVM 0.891 0.853-0.922 <0.001 0.606 0.552-0.658 0.001 Yes ( P < 0.001) Abbreviations: AUC, areas under the receiver operating characteristic curves; CI, confidence interval; LDA, confidence interval; Linear discriminant analysis; KNN, k-nearest neighbor; LR, logistic regression; NB, naïve Bayes; RF, random forest; SVM, support vector machine. Table 4. Model performance comparison in the training, internal validation and external validation datasets. Dataset Model AUC (95% CI) P-value SEN SPE PPV NPV Training Clinical 0.777 (0.719-0.828) 0.007 75.6% 74.7% 41.0% 93.0% Radiomics 0.765 (0.706-0.817) 0.010 66.7% 71.7% 35.3% 90.3% Combined 0.836 (0.783-0.881) reference 77.8% 74.2% 41.2% 93.5% Internal validation Clinical 0.719 (0.621-0.804) 0.030 52.6% 72.0% 30.3% 86.8% Radiomics 0.724 (0.626-0.808) 0.048 57.9% 74.4% 34.4% 88.4% Combined 0.808 (0.718-0.880) reference 73.7% 74.4% 40.0% 92.4% External validation Clinical 0.707 (0.644-0.765) 0.001 60.9% 72.4% 35.4% 88.2% Radiomics 0.709 (0.646-0.767) 0.042 52.2% 71.4% 31.2% 85.7% Combined 0.786 (0.728-0.837) reference 69.6% 78.9% 45.1% 91.3% Abbreviations: AUC, areas under the receiver operating characteristic curves; CI, confidence interval; SEN, sensitivity; SPE, specificity; PPV, positive predictive value; NPV, negative predictive value. Additional Declarations No competing interests reported. Supplementary Files ElectronicSupplementaryMaterial.docx Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 30 Mar, 2026 Reviews received at journal 22 Mar, 2026 Reviewers agreed at journal 16 Mar, 2026 Reviews received at journal 21 Dec, 2025 Reviewers agreed at journal 15 Dec, 2025 Reviewers agreed at journal 14 Dec, 2025 Reviewers invited by journal 01 Dec, 2025 Editor invited by journal 21 Nov, 2025 Editor assigned by journal 19 Nov, 2025 Submission checks completed at journal 19 Nov, 2025 First submitted to journal 18 Nov, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8145288","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":554282791,"identity":"235a62c7-eb4e-4bc3-b944-414332673dad","order_by":0,"name":"Wenzhang He","email":"","orcid":"","institution":"Chongqing University","correspondingAuthor":false,"prefix":"","firstName":"Wenzhang","middleName":"","lastName":"He","suffix":""},{"id":554282792,"identity":"0dfca609-1a8d-430a-82e5-b7c20d772bc2","order_by":1,"name":"Yulin Xiong","email":"","orcid":"","institution":"Chongqing University","correspondingAuthor":false,"prefix":"","firstName":"Yulin","middleName":"","lastName":"Xiong","suffix":""},{"id":554282793,"identity":"7d15f1c6-781e-4966-b16d-83888444fca3","order_by":2,"name":"Xuan Huang","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Xuan","middleName":"","lastName":"Huang","suffix":""},{"id":554282794,"identity":"74e046e7-0f4f-4c06-9b6b-2f2dac30fdf4","order_by":3,"name":"Haoran Luo","email":"","orcid":"","institution":"Chongqing University","correspondingAuthor":false,"prefix":"","firstName":"Haoran","middleName":"","lastName":"Luo","suffix":""},{"id":554282795,"identity":"c66fb4b1-1260-464d-9b08-c0d782b840a2","order_by":4,"name":"Shaoquan Zhou","email":"","orcid":"","institution":"Chongqing University","correspondingAuthor":false,"prefix":"","firstName":"Shaoquan","middleName":"","lastName":"Zhou","suffix":""},{"id":554282796,"identity":"692e97f0-2fa2-4c43-8766-6a0abd889f1f","order_by":5,"name":"Yongbo Tu","email":"","orcid":"","institution":"Chongqing University","correspondingAuthor":false,"prefix":"","firstName":"Yongbo","middleName":"","lastName":"Tu","suffix":""},{"id":554282797,"identity":"edcce97e-6fc1-4f19-b4fd-7c29bc34e153","order_by":6,"name":"Tie Deng","email":"","orcid":"","institution":"Chongqing University Central Hospital, Chongqing University","correspondingAuthor":false,"prefix":"","firstName":"Tie","middleName":"","lastName":"Deng","suffix":""},{"id":554282799,"identity":"94c94c5d-8cd2-454e-b003-6c03d7001048","order_by":7,"name":"Min Gu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3UlEQVRIiWNgGAWjYJCCAwwGNvX97A0kaSlIY5zZc4Akez4cZtxwI4FIxfIR2YmHCwwOMxvcfGP44ccfBnn+BuZnD/BpMbyRu+HwDIN0NsnbOcaSvW0MhjMOsJkb4NUyA6iFx8Cah+92jhkDbwMD4wYGHjYJIrQwSzDcPGPG+OcPgz1BLfISYC3OBgI3eMyYedgYEglqMeB5C9KSliDZk1YsLdsmkTzjMJsZflvaczd/5vljk8DPfnjjxzd/bGz725uf4bflACofqJgZn3qQLQ0EFIyCUTAKRsEoYAAAxQBGZPDMhDAAAAAASUVORK5CYII=","orcid":"","institution":"Chongqing University","correspondingAuthor":true,"prefix":"","firstName":"Min","middleName":"","lastName":"Gu","suffix":""}],"badges":[],"createdAt":"2025-11-18 12:23:08","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8145288/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8145288/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97370876,"identity":"bb619933-38c2-44fc-8e04-586f75037edb","added_by":"auto","created_at":"2025-12-03 16:28:05","extension":"tif","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342992,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/593633237e682f3191093ac7.tif"},{"id":97346273,"identity":"499efb87-0a33-4a77-8885-3563bca6a1ff","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102311,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/1624ec925d406e716becd81e.docx"},{"id":97370207,"identity":"d1758e69-b6eb-4408-a688-86c40be75bac","added_by":"auto","created_at":"2025-12-03 16:26:53","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":560144,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/56a3900b9e4c38bcca57cc8a.tif"},{"id":97371145,"identity":"070912f5-b62b-4f32-9c35-aff08d802e7c","added_by":"auto","created_at":"2025-12-03 16:28:27","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":32799,"visible":true,"origin":"","legend":"","description":"","filename":"Table.docx","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/db9b64ea149c68abde76610c.docx"},{"id":97371464,"identity":"1d35aa1a-888c-4ca6-bea5-507d874444e1","added_by":"auto","created_at":"2025-12-03 16:29:01","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":453472,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/b46d2000a886269995776ee8.tif"},{"id":97370155,"identity":"d5bffd41-8049-4941-b6e8-fc053d507f81","added_by":"auto","created_at":"2025-12-03 16:26:50","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":306164,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/e94a1a2e4366900db98fe1f5.tif"},{"id":97369917,"identity":"208cec05-9c23-4899-8ef7-e3a481a181bc","added_by":"auto","created_at":"2025-12-03 16:26:06","extension":"tif","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342096,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/07c3c3261f3400687523e02d.tif"},{"id":97370826,"identity":"276af57d-e688-4201-aa21-b35b9a9ca9f5","added_by":"auto","created_at":"2025-12-03 16:27:59","extension":"tif","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":233506,"visible":true,"origin":"","legend":"","description":"","filename":"Figure6.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/d80fe183b2540061d4dd6eab.tif"},{"id":97346276,"identity":"39f88b41-e932-41bd-b9a2-208cc2400576","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"json","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9582,"visible":true,"origin":"","legend":"","description":"","filename":"f6d60285bd29459daeb7647576c70f03.json","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/7a609df636c1d914f71a8315.json"},{"id":97346291,"identity":"1d9258a4-2f33-4294-806d-113a66c375df","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"docx","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":796981,"visible":true,"origin":"","legend":"","description":"","filename":"ElectronicSupplementaryMateriald1.docx","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/c27aeb17eaadafc9df5e4228.docx"},{"id":97346281,"identity":"b3a7c752-ceff-4586-b49f-28fce63cbdc5","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128424,"visible":true,"origin":"","legend":"","description":"","filename":"f6d60285bd29459daeb7647576c70f031enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/766fae9694ef789069fc6c66.xml"},{"id":97346288,"identity":"c2207900-e00f-4d6e-b8c8-ea6261f3c62a","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"tif","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342992,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/c5fc3c0efb33c99d00b9cf94.tif"},{"id":97346285,"identity":"80c4c958-7258-4c24-9c1b-14a79b986d25","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"tif","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":560144,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/bc2bae95127220e1dc598691.tif"},{"id":97371236,"identity":"a3d2edef-be11-4408-b008-82a1c4281a97","added_by":"auto","created_at":"2025-12-03 16:28:33","extension":"tif","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":453472,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/bcd44c9d46975fc855f918cf.tif"},{"id":97346283,"identity":"90818015-d1ea-486f-a848-3fcd2e37b41e","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"tif","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":306164,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/cb2acd2b76dfec9814b9d074.tif"},{"id":97346297,"identity":"e03e2d3b-4aa1-4780-a99b-ee8820a9238d","added_by":"auto","created_at":"2025-12-03 11:48:12","extension":"tif","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342096,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/c9afca658e0c62221b3290ff.tif"},{"id":97346280,"identity":"8a1803e0-df72-4e27-a38e-33cf1943919f","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"tif","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":233506,"visible":true,"origin":"","legend":"","description":"","filename":"Figure6.tif","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/2ff65a18a9fd1167305b57a5.tif"},{"id":97346296,"identity":"a4efb3da-5991-461b-8243-bb1ed2fc3e32","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":166803,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/80f53eec950049bd2650eb0f.png"},{"id":97346293,"identity":"d6165d0a-666e-4735-8b72-6016733dec17","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":97764,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/edaef80105425e8fb99327d2.png"},{"id":97371240,"identity":"7396f575-a95f-4071-8707-b633e2813f29","added_by":"auto","created_at":"2025-12-03 16:28:34","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66129,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/5f27e61d4a2bdfeb52f4203e.png"},{"id":97346290,"identity":"447ef33b-ec8f-4e05-9851-3d63abbacb99","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":48046,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure4.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/16c30c63b66caff0178e9f2e.png"},{"id":97370671,"identity":"9ee546e3-8c08-4453-8c72-66ae71efc9d9","added_by":"auto","created_at":"2025-12-03 16:27:45","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":56272,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure5.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/68d2041b2de761bf6921716a.png"},{"id":97346292,"identity":"28098466-362c-4949-868c-e38d9ba2d9c9","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":31968,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure6.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/bf816c57c9a3337b89ec9da9.png"},{"id":97370545,"identity":"39d190af-3efa-41da-bddc-169cf3310fc5","added_by":"auto","created_at":"2025-12-03 16:27:35","extension":"xml","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128819,"visible":true,"origin":"","legend":"","description":"","filename":"f6d60285bd29459daeb7647576c70f031structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/7688ef4d45ea43b8be5cafc8.xml"},{"id":97369872,"identity":"283ac9f1-5fd4-48ee-a188-6da2b80ec7da","added_by":"auto","created_at":"2025-12-03 16:25:57","extension":"html","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140677,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/d91383cd7ee975a4c8859bca.html"},{"id":97346267,"identity":"7d5f5397-5c76-4dac-bd80-bbdd0d8f0385","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":557230,"visible":true,"origin":"","legend":"\u003cp\u003ePatient enrollment flowchart.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/94b94ace6eb01851907457de.png"},{"id":97369886,"identity":"d3e61790-bf1d-4488-8f0e-3f0d42a2fc43","added_by":"auto","created_at":"2025-12-03 16:26:01","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":932386,"visible":true,"origin":"","legend":"\u003cp\u003eSelection of radiomics features using the least absolute shrinkage and selection operator regression.\u003c/p\u003e\n\u003cp\u003eA. Selection of the optimal tuning parameter lambda.\u003c/p\u003e\n\u003cp\u003eB. The coefficient profile plot of 6 non-zero coefficients against the optimal log(lambda) sequence.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/70de52fc14e4b2c9b01c3641.png"},{"id":97346272,"identity":"43ccd221-4dd8-4c9a-8e4f-abecba8d32ee","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":757783,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of different machine learning classifiers in the development cohort via 4-fold cross-validation.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/3968f1dd33e5df9044b13321.png"},{"id":97371002,"identity":"48caec85-deef-4c76-b3f2-a6c71fdfcf8c","added_by":"auto","created_at":"2025-12-03 16:28:15","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":502568,"visible":true,"origin":"","legend":"\u003cp\u003eROC analysis of the predictive models in (A) the training dataset, (B) the internal validation dataset and (C) the external validation dataset.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/ec8593c2737ec26d225d5a85.png"},{"id":97346279,"identity":"6d7bc118-d794-4ca3-baac-fb5523180fbb","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":163331,"visible":true,"origin":"","legend":"\u003cp\u003eDecision curve analysis of the clinical model, radiomics model and combined model. (A) In the internal validation dataset , (B) in the external validation dataset.\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/223e4c9e992ead377c3ef3e6.jpg"},{"id":97346286,"identity":"6255ec49-5c95-41d1-8ec9-f732a53ffb8f","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":352649,"visible":true,"origin":"","legend":"\u003cp\u003eSHAP analysis of the combined model.\u003c/p\u003e\n\u003cp\u003eA. The detailed contribution to the class 0 model prediction, where the colors represent the magnitude of the feature values.\u003c/p\u003e\n\u003cp\u003eB. The average SHAP value of each feature in the combined model.\u003c/p\u003e\n\u003cp\u003eAbbreviations: SHAP, SHapley Additive exPlanations.\u003c/p\u003e","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/9fb1e2d6a572dcc8072001c8.png"},{"id":97664538,"identity":"32cb7635-1785-4666-8729-3a25f7a2fc8f","added_by":"auto","created_at":"2025-12-08 09:09:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4857635,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/2f037591-1495-483f-bee8-2351b2b858dd.pdf"},{"id":97346278,"identity":"d182a024-2dfe-41e6-a9dc-1f879d0eabd7","added_by":"auto","created_at":"2025-12-03 11:48:11","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":796981,"visible":true,"origin":"","legend":"","description":"","filename":"ElectronicSupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-8145288/v1/9be2a892f8f3c12b7dd1450a.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"An interpretable radiomics-machine learning model for diagnosing invasive fungal infections in community-acquired pneumonia: multicenter study","fulltext":[{"header":"Key Points","content":"\u003cp\u003e1. The study addresses the challenge of diagnosing invasive fungal infections in patients with community-acquired pneumonia.\u003c/p\u003e\u003cp\u003e2. The developed radiomics-machine learning model, combining clinical variables and HRCT-derived radiomics features.\u003c/p\u003e\u003cp\u003e3. The model offers a non-invasive, interpretable tool for early invasive fungal infections diagnosis in community-acquired pneumonia patients, potentially reducing diagnostic delays.\u003c/p\u003e"},{"header":"1. Introduction","content":"\u003cp\u003eCommunity-acquired pneumonia (CAP) represents a global health priority, ranking among the top five causes of mortality worldwide and imposing substantial socioeconomic burdens (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). CAP exhibits considerable etiological diversity, with potential involvement of either singular or polymicrobial pathogens, resulting in marked heterogeneity in both clinical course and patient outcomes (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). A subset of patients, especially immunodeficient individuals, frequently present with co-occurring invasive fungal infection (IFI) which is associated with substantial morbidity, having a mortality rate of 40% to 50% (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). The current diagnostic workflow, dependent on culture methods with suboptimal sensitivity (30\u0026ndash;50%) and serological tests requiring 6\u0026ndash;8 hours processing time, demonstrates significant limitations (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). This diagnostic inadequacy creates a critical 72\u0026ndash;96 hours window of uncertainty, during which mortality risk escalates (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). This underscores the urgent need for advanced diagnostic frameworks capable of rapid, precise pathogen differentiation.\u003c/p\u003e\u003cp\u003eHigh-resolution computed tomography (HRCT) plays a pivotal role in evaluating pulmonary infections due to its ability to provide detailed morphological information (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Characteristic CT manifestations of invasive fungal pneumonia typically include multiple nodules accompanied by peripheral ground-glass opacities and/or wedge-shaped consolidations (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). Nevertheless, these radiographic features lack specificity and often demonstrate significant overlap with other pulmonary infectious processes. In polymicrobial infections, the radiological manifestations of co-existing pathogens can mask distinctive fungal signatures, resulting in reduced diagnostic recognition.\u003c/p\u003e\u003cp\u003eRadiomics, an emerging field in medical imaging, offers a promising solution by extracting high-dimensional quantitative features from medical images and analyzing them by using machine learning algorithms (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). Radiomic features invisible to the naked eye can reveal differences at the protein, cellular, and tissue levels (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). Yan et al. established CT radiomics as a viable predictor of pulmonary invasive fungal infections in immunocompromised hosts (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). In another research, Yang et al. demonstrated that machine learning-enhanced radiomics improves severe CAP identification (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). Subsequent multicenter studies have validated radiomics' diagnostic utility in distinguishing pathogens causing pulmonary infections (\u003cspan additionalcitationids=\"CR17\" citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). However, no machine learning-based system integrating imaging biomarkers with clinical parameters has yet been developed for invasive fungal infection identification and clinical decision support in large, multicenter CAP cohort. By leveraging advanced computational techniques, we seek to identify robust imaging biomarkers that can identify IFI with high precision.\u003c/p\u003e\u003cp\u003eThis study aims to develop and validate a radiomics-based model for diagnosing IFI in CAP patients based on multicenter datasets. In addition, the study aimed to explore the interpretability of the radiomics model by SHapley Additive exPlanations analysis.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cp\u003e The study was in line with the Declaration of Helsinki and was approved by the Institutional Review Board of Chongqing general hospital, Chongqing University [No. KYS2022-009-01].\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1. Patients\u003c/h2\u003e\u003cp\u003eThe inclusion criteria were as follows: (I) age 18 years and over; (II) meeting the diagnostic criteria for CAP; (III) CT scan with thin section (slice thickness less than 1.5 mm); (IV) definitive evidence of pulmonary infiltration/consolidation on HRCT; and (V) pathogen detection tests on the day of hospital admission (including identifying pathogens by deep sputum culture, blood culture, bronchoalveolar lavage culture, histopathology, and macrogenomic sequencing). A total of 1212 CAP patients who underwent HRCT examination from July 2022 to August 2024 in Chongqing General Hospital (Center 1) and Chongqing University Central Hospital (Center 2) were initially recruited. The exclusion criteria were as follows: (I) the interval between CT examination and pathogen specimen testing exceeds 1 days (n\u0026thinsp;=\u0026thinsp;276); (II) patient data missing exceeds 30% (n\u0026thinsp;=\u0026thinsp;27); (III) patients who received antibiotic therapy (n\u0026thinsp;=\u0026thinsp;265); (IV) CT image quality is insufficient for clinical diagnosis (n\u0026thinsp;=\u0026thinsp;74). The patient selection process is depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Ultimately, 340 patients from Center 1 comprised the development cohort. These patients were randomly divided into the training dataset (n\u0026thinsp;=\u0026thinsp;239) and the internal validation dataset (n\u0026thinsp;=\u0026thinsp;101) in a 7:3 ratio. An independent cohort of 230 patients from Center 2 served as the external validation dataset.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2. Volume segmentation of the region of interest (ROI)\u003c/h2\u003e\u003cp\u003eThe detailed HRCT acquisition information is provided in \u003cb\u003eS1\u003c/b\u003e. For lesion segmentation related to pneumonia, a vb-net pneumonia automatic segmentation algorithm was employed, which was developed by United Imaging Intelligence's one-stop research platform (uAI Research Portal, V20230515, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://urp.united-imaging.com/\u003c/span\u003e\u003cspan address=\"https://urp.united-imaging.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). Previously, in a pneumonia infection segmentation task, it was proven to highly overlap with the manual sketch, and the similarity coefficient of dice was 91.6%. To secure segmentation accuracy, two chest radiologists manually reviewed and adjusted algorithm-generated segmentations (with 9 and 25 years of experience in chest imaging diagnosis, respectively). The experts based their annotations of the boundaries and regions of disease lesions on their clinical skills and medical knowledge. \u003cb\u003eFigure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e shows the ROI segmentation example.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3. Radiomics features extraction and selection\u003c/h2\u003e\u003cp\u003eRadiomics feature extraction method and feature details are presented in \u003cb\u003eS2\u003c/b\u003e. A total of 2,264 radiomic features were extracted. To mitigate overfitting and reduce model complexity, initial feature selection was performed using the Mann\u0026ndash;Whitney U test, retaining only radiomic features that demonstrated a significant association with the endpoint outcome (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Subsequently, Pearson correlation coefficient analysis was conducted to assess feature redundancy; for pairs of features exhibiting high correlation (|r| \u0026gt;0.95), the feature with the higher P value from the Mann\u0026ndash;Whitney U test was excluded. Finally, the most informative radiomic features for predicting the endpoint outcome were identified using the least absolute shrinkage and selection operator regression. The optimal penalty parameter (lambda) was determined through 10-fold cross-validation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4. Selection of the clinical variables\u003c/h2\u003e\u003cp\u003eClinical variables for each patient were extracted from the electronic medical record system, including radiological features (ROI Volume and ROI Average HU Value), epidemiologic factors (sex, age, non-productive cough, hemoptysis, dyspnea, chest pain, confusion, chest tightness, myalgia, duration of symptoms, and fever), and laboratory findings (white blood cell count, neutrophil count, lymphocyte count, neutrophil/lymphocyte ratio, monocyte count, eosinophil count, basophil count, red blood cell count, hemoglobin level, and C-reactive protein level). Univariate logistic regression analysis was performed to evaluate the association between each clinical variable and fungal infection status. Variables demonstrating a significant association (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were subsequently included in the multivariate regression analysis, and variables with \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 in the multivariate model were selected for further analysis.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5. Model development and evaluation\u003c/h2\u003e\u003cp\u003ePrior to model development, multiple well-established and widely applied machine learning algorithms were evaluated, including linear discriminant analysis (LDA), k-nearest neighbor, logistic regression, na\u0026iuml;ve Bayes, random forest, and support vector machine classifiers. These models were compared within the development cohort using 4-fold cross-validation to ensure robustness and generalizability. Selection of the optimal classifier was based on the highest binary classification accuracy observed in the validation folds, while also considering the extent of overfitting, which was assessed by comparing model performance between the training and validation datasets.\u003c/p\u003e\u003cp\u003eBased on the selected machine learning classifier, three distinct predictive models were constructed in the training dataset: a clinical model utilizing only the selected clinical variables, a radiomics model incorporating solely the selected radiomic features, and a combined model that integrated both clinical and radiomic features. The performance of these models was evaluated and compared in both the internal and external validation datasets through receiver operating characteristic analysis with respect to the area under the curve (AUC). Additionally, the detailed sensitivity, specificity, positive predictive value, and negative predictive value for each model were computed at the optimal cutoff point, which was determined using the maximum Youden index criterion.\u003c/p\u003e\u003cp\u003eModel calibration was assessed using the Brier score, which quantifies the mean squared difference between the predicted probabilities and the actual endpoints (i.e., fungal infection status), with lower values indicating better calibration. Clinical utility was evaluated using decision curve analysis by comparing the net benefits of the models across a range of threshold probabilities.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e2.6. SHAP analysis\u003c/h2\u003e\u003cp\u003eIn order to assess the contribution of each feature in a predictive model, the SHapley Additive exPlanations (SHAP) analysis was performed to enhance model interpretability. Based on the concept of Shapley values from cooperative game theory, SHAP quantifies the impact of each feature on model predictions by decomposing the output into additive contributions of individual variables. This approach provides both global insights into overall feature importance and local explanations of specific predictions, offering a transparent, visual interpretation of model behavior.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e2.7. Statistical analysis\u003c/h2\u003e\u003cp\u003eAll statistical analyses were performed using SPSS software (version 21.0). One-way analysis of variance was used to compare quantitative variables across groups, and the chi-square test was used for categorical variables. Differences between two AUCs were assessed using the DeLong test. All statistical tests were two-sided, and a \u003cem\u003eP\u003c/em\u003e value less than 0.05 was considered indicative of statistical significance. Model development and validation were conducted on the InferScholar platform (version 3.5, Infervision). SHAP analysis and decision curve analysis were conducted using R software (version 3.6.3) with the shapviz and rmda packages, respectively.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cstrong\u003e3.1. Patient characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe prevalence of invasive fungal infection was 18.8% (45 of 239) in the training dataset, 18.8% (19 of 101) in the internal validation dataset, and 20.0% (46 of 230) in the external validation dataset, with no significant differences in fungal infection prevalence observed across the three datasets. As summarized in \u003cstrong\u003eTable 1\u003c/strong\u003e, the external validation dataset demonstrated significantly higher values for ROI volume, male prevalence, age, prevalence of productive cough, white blood cell count, and neutrophil-to-lymphocyte ratio compared with the training and internal validation datasets. In contrast, the prevalence of nonproductive cough, chest tightness, myalgia, fever, duration of symptoms, hemoglobin level, and lymphocyte count were significantly lower in the external validation dataset than in the training and internal validation datasets. No significant differences were observed in the remaining clinical variables across the datasets.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2. Selection of radiomics features\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFollowing significance testing and pearson correlation analysis, 129 radiomics features were retained for least absolute shrinkage and selection operator regression analysis. As shown in \u003cstrong\u003eFigure 2\u003c/strong\u003e, a total of six radiomics features associated with fungal infection, each with a nonzero coefficient, were selected for model development using the optimal tuning parameter (lambda = 0.0335). Heatmap of the selected radiomics features, based on standardized feature values, was presented in \u003cstrong\u003eFigure S2\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3. Selection of clinical variables\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 24 clinical variables were collected and analyzed using univariate logistic regression. As shown in \u003cstrong\u003eTable 2\u003c/strong\u003e, ROI volume, dyspnea, age, and sex were significantly associated with invasive fungal infection in the multivariate regression analysis and were subsequently incorporated into the development of the clinical and combined models.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4. Machine learning classifier selection \u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe receiver operating characteristic (ROC) analysis of the radiomics models based on LDA, k-nearest neighbor, logistic regression, na\u0026iuml;ve Bayes, random forest, and support vector machine classifiers in four-fold cross-validation was presented in \u003cstrong\u003eFigure 3\u003c/strong\u003e. The corresponding AUCs for each model were summarized in\u003cstrong\u003e Table 3\u003c/strong\u003e. Model robustness was assessed by comparing AUCs between the training and validation datasets, and only the LDA- and LR-based models demonstrated good robustness without significant overfitting. Notably, the LDA-based model achieved the highest AUC in the validation dataset. Accordingly, the LDA classifier was selected as the optimal machine learning algorithm for constructing the clinical, radiomics, and combined models in this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5. Model performance evaluation \u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs shown in\u003cstrong\u003e Figure 4\u003c/strong\u003e, receiver operating characteristic (ROC) analysis demonstrated that the areas under the curve (AUCs) for the clinical, radiomics, and combined models were 0.777 (95% CI: 0.719\u0026ndash;0.828), 0.765 (95% CI: 0.706\u0026ndash;0.817), and 0.836 (95% CI: 0.783\u0026ndash;0.881), respectively, in the training dataset; 0.719 (95% CI: 0.621\u0026ndash;0.804), 0.724 (95% CI: 0.626\u0026ndash;0.808), and 0.808 (95% CI: 0.718\u0026ndash;0.880), respectively, in the internal validation dataset; and 0.707 (95% CI: 0.644\u0026ndash;0.765), 0.709 (95% CI: 0.646\u0026ndash;0.767), and 0.786 (95% CI: 0.728\u0026ndash;0.837), respectively, in the external validation dataset. Across all datasets, the combined model demonstrated a significantly higher AUC than both the clinical and radiomics models; however, no significant difference was observed between the clinical and radiomics models. Sensitivity, specificity, positive predictive value, and negative predictive value for the models in the training and validation datasets are summarized in \u003cstrong\u003eTable 4\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.6. Calibration and decision curve analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGood calibration was observed for all models. The Brier scores for the clinical, radiomics, and combined models were 0.168, 0.166, and 0.152, respectively, in the internal validation dataset, and 0.162, 0.154, and 0.134, respectively, in the external validation dataset. Decision curve analysis in the validation datasets demonstrated that the combined model provided higher net benefit across nearly the entire range of threshold probabilities compared with both the clinical and radiomics models (\u003cstrong\u003eFigure 5\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.7. SHAP analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe feature importance and individual contributions to positive and negative invasive fungal infection predictions in the combined model are illustrated in the summary plot, with the mean SHAP value for each feature also displayed (\u003cstrong\u003eFigure 6\u003c/strong\u003e). Age exerted the greatest influence on the predictions of the combined model, while the three most important radiomics features were all higher-order texture features, which are not readily appreciable by visual assessment.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study developed an integrated machine learning model combining HRCT radiomics features and clinical variables to enable early diagnosis of invasive fungal infection (IFI) in community-acquired pneumonia (CAP). Comparative evaluation of six machine learning classifiers identified the Linear Discriminant Analysis algorithm as demonstrating optimal diagnosing performance and robustness, leading to its selection as our primary modeling framework. Through SHapley Additive exPlanations (SHAP) interpretability analysis, we determined that the model's decision-making was predominantly driven by four key predictors: patient age along with three high-order radiomics features.\u003c/p\u003e\u003cp\u003eOur combined model, including clinical variables and radiomics features, showed robust and consistent performance in both internal and external validation datasets. Current gold-standard diagnostic methods for IFI, including histopathological examination and microbial cultures, typically require 3\u0026ndash;7 days to yield results (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Such diagnostic delays carry grave consequences, as mortality risk increases by 33% for every 24-hour delay in initiating appropriate antifungal therapy (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e). Notably, invasive aspergillosis complicates 19% of severe influenza cases, with associated mortality rates doubling compared to controls (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). Current guidelines emphasize the importance of diagnosis-driven therapy for high-risk patients with evidence of IFI (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e). However, conventional CT interpretation faces limitations, radiologist agreement for characteristic signs like the halo sign remains suboptimal, and most early IFI cases lack pathognomonic imaging features (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). Our model overcomes these challenges by extracting and quantifying subvisual biological information from HRCT images that reflect lesion heterogeneity. Importantly, while polymicrobial infections are common, HRCT provides direct visualization of pulmonary pathology. Our approach leverages existing HRCT data without requiring additional examinations, thereby enabling early IFI diagnosis without imposing additional economic or physiological burdens on patients, while simultaneously avoiding diagnostic delays inherent to conventional methods.\u003c/p\u003e\u003cp\u003eRecent retrospective studies have highlighted the diagnostic potential of radiomics in detecting IFI in pulmonary diseases. Most research supports integrated models that combine clinical, radiological, and radiomic features, demonstrating their incremental diagnostic value. For instance, Yan et al. showed that a combined deep learning and radiomics approach effectively differentiated IFIs from bacterial infections in patients with hematologic malignancies, achieving an AUC of 0.844 in internal validation which significantly outperforming radiomics (AUC of 0.767) and clinical models (AUC of 0.696) (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). Similarly, Gong et al. reported that radiomics could discriminate IFI from bacterial pneumonia with high accuracy (AUC of 0.911 and accuracy of 0.837) in an internal validation cohort (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). Zhang et al. further validated the efficacy of a combined model integrating clinical, CT radiomic, and deep learning features for predicting invasive pulmonary aspergillosis (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). In our study, a combined clinical-radiological model exhibited robust diagnostic performance for IFI in CAP. Furthermore, validation in a large external cohort confirmed the model\u0026rsquo;s strong generalizability and reliability.\u003c/p\u003e\u003cp\u003eAdvanced age constitutes a major independent risk factor for IFI. Elderly patients typically present with a greater comorbidity burden, including chronic obstructive pulmonary disease, diabetes mellitus, and prolonged glucocorticoid/immunosuppressant use, all of which collectively compromise host defenses against fungal pathogens (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). At the immunological level, aging is characterized by progressive T-cell dysfunction (evidenced by CD4+/CD8\u0026thinsp;+\u0026thinsp;ratio disturbances) and diminished neutrophil capacity (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e). Clinically, these vulnerabilities translate to more severe disease presentations, as reflected by larger pulmonary infiltrates and increased dyspnea frequency. An intriguing gender disparity exists in IFI epidemiology, the incidence of IFI is higher in male patients (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). Current evidence suggests estrogen-mediated protection in females through augmentation of innate immune cell (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eSHAP analysis is an interpretable machine learning method based on game theory (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). To address the \"black box\" nature of machine learning models, we employed SHAP interpretability analysis to evaluate features in our model and quantify their respective SHAP values. This approach offers two key clinical advantages: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) it enhances physicians' understanding of the model's decision-making logic, thereby mitigating concerns about algorithmic opacity; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) for individual patients, it provides intuitive visualization of personalized risk factor combinations.\u003c/p\u003e\u003cp\u003eThis study had some limitations. Firstly, there may be data selection bias in this retrospective study. Although it was a multicenter study and the model was externally validated, data collection was conducted by professional radiologists to minimize bias. However, in order to provide more reliable models for clinical use, further prospective analysis is necessary in the future. Second, our study may be constrained by sample size limitations. While the inclusion of 579 patients provides meaningful data, the heterogeneous etiological spectrum of community-acquired pneumonia (CAP) warrants validation in more extensive multicenter cohorts to ensure generalizability. Third, microbiological confirmation posed challenges in our cohort. Although we rigorously excluded invasive fungal infections (IFIs) through standardized diagnostic criteria, approximately [X]% of cases lacked definitive pathogen identification. This diagnostic gap may reflect either technical limitations in conventional microbiological methods or infections caused by fastidious or atypical pathogens. A significant methodological consideration involves the exclusion of COPD as a covariate in our predictive model. This decision was based on two factors: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) the predominance of first-episode pneumonia cases in our cohort with incomplete documentation of pre-existing COPD diagnoses in medical records.\u003c/p\u003e"},{"header":"5. Conclusions","content":"\u003cp\u003eIn conclusion, a radiomics-based machine learning model can effectively diagnose IFI in CAP patients, demonstrating favorable interpretability. The performance of diagnosing IFI is further improved compared to the radiomics and the clinical models. This model is further analyzed by SHAP analysis, providing an interpretable and reliable guidance tool for a personalized clinical assessment of IFI risk in CAP patients early.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eCommunity-acquired pneumonia (CAP); invasive fungal infection (IFI); High-resolution computed tomography (HRCT); region of interest (ROI); linear discriminant analysis (LDA); SHapley Additive exPlanations (SHAP); area under the curve (AUC).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u0026nbsp;\u003c/strong\u003eThe study was in line with the Declaration of Helsinki and was approved by the Institutional Review Board of Chongqing general hospital, Chongqing University [No. KYS2022-009-01]. In view of the fact that the study was retrospective, the Medical Ethical Committee waived the requirement for written informed consent from the patients.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e Not applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material:\u003c/strong\u003e The data that support the findings of this study are available from the corresponding author, upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u0026nbsp;\u003c/strong\u003eNo conflict of interest exists in the submission of this manuscript, and the manuscript is approved by all authors for publication and not under consideration for publication elsewhere, in whole or in part.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e This study was supported by grants from the Chongqing Science and Health Joint Medical Research Project [No. 2022MSXM037], the Basic Research and Frontier Exploration Project of Science and Technology Commission of Yuzhong District, Chongqing [No. 20190126].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contribution:\u0026nbsp;\u003c/strong\u003eWenzhang He, Tie Deng, and Min Gu designed the study. Wenzhang He, Yulin Xiong, and Xuan Huang integrated and analyzed the data. Wenzhang He and Yulin Xiong collected the clinical and CT data and analysis. Min Gu obtained funding. Wenzhang He, Yulin Xiong, Xuan Huang, Haoran Luo, Shaoquan Zhou, and YongboTu prepared the figures. Wenzhang He, Yulin Xiong, Xuan Huang, Haoran Luo, Shaoquan Zhou, YongboTu, Tie Deng, and Min Gu revised the manuscript. Tie Deng and Min Gu supervised the study. All authors read and approved the final manuscript\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u0026nbsp;\u003c/strong\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStudy subjects or cohorts overlap:\u003c/strong\u003e None.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethodology\u003c/strong\u003e\u003c/p\u003e\n\u003cul start=\"50\"\u003e\n \u003cli\u003eSingle-center study\u003c/li\u003e\n \u003cli\u003eRetrospective\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eDiagnostic or prognostic study\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSalluh JIF, P\u0026oacute;voa P, Beane A, Kalil A, Sendagire C, Sweeney DA, et al. Challenges for a broad international implementation of the current severe community-acquired pneumonia guidelines. Intensive Care Med. 2024;50(4):526-38.\u003c/li\u003e\n\u003cli\u003eAliberti S, Dela Cruz CS, Amati F, Sotgiu G, Restrepo MI. Community-acquired pneumonia. Lancet. 2021;398(10303):906-19.\u003c/li\u003e\n\u003cli\u003eNiederman MS, Torres A. Severe community-acquired pneumonia. Eur Respir Rev. 2022;31(166).\u003c/li\u003e\n\u003cli\u003eChean D, Windsor C, Lafarge A, Dupont T, Nakaa S, Whiting L, et al. Severe Community-Acquired Pneumonia in Immunocompromised Patients. Semin Respir Crit Care Med. 2024;45(2):255-65.\u003c/li\u003e\n\u003cli\u003eArvanitis M, Anagnostou T, Fuchs BB, Caliendo AM, Mylonakis E. Molecular and nonmolecular diagnostic methods for invasive fungal infections. Clin Microbiol Rev. 2014;27(3):490-526.\u003c/li\u003e\n\u003cli\u003eRichardson M, Page I. Role of Serological Tests in the Diagnosis of Mold Infections. Curr Fungal Infect Rep. 2018;12(3):127-36.\u003c/li\u003e\n\u003cli\u003eGuarner J, Brandt ME. Histopathologic diagnosis of fungal infections in the 21st century. Clin Microbiol Rev. 2011;24(2):247-80.\u003c/li\u003e\n\u003cli\u003eChen Z, Fan H, Cai J, Li Y, Wu B, Hou Y, et al. High-resolution computed tomography manifestations of COVID-19 infections in patients of different ages. Eur J Radiol. 2020;126:108972.\u003c/li\u003e\n\u003cli\u003edu Plessis AM, Andronikou S, Machemedze T, Griffith-Richards S, Myer L, Mahtab S, et al. High-resolution computed tomography features of lung disease in perinatally HIV-infected adolescents on combined antiretroviral therapy. Pediatr Pulmonol. 2019;54(11):1765-73.\u003c/li\u003e\n\u003cli\u003eChen W, Xiong X, Xie B, Ou Y, Hou W, Du M, et al. Pulmonary invasive fungal disease and bacterial pneumonia: a comparative study with high-resolution CT. Am J Transl Res. 2019;11(7):4542-51.\u003c/li\u003e\n\u003cli\u003eLambin P, Leijenaar RTH, Deist TM, Peerlings J, de Jong EEC, van Timmeren J, et al. Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol. 2017;14(12):749-62.\u003c/li\u003e\n\u003cli\u003eSun R, Limkin EJ, Vakalopoulou M, Dercle L, Champiat S, Han SR, et al. A radiomics approach to assess tumour-infiltrating CD8 cells and response to anti-PD-1 or anti-PD-L1 immunotherapy: an imaging biomarker, retrospective multicohort study. Lancet Oncol. 2018;19(9):1180-91.\u003c/li\u003e\n\u003cli\u003eBitencourt AGV, Gibbs P, Rossi Saccarelli C, Daimiel I, Lo Gullo R, Fox MJ, et al. MRI-based machine learning radiomics can predict HER2 expression level and pathologic response after neoadjuvant therapy in HER2 overexpressing breast cancer. EBioMedicine. 2020;61:103042.\u003c/li\u003e\n\u003cli\u003eYan C, Hao P, Wu G, Lin J, Xu J, Zhang T, et al. Machine learning-based combined nomogram for predicting the risk of pulmonary invasive fungal infection in severely immunocompromised patients. Ann Transl Med. 2022;10(9):514.\u003c/li\u003e\n\u003cli\u003eYang T, Zhang L, Sun S, Yao X, Wang L, Ge Y. Identifying severe community-acquired pneumonia using radiomics and clinical data: a machine learning approach. Sci Rep. 2024;14(1):21884.\u003c/li\u003e\n\u003cli\u003eGong M, Xu J, Li K, Li K, Xia Y, Jing Y, et al. A CT-based nomogram for differentiating invasive fungal disease of the lung from bacterial pneumonia. BMC Med Imaging. 2022;22(1):172.\u003c/li\u003e\n\u003cli\u003eHu Y, Zhong L, Liu H, Ding W, Wang L, Xing Z, et al. Lung CT-based multi-lesion radiomic model to differentiate between nontuberculous mycobacteria and Mycobacterium tuberculosis. Med Phys. 2025;52(2):1086-95.\u003c/li\u003e\n\u003cli\u003eLi P, Wang J, Tang M, Li M, Han R, Zhou S, et al. A CT-based radiomics predictive nomogram to identify pulmonary tuberculosis from community-acquired pneumonia: a multicenter cohort study. Front Cell Infect Microbiol. 2024;14:1388991.\u003c/li\u003e\n\u003cli\u003eWu J, Xia Y, Wang X, Wei Y, Liu A, Innanje A, et al. uRP: An integrated research platform for one-stop analysis of medical images. Front Radiol. 2023;3:1153784.\u003c/li\u003e\n\u003cli\u003eAlbrich WC, Lamoth F. Viral-associated Pulmonary Aspergillosis: Have We Finally Overcome the Debate of Colonization versus Infection? Am J Respir Crit Care Med. 2023;208(3):230-1.\u003c/li\u003e\n\u003cli\u003eHeylen J, Vanbiervliet Y, Maertens J, Rijnders B, Wauters J. Acute Invasive Pulmonary Aspergillosis: Clinical Presentation and Treatment. Semin Respir Crit Care Med. 2024;45(1):69-87.\u003c/li\u003e\n\u003cli\u003eKo BS, Chen WT, Kung HC, Wu UI, Tang JL, Yao M, et al. 2016 guideline strategies for the use of antifungal agents in patients with hematological malignancies or hematopoietic stem cell transplantation recipients in Taiwan. J Microbiol Immunol Infect. 2018;51(3):287-301.\u003c/li\u003e\n\u003cli\u003eWang JW, Yang FF, Zhang CY, Lin JZ, Wang HX, Xu WJ. Imaging Characteristics of Invasive Pulmonary Fungal Infection Secondary to Hematological Diseases and Comparison before and after Treatment. J Healthc Eng. 2021;2021:3736108.\u003c/li\u003e\n\u003cli\u003eObmann VC, Bickel F, Hosek N, Ebner L, Huber AT, Damonti L, et al. Radiological CT Patterns and Distribution of Invasive Pulmonary Aspergillus, Non-Aspergillus, Cryptococcus and Pneumocystis Jirovecii Mold Infections - A Multicenter Study. Rofo. 2021;193(11):1304-14.\u003c/li\u003e\n\u003cli\u003eZhang K, Zhao G, Liu Y, Huang Y, Long J, Li N, et al. Clinic, CT radiomics, and deep learning combined model for the prediction of invasive pulmonary aspergillosis. BMC Med Imaging. 2024;24(1):264.\u003c/li\u003e\n\u003cli\u003eChristenson SA, Smith BM, Bafadhel M, Putcha N. Chronic obstructive pulmonary disease. Lancet. 2022;399(10342):2227-42.\u003c/li\u003e\n\u003cli\u003eMartinez-Gomez D, Ortega FB, Hamer M, Lopez-Garcia E, Struijk E, Sadarangani KP, et al. Physical Activity and Risk of Metabolic Phenotypes of Obesity: A Prospective Taiwanese Cohort Study in More Than 200,000 Adults. Mayo Clin Proc. 2019;94(11):2209-19.\u003c/li\u003e\n\u003cli\u003eGarrido-Rodr\u0026iacute;guez V, Herrero-Fern\u0026aacute;ndez I, Castro MJ, Castillo A, Rosado-S\u0026aacute;nchez I, Galv\u0026aacute; MI, et al. Immunological features beyond CD4/CD8 ratio values in older individuals. Aging (Albany NY). 2021;13(10):13443-59.\u003c/li\u003e\n\u003cli\u003eGoyani P, Christodoulou R, Vassiliou E. Immunosenescence: Aging and Immune System Decline. Vaccines (Basel). 2024;12(12).\u003c/li\u003e\n\u003cli\u003eEgger M, Hoenigl M, Thompson GR, 3rd, Carvalho A, Jenks JD. Let\u0026apos;s talk about sex characteristics-As a risk factor for invasive fungal diseases. Mycoses. 2022;65(6):599-612.\u003c/li\u003e\n\u003cli\u003eZhao L, Huang S, Mei S, Yang Z, Xu L, Zhou N, et al. Pharmacological activation of estrogen receptor beta augments innate immunity to suppress cancer metastasis. Proc Natl Acad Sci U S A. 2018;115(16):E3673-e81.\u003c/li\u003e\n\u003cli\u003ePonce-Bobadilla AV, Schmitt V, Maier CS, Mensing S, Stodtmann S. Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development. Clin Transl Sci. 2024;17(11):e70056.\u003c/li\u003e\n\u003cli\u003eGhasemi A, Hashtarkhani S, Schwartz DL, Shaban-Nejad A. Explainable artificial intelligence in breast cancer detection and risk prediction: A systematic scoping review. Cancer Innov. 2024;3(5):e136.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"727\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" style=\"width: 100%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 1. Comparison of patient characteristics across the training, internal validation and external validation datasets.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eClinical variables\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTraining dataset\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(n=239)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eInternal validation\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003edataset (n=101)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eExternal validation\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003edataset (n=230)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP-value\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003eRadiological feature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; ROI Volume (Mean \u0026plusmn; SD, cm\u003csup\u003e3\u003c/sup\u003e)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e218.3 \u0026plusmn; 282.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e157.5 \u0026plusmn; 191.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e340.8 \u0026plusmn; 341.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; ROI Average HU Value (Mean \u0026plusmn; SD, HU)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e-508.1 \u0026plusmn; 115.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e-497.2 \u0026plusmn; 131.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e-499.3 \u0026plusmn; 149.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.722\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003eEpidemiologic factors\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Gender, male, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e134 (56.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e44 (43.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e153 (66.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; Age (Mean \u0026plusmn; SD, years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e62.1 \u0026plusmn; 19.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e57.5 \u0026plusmn; 17.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e68.4 \u0026plusmn; 15.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;non-productive cough, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e30 (12.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e12 (11.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e7 (3.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;productive cough, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e193 (80.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e84 (83.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e223 (97.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Hemoptysis, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e22 (9.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e9 (8.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e17 (7.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.764\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Dyspnea, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e115 (48.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e46 (39.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e74 (32.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; Chest pain, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e35 (16.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e15 (14.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e26 (11.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.503\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Confusion, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e2 (0.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e0 (0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e1 (0.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.603\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; Chest tightness, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e203/36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e79/22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e220/10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Myalgia, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e190/49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e79/22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e227/3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Duration of Symptoms (Median, Q1, Q3, days)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e7.0 (4.0, 15.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e7.0 (4.0, 15.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e5.0 (3.0, 10.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; Fever, yes, n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e139 (58.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e61 (60.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e102 (44.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.003\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003eLaboratory findings\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; White Blood Cell (Median, Q1, Q3, 10^9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e7.50 (5.59, 9.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e7.44 (5.51, 10.04)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e8.36 (6.49, 11.24)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.022\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Neutrophil (Median, Q1, Q3, 10^9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e5.57 (4.00, 8.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e5.49 (3.76, 7.77)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e6.48 (4.65, 9.27)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.003\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Lymphocyte (Median, Q1, Q3, 10^9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e1.16 (0.85, 1.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e1.19 (0.89, 1.63)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e0.99 (0.59, 1.42)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.011\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Monocyte (Median, Q1, Q3, 10^9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e0.53 (0.38, 0.69)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e0.50 (0.35, 0.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e0.52 (0.35, 0.73)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.452\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Eosinophil (Median, Q1, Q3, 10^9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e0.05 (0.01, 0.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e0.06 (0.01, 0.15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e0.04 (0.01, 0.15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.627\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Basophil (Median, Q1, Q3, 10^9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e0.02 (0.01, 0.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e0.02 (0.01, 0.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e0.02 (0.01, 0.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.767\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; Red Blood Cell (Median, Q1, Q3, 10^12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e4.38 (3.95, 4.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e4.44 (4.18, 4.77)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e4.18 (3.70, 4.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.917\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; Hemoglobin (Median, Q1, Q3, g/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e131.0 (118.0, 144.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e132.0 (120.0, 141.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e125.0 (109.0, 139.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;C-reactive Protein (Median, Q1, Q3, mg/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e31.9 (10.8, 62.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e31.9 (9.5, 73.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e34.7 (8.8, 83.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e0.665\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 33.7001%;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;Neutrophil/Lymphocyte Ratio (Median, Q1, Q3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.5323%;\"\u003e\n \u003cp\u003e4.82 (3.16, 7.60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 18.5695%;\"\u003e\n \u003cp\u003e4.75 (2.90, 6.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19.1197%;\"\u003e\n \u003cp\u003e6.49 (3.57, 13.26)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 9.0784%;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" style=\"width: 100%;\"\u003e\n \u003cp\u003eAbbreviations: ROI, Region of interest; HU, Hounsfield Unit.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"647\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\" style=\"width: 601px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 2. Univariate and multivariate analysis of the clinical variables.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 148px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariable\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 248px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUnivariate\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 232px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMultivariate\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOdds\u0026nbsp;ratio\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95%\u0026nbsp;CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOdds\u0026nbsp;ratio\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95%\u0026nbsp;CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eROI Average HU Value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.998-1.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.746\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eROI Volume\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e3.137\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e1.829-5.382\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e2.337\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e1.316-4.148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.006\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eChest pain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.687\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.367-1.291\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.244\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eChest tightness\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.955\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.522-1.748\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.882\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eConfusion\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.000-0.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.998\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eDuration of Symptoms\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e1.001-1.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.028\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e1.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e0.999-1.022\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.062\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eDyspnea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.867\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e1.270-2.745\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.002\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e1.775\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e1.178-2.675\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.006\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eFever\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.826\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.587-1.163\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.274\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e1.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e0.995-1.085\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.081\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e1.014-1.039\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e1.016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e1.001-1.030\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.032\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eHemoptysis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.787\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.372-1.663\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.530\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eMyalgia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.290\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.123-0.684\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.005\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.474\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e0.194-1.157\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.101\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eNon-productive cough\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.382\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.149-0.983\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.046\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.604\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e0.226-1.617\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.316\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.887\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e1.248-2.853\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.003\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e1.588\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e1.011-2.494\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003e0.045\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eProductive cough\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.594\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.813-3.123\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.175\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eBasophil Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.831\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.165-20.347\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.623\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eC reactive Protein\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.165\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.739-1.836\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.510\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eEosinophil Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.805\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.553-1.173\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.259\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eHemoglobin\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.574\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.391-0.842\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.005\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eLymphocyte Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.778\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.531-1.141\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.199\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eMonocyte Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.026\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.683-1.539\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.903\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eNeutrophil Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.318\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.918-1.892\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.134\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eRed Blood Cell Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e0.556\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.385-0.801\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.002\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e0.839\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e0.619-1.138\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.259\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eWhite Blood Cell Count\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.176\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.831-1.664\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.360\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 148px;\"\u003e\n \u003cp\u003eNLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 75px;\"\u003e\n \u003cp\u003e1.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 123px;\"\u003e\n \u003cp\u003e0.995-1.020\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 50px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.260\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 19px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 110px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 46px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"8\" style=\"width: 647px;\"\u003e\n \u003cp\u003eAbbreviations: ROI, Region of interest; HU, Hounsfield Unit; CI, confidence interval; NLR, neutrophil-to-lymphocyte ratio.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"604\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\" style=\"width: 604px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 3. Detailed performance of the machine learning classifiers for invasive fungal infection identification in the development cohort through cross-validation.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eClassifier\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 197px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTraining\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"3\" style=\"width: 197px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eValidation\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 120px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSignificant Overfitting\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP-value\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e95% CI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP-value\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eLDA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.758\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.709-0.803\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003ereference\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.741\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.691-0.787\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003ereference\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eNo (\u003cem\u003eP\u003c/em\u003e = 0.066)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eKNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.906\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.870-0.935\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.656\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.603-0.707\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.018\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eYes (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eLR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.723\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.672-0.770\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.055\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.645\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.591-0.696\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.004\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eNo (\u003cem\u003eP\u003c/em\u003e = 0.100)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eNB\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.748\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.698-0.793\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.639\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.711\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.660-0.759\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.220\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eYes (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eRF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.989-1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.707\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.655-0.755\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.194\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eYes (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eSVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.891\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.853-0.922\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e\u0026lt;0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 15px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 45px;\"\u003e\n \u003cp\u003e0.606\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 91px;\"\u003e\n \u003cp\u003e0.552-0.658\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 61px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 120px;\"\u003e\n \u003cp\u003eYes (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\" style=\"width: 604px;\"\u003e\n \u003cp\u003eAbbreviations: AUC, areas under the receiver operating characteristic curves; CI, confidence interval; LDA, confidence interval; Linear discriminant analysis;\u0026nbsp;KNN,\u0026nbsp;k-nearest neighbor;\u0026nbsp;LR,\u0026nbsp;logistic regression;\u0026nbsp;NB,\u0026nbsp;na\u0026iuml;ve Bayes;\u0026nbsp;RF,\u0026nbsp;random forest;\u0026nbsp;SVM,\u0026nbsp;support vector machine.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"579\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"8\" style=\"width: 579px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTable 4. Model performance comparison in the training, internal validation and external validation datasets.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDataset\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAUC (95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eP-value\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSEN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSPE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 76px;\"\u003e\n \u003cp\u003eTraining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eClinical\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.777 (0.719-0.828)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.007\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e75.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e74.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e41.0%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e93.0%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eRadiomics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.765 (0.706-0.817)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.010\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e66.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e71.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e35.3%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e90.3%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eCombined\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.836 (0.783-0.881)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003ereference\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e77.8%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e74.2%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e41.2%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e93.5%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 76px;\"\u003e\n \u003cp\u003eInternal validation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eClinical\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.719 (0.621-0.804)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.030\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e52.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e72.0%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e30.3%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e86.8%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eRadiomics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.724 (0.626-0.808)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.048\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e57.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e74.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e34.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e88.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eCombined\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.808 (0.718-0.880)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003ereference\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e73.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e74.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e40.0%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e92.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 76px;\"\u003e\n \u003cp\u003eExternal validation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eClinical\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.707 (0.644-0.765)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.001\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e60.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e72.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e35.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e88.2%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eRadiomics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.709 (0.646-0.767)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003e0.042\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e52.2%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e71.4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e31.2%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e85.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 76px;\"\u003e\n \u003cp\u003eCombined\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 135px;\"\u003e\n \u003cp\u003e0.786 (0.728-0.837)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 64px;\"\u003e\n \u003cp\u003e\u003cem\u003ereference\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e69.6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e78.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e45.1%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e91.3%\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"8\" style=\"width: 579px;\"\u003e\n \u003cp\u003eAbbreviations: AUC, areas under the receiver operating characteristic curves; CI, confidence interval; SEN, \u0026nbsp; \u0026nbsp; \u0026nbsp; sensitivity; SPE, \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;specificity; PPV, \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;positive predictive value; NPV, negative predictive value.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Machine learning, Invasive fungal infections, Community-acquired pneumonia, Radiomics, Interpretability","lastPublishedDoi":"10.21203/rs.3.rs-8145288/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8145288/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjectives\u003c/h2\u003e\u003cp\u003eTo develop a high-resolution computed tomography (HRCT) radiomics-based interpretable machine learning model for diagnosing invasive fungal infection (IFI) in community-acquired pneumonia (CAP) patient.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eA total of 570 CAP patients who underwent HRCT from July 2022 to August 2024 in Center 1 and Center 2 were recruited. A vb-net pneumonia automatic segmentation algorithm was employed. Three models, a radiomics model (HRCT-derived radiomics features), a clinical model (clinical variables), and a combined model (integrating both), were developed. The performance of these models was evaluated through receiver operating characteristic analysis with respect to the area under the curve (AUC). Clinical utility was evaluated by using decision curve analysis. The Shapley Additive Explanation tool was employed.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003e239 (mean age: 62.1\u0026thinsp;\u0026plusmn;\u0026thinsp;19.3 years; 134 male), 101 (mean age: 57.5\u0026thinsp;\u0026plusmn;\u0026thinsp;17.3 years; 44 male), and 230 (mean age: 68.4\u0026thinsp;\u0026plusmn;\u0026thinsp;15.3 years; 153 male) patients were included in the training, internal validation, and external validation datasets. Based on linear discriminant analysis classifier, the AUCs of the clinical, radiomics, and combined models were 0.719, 0.724, and 0.808, respectively, in the internal validation dataset; and 0.707, 0.709, and 0.786, respectively, in the external validation dataset. The combined model yielded a superior net benefit relative to both the clinical and radiomics models alone. Age exerted the greatest influence on the predictions of the combined model, while the three most important radiomics features were all higher-order texture features.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eA radiomics-based machine learning model can effectively diagnose IFI in CAP patients, demonstrating favorable interpretability.\u003c/p\u003e\u003ch2\u003eClinical relevance statement:\u003c/h2\u003e\u003cp\u003eThe radiomics-based interpretable model developed in this study has significant clinical relevance as it offers a non-invasive method for diagnosing invasive fungal infection in community-acquired pneumonia patients and holds promise as an early diagnostic tool.\u003c/p\u003e","manuscriptTitle":"An interpretable radiomics-machine learning model for diagnosing invasive fungal infections in community-acquired pneumonia: multicenter study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-03 11:48:06","doi":"10.21203/rs.3.rs-8145288/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-30T10:05:15+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-22T20:25:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"334265944940614421027601050898293542050","date":"2026-03-16T10:15:00+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-22T02:45:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"332406497903230843023532858797494850851","date":"2025-12-15T15:07:16+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"227798695721617292432278654323970605950","date":"2025-12-14T23:42:12+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-12-01T21:50:18+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-11-21T08:13:12+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-19T13:21:18+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-19T13:18:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-11-18T12:13:53+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"737405fa-caaa-4fec-9927-c49f0c78c3cf","owner":[],"postedDate":"December 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":58986800,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":58986801,"name":"Health sciences/Diseases"},{"id":58986802,"name":"Health sciences/Medical research"}],"tags":[],"updatedAt":"2026-05-16T05:08:15+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-03 11:48:06","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8145288","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8145288","identity":"rs-8145288","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}