Development and Application of a T1-Weighted Enhanced MRI Radiomics Nomogram for Differentiating Cerebral Tuberculoma from Brain Metastasis

Development and Application of a T1-Weighted Enhanced MRI Radiomics Nomogram for Differentiating Cerebral Tuberculoma from Brain Metastasis | 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 Research Article Development and Application of a T1-Weighted Enhanced MRI Radiomics Nomogram for Differentiating Cerebral Tuberculoma from Brain Metastasis Ruiyun Liang, Quhua Yin, Chengcheng Li, Huili Ren, Huiru Li, Min Song, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9164706/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Objective Cerebral tuberculoma and brain metastasis are often indistinguishable on MRI, leading to misdiagnosis. This study develops a T1WI-enhanced MRI radiomics-based nomogram to accurately differentiate them. Methods A total of 482 patients (290 males, 192 females; mean age 56.08 ± 12.96 years) including 254 with cerebral tuberculoma and 228 with brain metastasis were enrolled from two centers. Radiomic features were extracted, standardized, and reduced via significance testing, Pearson correlation, and LASSO regression. A radiomic model was built, then integrated with clinical predictors into a nomogram. Performance was assessed across training, internal validation, and external validation cohorts using ROC curves, calibration, and decision curve analysis (DCA). Results A total of 20 radiomic features were selected for model development. The radiomic model demonstrated areas under the curve (AUC) of 0.907, 0.891and 0.886 in training, internal validation, and independent external validation cohorts, respectively, significantly outperforming models based solely on clinical factors, which achieved AUCs of 0.781, 0.712, and 0.828 ( p < 0.05). The nomogram, which integrated the radiomic scores, age, ring enhancement, meningeal thickening, and carcinoembryonic antigen (CEA) levels, exhibited superior performance with AUCs of 0.951, 0.937, and 0.940 across the three cohorts. DCA demonstrated that the nomogram provided greater clinical net benefits across a range of threshold probabilities, and calibration curves indicated strong consistency between predicted probabilities and actual outcomes. Conclusion A nomogram combining clinical factors and T1WI-enhanced MRI radiomics accurately differentiates cerebral tuberculoma from brain metastasis, offering robust decision support for individualized diagnosis and treatment. cerebral tuberculoma brain metastasis brain MRI radiomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction C erebral tuberculoma and brain metastasis are pathologically distinct—one arising from Mycobacterium tuberculosis infection, the other from hematogenous spread of solid tumors [ 1 ][ 2 ]. Yet both cause severe neurological deficits and increasingly overlap in incidence due to shifting TB epidemiology and improved cancer survival [ 3 – 6 ]. Clinically and radiologically, they present with similar symptoms (headaches, seizures) and nonspecific imaging features, making reliable non-invasive differentiation difficult. However, their treatment and prognosis diverge sharply, underscoring the critical need for accurate preoperative distinction [ 7 ][ 8 ]. Conventional MRI remains essential for evaluating intracranial lesions, relying on features such as signal intensity, morphology, margins, and contrast enhancement patterns. However, both tuberculomas and metastases frequently present as round or irregular nodules with overlapping nodular or ring-like enhancement [ 9 , 10 ], limiting diagnostic specificity. Such nonspecific imaging features lead to subjective interpretation and often necessitate invasive procedures like biopsy or lumbar puncture for definitive diagnosis—particularly challenging and risky for deep-seated lesions [ 11 ]. These limitations underscore the urgent need for non-invasive, objective diagnostic tools. R adiomics enables non-invasive detection of subtle pathological changes invisible to the naked eye and has proven valuable in tumor grading and infectious central nervous system(CNS) diseases [ 12 ][ 13 ]. However, individual radiomic features are often unstable due to technical variability, while clinical parameters alone offer limited predictive power [ 14 – 17 ]. Integrating both data types has therefore emerged as a promising strategy, yet comprehensive models specifically designed to differentiate cerebral tuberculoma from brain metastasis remain lacking [ 18 ][ 19 ]. T o address this gap, we aim to develop and validate a nomogram that integrates T1WI-enhanced MRI radiomic features with key clinical indicators using multi-center data. Through rigorous feature selection, model optimization, and validation across internal and independent external cohorts, we seek to establish a robust, generalizable tool for early and accurate differential diagnosis—supporting individualized treatment decisions without recourse to invasive procedures. Materials and Methods Study Population This retrospective study, approved by the Ethics Committees of Guangzhou Chest Hospital (KYSB-2024-035) and Hunan Chest Hospital (LS2026011201) in accordance with the Declaration of Helsinki, waived the requirement for written informed consent. The initial cohort included patients who underwent radiological assessments at both institutions from January 2021 to December 2024. Inclusion criteria were: (1) meeting the diagnostic criteria for cerebral tuberculoma or brain metastasis, with complete clinical and T1WI enhanced MRI data; (2) possessing at least one identifiable lesion in the brain parenchyma on T1WI enhanced MRI scans; (3) histological confirmation of diagnosis via surgical resection or biopsy; and (4) a molecular biological diagnosis validated by cerebrospinal fluid obtained through lumbar puncture.Patients were excluded based on the following criteria: (1) unclear clinical diagnosis, with ineffective anti-tuberculosis or anti-tumor treatments; (2) unclear pathological diagnosis attributed to insufficient tissue samples; (3) history of other intracranial malignancies or co-infections with other pathogens (e.g., viral, bacterial, or fungal infections); and (4) poor image quality due to motion artifacts. The reference standard for the final diagnosis of cerebral tuberculoma or brain metastasis was established by histopathological examination of surgically resected or biopsied tissue, or by molecular biological analysis of cerebrospinal fluid (CSF) samples.Ultimately, the study included 482 patients, comprising 406 from Guangzhou Chest Hospital (204 diagnosed with cerebral tuberculoma and 202 with brain metastasis), randomly stratified into a training set (n = 284) and an internal validation set (n = 122). A total of 76 patients from Hunan Chest Hospital were allocated to the independent external validation cohort (n = 76). A detailed workflow diagram is presented in Fig. 1 . Image Segmentation and Feature Extraction Intra- and inter-observer reproducibility was assessed using intraclass correlation coefficients (ICCs). Two radiologists randomly selected 60 T1WI enhanced MRI images (30 cerebral tuberculomas and 30 brain metastases) and delineated volumes of interest (VOIs). Radiologist 1 repeated the segmentation after two weeks. An ICC threshold of > 0.80 was considered indicative of good consistency. Following reliability assessment, Radiologist 1 performed VOI segmentation for the remaining cases. Only radiomic features with ICCs > 0.80 were retained for subsequent analysis. Using 3D Slicer software (version 5.4.0, https://www.slicer.org/ ), regions of interest (ROI) were manually delineated on T1-weighted enhanced MRI images along the interface between each enhanced lesion and the surrounding brain tissue. Representative segmentation examples are shown in Fig. 2 . From each ROI, 1,746 candidate radiomic features were extracted using the ImaScholar platform (Version 3.5), including first-order statistics, gray level size zone matrix (GLSZM), neighboring gray tone difference matrix (NGTDM), and gray level dependence matrix (GLDM) features (Supplementary data 1). Dimensionality reduction was performed using significance testing ( p 0.85). The most informative features were then selected via Least Absolute Shrinkage and Selection Operator(LASSO) regression with ten-fold cross-validation, yielding an optimal subset of 20 features. These were linearly combined using their respective coefficients to generate a radiomic score (Radscore) for each patient, thereby establishing the radiomic model. The index test was T1-weighted enhanced MRI radiomics analysis, which involved extracting radiomic features from preprocessed images. Model Construction, Evaluation, and Nomogram Development We developed clinical, radiomic, and combined models for distinguishing cerebral tuberculoma from brain metastasis. Following feature selection, the radiomic model was constructed using the ImaScholar platform (Version 3.5). The workflow for feature extraction and model development is illustrated in Fig. 3 . For the clinical model, univariate logistic regression was used to identify significant risk factors and then select independent predictors via multivariate analysis. Significant variables from both univariate and multivariate analyses were subsequently incorporated into the combined model. A nomogram was developed to visualize the combined model, enabling assessment of variable contributions and facilitating prediction of individual risk. Model discrimination was compared using the DeLong test (AUC, sensitivity, specificity, accuracy) across all cohorts. Nomogram performance was further assessed via calibration (Hosmer-Lemeshow) and decision curve analyses. Statistical Analysis Statistical analyses were performed using R software (version 4.4.2). Throughout the analysis, various specialized statistical packages—including glmnet, cmprsk, rms, rmda, and devtools—were employed. Continuous data were presented as mean±standard deviation. Univariate and multivariate logistic regression models were utilized to identify independent predictors of outcomes based on collected clinical variables, with p < 0.05 considered statistically significant. The discriminative performance of each model was appraised through the construction of receiver operating characteristic (ROC) curves. Statistical inference was conducted via two-tailed tests, with significance set at p < 0.05. Sample size was calculated based on a power of 80% and alpha of 0.05 to detect a minimum AUC difference of 0.1 between models.Author Liang served as the statistical guarantor, responsible for the transparency of statistical analyses. By following these methodological approaches, we aim to develop a robust and interpretable model capable of distinguishing between cerebral tuberculoma and brain metastasis, thus enhancing the diagnostic accuracy and facilitating personalized treatment strategies for patients. Results Baseline Characteristics of Patients This study analyzed 482 patients (290 males, 192 females; mean age 56.08 ± 12.96 years), including 254 with cerebral tuberculoma and 228 with brain metastasis. Of these, 406 patients were from Guangzhou Chest Hospital (204 tuberculoma, 202 metastasis), and 76 from Hunan Chest Hospital (50 tuberculoma, 26 metastasis). The Guangzhou cohort was randomly split into a training set (142 tuberculoma, 142 metastasis) and an internal validation set (62 tuberculoma, 60 metastasis). The Hunan cohort served as an independent external validation set (50 tuberculoma, 26 metastasis). Table 1 summarizes the demographic and clinical imaging characteristics of the study population. Reproducibility of Feature Extraction The process of radiomic feature extraction demonstrated robust reproducibility. The intra-observer intraclass correlation coefficients (ICCs) ranged from 0.823 to 0.935, while the inter-observer ICCs fell between 0.802 and 0.896, with all values surpassing the threshold of > 0.80, signifying stable and dependable feature extraction results. Feature Selection and Radscore Distribution Our analysis identified a total of 1,746 radiomic features extracted from each region of interest (VOI) on T1-weighted enhanced MRI images. These features encompassed first-order statistics, gray level size zone matrix (GLSZM), neighboring gray tone difference matrix (NGTDM), and gray level dependence matrix (GLDM) (refer to Supplementary Data 1). All extracted radiomic features underwent preprocessing and standardization via the Z-score method. Through rigorous significance testing, we selected features that achieved statistical significance ( p < 0.05). Furthermore, we eliminated features exhibiting a Pearson correlation coefficient greater than 0.85 between the conditions, narrowing down to 237 radiomic features. We then employed LASSO regression to achieve dimensionality reduction, ultimately isolating a subset of 20 predictive radiomic features to construct our radiomic model (as illustrated in Figs. 3A-C). The radiomic feature label was constructed by linearly combining these features, weighted according to their respective coefficients (Fig. 4 presents the waterfall plot of Radscore). The waterfall plot effectively visualizes the Radscore distribution characteristics for cerebral tuberculoma and brain metastasis across the training, internal validation, and external validation cohorts. Construction of Clinical Model, Radiomic Model, and Combined Model Statistical analysis, specifically the Wilcoxon test, revealed significant differences in Radscore between the cerebral tuberculoma and brain metastasis groups ( p < 0.05). In contrast, significant differences were detected in factors such as age, ring enhancement, meningeal thickening, and levels of biological markers (CEA, CA199, IGRAs) ( p < 0.05). Through multivariate regression analysis, we identified age ( p = 0.002), CEA ( p < 0.001), meningeal thickening ( p = 0.002), ring enhancement ( p = 0.02),and IGRAs( p = 0.02) as independent predictive factors for lesion classification, leading us to incorporate these variables into the clinical model (Table 2). Ultimately, we constructed a combined model integrating both Radscore and the independent predictive factors identified. Table 2 Univariate and Multivariate Logistic Regression Analysis Variable Univariable analysis Multivariable analysis (clinical model parameters) Multivariable analysis (combined model parameters) OR (95%CI) p -value OR (95%CI) p -value OR (95%CI) p -value Sex 0.943(0.586–1.516) 0.81 — — — — Age(years, mean ± SD) 1.027(1.007–1.048) 0.01 1.038 (1.014–1.062) 0.002 1.041 (1.007–1.076) 0.02 Emphysema 1.328(0.788–2.249) 0.29 — — — — Diabetes Mellitus 0.764(0.441–1.316) 0.33 — — — — Whole-nodule enhancement 0.826(0.502–1.355) 0.45 — — — — ring enhancement 0.482(0.29–0.792) 0.004 0.504 (0.286–0.891) 0.02 0.356 (0.148–0.858) 0.02 meningeal thickening 0.578(0.358–0.926) 0.02 0.416 (0.237–0.730) 0.002 0.292 (0.125–0.686) 0.01 Hydrocephalus 0.69(0.412–1.148) 0.16 — — — — CA125(u/ml) 0.999(0.996–1.001) 0.30 — — — — CEA(ng/ml) 1.017(1.011–1.026) < 0.001 1.021 (1.013–1.030) < 0.001 1.030 (1.017–1.042) 10 1.109(0.662–1.861) 0.70 — — — — 1 1.776(0.844–3.887) 0.14 — — — — 2 ~ 10 1.602(0.995–2.593) 0.05 — — — — IGRAs 0.557(0.335–0.92) 0.02 0.508 (0.283–0.911) 0.02 — — rad_score 0.893(0.867–0.915) < 0.001 — — 0.876 (0.846–0.906) 10), resulting in a reduced AIC value after exclusion. Performance of Clinical Model, Radiomic Model, and Combined Model Receiver operating characteristic (ROC) curves were used to evaluate the performance of the clinical, radiomic, and combined models across the training, internal validation, and external validation cohorts. The area under the curve (AUC), sensitivity, specificity, and accuracy for each model are summarized in Fig. 5 and Table 3. Table 3 Comparison of diagnostic performance of clinical models, radiomics models, and combined models in training cohorts, internal validation cohorts, and external validation cohorts Model AUC (95%CI) Accuracy(%) Sensitivity(%) Specificity(%) PPV (%) NPV (%) Clinical model Training cohort 0.781 (0.729–0.834) 71.83 67.61 76.06 73.85 70.13 Internal validation cohort 0.712(0.621–0.802) 66.39 55.00 77.42 70.21 64.00 External validation cohort 0.828 (0.735–0.920) 76.32 76.92 76.00 62.50 86.36 Radiomics model Training cohort 0.907(0.873–0.941) 84.86 79.58 90.14 88.98 81.53 Internal validation cohort 0.891 (0.836–0.947) 81.97 93.33 70.97 75.68 91.67 External validation cohort 0.886 (0.813–0.959) 76.32 92.31 68.00 60.00 94.44 Combined model Training cohort 0.951 (0.929–0.973) 88.03 84.51 91.55 90.91 85.53 Internal validation cohort 0.937 (0.894–0.98) 90.16 86.67 93.55 92.86 87.88 External validation cohort 0.940 (0.890–0.989) 86.84 80.77 90.00 80.77 90.00 Note: AUC (area under the curve), CI (confidence interval), PPV (positive predictive value), NPV (negative predictive value) All three models demonstrated good discriminative ability. The clinical model achieved AUCs of 0.781, 0.712, and 0.828 in the training, internal validation, and external validation cohorts, respectively. The radiomic model showed excellent performance, with corresponding AUCs of 0.907, 0.891, and 0.886. The combined model outperformed both, achieving AUCs of 0.951, 0.937, and 0.940 across the three cohorts. The DeLong test confirmed that the combined model's AUC was significantly higher than those of the clinical and radiomic models in all cohorts ( p < 0.05). Significant differences were also observed between the clinical and radiomic models in the training cohort (0.781 vs. 0.907, p < 0.001) and internal validation cohort (0.712 vs. 0.891, p = 0.002), as well as between the clinical and combined models in the external validation cohort (0.828 vs. 0.937, p = 0.02). Nomogram and Its Validation A nomogram was developed to visualize the combined model and its contributing variables (Fig. 6A). Calibration, assessed by the Hosmer-Lemeshow test, showed no significant deviation from ideal predictions in the training p = 0.792 internal validation p = 0.200, or external validation p = 0.2775 cohorts (Fig. 6B). Decision curve analysis indicated that the nomogram offers a high clinical net benefit for differential diagnosis across threshold probabilities from 0.1 to 0.9 (Fig. 6C). Discussion Differentiating cerebral tuberculoma from brain metastasis remains clinically challenging due to overlapping imaging features, yet treatment and prognosis differ fundamentally. We developed and validated a nomogram integrating T1WI-enhanced MRI radiomic features with clinical predictors, achieving excellent discrimination and calibration across multi-center cohorts. This non-invasive tool provides robust quantitative support for personalized differential diagnosis. We identified 20 core radiomic features for further analysis, encompassing a variety of categories such as wavelet transforms, first-order statistics, gray-level co-occurrence matrices (GLCM), and gray-level run length matrices (GLRLM), among others. These features elucidate the quantitative morphological differences between tuberculoma and brain metastasis on MRI, particularly highlighting distinctions in tumor boundary smoothness and tissue density distributions. Existing literature suggests that tuberculomas, as a result of granulomatous inflammation induced by Mycobacterium tuberculosis, typically display relatively well-defined margins and ring enhancement. This characteristic effectively distinguishes them from the infiltrative boundaries that are often observed in various malignant tumors [ 1 ][ 20 ]. In contrast, brain metastasis generally exhibit a high degree of cellular heterogeneity and angiogenesis, resulting in blurred margins and heterogeneous internal enhancement patterns [ 21 ]. Mechanistically, our findings indicate that the radiomic features we analyzed reflect notable differences in the tissue microenvironment associated with each condition. The pathological characteristics inherent to tuberculomas lead to first-order features exhibiting lower overall density and a more symmetrical gray-level distribution. Additionally, GLCM features reveal an enhanced texture uniformity, while shape features indicate greater flatness due to the tumors' regular morphologies. On the other hand, brain metastasis comprise densely populated cancerous regions and areas of necrosis, characterized by a rich vascular supply and uneven necrotic patches. T1WI-enhanced MRI often reveals a heterogeneous ring enhancement consistent with significant pathological variability. The corresponding radiomic features associated with brain metastasis include first-order statistics showing higher mean values due to abundant tumor blood supply, coupled with positive skewness in regions dense with cancer cells. Furthermore, GLCM features exhibit lower maximum correlation coefficients, while wavelet transform features identify high-frequency variability at the tumor infiltration edges [ 22 ][ 23 ]. This study further delineates the capacity of high-dimensional feature analysis to uncover distinctions that extend beyond traditional imaging observations. Advanced parameters such as GLCM and texture heterogeneity effectively differentiate between cerebral tuberculoma and brain metastasis. In contrast to previous methodologies that largely relied on subjective imaging evaluations or limited low-dimensional quantitative metrics, our research employs a multi-parameter, high-throughput feature mining approach. This strategy not only reveals subtle differences in imaging phenotypic mechanisms but also enriches the understanding of tissue-level heterogeneity between cerebral tuberculosis infections and tumor lesions. Overall, the integration of advanced radiomic analysis into clinical practice represents a promising avenue for enhancing diagnostic accuracy and tailoring treatment modalities for patients with intracranial lesions. Future studies should aim to further refine these models and explore their applicability across diverse patient populations and imaging modalities. In this study, we extracted 20 core radiomic features from T1-weighted imaging (T1WI) enhanced magnetic resonance imaging (MRI) of 406 patients to develop a comprehensive model that integrates these features with clinical indicator data for the differentiation between tuberculoma and brain metastasis. Notably, our model identified age, carcinoembryonic antigen (CEA) levels, ring enhancement, and meningeal thickening on T1WI-enhanced MRI as independent risk factors for distinguishing between these two conditions. The combined model demonstrated remarkable performance, achieving an area under the curve (AUC) of 0.951 in the training cohort, 0.937 in the internal validation cohort, and 0.940 in the independent external validation cohort. Previous research corroborates the effectiveness of combined models in differentiating tuberculoma from malignant tumors. For instance, a study evaluating a deep learning model based on computed tomography (CT) for distinguishing tuberculous granulomas from lung adenocarcinoma reported AUC values of 0.903, 0.933, and 0.914 across its training, internal validation, and external validation sets, respectively—figures comparable to those observed in our study. Despite differences in the anatomy and pathological contexts of the studied organs, our findings suggest that the combined model offers superior discriminative capability in distinguishing tuberculoma from malignant tumors, as evidenced by its enhanced overall performance. The strength of our model lies in its comprehensive integration of clinical indicators, lesion morphology, texture features, and high-dimensional implicit information during feature extraction and model construction. This allows for precise identification and classification of lesions using advanced machine learning algorithms. Unlike traditional imaging diagnostic approaches that rely predominantly on morphological and signal change assessments, our model significantly improves the detection of subtle differences between tuberculoma and metastases, thereby enhancing both the repeatability and objectivity of the diagnostic results. To facilitate individualized classification and aid radiologists in distinguishing between tuberculoma and brain metastasis, we also developed a bar graph visualization tool based on our combined model. Compared to conventional diagnostic methods, the classification predictions generated by this visualization tool are faster, more user-friendly, and more accurate, ultimately assisting radiologists in making precise diagnoses. This study systematically assessed the sensitivity and specificity of the combined model through cross-validation, thereby validating the model's stability across diverse data distributions. Unlike traditional single-center or retrospective studies with small sample sizes, cross-validation provides a more realistic reflection of the model's performance in actual clinical scenarios. Existing literature indicates that radiomic models often exhibit decreased performance in independent external cohorts compared to their initial training sets, which can be attributed to factors such as sample distribution, imaging acquisition protocols, and patient heterogeneity. In our analysis, utilizing an external validation cohort, we found that the combined model achieved a sensitivity of 0.8077 and specificity of 0.9000, outperforming the clinical model (sensitivity: 0.7692, specificity: 0.7600). Although the sensitivity of the combined model was lower than that of the radiomics model (0.9231), its specificity was superior (0.6800). The combined model achieved a balanced trade-off among sensitivity, specificity, and positive predictive value (PPV), with stable and accurate performance in the independent external cohort. These findings indicate that the model maintains strong predictive capability when applied to unseen datasets. Through multi-fold cross-validation and data stratification, the study reduced the model's dependence on sample distribution and demonstrated its robustness on boundary samples. This approach ensures the model captures consistent imaging features across varying parameters and patient populations, enhancing its clinical applicability. Unlike studies reporting only single performance metrics, our comprehensive evaluation of both sensitivity and specificity shows that the model effectively identifies tuberculoma while minimizing the risk of misdiagnosing metastatic lesions. Overall, the cross-validation framework provides a solid foundation for clinical translation, reinforcing the model's utility in real-world practice. This study elucidates the relationship between radiomic features and clinical molecular markers, particularly carcinoembryonic antigen (CEA). Our findings indicate a significant difference in CEA levels between tuberculomas and brain metastasis ( p < 0.001), corroborating previously reported studies [ 27 ]. Notably, CEA values tend to be elevated in brain metastasis compared to those in tuberculomas. Although CEA is elevated in tuberculomas, the increase is not statistically significant, potentially reflecting the disease stage of the patients included in our analysis. Prior research has indicated that the serum CEA positivity rate is notably higher in patients with EGFR mutant lung adenocarcinoma compared to those with wild-type status, suggesting that CEA levels may serve as a useful biomarker for monitoring the efficacy of EGFR-TKI targeted therapies [ 28 ]. Emerging evidence indicates that tumor molecular subtypes can profoundly influence imaging phenotypes. For example, brain metastases from EGFR-mutant lung cancer typically present as well-defined, high-density lesions, whereas those harboring KRAS mutations exhibit greater imaging heterogeneity [ 27 , 28 ]. Mechanistically, these differences may reflect underlying molecular pathways that regulate tumor proliferation, angiogenesis, and stromal remodeling—processes that shape imaging structure and texture. Despite such associations, significant phenotypic overlap across subtypes persists, highlighting the need for multi-omics integration to enhance diagnostic discrimination [ 21 , 29 ]. Our systematic analysis reinforces the potential of radiomic features not only to differentiate brain lesion types but also to serve as non-invasive imaging biomarkers for molecular classification, offering promising avenues for targeted diagnostics and therapy. This study has several limitations. First, the sample size—though multi-center—remains limited, potentially restricting generalizability. Second,its retrospective design introduces inherent selection bias. Third, analysis was confined to contrast-enhanced T1WI; other sequences (T2WI, DWI) were not assessed. Additionally, dataset heterogeneity may affect feature-generalizability, and external validation was performed on a single cohort. Future work should incorporate larger, prospective cohorts, multi-modal imaging, and molecular data to further enhance model robustness and clinical utility. In conclusion,We developed and validated a nomogram integrating clinical factors with T1WI-enhanced MRI radiomic features, which accurately distinguishes cerebral tuberculoma from brain metastasis—outperforming traditional clinical models. This non-invasive tool provides objective, quantitative decision support for early differentiation and individualized treatment planning. Abbreviations AUC Area under the curve DCA Decision curve analysis LASSO Least absolute shrinkage and selection operator ROI Region of interest IGRAs Interferon-Gamma Release Assays CEA Carcinoembryonic Antigen CA125 Cancer Antigen 125 CA199 Cancer Antigen 19-9 Declarations Acknowledgements DCPM (V5.49,Jingding Medical Technology Co., Ltd.) Author contributions Ruiyun Liang conceived and designed the study. Chengcheng Li, Quhua Yin, Min Song, and Hui Zhang performed the experiments and collected the data. Ruiyun Liang, Chengcheng Li, Yuanyuan Han, Quan Guan, and Xiang Zhang manually delineated the regions of interest (ROIs). Jingqiang Wu and Wanying Ren conducted the data analysis and results visualization. Huili Ren and Huiru Li provided critical experimental resources and technical support. Weijun Fang and Qi Wang supervised and guided the research. Ruiyun Liang drafted the initial manuscript. Weijun Fang is the responsible corresponding author for all editorial communications with the journal. Qi Wang and Weijun Fang are co-corresponding authors and contributed equally to this work.All authors reviewed and revised the manuscript and approved the final version for publication. Funding This study was supported by the Guangzhou Science and Technology Program (Grant Nos. 2025A03J3606, 2024A03J0583, 2024A03J0511) and the Project of Guangdong Provincial Bureau of Traditional Chinese Medicine (Grant No. 20251291). Availability of data and materials All data generated or analyzed during this study are included in this article and its additional files. Ethics approval and consent to participate This study was approved by the Ethics Committee of Guangzhou Chest Hospital (approval number: KYSB-2024-035) and the Ethics Committee of Hunan Chest Hospital (approval number: LS2026011201). All procedures performed in this study involving human participants were in accordance with the Declaration of Helsinki (revised 2013) and the relevant ethical guidelines of the above-mentioned institutions. The requirement for written informed consent from participants was waived by the ethics committees due to the retrospective nature of the study, and all clinical and imaging data were anonymized to protect patient privacy. Consent for publication No individual participant data is reported that would require consent to publish from the participant (or legal parent or guardian for children). Competing interests The authors declare that they have no competing interests. Author details 1.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail： [email protected] :+8613763394065. 1.Department of Radiology, Hunan Institute for Tuberculosis Control, Hunan Chest Hospital, Changsha 410013, China.E-mail: [email protected] :+8613763394065. 2.Department of Internal Medicine , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china..E-mail: [email protected] :+8613824462125. 3.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8613316288660. 4.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china..E-mail: [email protected] :+8613822215503. 5.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8618122262446. 6.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8613826475056. 7.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8613763394065. 8.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8613763394065. 9.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8613763394065. 10.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8613763394065. 11.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail: [email protected] :+8613763394065. 12. Institute of Pulmonary Diseases, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected] . 13.Department of Radiology, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected] :+8619011439120 *Responsible corresponding author*: Weijun Fang, Department of Radiology, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected] . *Co-corresponding author*: Qi Wang, Institute of Pulmonary Diseases, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected] . References Khairy S, Alkhaibary A, AlQahtani S, Alkhani A, Cerebral tuberculoma. IDCases. 2021;26:e01319. https://doi.org/10.1016/j.idcr.2021.e01319.do . Accessed November 1, 2025. Huntoon K, Musgrave N, Shaikhouni A, Elder J. Frequency of seizures in patients with metastatic brain tumors. Neurol Sci. 2023;44(7):2501–2507. https://doi.org/10.1007/s10072-023-06695-y.do . Accessed November 2, 2025. Lee RD, Atencio ACNS, tuberculomas. J Gen Intern Med. 2023;38(11):2621. https://doi.org/10.1007/s11606-023-08264-7.do . Accessed November 5, 2025. Kokkali S, Andriotis E, Katsarou E, et al. Cerebral metastasis from osteosarcoma: Bone in the brain. Radiol Case Rep. 2020;15(6):780–3. https://doi.org/10.1016/j.radcr.2020.03.020 . Sidow NO. Large cerebral tuberculoma. Clin Case Rep. 2024;12(5):e8827. https://doi.org/10.1002/ccr3.8827.do . Accessed November 14, 2025. Vargas-Urbina J, Martinez-Silva R, Rojas-Panta G, Ponce-Manrique G, Flores-Castillo J, Anicama-Lima. W.Unusual brain metastasis from colon cancer. Surg Neurol Int.2025;16:5. https://doi.org/10.25259/SNI_636_2024.do . Accessed November 14, 2025. Mackel CE, Rosenberg H, Varma H, Uhlmann EJ, Vega RA, Alterman RL. Intracranial metastasis of extracranial chondrosarcoma: systematic review with illustrative case. Brain Tumor Res Treat. 2023;11(2):103–113. https://doi.org/10.14791/btrt.2023.0003.do . Accessed November 4, 2025. Papadimitriou K, Kiss-Bodolay D, Hedjoudje A et al. Late metachronous cerebral metastasis of pancreatic adenocarcinoma of the tail of the pancreas: a case report. J Med Case Rep. 2022;16(1):144. https://doi.org/10.1186/s13256-022-03314-w.do . Accessed November 16, 2025. Mertiri L, Freiling JT, Desai NK, Huisman TAGM. Pediatric and adult meningeal, parenchymal, and spinal tuberculosis: a neuroimaging review. J Neuroimaging. 2024;34(2):179–194. https://doi.org/10.1111/jon.13177.do . Accessed November 17, 2025. Chinthala AS, Obeng-Gyasi B, Virgin KL, Mao G. Brain metastasis mimicking brain abscess: illustrative case and systematic review. J Neurosurg Case Lessons. 2025;10(16):CASE25528. https://doi.org/10.3171/CASE25528.do . Accessed November 21, 2025. Dahal P, Parajuli S. Magnetic resonance imaging findings in central nervous system tuberculosis: a pictorial review. Heliyon. 2024;10(8):e29779. https://doi.org/10.1016/j.heliyon.2024.e29779.do . Accessed November 22, 2025. Ma Q, Yi Y, Liu T, Shi Y. MRI-based radiomics signature for identification of invisible basal cisterns changes in tuberculous meningitis: a preliminary multicenter study. Eur Radiol. 2022;32(12):8659–8669. https://doi.org/10.1007/s00330-022-08911-3.do . Accessed November 23, 2025. Collie D, Chang Z, Meehan J et al. (2022). Radiomic feature characteristics of ovine pulmonary adenocarcinoma. Vet Sci. 2022;12(5):400. https://doi.org/10.3390/vetsci12050400 Demircioğlu A. Benchmarking feature selection methods in radiomics. Invest Radiol. 2022;57(7):433–443. https://doi.org/10.1097/RLI.0000000000000855.do . Accessed November 26, 2025. Demircioğlu A. Benchmarking feature projection methods in radiomics. Sci Rep. 2025;15(1):32368. https://doi.org/10.1038/s41598-025-16070-w.do . Accessed November 26, 2025. Fellner A, Gavriel H, Pitaro J, Muallem Kalmovich L. Clinical parameters predicting tonsillar malignancy. Eur Arch Otorhinolaryngol. 2020;277(6):1779–1783. https://doi.org/10.1007/s00405-020-05873-4.do . Accessed November 27, 2025. Alam T, AlShahrani I, Assiri KI, Almoammar S, Togoo RA, Luqman M. Evaluation of clinical and radiographic parameters as dental indicators for postmenopausal osteoporosis. Oral Health Prev Dent. 2020;18(3):499–504. https://doi.org/10.3290/j.ohpd.a44688.do . Accessed December 7, 2025. Zhu D, Zhang M, Li Q, Liu J, Zhuang Y, Chen Q et al. Can perihaematomal radiomics features predict haematoma expansion? Clin Radiol. 2021;76(8):629.e1–629.e9. https://doi.org/10.1016/j.crad.2021.03.003.do . Accessed December 11, 2025. Kobayashi T. RadiomicsJ: a library to compute radiomic features. Radiol Phys Technol.2022;15(3):255–263. https://doi.org/10.1007/s12194-022-00664-4.do . Accessed December 15, 2025. Agrawal P, Phuyal S, Panth R, Shrestha P, Lamsal R. Giant cerebral tuberculoma masquerading as malignant brain tumor — a report of two cases. Cureus. 2020;12(9):e10546. https://doi.org/10.7759/cureus.10546.do . Accessed December 21, 2025. Sanvito F, Castellano A, Falini A. Advancements in neuroimaging to unravel biological and molecular features of brain tumors. Cancers. 2021;13(3):424. https://doi.org/10.3390/cancers13030424.do . Accessed December 21, 2025. Bravo-Tsri AEB, Konaté I, Kouassi KPB et al. Meningeal tuberculoma mimicking a brain tumor. Radiol Case Rep. 2020;16(2):284–288. https://doi.org/10.1016/j.radcr.2020.11.028.do . Accessed December 22, 2025. McMahon P, Pisapia DJ, Schweitzer AD, Heier L, Souweidane MM, Roytman M. Central nervous system tuberculoma mimicking a brain tumor: a case report. Radiol Case Rep. 2023;19(1):414–417. https://doi.org/10.1016/j.radcr.2023.10.042.do . Accessed December 23, 2025. Yang LP, Jiang ZY, Tong JL, Li N, Dong Q, Wang KZ. Development and validation of a preoperative CT-based radiomics nomogram to differentiate tuberculosis granulomas from lung adenocarcinomas: an external validation study. BMC Cancer. 2024;24(1):670. https://doi.org/10.1186/s12885-024-12422-3.do . Accessed December 27, 2025. Cabot JH, Ross EG. Evaluating prediction model performance. Surgery. 2023;174(3):723–726. https://doi.org/10.1016/j.surg.2023.05.023.do . Accessed December 28, 2025. Teng X, Zhang J, Zwanenburg A, Sun J, Huang Y, Lam S et al. Building reliable radiomic models using image perturbation. Sci Rep. 2022;12(1):10035. https://doi.org/10.1038/s41598-022-14178-x.do . Accessed December 29, 2025. Di Picca SAL, Saragoussi A, Bielle E, Ducray F, Villa F et al. CClinical, molecular, and radiomic profile of gliomas with FGFR3-TACC3 fusions. Neuro-oncology. 2020;22(11):1614–1624. https://doi.org/10.1093/neuonc/noaa121.do . Accessed January 4, 2026. Xu J, Yang Y, Gao Z, Song T, Ma Y, Yu X et al. Distinguishing EGFR mutation molecular subtypes based on MRI radiomics features of lung adenocarcinoma brain metastases. Clin Neurol Neurosurg. 2024;240:108258. https://doi.org/10.1016/j.clineuro.2024.108258.do . Accessed January 9, 2026. Liu Q, Hu P. Radiogenomic association of deep MR imaging features with genomic profiles and clinical characteristics in breast cancer. Biomark Res. 2023;11(1):9. https://doi.org/10.1186/s40364-023-00455-y.do . Accessed January 21, 2026. Table 1 Table 1 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table1.docx Cite Share Download PDF Status: Posted Version 1 posted 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-9164706","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":629586969,"identity":"308368bd-be84-4e6d-a06c-beeb36e218a4","order_by":0,"name":"Ruiyun Liang","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ruiyun","middleName":"","lastName":"Liang","suffix":""},{"id":629586973,"identity":"e3fb3a8e-1720-4306-9b22-4101133f7ca8","order_by":1,"name":"Quhua Yin","email":"","orcid":"","institution":"Hunan Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Quhua","middleName":"","lastName":"Yin","suffix":""},{"id":629586975,"identity":"f57a460e-7c34-4f76-ba58-fc92cce889b5","order_by":2,"name":"Chengcheng Li","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chengcheng","middleName":"","lastName":"Li","suffix":""},{"id":629586983,"identity":"b30b29cc-63b9-4a6d-8c25-94870ca6c383","order_by":3,"name":"Huili Ren","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Huili","middleName":"","lastName":"Ren","suffix":""},{"id":629586984,"identity":"ae4a3683-c678-459c-aaae-c898976c57cc","order_by":4,"name":"Huiru Li","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Huiru","middleName":"","lastName":"Li","suffix":""},{"id":629586987,"identity":"92dbf7a5-0ccd-4551-9a7d-713c42ebbe4b","order_by":5,"name":"Min Song","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"Song","suffix":""},{"id":629586997,"identity":"fcf95572-bf1a-402a-85fb-96afff2867c0","order_by":6,"name":"Hui Zhang","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hui","middleName":"","lastName":"Zhang","suffix":""},{"id":629586998,"identity":"727cc9e4-3fc3-479c-93b7-aafe3a72ab04","order_by":7,"name":"Yuanyuan Han","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yuanyuan","middleName":"","lastName":"Han","suffix":""},{"id":629587003,"identity":"39189b8a-b765-463b-b227-20b5533af5c0","order_by":8,"name":"Wanying Ren","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Wanying","middleName":"","lastName":"Ren","suffix":""},{"id":629587004,"identity":"7b9ef58f-c7e4-442c-8cce-b66f5fd5d3ac","order_by":9,"name":"Xiang Zhang","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Xiang","middleName":"","lastName":"Zhang","suffix":""},{"id":629587007,"identity":"63c85221-3195-4d03-ba26-8e03e959f033","order_by":10,"name":"Jingqiang Wu","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jingqiang","middleName":"","lastName":"Wu","suffix":""},{"id":629587008,"identity":"c247bc5f-dc2f-481d-b974-169d497dc9f6","order_by":11,"name":"Quan Guan","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Quan","middleName":"","lastName":"Guan","suffix":""},{"id":629587009,"identity":"966cb952-61e7-4f7c-bf9f-586147e2c2d3","order_by":12,"name":"Qi Wang","email":"","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qi","middleName":"","lastName":"Wang","suffix":""},{"id":629587010,"identity":"2ea10df4-e58d-42e1-a552-09a9d41a5cd5","order_by":13,"name":"Weijun Fang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4UlEQVRIie3RPQrCMBiA4a8EUodg1wi14g0KhU4eJkHIVMGxQweLkg7WK+gVHB1bCp3i7lh3FzcnMc5KWzeHvNM35CF/ACbTH4btdFUwknh6SBsWJ91kSKqyaNw60EPlN6ruJh4V87KZIb6nQoyuG9TjYDTyCxZhLiEKY77C4GRb1k7ITRPlBhJUeOEnF6g6H9uJrXfhOR5LK9dEYfDpooPAmzyRJREJl1yiHmQgmH5kNJUYC+hHSFVoUgeYoIoyVZPOu0yydH1/6K+cHK7p/REnnpPt2slH5LflJpPJZPraC3XETCxrN+7vAAAAAElFTkSuQmCC","orcid":"","institution":"Guangzhou Chest Hospital","correspondingAuthor":true,"prefix":"","firstName":"Weijun","middleName":"","lastName":"Fang","suffix":""}],"badges":[],"createdAt":"2026-03-19 04:10:07","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9164706/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9164706/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108493297,"identity":"c9283ac0-ad92-4e48-b13f-49fe6dabaf24","added_by":"auto","created_at":"2026-05-05 09:59:51","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":246790,"visible":true,"origin":"","legend":"\u003cp\u003ePatient selection flowchart.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/f79c0fd64c31ec2024f54400.png"},{"id":108383876,"identity":"c92cccd4-d1ce-4737-9c2c-468a8495bc4a","added_by":"auto","created_at":"2026-05-04 05:48:56","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":563494,"visible":true,"origin":"","legend":"\u003cp\u003eRepresentative MRI findings of cerebral tuberculoma and brain metastasis. (A and B) MRI images obtained by scanning the perituberculoma region and manually delineated ROI. (C and D) MRI images obtained by scanning the perimetastatic region and manually delineated ROI.ROI:region of interest\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/a34d160578ac8d39601f7e1f.png"},{"id":108492650,"identity":"3ed4fc55-41c0-4f8d-a005-5e73fefa9c4c","added_by":"auto","created_at":"2026-05-05 09:58:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":486445,"visible":true,"origin":"","legend":"\u003cp\u003eSelection of radiomics features using the LASSO algorithm.\u003c/p\u003e\n\u003cp\u003e(A) Radiomic Feature Coefficient Trajectories During LASSO Feature Selection. (B)Regularization Path of Binomial Deviance in LASSO Regression. In the model, optimization of omics parameters ultimately retained the final set containing 20 radiomics features. (C) Bar chart of radiomics feature coefficients: The Y-axis represents the selected 20 radiomics features, and the X-axis represents the coefficients of radiomics features.LASSO:least absolute shrinkage and selection operator.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/44f39a16fcb57ff015ab8aba.png"},{"id":108492876,"identity":"3dac7c49-3745-432c-a008-7fcf74bf371c","added_by":"auto","created_at":"2026-05-05 09:58:51","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":83982,"visible":true,"origin":"","legend":"\u003cp\u003e(A) Waterfall plot of Rad_score for the training set. (B) Waterfall plot of Rad_score for the internal validation set. (C) Waterfall plot of Rad_score for the independent external validation set. The waterfall plot is ordered in descending order of Rad_score.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/48dee83458bbb2ff922e789c.png"},{"id":108383878,"identity":"e818b96c-e1f6-40b5-9bb3-1bba0c7e1065","added_by":"auto","created_at":"2026-05-04 05:48:56","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":178712,"visible":true,"origin":"","legend":"\u003cp\u003eROC curves of radiomics model, clinical model, and combined model in predicting cerebral tuberculoma and brain metastasis in the training cohort (A), internal validation cohort (B), and external validation cohort (C).\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/03c49012f80516a63e752b1c.png"},{"id":108492731,"identity":"53e09176-fc6c-4a9f-8a0d-9a178243417b","added_by":"auto","created_at":"2026-05-05 09:58:28","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":605005,"visible":true,"origin":"","legend":"\u003cp\u003eNomogram and calibration curve model evaluation for distinguishing cerebral tuberculoma from brain metastasis based on training cohort. (A) Nomogram of a combined model based on clinical indicators and radiomics scores. (B) Calibration curve: Evaluating the consistency between the predicted probabilities from the nomogram and actual clinical observations in training cohorts, internal validation cohorts, and external validation cohorts. (C) Decision curve analysis(DCA) of each model for classifying cerebral tuberculoma versus brain metastasis. The x-axis represents threshold probabilities, and the y-axis represents net benefit. The blue curve represents the combined model, the green curve represents the radiomics model, and the yellow curve represents the clinical model.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/50e1565ff4319d21fb1c89de.png"},{"id":109296584,"identity":"2051c780-7af4-48d8-a7f4-a3a5c95fbae7","added_by":"auto","created_at":"2026-05-15 08:48:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2317680,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/3c973585-d82d-4c62-a156-7c4bda725d05.pdf"},{"id":108383874,"identity":"eedb29e1-8eb4-4215-8e54-585f6d5636db","added_by":"auto","created_at":"2026-05-04 05:48:56","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":22113,"visible":true,"origin":"","legend":"","description":"","filename":"Table1.docx","url":"https://assets-eu.researchsquare.com/files/rs-9164706/v1/5ee51f7559608af9032988c9.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eDevelopment and Application of a T1-Weighted Enhanced MRI Radiomics Nomogram for Differentiating Cerebral Tuberculoma from Brain Metastasis\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003e \u003cb\u003eC\u003c/b\u003eerebral tuberculoma and brain metastasis are pathologically distinct\u0026mdash;one arising from Mycobacterium tuberculosis infection, the other from hematogenous spread of solid tumors [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e][\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Yet both cause severe neurological deficits and increasingly overlap in incidence due to shifting TB epidemiology and improved cancer survival [\u003cspan additionalcitationids=\"CR4 CR5\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Clinically and radiologically, they present with similar symptoms (headaches, seizures) and nonspecific imaging features, making reliable non-invasive differentiation difficult. However, their treatment and prognosis diverge sharply, underscoring the critical need for accurate preoperative distinction [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e][\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eConventional MRI remains essential for evaluating intracranial lesions, relying on features such as signal intensity, morphology, margins, and contrast enhancement patterns. However, both tuberculomas and metastases frequently present as round or irregular nodules with overlapping nodular or ring-like enhancement [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], limiting diagnostic specificity. Such nonspecific imaging features lead to subjective interpretation and often necessitate invasive procedures like biopsy or lumbar puncture for definitive diagnosis\u0026mdash;particularly challenging and risky for deep-seated lesions [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. These limitations underscore the urgent need for non-invasive, objective diagnostic tools.\u003c/p\u003e \u003cp\u003e \u003cb\u003eR\u003c/b\u003eadiomics enables non-invasive detection of subtle pathological changes invisible to the naked eye and has proven valuable in tumor grading and infectious central nervous system(CNS) diseases [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e][\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. However, individual radiomic features are often unstable due to technical variability, while clinical parameters alone offer limited predictive power [\u003cspan additionalcitationids=\"CR15 CR16\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Integrating both data types has therefore emerged as a promising strategy, yet comprehensive models specifically designed to differentiate cerebral tuberculoma from brain metastasis remain lacking [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e][\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e \u003cb\u003eT\u003c/b\u003eo address this gap, we aim to develop and validate a nomogram that integrates T1WI-enhanced MRI radiomic features with key clinical indicators using multi-center data. Through rigorous feature selection, model optimization, and validation across internal and independent external cohorts, we seek to establish a robust, generalizable tool for early and accurate differential diagnosis\u0026mdash;supporting individualized treatment decisions without recourse to invasive procedures.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003eStudy Population\u003c/p\u003e \u003cp\u003e This retrospective study, approved by the Ethics Committees of Guangzhou Chest Hospital (KYSB-2024-035) and Hunan Chest Hospital (LS2026011201) in accordance with the Declaration of Helsinki, waived the requirement for written informed consent. The initial cohort included patients who underwent radiological assessments at both institutions from January 2021 to December 2024. Inclusion criteria were: (1) meeting the diagnostic criteria for cerebral tuberculoma or brain metastasis, with complete clinical and T1WI enhanced MRI data; (2) possessing at least one identifiable lesion in the brain parenchyma on T1WI enhanced MRI scans; (3) histological confirmation of diagnosis via surgical resection or biopsy; and (4) a molecular biological diagnosis validated by cerebrospinal fluid obtained through lumbar puncture.Patients were excluded based on the following criteria: (1) unclear clinical diagnosis, with ineffective anti-tuberculosis or anti-tumor treatments; (2) unclear pathological diagnosis attributed to insufficient tissue samples; (3) history of other intracranial malignancies or co-infections with other pathogens (e.g., viral, bacterial, or fungal infections); and (4) poor image quality due to motion artifacts. The reference standard for the final diagnosis of cerebral tuberculoma or brain metastasis was established by histopathological examination of surgically resected or biopsied tissue, or by molecular biological analysis of cerebrospinal fluid (CSF) samples.Ultimately, the study included 482 patients, comprising 406 from Guangzhou Chest Hospital (204 diagnosed with cerebral tuberculoma and 202 with brain metastasis), randomly stratified into a training set (n\u0026thinsp;=\u0026thinsp;284) and an internal validation set (n\u0026thinsp;=\u0026thinsp;122). A total of 76 patients from Hunan Chest Hospital were allocated to the independent external validation cohort (n\u0026thinsp;=\u0026thinsp;76). A detailed workflow diagram is presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eImage Segmentation and Feature Extraction\u003c/p\u003e \u003cp\u003eIntra- and inter-observer reproducibility was assessed using intraclass correlation coefficients (ICCs). Two radiologists randomly selected 60 T1WI enhanced MRI images (30 cerebral tuberculomas and 30 brain metastases) and delineated volumes of interest (VOIs). Radiologist 1 repeated the segmentation after two weeks. An ICC threshold of \u0026gt;\u0026thinsp;0.80 was considered indicative of good consistency. Following reliability assessment, Radiologist 1 performed VOI segmentation for the remaining cases. Only radiomic features with ICCs\u0026thinsp;\u0026gt;\u0026thinsp;0.80 were retained for subsequent analysis.\u003c/p\u003e \u003cp\u003eUsing 3D Slicer software (version 5.4.0, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.slicer.org/\u003c/span\u003e\u003cspan address=\"https://www.slicer.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), regions of interest (ROI) were manually delineated on T1-weighted enhanced MRI images along the interface between each enhanced lesion and the surrounding brain tissue. Representative segmentation examples are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. From each ROI, 1,746 candidate radiomic features were extracted using the ImaScholar platform (Version 3.5), including first-order statistics, gray level size zone matrix (GLSZM), neighboring gray tone difference matrix (NGTDM), and gray level dependence matrix (GLDM) features (Supplementary data 1). Dimensionality reduction was performed using significance testing (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) and Spearman correlation analysis (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.85). The most informative features were then selected via Least Absolute Shrinkage and Selection Operator(LASSO) regression with ten-fold cross-validation, yielding an optimal subset of 20 features. These were linearly combined using their respective coefficients to generate a radiomic score (Radscore) for each patient, thereby establishing the radiomic model. The index test was T1-weighted enhanced MRI radiomics analysis, which involved extracting radiomic features from preprocessed images.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eModel Construction, Evaluation, and Nomogram Development\u003c/p\u003e \u003cp\u003eWe developed clinical, radiomic, and combined models for distinguishing cerebral tuberculoma from brain metastasis. Following feature selection, the radiomic model was constructed using the ImaScholar platform (Version 3.5). The workflow for feature extraction and model development is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. For the clinical model, univariate logistic regression was used to identify significant risk factors and then select independent predictors via multivariate analysis. Significant variables from both univariate and multivariate analyses were subsequently incorporated into the combined model. A nomogram was developed to visualize the combined model, enabling assessment of variable contributions and facilitating prediction of individual risk.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eModel discrimination was compared using the DeLong test (AUC, sensitivity, specificity, accuracy) across all cohorts. Nomogram performance was further assessed via calibration (Hosmer-Lemeshow) and decision curve analyses.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eStatistical analyses were performed using R software (version 4.4.2). Throughout the analysis, various specialized statistical packages\u0026mdash;including glmnet, cmprsk, rms, rmda, and devtools\u0026mdash;were employed. Continuous data were presented as mean\u0026plusmn;standard deviation. Univariate and multivariate logistic regression models were utilized to identify independent predictors of outcomes based on collected clinical variables, with \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05 considered statistically significant. The discriminative performance of each model was appraised through the construction of receiver operating characteristic (ROC) curves. Statistical inference was conducted via two-tailed tests, with significance set at \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Sample size was calculated based on a power of 80% and alpha of 0.05 to detect a minimum AUC difference of 0.1 between models.Author Liang served as the statistical guarantor, responsible for the transparency of statistical analyses.\u003c/p\u003e \u003cp\u003eBy following these methodological approaches, we aim to develop a robust and interpretable model capable of distinguishing between cerebral tuberculoma and brain metastasis, thus enhancing the diagnostic accuracy and facilitating personalized treatment strategies for patients.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eBaseline Characteristics of Patients\u003c/p\u003e\n\u003cp\u003eThis study analyzed 482 patients (290 males, 192 females; mean age 56.08 ± 12.96 years), including 254 with cerebral tuberculoma and 228 with brain metastasis. Of these, 406 patients were from Guangzhou Chest Hospital (204 tuberculoma, 202 metastasis), and 76 from Hunan Chest Hospital (50 tuberculoma, 26 metastasis). The Guangzhou cohort was randomly split into a training set (142 tuberculoma, 142 metastasis) and an internal validation set (62 tuberculoma, 60 metastasis). The Hunan cohort served as an independent external validation set (50 tuberculoma, 26 metastasis). Table\u0026nbsp;1 summarizes the demographic and clinical imaging characteristics of the study population.\u003c/p\u003e\n\u003cdiv\u003e\n\u003c/div\u003e\n\u003cp\u003eReproducibility of Feature Extraction\u003c/p\u003e\n\u003cp\u003eThe process of radiomic feature extraction demonstrated robust reproducibility. The intra-observer intraclass correlation coefficients (ICCs) ranged from 0.823 to 0.935, while the inter-observer ICCs fell between 0.802 and 0.896, with all values surpassing the threshold of \u0026gt; 0.80, signifying stable and dependable feature extraction results.\u003c/p\u003e\n\u003cp\u003eFeature Selection and Radscore Distribution\u003c/p\u003e\n\u003cp\u003eOur analysis identified a total of 1,746 radiomic features extracted from each region of interest (VOI) on T1-weighted enhanced MRI images. These features encompassed first-order statistics, gray level size zone matrix (GLSZM), neighboring gray tone difference matrix (NGTDM), and gray level dependence matrix (GLDM) (refer to Supplementary Data 1). All extracted radiomic features underwent preprocessing and standardization via the Z-score method. Through rigorous significance testing, we selected features that achieved statistical significance (\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05). Furthermore, we eliminated features exhibiting a Pearson correlation coefficient greater than 0.85 between the conditions, narrowing down to 237 radiomic features. We then employed LASSO regression to achieve dimensionality reduction, ultimately isolating a subset of 20 predictive radiomic features to construct our radiomic model (as illustrated in Figs.\u0026nbsp;3A-C). The radiomic feature label was constructed by linearly combining these features, weighted according to their respective coefficients (Fig.\u0026nbsp;4 presents the waterfall plot of Radscore). The waterfall plot effectively visualizes the Radscore distribution characteristics for cerebral tuberculoma and brain metastasis across the training, internal validation, and external validation cohorts.\u003c/p\u003e\n\u003cp\u003eConstruction of Clinical Model, Radiomic Model, and Combined Model\u003c/p\u003e\n\u003cp\u003eStatistical analysis, specifically the Wilcoxon test, revealed significant differences in Radscore between the cerebral tuberculoma and brain metastasis groups (\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05). In contrast, significant differences were detected in factors such as age, ring enhancement, meningeal thickening, and levels of biological markers (CEA, CA199, IGRAs) (\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05). Through multivariate regression analysis, we identified age (\u003cem\u003ep\u003c/em\u003e = 0.002), CEA (\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001), meningeal thickening (\u003cem\u003ep\u003c/em\u003e = 0.002), ring enhancement (\u003cem\u003ep\u003c/em\u003e = 0.02),and IGRAs(\u003cem\u003ep\u003c/em\u003e = 0.02) as independent predictive factors for lesion classification, leading us to incorporate these variables into the clinical model (Table\u0026nbsp;2). Ultimately, we constructed a combined model integrating both Radscore and the independent predictive factors identified.\u003c/p\u003e\n\u003cdiv\u003e \u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv\u003eTable 2\u003c/div\u003e\n \u003cdiv\u003e\n \u003cp\u003eUnivariate and Multivariate Logistic Regression Analysis\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"9\"\u003e\u003c/colgroup\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\n \u003cp\u003eUnivariable analysis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\n \u003cp\u003eMultivariable analysis (clinical model parameters)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"2\" nameend=\"c9\" namest=\"c8\"\u003e\n \u003cp\u003eMultivariable analysis (combined model parameters)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eOR (95%CI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003eOR (95%CI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003eOR (95%CI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e-value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eSex\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.943(0.586–1.516)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\" morerows=\"16\" rowspan=\"17\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\" morerows=\"16\" rowspan=\"17\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge(years, mean ± SD)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.027(1.007–1.048)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e1.038 (1.014–1.062)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e1.041 (1.007–1.076)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eEmphysema\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.328(0.788–2.249)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eDiabetes Mellitus\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.764(0.441–1.316)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eWhole-nodule enhancement\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.826(0.502–1.355)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003ering enhancement\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.482(0.29–0.792)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0.504 (0.286–0.891)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e0.356 (0.148–0.858)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003emeningeal thickening\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.578(0.358–0.926)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0.416 (0.237–0.730)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e0.292 (0.125–0.686)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eHydrocephalus\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.69(0.412–1.148)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eCA125(u/ml)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.999(0.996–1.001)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eCEA(ng/ml)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.017(1.011–1.026)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e1.021 (1.013–1.030)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e1.030 (1.017–1.042)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eCA199(U/mL)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.005(1.002–1.009)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eNumber of lesions(n)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u0026gt; 10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.109(0.662–1.861)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.776(0.844–3.887)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e2 ~ 10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.602(0.995–2.593)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eIGRAs\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.557(0.335–0.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0.508 (0.283–0.911)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003erad_score\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.893(0.867–0.915)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e—\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c8\"\u003e\n \u003cp\u003e0.876 (0.846–0.906)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c9\"\u003e\n \u003cp\u003e\u0026lt; 0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\"\u003eNote: Reasons for excluding IGRAs in the joint model: Collinearity with rad_score (VIF \u0026gt; 10), resulting in a reduced AIC value after exclusion.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003ePerformance of Clinical Model, Radiomic Model, and Combined Model\u003c/p\u003e\n\u003cp\u003eReceiver operating characteristic (ROC) curves were used to evaluate the performance of the clinical, radiomic, and combined models across the training, internal validation, and external validation cohorts. The area under the curve (AUC), sensitivity, specificity, and accuracy for each model are summarized in Fig.\u0026nbsp;5 and Table\u0026nbsp;3.\u003c/p\u003e\n\u003cdiv\u003e \u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv\u003eTable 3\u003c/div\u003e\n \u003cdiv\u003e\n \u003cp\u003eComparison of diagnostic performance of clinical models, radiomics models, and combined models in training cohorts, internal validation cohorts, and external validation cohorts\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"7\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eModel\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eAUC (95%CI)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003eAccuracy(%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003eSensitivity(%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003eSpecificity(%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003ePPV (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003eNPV (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eClinical model\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"6\" nameend=\"c7\" namest=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTraining cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.781 (0.729–0.834)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e71.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e67.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e76.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e73.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e70.13\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eInternal validation cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.712(0.621–0.802)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e66.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e55.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e77.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e70.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e64.00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eExternal validation cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.828 (0.735–0.920)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e76.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e76.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e76.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e62.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e86.36\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eRadiomics model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"6\" nameend=\"c7\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTraining cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.907(0.873–0.941)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e84.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e79.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e90.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e88.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e81.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eInternal validation cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.891 (0.836–0.947)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e81.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e93.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e70.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e75.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e91.67\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eExternal validation cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.886 (0.813–0.959)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e76.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e92.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e68.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e60.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e94.44\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eCombined model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colspan=\"6\" nameend=\"c7\" namest=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTraining cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.951 (0.929–0.973)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e88.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e84.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e91.55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e90.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e85.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eInternal validation cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.937 (0.894–0.98)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e90.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e86.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e93.55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e92.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e87.88\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eExternal validation cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.940 (0.890–0.989)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e86.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e80.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e90.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e80.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e90.00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003eNote: AUC (area under the curve), CI (confidence interval), PPV (positive predictive value), NPV (negative predictive value)\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eAll three models demonstrated good discriminative ability. The clinical model achieved AUCs of 0.781, 0.712, and 0.828 in the training, internal validation, and external validation cohorts, respectively. The radiomic model showed excellent performance, with corresponding AUCs of 0.907, 0.891, and 0.886. The combined model outperformed both, achieving AUCs of 0.951, 0.937, and 0.940 across the three cohorts.\u003c/p\u003e\n\u003cp\u003eThe DeLong test confirmed that the combined model's AUC was significantly higher than those of the clinical and radiomic models in all cohorts (\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05). Significant differences were also observed between the clinical and radiomic models in the training cohort (0.781 vs. 0.907, \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001) and internal validation cohort (0.712 vs. 0.891, \u003cem\u003ep\u003c/em\u003e = 0.002), as well as between the clinical and combined models in the external validation cohort (0.828 vs. 0.937, \u003cem\u003ep\u003c/em\u003e = 0.02).\u003c/p\u003e\n\u003cp\u003eNomogram and Its Validation\u003c/p\u003e\n\u003cp\u003eA nomogram was developed to visualize the combined model and its contributing variables (Fig.\u0026nbsp;6A). Calibration, assessed by the Hosmer-Lemeshow test, showed no significant deviation from ideal predictions in the training \u003cem\u003ep\u003c/em\u003e = 0.792 internal validation \u003cem\u003ep\u003c/em\u003e = 0.200,\u003c/p\u003e\n\u003cp\u003eor external validation \u003cem\u003ep\u003c/em\u003e = 0.2775 cohorts (Fig.\u0026nbsp;6B). Decision curve analysis indicated that the nomogram offers a high clinical net benefit for differential diagnosis across threshold probabilities from 0.1 to 0.9 (Fig.\u0026nbsp;6C).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eDifferentiating cerebral tuberculoma from brain metastasis remains clinically challenging due to overlapping imaging features, yet treatment and prognosis differ fundamentally. We developed and validated a nomogram integrating T1WI-enhanced MRI radiomic features with clinical predictors, achieving excellent discrimination and calibration across multi-center cohorts. This non-invasive tool provides robust quantitative support for personalized differential diagnosis.\u003c/p\u003e \u003cp\u003eWe identified 20 core radiomic features for further analysis, encompassing a variety of categories such as wavelet transforms, first-order statistics, gray-level co-occurrence matrices (GLCM), and gray-level run length matrices (GLRLM), among others. These features elucidate the quantitative morphological differences between tuberculoma and brain metastasis on MRI, particularly highlighting distinctions in tumor boundary smoothness and tissue density distributions. Existing literature suggests that tuberculomas, as a result of granulomatous inflammation induced by Mycobacterium tuberculosis, typically display relatively well-defined margins and ring enhancement. This characteristic effectively distinguishes them from the infiltrative boundaries that are often observed in various malignant tumors [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e][\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. In contrast, brain metastasis generally exhibit a high degree of cellular heterogeneity and angiogenesis, resulting in blurred margins and heterogeneous internal enhancement patterns [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eMechanistically, our findings indicate that the radiomic features we analyzed reflect notable differences in the tissue microenvironment associated with each condition. The pathological characteristics inherent to tuberculomas lead to first-order features exhibiting lower overall density and a more symmetrical gray-level distribution. Additionally, GLCM features reveal an enhanced texture uniformity, while shape features indicate greater flatness due to the tumors' regular morphologies. On the other hand, brain metastasis comprise densely populated cancerous regions and areas of necrosis, characterized by a rich vascular supply and uneven necrotic patches. T1WI-enhanced MRI often reveals a heterogeneous ring enhancement consistent with significant pathological variability. The corresponding radiomic features associated with brain metastasis include first-order statistics showing higher mean values due to abundant tumor blood supply, coupled with positive skewness in regions dense with cancer cells. Furthermore, GLCM features exhibit lower maximum correlation coefficients, while wavelet transform features identify high-frequency variability at the tumor infiltration edges [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e][\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis study further delineates the capacity of high-dimensional feature analysis to uncover distinctions that extend beyond traditional imaging observations. Advanced parameters such as GLCM and texture heterogeneity effectively differentiate between cerebral tuberculoma and brain metastasis. In contrast to previous methodologies that largely relied on subjective imaging evaluations or limited low-dimensional quantitative metrics, our research employs a multi-parameter, high-throughput feature mining approach. This strategy not only reveals subtle differences in imaging phenotypic mechanisms but also enriches the understanding of tissue-level heterogeneity between cerebral tuberculosis infections and tumor lesions.\u003c/p\u003e \u003cp\u003eOverall, the integration of advanced radiomic analysis into clinical practice represents a promising avenue for enhancing diagnostic accuracy and tailoring treatment modalities for patients with intracranial lesions. Future studies should aim to further refine these models and explore their applicability across diverse patient populations and imaging modalities.\u003c/p\u003e \u003cp\u003eIn this study, we extracted 20 core radiomic features from T1-weighted imaging (T1WI) enhanced magnetic resonance imaging (MRI) of 406 patients to develop a comprehensive model that integrates these features with clinical indicator data for the differentiation between tuberculoma and brain metastasis. Notably, our model identified age, carcinoembryonic antigen (CEA) levels, ring enhancement, and meningeal thickening on T1WI-enhanced MRI as independent risk factors for distinguishing between these two conditions. The combined model demonstrated remarkable performance, achieving an area under the curve (AUC) of 0.951 in the training cohort, 0.937 in the internal validation cohort, and 0.940 in the independent external validation cohort.\u003c/p\u003e \u003cp\u003ePrevious research corroborates the effectiveness of combined models in differentiating tuberculoma from malignant tumors. For instance, a study evaluating a deep learning model based on computed tomography (CT) for distinguishing tuberculous granulomas from lung adenocarcinoma reported AUC values of 0.903, 0.933, and 0.914 across its training, internal validation, and external validation sets, respectively\u0026mdash;figures comparable to those observed in our study. Despite differences in the anatomy and pathological contexts of the studied organs, our findings suggest that the combined model offers superior discriminative capability in distinguishing tuberculoma from malignant tumors, as evidenced by its enhanced overall performance.\u003c/p\u003e \u003cp\u003eThe strength of our model lies in its comprehensive integration of clinical indicators, lesion morphology, texture features, and high-dimensional implicit information during feature extraction and model construction. This allows for precise identification and classification of lesions using advanced machine learning algorithms. Unlike traditional imaging diagnostic approaches that rely predominantly on morphological and signal change assessments, our model significantly improves the detection of subtle differences between tuberculoma and metastases, thereby enhancing both the repeatability and objectivity of the diagnostic results.\u003c/p\u003e \u003cp\u003eTo facilitate individualized classification and aid radiologists in distinguishing between tuberculoma and brain metastasis, we also developed a bar graph visualization tool based on our combined model. Compared to conventional diagnostic methods, the classification predictions generated by this visualization tool are faster, more user-friendly, and more accurate, ultimately assisting radiologists in making precise diagnoses.\u003c/p\u003e \u003cp\u003eThis study systematically assessed the sensitivity and specificity of the combined model through cross-validation, thereby validating the model's stability across diverse data distributions. Unlike traditional single-center or retrospective studies with small sample sizes, cross-validation provides a more realistic reflection of the model's performance in actual clinical scenarios. Existing literature indicates that radiomic models often exhibit decreased performance in independent external cohorts compared to their initial training sets, which can be attributed to factors such as sample distribution, imaging acquisition protocols, and patient heterogeneity. In our analysis, utilizing an external validation cohort, we found that the combined model achieved a sensitivity of 0.8077 and specificity of 0.9000, outperforming the clinical model (sensitivity: 0.7692, specificity: 0.7600). Although the sensitivity of the combined model was lower than that of the radiomics model (0.9231), its specificity was superior (0.6800).\u003c/p\u003e \u003cp\u003eThe combined model achieved a balanced trade-off among sensitivity, specificity, and positive predictive value (PPV), with stable and accurate performance in the independent external cohort. These findings indicate that the model maintains strong predictive capability when applied to unseen datasets. Through multi-fold cross-validation and data stratification, the study reduced the model's dependence on sample distribution and demonstrated its robustness on boundary samples. This approach ensures the model captures consistent imaging features across varying parameters and patient populations, enhancing its clinical applicability.\u003c/p\u003e \u003cp\u003eUnlike studies reporting only single performance metrics, our comprehensive evaluation of both sensitivity and specificity shows that the model effectively identifies tuberculoma while minimizing the risk of misdiagnosing metastatic lesions. Overall, the cross-validation framework provides a solid foundation for clinical translation, reinforcing the model's utility in real-world practice.\u003c/p\u003e \u003cp\u003eThis study elucidates the relationship between radiomic features and clinical molecular markers, particularly carcinoembryonic antigen (CEA). Our findings indicate a significant difference in CEA levels between tuberculomas and brain metastasis (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), corroborating previously reported studies [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. Notably, CEA values tend to be elevated in brain metastasis compared to those in tuberculomas. Although CEA is elevated in tuberculomas, the increase is not statistically significant, potentially reflecting the disease stage of the patients included in our analysis. Prior research has indicated that the serum CEA positivity rate is notably higher in patients with EGFR mutant lung adenocarcinoma compared to those with wild-type status, suggesting that CEA levels may serve as a useful biomarker for monitoring the efficacy of EGFR-TKI targeted therapies [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eEmerging evidence indicates that tumor molecular subtypes can profoundly influence imaging phenotypes. For example, brain metastases from EGFR-mutant lung cancer typically present as well-defined, high-density lesions, whereas those harboring KRAS mutations exhibit greater imaging heterogeneity [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Mechanistically, these differences may reflect underlying molecular pathways that regulate tumor proliferation, angiogenesis, and stromal remodeling\u0026mdash;processes that shape imaging structure and texture. Despite such associations, significant phenotypic overlap across subtypes persists, highlighting the need for multi-omics integration to enhance diagnostic discrimination [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Our systematic analysis reinforces the potential of radiomic features not only to differentiate brain lesion types but also to serve as non-invasive imaging biomarkers for molecular classification, offering promising avenues for targeted diagnostics and therapy.\u003c/p\u003e \u003cp\u003eThis study has several limitations. First, the sample size\u0026mdash;though multi-center\u0026mdash;remains limited, potentially restricting generalizability. Second,its retrospective design introduces inherent selection bias. Third, analysis was confined to contrast-enhanced T1WI; other sequences (T2WI, DWI) were not assessed. Additionally, dataset heterogeneity may affect feature-generalizability, and external validation was performed on a single cohort. Future work should incorporate larger, prospective cohorts, multi-modal imaging, and molecular data to further enhance model robustness and clinical utility.\u003c/p\u003e \u003cp\u003eIn conclusion,We developed and validated a nomogram integrating clinical factors with T1WI-enhanced MRI radiomic features, which accurately distinguishes cerebral tuberculoma from brain metastasis\u0026mdash;outperforming traditional clinical models. This non-invasive tool provides objective, quantitative decision support for early differentiation and individualized treatment planning.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAUC \u0026nbsp; Area under the curve\u003c/p\u003e\n\u003cp\u003eDCA \u0026nbsp; Decision curve analysis\u003c/p\u003e\n\u003cp\u003eLASSO \u0026nbsp;Least absolute shrinkage and selection operator\u003c/p\u003e\n\u003cp\u003eROI \u0026nbsp;Region of interest\u003c/p\u003e\n\u003cp\u003eIGRAs Interferon-Gamma Release Assays\u003c/p\u003e\n\u003cp\u003eCEA Carcinoembryonic Antigen\u003c/p\u003e\n\u003cp\u003eCA125 Cancer Antigen 125\u003c/p\u003e\n\u003cp\u003eCA199 Cancer Antigen 19-9\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eDCPM (V5.49,Jingding Medical Technology Co., Ltd.)\u003c/p\u003e\n\u003cp\u003eAuthor contributions\u003c/p\u003e\n\u003cp\u003eRuiyun Liang conceived and designed the study. Chengcheng Li, Quhua Yin, Min Song, and Hui Zhang performed the experiments and collected the data. Ruiyun Liang, Chengcheng Li, Yuanyuan Han, Quan Guan, and Xiang Zhang manually delineated the regions of interest (ROIs). Jingqiang Wu and Wanying Ren conducted the data analysis and results visualization. Huili Ren and Huiru Li provided critical experimental resources and technical support. Weijun Fang and Qi Wang supervised and guided the research. Ruiyun Liang drafted the initial manuscript. Weijun Fang is the responsible corresponding author for all editorial communications with the journal. Qi Wang and Weijun Fang are co-corresponding authors and contributed equally to this work.All authors reviewed and revised the manuscript and approved the final version for publication.\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Guangzhou Science and Technology Program (Grant Nos. 2025A03J3606, 2024A03J0583, 2024A03J0511) and the Project of Guangdong Provincial Bureau of Traditional Chinese Medicine (Grant No. 20251291).\u003c/p\u003e\n\u003cp\u003eAvailability of data and materials\u003c/p\u003e\n\u003cp\u003eAll data generated or analyzed during this study are included in this article and its additional files.\u003c/p\u003e\n\u003cp\u003eEthics approval and consent to participate\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Ethics Committee of Guangzhou Chest Hospital (approval number: KYSB-2024-035) and the Ethics Committee of Hunan Chest Hospital (approval number: LS2026011201). All procedures performed in this study involving human participants were in accordance with the Declaration of Helsinki (revised 2013) and the relevant ethical guidelines of the above-mentioned institutions. The requirement for written informed consent from participants was waived by the ethics committees due to the retrospective nature of the study, and all clinical and imaging data were anonymized to protect patient privacy.\u003c/p\u003e\n\u003cp\u003eConsent for publication\u003c/p\u003e\n\u003cp\u003eNo individual participant data is reported that would require consent to publish from the participant (or legal parent or guardian for children).\u003c/p\u003e\n\u003cp\u003eCompeting interests\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003eAuthor details\u003c/p\u003e\n\u003cp\u003e1.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail：[email protected]:+8613763394065.\u003c/p\u003e\n\u003cp\u003e1.Department of Radiology, Hunan Institute for Tuberculosis Control, Hunan Chest Hospital, Changsha 410013, China.E-mail:[email protected]:+8613763394065.\u003c/p\u003e\n\u003cp\u003e2.Department of Internal Medicine , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china..E-mail:[email protected]:+8613824462125.\u003c/p\u003e\n\u003cp\u003e3.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8613316288660.\u003c/p\u003e\n\u003cp\u003e4.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china..E-mail:[email protected]:+8613822215503.\u003c/p\u003e\n\u003cp\u003e5.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8618122262446.\u003c/p\u003e\n\u003cp\u003e6.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8613826475056.\u003c/p\u003e\n\u003cp\u003e7.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8613763394065.\u003c/p\u003e\n\u003cp\u003e8.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8613763394065.\u003c/p\u003e\n\u003cp\u003e9.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8613763394065.\u003c/p\u003e\n\u003cp\u003e10.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8613763394065.\u003c/p\u003e\n\u003cp\u003e11.Department of Radiology , Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease，Guangzhou 510095，china.E-mail:[email protected]:+8613763394065.\u003c/p\u003e\n\u003cp\u003e12. Institute of Pulmonary Diseases, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected].\u003c/p\u003e\n\u003cp\u003e13.Department of Radiology, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected]:+8619011439120\u003c/p\u003e\n\u003cp\u003e*Responsible corresponding author*: Weijun Fang, Department of Radiology, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected]. *Co-corresponding author*: Qi Wang, Institute of Pulmonary Diseases, Guangzhou Chest Hospital, State Key Laboratory of Respiratory Disease, Guangzhou 510095, China; Email: [email protected].\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eKhairy S, Alkhaibary A, AlQahtani S, Alkhani A, Cerebral tuberculoma. IDCases. 2021;26:e01319. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.idcr.2021.e01319.do\u003c/span\u003e\u003cspan address=\"https://doi.org/10.1016/j.idcr.2021.e01319.do\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 1, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuntoon K, Musgrave N, Shaikhouni A, Elder J. Frequency of seizures in patients with metastatic brain tumors. Neurol Sci. 2023;44(7):2501\u0026ndash;2507. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10072-023-06695-y.do\u003c/span\u003e\u003cspan address=\"10.1007/s10072-023-06695-y.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 2, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee RD, Atencio ACNS, tuberculomas. J Gen Intern Med. 2023;38(11):2621. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11606-023-08264-7.do\u003c/span\u003e\u003cspan address=\"10.1007/s11606-023-08264-7.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 5, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKokkali S, Andriotis E, Katsarou E, et al. Cerebral metastasis from osteosarcoma: Bone in the brain. Radiol Case Rep. 2020;15(6):780\u0026ndash;3. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.radcr.2020.03.020\u003c/span\u003e\u003cspan address=\"10.1016/j.radcr.2020.03.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSidow NO. Large cerebral tuberculoma. Clin Case Rep. 2024;12(5):e8827. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/ccr3.8827.do\u003c/span\u003e\u003cspan address=\"10.1002/ccr3.8827.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 14, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVargas-Urbina J, Martinez-Silva R, Rojas-Panta G, Ponce-Manrique G, Flores-Castillo J, Anicama-Lima. W.Unusual brain metastasis from colon cancer. Surg Neurol Int.2025;16:5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.25259/SNI_636_2024.do\u003c/span\u003e\u003cspan address=\"10.25259/SNI_636_2024.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 14, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMackel CE, Rosenberg H, Varma H, Uhlmann EJ, Vega RA, Alterman RL. Intracranial metastasis of extracranial chondrosarcoma: systematic review with illustrative case. Brain Tumor Res Treat. 2023;11(2):103\u0026ndash;113. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.14791/btrt.2023.0003.do\u003c/span\u003e\u003cspan address=\"10.14791/btrt.2023.0003.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 4, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePapadimitriou K, Kiss-Bodolay D, Hedjoudje A et al. Late metachronous cerebral metastasis of pancreatic adenocarcinoma of the tail of the pancreas: a case report. J Med Case Rep. 2022;16(1):144. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13256-022-03314-w.do\u003c/span\u003e\u003cspan address=\"10.1186/s13256-022-03314-w.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 16, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMertiri L, Freiling JT, Desai NK, Huisman TAGM. Pediatric and adult meningeal, parenchymal, and spinal tuberculosis: a neuroimaging review. J Neuroimaging. 2024;34(2):179\u0026ndash;194. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/jon.13177.do\u003c/span\u003e\u003cspan address=\"10.1111/jon.13177.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 17, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChinthala AS, Obeng-Gyasi B, Virgin KL, Mao G. Brain metastasis mimicking brain abscess: illustrative case and systematic review. J Neurosurg Case Lessons. 2025;10(16):CASE25528. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3171/CASE25528.do\u003c/span\u003e\u003cspan address=\"10.3171/CASE25528.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 21, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDahal P, Parajuli S. Magnetic resonance imaging findings in central nervous system tuberculosis: a pictorial review. Heliyon. 2024;10(8):e29779. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.heliyon.2024.e29779.do\u003c/span\u003e\u003cspan address=\"https://doi.org/10.1016/j.heliyon.2024.e29779.do\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 22, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa Q, Yi Y, Liu T, Shi Y. MRI-based radiomics signature for identification of invisible basal cisterns changes in tuberculous meningitis: a preliminary multicenter study. Eur Radiol. 2022;32(12):8659\u0026ndash;8669. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00330-022-08911-3.do\u003c/span\u003e\u003cspan address=\"10.1007/s00330-022-08911-3.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 23, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollie D, Chang Z, Meehan J et al. (2022). Radiomic feature characteristics of ovine pulmonary adenocarcinoma. Vet Sci. 2022;12(5):400. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/vetsci12050400\u003c/span\u003e\u003cspan address=\"10.3390/vetsci12050400\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemircioğlu A. Benchmarking feature selection methods in radiomics. Invest Radiol. 2022;57(7):433\u0026ndash;443. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1097/RLI.0000000000000855.do\u003c/span\u003e\u003cspan address=\"10.1097/RLI.0000000000000855.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 26, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemircioğlu A. Benchmarking feature projection methods in radiomics. Sci Rep. 2025;15(1):32368. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-025-16070-w.do\u003c/span\u003e\u003cspan address=\"10.1038/s41598-025-16070-w.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 26, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFellner A, Gavriel H, Pitaro J, Muallem Kalmovich L. Clinical parameters predicting tonsillar malignancy. Eur Arch Otorhinolaryngol. 2020;277(6):1779\u0026ndash;1783. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s00405-020-05873-4.do\u003c/span\u003e\u003cspan address=\"10.1007/s00405-020-05873-4.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed November 27, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlam T, AlShahrani I, Assiri KI, Almoammar S, Togoo RA, Luqman M. Evaluation of clinical and radiographic parameters as dental indicators for postmenopausal osteoporosis. Oral Health Prev Dent. 2020;18(3):499\u0026ndash;504. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3290/j.ohpd.a44688.do\u003c/span\u003e\u003cspan address=\"10.3290/j.ohpd.a44688.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 7, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu D, Zhang M, Li Q, Liu J, Zhuang Y, Chen Q et al. Can perihaematomal radiomics features predict haematoma expansion? Clin Radiol. 2021;76(8):629.e1\u0026ndash;629.e9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.crad.2021.03.003.do\u003c/span\u003e\u003cspan address=\"10.1016/j.crad.2021.03.003.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 11, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKobayashi T. RadiomicsJ: a library to compute radiomic features. Radiol Phys Technol.2022;15(3):255\u0026ndash;263. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s12194-022-00664-4.do\u003c/span\u003e\u003cspan address=\"10.1007/s12194-022-00664-4.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 15, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAgrawal P, Phuyal S, Panth R, Shrestha P, Lamsal R. Giant cerebral tuberculoma masquerading as malignant brain tumor \u0026mdash; a report of two cases. Cureus. 2020;12(9):e10546. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7759/cureus.10546.do\u003c/span\u003e\u003cspan address=\"10.7759/cureus.10546.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 21, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanvito F, Castellano A, Falini A. Advancements in neuroimaging to unravel biological and molecular features of brain tumors. Cancers. 2021;13(3):424. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/cancers13030424.do\u003c/span\u003e\u003cspan address=\"10.3390/cancers13030424.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 21, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBravo-Tsri AEB, Konat\u0026eacute; I, Kouassi KPB et al. Meningeal tuberculoma mimicking a brain tumor. Radiol Case Rep. 2020;16(2):284\u0026ndash;288. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.radcr.2020.11.028.do\u003c/span\u003e\u003cspan address=\"10.1016/j.radcr.2020.11.028.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 22, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcMahon P, Pisapia DJ, Schweitzer AD, Heier L, Souweidane MM, Roytman M. Central nervous system tuberculoma mimicking a brain tumor: a case report. Radiol Case Rep. 2023;19(1):414\u0026ndash;417. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.radcr.2023.10.042.do\u003c/span\u003e\u003cspan address=\"10.1016/j.radcr.2023.10.042.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 23, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang LP, Jiang ZY, Tong JL, Li N, Dong Q, Wang KZ. Development and validation of a preoperative CT-based radiomics nomogram to differentiate tuberculosis granulomas from lung adenocarcinomas: an external validation study. BMC Cancer. 2024;24(1):670. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s12885-024-12422-3.do\u003c/span\u003e\u003cspan address=\"10.1186/s12885-024-12422-3.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 27, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCabot JH, Ross EG. Evaluating prediction model performance. Surgery. 2023;174(3):723\u0026ndash;726. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.surg.2023.05.023.do\u003c/span\u003e\u003cspan address=\"10.1016/j.surg.2023.05.023.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 28, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeng X, Zhang J, Zwanenburg A, Sun J, Huang Y, Lam S et al. Building reliable radiomic models using image perturbation. Sci Rep. 2022;12(1):10035. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41598-022-14178-x.do\u003c/span\u003e\u003cspan address=\"10.1038/s41598-022-14178-x.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed December 29, 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDi Picca SAL, Saragoussi A, Bielle E, Ducray F, Villa F et al. CClinical, molecular, and radiomic profile of gliomas with FGFR3-TACC3 fusions. Neuro-oncology. 2020;22(11):1614\u0026ndash;1624. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/neuonc/noaa121.do\u003c/span\u003e\u003cspan address=\"10.1093/neuonc/noaa121.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed January 4, 2026.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu J, Yang Y, Gao Z, Song T, Ma Y, Yu X et al. Distinguishing EGFR mutation molecular subtypes based on MRI radiomics features of lung adenocarcinoma brain metastases. Clin Neurol Neurosurg. 2024;240:108258. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.clineuro.2024.108258.do\u003c/span\u003e\u003cspan address=\"10.1016/j.clineuro.2024.108258.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed January 9, 2026.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Q, Hu P. Radiogenomic association of deep MR imaging features with genomic profiles and clinical characteristics in breast cancer. Biomark Res. 2023;11(1):9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40364-023-00455-y.do\u003c/span\u003e\u003cspan address=\"10.1186/s40364-023-00455-y.do\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Accessed January 21, 2026.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Table 1","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\u003e\n"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"cerebral tuberculoma, brain metastasis, brain, MRI, radiomics","lastPublishedDoi":"10.21203/rs.3.rs-9164706/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9164706/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjective\u003c/h2\u003e \u003cp\u003eCerebral tuberculoma and brain metastasis are often indistinguishable on MRI, leading to misdiagnosis. This study develops a T1WI-enhanced MRI radiomics-based nomogram to accurately differentiate them.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA total of 482 patients (290 males, 192 females; mean age 56.08\u0026thinsp;\u0026plusmn;\u0026thinsp;12.96 years) including 254 with cerebral tuberculoma and 228 with brain metastasis were enrolled from two centers. Radiomic features were extracted, standardized, and reduced via significance testing, Pearson correlation, and LASSO regression. A radiomic model was built, then integrated with clinical predictors into a nomogram. Performance was assessed across training, internal validation, and external validation cohorts using ROC curves, calibration, and decision curve analysis (DCA).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eA total of 20 radiomic features were selected for model development. The radiomic model demonstrated areas under the curve (AUC) of 0.907, 0.891and 0.886 in training, internal validation, and independent external validation cohorts, respectively, significantly outperforming models based solely on clinical factors, which achieved AUCs of 0.781, 0.712, and 0.828 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The nomogram, which integrated the radiomic scores, age, ring enhancement, meningeal thickening, and carcinoembryonic antigen (CEA) levels, exhibited superior performance with AUCs of 0.951, 0.937, and 0.940 across the three cohorts. DCA demonstrated that the nomogram provided greater clinical net benefits across a range of threshold probabilities, and calibration curves indicated strong consistency between predicted probabilities and actual outcomes.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eA nomogram combining clinical factors and T1WI-enhanced MRI radiomics accurately differentiates cerebral tuberculoma from brain metastasis, offering robust decision support for individualized diagnosis and treatment.\u003c/p\u003e","manuscriptTitle":"Development and Application of a T1-Weighted Enhanced MRI Radiomics Nomogram for Differentiating Cerebral Tuberculoma from Brain Metastasis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-04 05:48:44","doi":"10.21203/rs.3.rs-9164706/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c512df3f-0fbb-4fff-ac41-095149095f11","owner":[],"postedDate":"May 4th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Rejected","date":"2026-05-09T15:25:34+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-05-09T15:40:08+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-04 05:48:44","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9164706","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9164706","identity":"rs-9164706","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}