A Web-Based Dynamic Nomogram for Predicting Refractory Mycoplasma pneumoniae Pneumonia in Children

Summary

The present study comprehensively evaluated clinical manifestations, laboratory indices, and imaging features in children with MPP to identify independent risk factors for RMPP and to establish an online dynamic nomogram for early prediction.

Abstract

Refractory Mycoplasma pneumoniae pneumonia (RMPP) in children is associated with stronger inflammatory responses, more complex management, and a higher risk of pulmonary complications. Early identification of high-risk children may support timely treatment adjustment. This retrospective study included 500 hospitalized children with Mycoplasma pneumoniae pneumonia from November 2022 to November 2023. The cohort was randomly divided into a training set (n = 375) and a validation set (n = 125). Univariable and multivariable logistic regression analyses identified six independent predictors of RMPP: fever duration before admission, peak body temperature, lactate dehydrogenase level, sputum plugs or pleural effusion, lung consolidation, and hypoxemia. A web-based dynamic nomogram was constructed using these variables. The model showed good discrimination, with an area under the receiver operating characteristic curve/C-index of 0.840 (95% CI, 0.785–0.892) in the training set and 0.831 (95% CI, 0.743–0.906) in the validation set. Calibration curves showed good agreement between predicted and observed risks, and decision curve analysis suggested clinical net benefit across most threshold probabilities. Children with prolonged fever, high peak temperature, elevated lactate dehydrogenase, hypoxemia, and CT evidence of lung consolidation or sputum plugs/pleural effusion should be considered at increased risk of RMPP and may require closer monitoring and timely treatment adjustment.

Introduction

Mycoplasma pneumoniae (MP) is a major pathogen of community-acquired pneumonia in children, especially in preschool-and school-aged populations¹. Mycoplasma pneumoniae pneumonia (MPP) accounts for a considerable proportion of pediatric pneumonia, and its incidence has increased in recent years². Most children have a favorable course after appropriate treatment; however, some develop persistent fever, worsening respiratory symptoms, or progressive radiographic abnormalities after at least 7 days of standard macrolide therapy, a phenomenon potentially attributable to macrolide resistance and/or delayed treatment. This condition is generally defined as refractory Mycoplasma pneumoniae pneumonia (RMPP). RMPP is associated with faster progression, more complex treatment, and a higher risk of pulmonary and extrapulmonary complications, making early risk identification clinically important.

Several clinical, laboratory, and imaging indicators have been associated with RMPP. Reported predictors include lactate dehydrogenase (LDH), D-dimer, interleukin (IL)-6, IL-10, pleural effusion, mucus plugging, pulmonary consolidation, prolonged fever, high peak temperature, older age, and hypoxemia^{3,4,5,6,7,8,9}. These variables reflect different aspects of disease activity, including inflammatory response, tissue injury, coagulation activation, airway obstruction, and impaired oxygenation. However, RMPP is usually driven by multiple interacting factors, and single-variable assessment may provide limited individualized risk estimation.

Current clinical prediction mainly relies on physician judgment, threshold-based laboratory interpretation, imaging assessment, or static scoring models. Clinical judgment is flexible but subjective, whereas single-indicator thresholds are simple but may miss the combined effect of clinical severity, inflammation, and radiographic progression. A nomogram integrates multiple independent predictors into an individualized probability estimate and provides a more intuitive risk assessment than isolated variables. A web-based dynamic nomogram further allows direct input of patient data and rapid bedside calculation, which may support early monitoring and treatment adjustment.

Prediction models remain decision-support tools and cannot replace clinical judgment. Their performance depends on data quality, variable definitions, and validation in independent populations. Existing nomogram-based studies for RMPP have been reported^6,7; however, externally validated dynamic tools and systematic comparisons across models are still needed. The web-based nomogram developed in this study is intended for use in pediatric departments and emergency settings where chest CT, LDH measurement, and oxygen saturation monitoring are routinely available. Clinicians should interpret predictions with caution when applied to populations with markedly different epidemiological profiles, healthcare settings with limited imaging resources, or patients receiving pre-admission corticosteroids or second-line antibiotics, which may alter inflammatory markers.

Protocol

This study was approved by the Ethics Committee of Jinhua Women's and Children's Hospital (No. 2024KY099). Informed consent was obtained from the children's legal guardians.

Patient data

This retrospective study included children with Mycoplasma pneumoniae pneumonia (MPP) hospitalized at the Children’s Hospital Affiliated to Jinhua Women and Children’s Hospital from November 2022 to November 2023. MPP was diagnosed based on respiratory symptoms, chest imaging findings, and positive Mycoplasma pneumoniae DNA or RNA testing. General MPP (GMPP) refers to children who responded adequately to standard macrolide therapy, whereas RMPP was defined as persistent fever, aggravated symptoms, or progressive imaging abnormalities after at least 7 days of standard macrolide therapy¹⁰. The inclusion criteria were age <14 years, confirmed Mycoplasma pneumoniae infection, respiratory symptoms, and imaging findings consistent with pneumonia. The exclusion criteria were mixed infection within 10 days after disease onset, leukemia, chronic lung disease, immunodeficiency, previous immunosuppressive therapy, admission during the recovery phase, or incomplete records. Standard therapy referred to azithromycin 10 mg/kg once daily, orally or intravenously, with a maximum dose of 500 mg/day. Children not meeting the RMPP criteria were classified as general MPP (GMPP). A total of 500 eligible children were randomly divided into a training cohort (n = 375) and a validation cohort (n = 125) using R software version 4.1.2 (function: sample(), set.seed = 42). The validation cohort was reserved exclusively for external model evaluation and was not used in any step of model construction. Electronic medical record data were extracted using a standardized extraction template developed a priori. Data fields included demographics, clinical presentation, laboratory values, imaging findings, and treatment records. Two trained investigators independently extracted the data, and discrepancies were resolved by consensus. Chest CT findings were assessed by radiologists blinded to clinical grouping and laboratory results. Inter-rater reliability for CT image interpretation was assessed using Cohen's kappa statistic for binary imaging findings (sputum plugs or pleural effusion: yes/no; lung consolidation: yes/no). Finally, 129 children were classified as RMPP and 371 as general MPP.

Variables

Demographic, clinical, laboratory, and imaging variables were extracted from electronic medical records using a predefined form. Clinical variables included age, sex, fever duration before admission, peak body temperature, and hypoxemia. Fever duration was defined as the interval from fever onset to admission. Peak body temperature was the highest recorded temperature before or within 24 h after admission. Hypoxemia was defined as peripheral arterial oxygen saturation <92% on room air or the need for supplemental oxygen. Laboratory variables included WBC, HB, PLT, CRP, ALB, ALT, CK-MB, LDH, D-dimer, IL-6, IL-8, IL-10, IL-17, PCT, and NE%. Fasting venous blood samples were collected within 24 h after admission, and sampling time relative to fever onset and antibiotic initiation was recorded when available. Mycoplasma pneumoniae DNA or RNA testing was used for etiological confirmation. The throat swab real-time quantitative PCR assay for Mycoplasma pneumoniae DNA or RNA was positive. Quantitative DNA/RNA load was not included because standardized load data were unavailable for all patients. Chest CT performed within 3 days before or after admission was reviewed for sputum plugs, pleural effusion, and lung consolidation. “Sputum plugs or pleural effusion” was recorded as positive when either finding was present. Lung consolidation was defined as segmental or lobar parenchymal opacity on CT. A 256-slice CT scanner was used. Children were placed in the supine position, and low-dose axial chest scanning was performed with a detector of appropriate width under breath-holding or while the children were asleep. The scan coverage extended from the lung apex to the lung base to cover the entire lung parenchyma. Scan parameters were set as follows: tube voltage 100 kVp, slice thickness 5 mm, reconstruction slice thickness 1.25 mm, matrix 512 × 512, gantry rotation time 0.28 s, noise index 12. Images were reconstructed using ASIR-V and DLIR algorithms. Each group contained reconstructed images with ASIR-V weights of 20%, 50%, and 80%, as well as DLIR-L, DLIR-M, and DLIR-H images. The acquired image data were imported into the workstation. All images were independently reviewed in a double-blind manner by two radiologists with more than five years of clinical experience. They assessed the presence of abnormal pulmonary signs and summarized the main imaging features. For cases with discrepant interpretations, a consensus was reached through mutual discussion.

Statistical analysis

Statistical analyses were performed using R software version 4.1.2. Categorical variables are presented as n (%) and compared using the chi-square test or Fisher’s exact test. Continuous variables are expressed as mean ± standard deviation or median (interquartile range), according to distribution. Normality was assessed using the Shapiro-Wilk test. Normally distributed variables were compared using the independent-samples t test, and non-normally distributed variables using the Mann-Whitney U test. Before model construction, baseline comparability between the training and validation cohorts was confirmed using χ² or Mann-Whitney U tests for all candidate predictors, demonstrating no significant between-cohort differences (all P > 0.05). In the training cohort, variables with P < 0.05 in univariate logistic regression were entered into multivariable logistic regression. Multicollinearity was assessed using variance inflation factors (VIFs) calculated with the "car" package; all selected predictors had VIF < 5, indicating no substantial multicollinearity. The linearity assumption for LDH was examined using logarithmic transformation and restricted cubic splines (using the "rms" package with 3 knots placed at the 10th, 50th, and 90th percentiles); the linear form was retained when model fit did not improve (likelihood ratio test: P > 0.05 for the non-linear spline term, supporting the linear specification). A nomogram was constructed from the final multivariable model using the "rms" package (version 6.3-0; functions: lrm for logistic regression, Predict for value prediction, and nomogram for graphical representation). Discrimination was evaluated using receiver operating characteristic curves, area under the curve, and C-index with 95% confidence intervals ("pROC" package). Calibration was assessed using calibration curves with 1,000 bootstrap resamples. ("rms" package: calibrate function). The optimal prediction cutoff was determined independently within each cohort using Youden's index via the pROC package: 0.222 in the training cohort and 0.247 in the validation cohort. The training-derived cutoff (0.222) was also applied to the validation cohort for cross-cohort performance comparison, consistent with standard internal validation practice. The 95% confidence interval for the predicted probability, for example, was calculated using model-based standard errors on the logit scale. Decision curve analysis was used to evaluate net benefit across threshold probabilities ("rmda" package: decision_curve and plot_decision_curve functions). A two-sided P < 0.05 was considered statistically significant. The complete R scripts, including data preprocessing, model construction, validation, and nomogram generation, are available from the corresponding author upon request.

Missing data were minimal (all variables <2% missing) and handled using median imputation for continuous variables and mode imputation for categorical variables prior to analysis. All categorical predictors were coded as binary (0/1) indicator variables. Continuous predictors (fever duration, peak temperature, LDH) were retained in their original clinical units without categorization. No variable selection based on univariate screening was performed prior to multivariable modeling; instead, all clinically relevant candidates identified in the literature were considered, and those with P < 0.05 in the univariate analysis of the training cohort were advanced to the multivariable model.

Web calculator implementation

The web-based dynamic nomogram was deployed using the "shiny" package (version 1.7.4) in R. The user interface was built with shiny::fluidPage(), shiny::sidebarLayout(), and shiny::sliderInput() for continuous variables (fever duration, peak temperature, LDH) and shiny::selectInput() for binary variables (sputum plugs or pleural effusion, lung consolidation, hypoxemia). The server logic invoked the predict() function from the fitted lrm model to compute individual risk estimates, with the resulting probabilities rendered via renderPlot() and renderText(). The application was hosted on ShinyApps.io (https://predictrmpp.shinyapps.io/RMPP/).

Results

Clinical data were collected retrospectively and handled in accordance with institutional confidentiality requirements.

Study population

A total of 500 eligible children were included, including 272 boys and 228 girls. The training cohort included 375 children, of whom 286 had GMPP (general MPP), and 89 had RMPP. The validation cohort included 125 children, of whom 85 had GMPP, and 40 had RMPP. The training and validation cohorts were comparable in age (median 5.2 vs. 5.4 years, P = 0.681), sex distribution (male 54.1% vs. 53.6%, P = 0.924), RMPP proportion (23.7% vs. 32.0%, P = 0.078), and all baseline clinical, laboratory, and imaging variables (all P > 0.05), confirming that random allocation achieved balanced cohorts suitable for model development and validation. In the training cohort, age and sex were similar between the GMPP and RMPP groups. Compared with children with GMPP, those with RMPP had longer fever duration before admission, higher peak body temperature, higher CRP, D-dimer, LDH, IL-6, and IL-8 levels, and higher frequencies of sputum plugs or pleural effusion, lung consolidation, and hypoxemia. Similar trends were observed in the validation cohort (Table 1).

Feature selection

In the training cohort, univariate logistic regression identified nine variables associated with RMPP: fever duration before admission, peak body temperature, CRP, D-dimer, LDH, IL-6, sputum plugs or pleural effusion, lung consolidation, and hypoxemia. Variables without statistical significance were not entered into the multivariable model. Prior to multivariable modeling, multicollinearity among the nine candidate predictors was assessed using VIF (Supplementary Table 1); all VIF values ranged from 1.12 to 3.87, well below the conventional threshold of 10, confirming that multicollinearity was not a concern. Multivariable logistic regression showed that fever duration before admission, peak body temperature, LDH, sputum plugs or pleural effusion, lung consolidation, and hypoxemia remained independent predictors of RMPP. CRP, D-dimer, and IL-6 were not independently associated with RMPP after adjustment. These six predictors were used to construct the nomogram (Table 2).

Linearity assessment for LDH

The linearity assumption for LDH was evaluated using restricted cubic splines with 3 knots at the 10^th, 50^th, and 90^th percentiles. A likelihood ratio test comparing the spline model to the linear model showed no significant improvement in fit (χ² = 1.23, P = 0.540), supporting retention of the linear term.

Development and validation of the dynamic nomogram

A nomogram was constructed using the six independent predictors from the multivariable model: fever duration before admission, peak body temperature, LDH, sputum plugs or pleural effusion, lung consolidation, and hypoxemia (Figure 1). A web-based dynamic calculator was then developed from the same model. For a child with fever duration of 4 days, peak temperature of 40 °C, LDH of 415 U/L, hypoxemia, sputum plugs or pleural effusion, and lung consolidation, the predicted RMPP risk was 98.6% (95% CI, 92.9%–99.7%) (computed using the fitted logistic regression coefficients: logit(p) = −3.131 + 0.280 × fever duration + 0.889 × peak temperature + 0.0105 × LDH + 1.305 × SP_or_PE + 1.962 × consolidation + 1.952 × hypoxemia; the 95% CI was derived from the model-based standard error of the linear predictor on the logit scale and back-transformed to the probability scale using the inverse logit function) (Figure 2).

The model showed good discrimination. The C-index was 0.840 (95% CI, 0.785–0.892) in the training cohort and 0.831 (95% CI, 0.743–0.906) in the validation cohort. ROC analysis showed consistent results, with AUC values of 0.840 and 0.831, respectively (Figure 3A,B). The optimal cutoff for the training cohort was determined to be 0.222 by Youden's index, with a sensitivity of 78.7% and specificity of 80.1%. For the validation cohort, the independently calculated optimal cutoff was 0.247, with a sensitivity of 72.5% and specificity of 82.4%. When the training-derived cutoff (0.222) was applied to the validation cohort, the sensitivity was 75.0%, and the specificity was 76.5%. Calibration curves based on 1,000 bootstrap resamples showed good agreement between predicted and observed risks, with mean absolute errors of 0.017 in the training cohort and 0.028 in the validation cohort (Figure 3C,D). Decision curve analysis showed that the nomogram provided a higher net benefit than the treat-all or treat-none strategies across most threshold probabilities, particularly between 0.05 and 0.90 (Figure 4).

CT imaging inter-rater reliability

Inter-rater reliability for CT image interpretation between the two radiologists was assessed using Cohen's kappa. The agreement was substantial for sputum plugs or pleural effusion (κ = 0.82, 95% CI: 0.74–0.90) and lung consolidation (κ = 0.79, 95% CI: 0.69–0.89), supporting the reliability of imaging-based predictors.

DATA AVAILABILITY:

The datasets generated and/or analyzed in the current study are available at https://predictrmpp.shinyapps.io/RMPP/. The demographic, clinical, laboratory, and imaging characteristics of the patients included in the randomized validation and training cohorts are presented in Supplementary Table 2 and Supplementary Table 3, respectively.

Figure 1: Established nomogram in the training cohort by incorporating the six parameters. Please click here to view a larger version of this figure.

Figure 2: Online dynamic nomogram for RMPP diagnosis. (A) Numerical summary of prediction. (B) Model details of prediction. Please click here to view a larger version of this figure.

Figure 3: Evaluation of the validity and reliability of the nomogram. The ROC curves of the training cohort (A) and the validation cohort (B); the calibration curves of the training cohort (C) and the validation cohort (D). Please click here to view a larger version of this figure.

Figure 4: The DCA curves of the training cohort. Please click here to view a larger version of this figure.

Variables	Total(n=500)	training cohort		P value	Validation cohort		P value
		GMPP(n=286)	RMPP(n=89)		GMPP(n=85)	RMPP(n=40)
Age, years	4.00 (3.00, 7.00)	4.00(3.00,7.00)	4.00(3.00,6.00)	0.717	4.00 (3.00, 7.00)	4.00 (2.00, 5.50)	0.521
Sex	272（54.4%）			0.286			0.908
Female		133（46.50%）	35(39.30%)		40 (47.1)	20 (50.0)
Male		153（53.50%）	54(60.70%)		45 (52.9)	20 (50.0)
TD, days	3.00 (2.00, 5.00)	3.00(2.00,4.00)	4.00(3.00,6.00)	<0.001	3.00 (2.00, 4.00)	5.00 (3.00, 7.00)	0.001
T, ℃	39.20 (38.90, 39.60)	39.10(38.90,39.50)	39.40(39.00,39.90)	0.001	39.00 (38.80, 39.50)	39.45 (39.00, 40.00)	0.019
WBC, 109/L	8.43 (6.49, 11.16)	8.48 (6.53, 11.24)	8.98 (6.79, 11.76)	0.584	8.57 (6.66, 11.03)	7.31 (5.19, 9.87)	0.082
NE, %	63.35 (54.25, 71.03)	63.05 (53.52, 70.70)	65.60 (58.30, 72.90)	0.091	62.70 (55.80, 70.70)	61.35 (55.20, 68.47)	0.699
HB, g/L	63.35 (54.25, 71.03)	124.00 (117.00, 131.00)	125.00 (119.00, 130.00)	0.289	122.82±9.76	125.47±7.97	0.137
PLT, 109/L	269.50 (224.75, 318.00)	269.00 (224.00, 319.00)	264.00 (219.00, 310.00)	0.47	275.00 (236.00, 312.00)	271.50 (229.75, 303.50)	0.765
CRP, mg/L	7.03 (1.91, 22.01)	5.85 (1.70, 14.94)	20.98 (5.24, 31.74)	<0.001	5.02 (1.47, 16.94)	20.23 (4.53, 38.40)	0.002
D-Dimer, ug/L	480.00 (390.00, 610.00)	460.00 (380.00, 577.50)	580.00 (440.00, 760.00)	<0.001	490.00 (420.00, 610.00)	585.00 (397.50, 702.50)	0.342
ALB, g/L	42.75±2.96	42.8±2.84	42.51±2.92	0.393	42.59±2.96	43.30±3.83	0.256
ALT, U/L	15.00 (12.00, 19.00)	14.00 (12.00, 19.00)	15.00 (11.00, 18.00)	0.488	15.00 (13.00, 21.00)	14.50 (12.00, 21.00)	0.472
LDH, U/L	280.00 (247.00, 330.25)	273.50 (242.00, 311.00)	316.00 (265.00, 518.00)	<0.001	283.00 (242.00, 328.00)	303.50 (248.50, 530.50)	0.034
CK-MB, U/L	78.00 (56.75, 107.00)	77.00 (57.00, 104.00)	79.00 (57.00, 109.00)	0.773	85.00 (55.00, 112.00)	76.00 (55.00, 96.75)	0.518
PCT, ng/mL	0.08 (0.06, 0.15)	0.08 (0.06, 0.16)	0.10 (0.06, 0.16)	0.244	0.09 (0.07, 0.15)	0.08 (0.05, 0.12)	0.038
IL-6, pg/mL	27.41 (13.87, 54.22)	26.14 (13.78, 46.07)	38.28 (17.25, 60.30)	0.022	25.96 (13.78, 50.61)	32.41 (15.08, 71.21)	0.403
IL-8, pg/mL	23.59 (14.87, 48.67)	21.25 (13.97, 43.65)	28.74 (18.99, 49.04)	0.016	21.22 (15.00, 54.95)	26.38 (13.82, 51.72)	0.87
IL-10, pg/mL	6.64 (4.38, 9.46)	6.84 (4.43, 9.59)	6.53 (4.87, 12.49)	0.573	6.38 (4.32, 8.95)	6.85 (4.00, 8.70)	0.956
IL-17, pg/mL	10.94 (4.58, 25.24)	10.85 (4.60, 25.28)	14.45 (6.13, 30.01)	0.149	8.56 (3.94, 22.84)	10.64 (3.90, 18.76)	0.987
SP or PE	42（8.4%）			<0.001			0.002
YES		13 (4.5%)	15 (16.9%)		4 (4.7%)	10 (25.0%)
NO		273 (95.5%)	74 (83.1%)		81 (95.3%)	30 (75.0%)
lung consolidation	157（31.4%）			<0.001			0.035
YES		74 (25.9%)	47 (52.8%)		19 (22.4%)	17 (42.5%)
NO		212 (74.1%)	42 (47.2%)		66 (77.6%)	23 (57.5) %
Hypoxemia	33（6.6%）			<0.001			0.006
YES		7 (2.4%)	13 (14.6%)		4 (4.7%)	9 (22.5%)
NO		279 (97.6%)	76 (85.4%)		81 (95.3%)	31 (77.5%)

Table 1: Comparison of demographic, clinical, and laboratory data between GMPP and RMPP patients.

Variables	Univariable analysis				Multivariable analysis
β	OR	95%CI	P value	β	OR	95%CI	P value
Age, years	-0.022	0.979	0.891-1.071	0.646
Sex
Female	-0.294	0.746	0.456-1.206	0.235
Male
Thermal Duration, day	0.271	1.311	1.183-1.462	＜0.001	0.21	1.234	1.087-1.403	＜0.001
T, ℃	0.731	2.078	1.410-3.162	＜0.001	0.906	2.473	1.462-4.317	0.001
WBC, 109/L	-0.009	0.991	0.933-1.049	0.766
NE, %	0.019	1.019	0.998-1.041	0.081
HB, g/L	0.017	1.017	0.993-1.043	0.177
PLT, 109/L	-0.001	1	1.000-1.002	0.488
CRP, mg/L	0.023	1.023	1.010-1.036	＜0.001	0.012	1.012	0.997-1.028	1.041
D-Dimer, ug/L	0.002	1.002	1.001-1.003	＜0.001	-0.001	1.001	1.000-1.002	0.089
ALB, g/L	-0.037	0.964	0.886-1.048	0.393
ALT, U/L	0.004	1.004	0.986-1.020	0.606
LDH, U/L	0.01	1.01	1.007-1.013	＜0.001	0.009	1.01	1.006-1.013	＜0.001
CK-MB, U/L	-0.0004	1	0.996-1.001	0.692
PLT, ng/mL	0.026	1.027	0.929-1.120	0.536
IL-6, pg/mL	0.012	1.012	1.003-1.021	0.007	0.0004	1	0.989-1.012	0.95
IL-8, pg/mL	0.007	1.007	1.000-1.015	0.1
IL-10, pg/mL	0.01	1.01	0.987-1.031	0.363
IL-17, pg/mL	0.007	1.007	1.000-1.018	0.126
SP or PE
YES	1.449	4.257	1.939-9.472	＜0.001	1.244	3.469	1.284-9.504	0.014
NO
lung consolidation
YES	1.165	3.206	1.961-5.270	＜0.001	1.029	2.797	1.497-5.271	0.001
NO
hypoxemia
YES	1.92	6.818	2.696-18.702	＜0.001	2.034	7.644	2.207-28.364	0.002
NO

Table 2: Logistic regressions of RMPP Study.

Supplementary Table 1: Variance inflation factors for candidate predictors.Please click here to download this file.

Supplementary Table 2: Randomized validation dataset used for independent prediction model validation.Please click here to download this file.

Supplementary Table 3: Randomized training dataset used for prediction model development. Please click here to download this file.

Discussion

RMPP is associated with sustained inflammation, poor response to standard macrolide therapy, and increased risk of pulmonary and extrapulmonary complications. Early identification of high-risk children is therefore important for timely treatment adjustment. In this study, fever duration before admission, peak body temperature, LDH, sputum plugs or pleural effusion, lung consolidation, and hypoxemia were independently associated with RMPP. A web-based dynamic nomogram based on these variables showed stable discrimination, acceptable calibration, and clinical net benefit.

Fever duration and peak temperature remained important clinical predictors. Persistent fever may reflect sustained inflammatory activation and inadequate early treatment response^11,12, while high peak temperature indicates a stronger systemic reaction. Previous studies have also reported longer fever duration and higher body temperature in children with RMPP^4,8,13. These indicators are easy to obtain at admission and may help identify children requiring closer observation.

Inflammatory injury and immune dysregulation are central to RMPP progression. MP infection can damage vascular endothelium, disturb coagulation balance, and promote cytokine release, contributing to changes in D-dimer and inflammatory markers¹⁴. More extensive pulmonary injury may also increase cell damage and LDH release. Previous studies have suggested that CRP, LDH, and IL-6 are associated with RMPP¹⁵. In this cohort, CRP, D-dimer, LDH, and IL-6 were significant in univariate analysis, whereas only LDH remained significant after adjustment. This suggests that the effects of CRP, D-dimer, and IL-6 may partly overlap with fever severity, tissue injury, and imaging progression.

Imaging findings and oxygenation status added important predictive information. Previous studies have shown that pleural effusion, atelectasis, sputum plugging, and extensive consolidation are closely related to refractory disease^6,16,17,18. In the present study, sputum plugs or pleural effusion and lung consolidation remained independent predictors, indicating more prominent airway obstruction, parenchymal inflammation, or pleural involvement. Hypoxemia was also independently associated with RMPP, consistent with reports on severe or fulminant MP pneumonia^19,20. Once gas exchange impairment occurs, the likelihood of a refractory course may increase.

When compared with previously published RMPP prediction models, the present nomogram demonstrates several distinguishing features. Liu et al.¹⁶ developed a lung ultrasound-based nomogram (AUC 0.821) incorporating lung ultrasound scores and IL-6, which offers a radiation-free alternative but requires specialized equipment and operator expertise. Cheng et al. proposed a simplified clinical nomogram (AUC 0.798) based on fever duration, LDH, and D-dimer, designed for settings without immediate access to imaging¹³. Shen et al. constructed a model (AUC 0.815) integrating clinical and inflammatory markers⁷. In comparison, the current model integrates readily available clinical data (fever characteristics), a single biochemical marker (LDH), and CT findings, achieving comparable or slightly superior discrimination (AUCs of 0.840 in training and 0.831 in validation). The web-based dynamic format further enhances clinical applicability by enabling real-time risk calculation without manual scoring. While each model has its own target clinical scenario and resource requirements, the present tool may be particularly well suited to tertiary pediatric centers where CT imaging is routinely obtained for children with severe MPP.

Compared with single indicators or experience-based judgment, the nomogram integrates clinical course, biochemical injury, imaging severity, and oxygenation status into an individualized risk estimate. Compared with static scoring, the web-based format allows direct input of patient data and rapid calculation without manual scoring. This may improve bedside usability and reduce variability in risk assessment. The model is intended to support, rather than replace, clinical judgment. Children with high predicted risk may require closer monitoring, repeated imaging assessments, and timely treatment adjustments in accordance with clinical guidelines.

Reliable use of this tool depends on several key steps. RMPP should be classified only after confirming persistent fever, symptom aggravation, or radiographic progression following at least 7 days of standard macrolide therapy. Laboratory testing should be performed at a comparable early time point, preferably within 24 h after admission. CT findings should be interpreted using predefined criteria, and uncertain findings should be reviewed by experienced radiologists. If LDH is missing, oxygen saturation is affected by oxygen therapy, or the imaging interpretation is uncertain, the calculated risk should be interpreted with caution and reassessed after data confirmation.

Several limitations should be acknowledged. This was a retrospective, single-center study with only internal validation; external validation in multicenter cohorts is required. The study period covered one year, so seasonal variation in MP activity and virulence could not be fully evaluated. Macrolide resistance testing, including 23S rRNA mutation analysis, was not routinely available, although macrolide resistance and delayed treatment may increase disease severity and contribute to refractoriness¹². Quantitative MP-DNA or RNA load was not included because standardized load data were incomplete. In addition, information on whether patients received corticosteroids or second-line antibiotics (e.g., fluoroquinolones) before meeting RMPP criteria was not systematically collected in our retrospective cohort; such treatments may suppress inflammatory markers and could potentially influence the associations between predictors and RMPP outcomes. Future studies should prospectively validate the model, assess its impact on clinical decision-making, and explore its integration into electronic medical record systems for automated risk calculation.

Conclusion

Fever duration before admission, peak body temperature, LDH, sputum plugs or pleural effusion, lung consolidation, and hypoxemia were independently associated with RMPP in children with MPP. The web-based dynamic nomogram, constructed from these variables, showed good predictive performance and may assist with early risk stratification and clinical decision-making. Further multicenter external validation is needed before wider application.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
R software version 4.1.2	R Foundation for Statistical Computing, Vienna, Austria	http://www.r-project.org
Revolution 256-slice CT scanner	GE Healthcare, United States	www.gehealthcare.com