Development of a novel clinical decision support tool for determining the need for whole-body computed tomography in pediatric trauma patients: the WHOLE score Authors: Enes Çay, Gül Pamukçu Günaydın, Çağdaş Yıldırım, Alp Şener Published: 2026 ABSTRACT AbstractBackgroundWhole body computed tomography (WBCT) is widely used in the initial assessment of pediatric trauma patients to rapidly identify injuries; however, its routine use exposes children to substantial ionizing radiation. Clear guidance on which pediatric trauma patients truly benefit from WBCT is lacking, leading to practice variability and potential overuse in emergency departments. The proposed score is intended to complement existing pediatric imaging decision frameworks that emphasize selective and risk-based use of computed tomography.MethodsThis retrospective, single-center observational study included trauma patients under 18 years of age who underwent WBCT in a tertiary pediatric trauma center between June 2021 and June 2022. The primary outcome was the presence of traumatic injuries involving ≥ 2 anatomical regions on WBCT not necessarily the clinical need for WBCT. Candidate predictors were identified from the literature and clinical practice and evaluated using multivariable logistic regression. Continuous variables were dichotomized based on receiver operating characteristics (ROC) analysis. Independent predictors were used to develop a clinical decision support tool, the WHOLE score. Model performance was assessed using ROC analysis and diagnostic accuracy metrics.ResultsA total of 407 pediatric trauma patients who underwent WBCT were included; 37.6% (n = 153), had injuries involving two or more anatomical regions. Independent predictors of multi-region injury were age ≥ 133 months, tachycardia, hypoxia, positive physical examination findings in the head, thoracic, pelvic, or lower extremity regions, and aspartate aminotransferase (AST) ≥ 50 U/L. These variables were incorporated into the WHOLE score (range 0–8). The WHOLE score demonstrated good discrimination (AUC 0.84; 95% CI 0.80–0.88). Using a cut-off of ≥ 2 points, sensitivity was 96.7% (CI 92.5–98.9) and NPV was 94.8% for detecting multi-region injuries.ConclusionsThe WHOLE score is a simple clinical decision support tool designed to identify pediatric trauma patients at low risk for multi-region injuries, thereby supporting emergency physicians in safely withholding WBCT in appropriately selected cases. The score functions as a rule-out tool: a low score (0–1) may support a decision to forego WBCT, whereas a high score does not independently mandate it. By reducing unnecessary WBCT in low-risk patients, the score has the potential to decrease radiation exposure while maintaining diagnostic safety. This is a preliminary single-center derivation study, and external prospective validation is required before routine clinical implementation.Trial registrationThe study protocol was registered retrospectively at clinicaltrials.gov (Clinical Trials Identifier: NCT07411235, Date:11/02/2026).Keywords: Pediatric trauma, Whole-body computed tomography, Clinical decision support systems Background Trauma is the leading cause of mortality in childhood and young adulthood [1, 2]. Whole-body computed tomography (WBCT) facilitates clinical decision-making in trauma patients during the critical initial hours by providing rapid and accurate diagnostic information [3]. In the initial evaluation of pediatric trauma patients, physicians may prefer the use of WBCT due to uncertainty regarding the mechanism of injury and challenges in the assessment of physical examination findings [3]. Children are particularly vulnerable to the effects of ionizing radiation due to increased radiosensitivity of developing tissues and a longer remaining lifespan during which radiation-related malignancies may develop. Epidemiological studies have demonstrated an association between pediatric computed tomography exposure and an increased lifetime risk of cancer, with risk magnitude influenced by cumulative radiation dose and age at exposure [4]. Therefore, imaging strategies in pediatric trauma must carefully balance the risk of missed injuries against the potential long-term harms of radiation exposure [4–6]. Consequently, emergency medicine specialists caring for pediatric patients must carefully balance the risk of missed, potentially life-threatening injuries against the risk of radiation-induced malignancy [7]. Compared with WBCT, the use of selective computed tomography (SCT) is associated with reduced radiation exposure [8]. Although numerous studies have examined WBCT findings in adult trauma populations, the literature regarding pediatric patients with high-energy trauma remains limited [9]. Moreover, existing studies emphasize substantial health disparities in pediatric trauma care, demonstrating that sociodemographic factors—such as residence in rural versus urban areas and the type of treating facility (pediatric vs. adult trauma center)—significantly influence both computed tomography (CT) utilization rates and access to specialized trauma services [3, 10]. In pediatric populations, the results of studies comparing WBCT with SCT remain inconclusive. Caputo et al. reported that, among severely injured trauma patients, those who underwent WBCT were less likely to experience a fatal outcome [11]. In contrast, other studies demonstrated that, in hemodynamically stable children with blunt trauma, the use of SCT was not associated with worse clinical outcomes compared with WBCT, and no significant difference in mortality was observed [7, 12, 13]. Despite the availability of numerous pediatric trauma scoring systems, there is currently no consensus regarding which pediatric trauma patients should undergo WBCT, and the applied criteria vary considerably among trauma centers. Several international guidelines emphasize a selective and risk-based approach to imaging in pediatric trauma. The American Academy of Pediatrics recommends judicious use of advanced imaging to minimize unnecessary radiation exposure [5], while clinical decision instruments such as the Pediatric Emergency Care Applied Research Network (PECARN) rules and the National Institute for Health and Care Excellence (NICE) guidelines provide structured criteria for region-specific imaging [14, 15]. These frameworks prioritize clinical assessment and selective computed tomography rather than routine whole-body imaging. However, they primarily focus on individual anatomical regions and do not specifically address the identification of patients who may benefit from whole-body computed tomography (WBCT). In this context, a simple and clinically applicable tool to support decision-making for WBCT use in pediatric trauma may complement existing selective imaging strategies. The aim of this study was to investigate the association between injury mechanism, physical examination findings, laboratory parameters, and the presence of injuries involving multiple CT regions on WBCT in trauma patients under 18 years of age. Unlike existing decision rules that are primarily designed for region-specific imaging, the WHOLE score aims to support identification of patients at risk for multi-region injuries who may benefit from whole-body computed tomography. The score is specifically intended for use among patients in whom the treating physician is already considering WBCT and it serves as a second-step risk stratification tool to identify those at sufficiently low risk for multi-region injury that WBCT may be safely withheld, rather than a primary screening instrument for all pediatric trauma presentations. Additionally, this study sought to develop a clinical risk assessment score to assist emergency medicine specialists in decision-making, with the goal of reducing unnecessary WBCT utilization. Methods This study was conducted using data derived from a retrospective, cross-sectional, single-center observational cohort. The study included patients evaluated between June 1, 2021, and June 1, 2022. The study was conducted in accordance with the current version of the Declaration of Helsinki and was approved by the Ankara Bilkent City Hospital Clinical Research Ethics Committee No. 2 (approval date: 23.11.2022; decision number: E2-22-2872). This study adheres to the TRIPOD (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis) 2015 guidelines, ensuring transparent reporting of the development and, where applicable, validation of the predictive model. Participants The study was conducted at one of the two pediatric trauma centers located in a metropolitan area with a population of approximately 5.5 million. The center provides 24-hour coverage by emergency medicine specialists, pediatricians, pediatric surgeons, radiologists, and other trauma-related specialties. The emergency department has an annual visit volume of approximately 500,000 adult and pediatric patients. Inclusion criteria All trauma patients under 18 years of age who presented to the emergency department and underwent WBCT based on the decision of the attending emergency medicine specialist at the time of presentation were included in the study. Exclusion criteria Patients with missing data in the hospital information system or those who left the hospital for any reason before completion of the emergency department diagnostic and treatment process were excluded. No data from these patients were included in the analysis. Patients were classified as having undergone WBCT if they received CT of the head, chest, and abdomen/pelvis during their emergency department presentation. Patients who did not undergo all three scans but received at least one CT scan were classified as having undergone SCT. As this was a retrospective chart review, the conduct of the study did not influence the indications for CT imaging. Throughout the study period, the decision to perform WBCT in pediatric trauma patients presenting to the emergency department was made by the responsible emergency medicine specialist based on injury mechanism, vital signs, physical examination findings, laboratory results when indicated, and ultrasonography findings if performed. No formal risk stratification tool was used beyond clinical gestalt. All patients received standard pediatric trauma care in accordance with local protocols. As only patients who underwent WBCT were included, the study population represents a pre-selected cohort of higher-risk trauma patients. Outcome measures The primary outcome was defined as the detection of injuries involving more than one anatomical region on WBCT. Patients were considered to have positive imaging findings if any trauma-related injury was identified on CT. Incidental findings or abnormalities not related to trauma were not considered positive imaging findings. It should be noted that this outcome represents a radiological surrogate rather than a direct measure of clinical necessity for WBCT or requirement for therapeutic intervention. Physical examination and vital signs Physical examination findings were considered positive if any of the following were present. For the head and neck region, these included deterioration in level of consciousness, abrasions, ecchymosis, hematoma, lacerations, palpable fractures, cervical tenderness, restricted or painful neck movements, and soft tissue injuries of the cervical area. For the thoracic region, positive findings included chest wall ecchymosis, hematoma, abrasion, laceration, deformity, paradoxical respiratory movements, subcutaneous emphysema, absent or asymmetric breath sounds on auscultation, chest wall tenderness, tracheal deviation, and tenderness, pain, or soft tissue injuries over the thoracic vertebrae. Abdominal examination was considered positive in the presence of ecchymosis, abrasion, laceration, tenderness on palpation, guarding, rebound tenderness, or lumbar vertebral tenderness or soft tissue injury. Pelvic examination was considered positive if perineal or pelvic edema, scrotal or labial swelling, lacerations, deformity, pelvic hematoma, or tenderness on palpation were present. For the extremities, ecchymosis, abrasion, laceration, tenderness, restricted range of motion, deformity, or the presence of open fractures were recorded as positive findings. Vital signs at presentation were classified according to age-specific normal ranges based on Table 10.5 of the pediatric trauma section of the Advanced Trauma Life Support Course Student Manual, 10th edition [16]. Image acquisition During the study period, all WBCT examinations were performed using two CT scanners located within the emergency department. Both scanners were 128-detector systems (GE/Revolution EVO, GE Medical Systems, Milwaukee, WI, USA). Images were initially acquired with a slice thickness of 2.5 mm and subsequently reformatted to 1.25 mm. Facial CT scans were obtained with a slice thickness of 0.6 mm. Age-adjusted pediatric dose-reduction protocols were applied for all scans, using predefined age groups of 0–4 years, 5–9 years, and 10–16 years. Adult imaging protocols were applied for patients aged 17 years and older. Tube voltage settings were as follows: for patients aged 0–4 years, 100 kV for non-thoracic regions and 80 kV for thoracic imaging; for patients aged 5–9 years, 120 kV for cranial imaging and 100 kV for other regions; and for patients aged 10–16 years, 120 kV for cranial and cervical imaging and 100 kV for thoracic and abdominal imaging. For abdominal CT examinations, intravenous contrast material was administered at a dose of 1 mL/kg. All imaging studies were independently evaluated by two board-certified radiologists with experience in pediatric trauma imaging, both of whom routinely work in emergency radiology. Image interpretation was performed using a remote workstation. Image interpretation was performed by radiologists with knowledge of the patients’ physical examination findings; therefore, no blinding to clinical information was applied. Sample size The entire dataset collected over a one-year period was used for practical considerations. The sample size of the present study is comparable to those reported in previous studies addressing similar research questions. Missing data Patients with missing data in the collected records were excluded from the study. Accordingly, analyses were conducted using a complete-case approach, and no incomplete cases were included in the final analysis. Statistical analysis Statistical analyses were performed using IBM SPSS Statistics for Windows, Version 20.0 (IBM Corp., Armonk, NY, USA). Continuous variables are presented as mean ± standard deviation as well as median and interquartile range (IQR), while categorical variables are expressed as frequencies and percentages. The distribution of continuous variables was assessed using the Shapiro–Wilk test, histograms, and Q–Q plots. For comparisons between two independent groups, the Mann–Whitney U test was used for non-normally distributed variables, and results are presented as medians with 25th–75th percentiles. The Independent Samples t-test was applied for comparisons of normally distributed continuous variables. For comparisons involving more than two groups, the Kruskal–Wallis test was used for non-normally distributed variables, while one-way analysis of variance (ANOVA) was applied for normally distributed variables. Comparisons of categorical variables were performed using Pearson’s chi-square test or Fisher’s exact test, as appropriate, for contingency tables (e.g., 2 × 2, 3 × 2). To identify independent predictors of the outcome “pathology involving two or more anatomical regions on imaging,” variables demonstrating an association with the outcome at a significance level of p < 0.250 in univariable analyses were considered for inclusion in the multivariable logistic regression model. Collinearity was assessed, and correlated variables were excluded. The final regression model was constructed using a backward elimination approach. For continuous variables that remained statistically significant in the multivariable analysis, receiver operating characteristic (ROC) curve analysis was performed to determine optimal cut-off values and corresponding diagnostic performance metrics, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). A p value of < 0.05 was considered statistically significant for all comparative analyses. Based on the results of these analyses, a clinical risk assessment score was developed. No internal validation techniques, such as bootstrapping or cross-validation, were performed. Results Participants Between June 1, 2021, and June 1, 2022, a total of 2,562 pediatric patients (aged 0–18 years) presented to the study center with trauma and were assessed for eligibility. Of these patients, 816 underwent WBCT based on the decision of the attending emergency medicine specialist at presentation. After exclusion of 409 patients who met the predefined exclusion criteria, a total of 407 patients were included in the final analysis. The flow of participants through the study is shown in Fig. 1. Flow diagram of the study Baseline characteristics The demographic characteristics of the patients are presented in Table 1. The mean age of the study population was 119 ± 60 months (9.9 ± 5 years). The youngest patient was 3 months old, while the oldest patient was 210 months (17.5 years) of age. When stratified by age groups, the distribution of patients was as follows: infants (0–1 year) comprised 1.7% (n = 7), toddlers (1–3 years) 6.6% (n = 27), preschool children (3–6 years) 19.4% (n = 79), school-aged children (6–12 years) 32.7% (n = 133), and adolescents (12–18 years) 39.6% (n = 161). Table 1Comparison of all variables between two groups in terms of number of regions of pathologic radiological signsVariablesOverallRadiological signs- two or more region pathologyNoYesp-valuen (%- column)Med (25–75%)n (%- row)Med (25–75%)n (%- row)Med (25–75%)GenderFemale147 (36.1)88 (59.9)59 (40.1)0.426Erkek260 (63.9)166 (63.8)94 (36.2)Age (month)121 (63–171)110 (63–160)138 (70–188)0.022MechanismMotor vehicle injury110 (27)66 (60)44 (40)0.003Pedestrian injury150 (36.9)103 (68.7)47 (31.3)Fall (over 1.5 m)90 (22.1)46 (51.1)44 (48.9)Motorcycle or ATV17 (4.2)7 (41.2)10 (58.8)Other*40 (9.8)32 (80)8 (20)Body weight30 (20–50)29 (19–48)40 (21–56)0.008Vital signsHeart rate89 (85–100)89 (84–98)98 (86–117)< 0.001Tachycardic55 (13.5)13 (23.6)42 (76.4)< 0.001SBP113 (110–120)115 (110–120)111 (100–120)0.005DBP76 (70–80)78 (72–82)72 (60–78)< 0.001Hypotensive15 (3.7)1 (6.7)14 (93.3)< 0.001sO297 (94–99)98 (96–99)94 (89–98)< 0.001Hypoxic109 (26.8)31 (28.4)78 (71.6)< 0.001Physical exam- pathology exists396 (97.3)243 (61.4)153 (38.6)0.008Physical exam- pathology exists inHead288 (70.8)165 (57.3)123 (42.7)0.001*Cervical spine23 (5.7)13 (56.5)10 (43.5)0.549Thoracic spine25 (6.1)12 (48)13 (52)0.125Lumbar spine49 (12)30 (61.2)19 (38.8)0.855Thorax103 (25.3)29 (28.2)74 (71.8)< 0.001Abdomen112 (27.5)60 (53.6)52 (46.4)0.023Pelvic59 (14.5)23 (39)36 (61)< 0.001Upper extremity108 (26.5)60 (55.6)48 (44.4)0.086Lower extremity150 (36.9)76 (50.7)74 (49.3)< 0.001LaboratoryHb12.9 (12–13.7)12.9 (12.3–13.9)12.8 (11.6–13.6)0.032Hct38.8 (36.5–41.6)38.8 (36.9–41.5)38.9 (35.5–41.6)0.313Lactate1.72 (1.35–2.43)1.6 (1.28–2.1)2.15 (1.45–2.94)< 0.001Urea28 (24–34)28 (24–33)30 (26–34)0.021Cre0.5 (0.4–0.6)0.5 (0.4–0.6)0.5 (0.4–0.7)< 0.001ALT27 (20–51)23 (19–36)42 (25–94)< 0.001AST43 (29–83)36 (26–56)65 (42–161)< 0.001Amylase66 (49–85)64 (49–81)69 (51–90)0.048Lipase28 (24–36)28 (24–33)30 (25–44)< 0.001Microscopic hematuria73 (17.9)24 (32.9)49 (67.1)< 0.001ED outcomeDischarge168 (41.3)158 (94)10 (6)< 0.001Ward admission101 (24.8)58 (57.4)43 (42.6)ICU admission138 (33.9)38 (27.5)100 (72.5)Pearson Chi-squared test was used for categorical variables (Fisher’s Exact test was used for only the “Head” variable) and Mann Whitney-U test was used for continues variablesED outcome: All three groups different from each otherMechanism: Difference was originated from the differences between “fall” and “other” groups and also “Motorcycle or ATV” and “other” groups*Other mechanisms were: Fall from bed/stroller/lap, fall from bicycle, assault, stabbing, playground injury, firearm injury, occupational injury Comparison of all variables between two groups in terms of number of regions of pathologic radiological signs Pearson Chi-squared test was used for categorical variables (Fisher’s Exact test was used for only the “Head” variable) and Mann Whitney-U test was used for continues variables ED outcome: All three groups different from each other Mechanism: Difference was originated from the differences between “fall” and “other” groups and also “Motorcycle or ATV” and “other” groups *Other mechanisms were: Fall from bed/stroller/lap, fall from bicycle, assault, stabbing, playground injury, firearm injury, occupational injury Mechanism of injury The distribution of injury mechanisms according to age groups is also shown in Table 1. Falls from height were the most common mechanism of injury among infants (71.4%, n = 5) and toddlers (63.0%, n = 17). In contrast, traffic-related injuries (ADTK) were the most frequently observed mechanism in preschool children (44.3%, n = 35), school-aged children (48.1%, n = 64), and adolescents (31.1%, n = 50). Physical examination findings and imaging results When the distribution of positive physical examination findings was evaluated according to anatomical regions (1. head and neck, 2. thorax, 3. abdomen, 4. pelvis, and 5. extremities), 2.7% of patients (n = 11) were found to have no positive physical examination findings. The proportion of patients with positive physical examination findings in one region was 30.5% (n = 124), while 39.6% (n = 150) had findings in two regions, 21.4% (n = 87) in three regions, 7.6% (n = 31) in four regions, and 1.0% (n = 4) in all five regions. Evaluation of WBCT scans revealed that 33.2% of patients (n = 135) of patients had no traumatic injury identified in any anatomical system. Injuries involving a single region were detected in 29.2% of patients (n = 119). Injuries involving multiple regions were observed in 37.6% of patients (n = 153), including involvement of two regions in 22.4% (n = 91), three regions in 9.3% (n = 38), four regions in 4.9% (n = 20), and all five regions in 1.0% (n = 4) of patients. Model development: logistic regression analysis Patients were categorized into two groups based on WBCT findings: those with injuries involving 0–1 anatomical region and those with injuries involving two or more regions. All variables were evaluated for their ability to predict the presence of injuries in two or more regions on WBCT. Comparisons of all parameters between these two groups are summarized in Table 1. Vital signs with age-dependent cut-off values, including heart rate, blood pressure, and oxygen saturation, were categorized according to age-specific thresholds, and group differences were assessed accordingly. Multivariable logistic regression analysis was then performed. Based on this analysis, age, the presence of tachycardia or hypoxia, positive physical examination findings in the head, thoracic, pelvic, or lower extremity regions, and AST levels were identified as independent predictors of injuries involving two or more anatomical regions. ROC curve analysis was subsequently conducted for the continuous variables age and AST to determine optimal cut-off values based on diagnostic performance metrics. Accordingly, age ≥ 133 months and AST ≥ 50 were identified as optimal cut-off values, and these variables were dichotomized using these thresholds. Variables demonstrating an association with the outcome at a significance level of p < 0.250 in univariable analyses (Table 1) were included in the multivariable logistic regression analysis, with results presented in Table 2. The variables entered into the model at the initial step are detailed in the footnote of Table 2. The cumulative ROC curve of the final regression model is shown in Fig. 2. Table 2Multiple logistic regression analysis for the number of regions of pathologic radiological signsVariables in the equationBSig.Odds95% CI for OddsLowerUpperAge > = 133 months (1)0.6660.0151.9471.1363.338Tachycardia (1)1.2040.0043.3351.4667.587Hypoxia (1)0.9780.0022.6591.4344.930Head (1)0.8960.0032.4511.3444.470Thorax (1)1.524< 0.0014.5902.4428.627Pelvic (1)1.2310.0013.4231.6297.193Lower extremity (1)0.962< 0.0012.6161.5324.465AST > = 50 U/L (1)0.977< 0.0012.6561.5474.558Constant0.9450.0012.572a Logit(p) = 0.945 + 0.666If age>=133 months + 1.204If tachycardia exists + 0.978If hypoxia exists + 0.896If head exam pathology exists + 1.524If thorax exam pathology exists + 1.231If pelvic exam pathology exists + 0.962If lower extremity exam pathology exists + 0.977If AST>= 50 U/L Multiple logistic regression analysis for the number of regions of pathologic radiological signs a Logit(p) = 0.945 + 0.666If age>=133 months + 1.204If tachycardia exists + 0.978If hypoxia exists + 0.896If head exam pathology exists + 1.524If thorax exam pathology exists + 1.231If pelvic exam pathology exists + 0.962If lower extremity exam pathology exists + 0.977If AST>= 50 U/L ROC curve for the final regression model Using these dichotomized variables, the model was reconstructed to include only categorical parameters (Table 2). As a result, older age (≥ 133 months), the presence of tachycardia or hypoxia, positive physical examination findings in the head, thoracic, pelvic, or lower extremity regions, and elevated AST levels (≥ 50) were identified as independent categorical predictors of the outcome “pathology involving two or more anatomical regions on imaging.” The cumulative ROC curve of this final model is shown in Fig. 2. The variables included in the initial step of the regression model are detailed in the footnote of Table 2. Based on the independent predictors identified in this model—age ≥ 133 months, presence of tachycardia, presence of hypoxia, positive physical examination findings in the head, thoracic, pelvic, or lower extremity regions, and AST ≥ 50 U/L—a scoring system (the WHOLE score) was developed. Considering the regression coefficients and prioritizing ease of clinical use, each predictor was assigned 1 point, resulting in a total score ranging from 0 to 8. Given the comparable magnitude of regression coefficients, all predictors were equally weighted to enhance clinical usability. The distribution of patients according to WHOLE score values, along with corresponding frequencies, percentages, and rates of the outcome “pathology involving two or more anatomical regions on imaging,” is presented in Table 3. Table 3Distribution of patients according to the WHOLE score pointsOverallRadiological signs- two or more region pathologyNoYesn%- columnn%-CountRow N %Score092.2%9100.0%00.0%18721.4%8294.3%55.7%211929.2%9378.2%2621.8%38320.4%5060.2%3339.8%44611.3%1226.1%3473.9%5338.1%618.2%2781.8%6174.2%15.9%1694.1%7122.9%18.3%1191.7%810.2%00.0%1100.0%Score grouped092.2%9100.0%00.0%18721.4%8294.3%55.7%>=231176.4%16352.4%14847.6%Score grouped 2092.2%9100.0%00.0%1–220650.6%17585.0%3115.0%>=319247.2%7036.5%12263.5% Distribution of patients according to the WHOLE score points Score group cut-offs for the WHOLE score were defined as post hoc, informed by the distribution of patients and outcome frequencies across score levels. ROC analysis of the WHOLE score was performed to predict the outcome “pathology involving two or more anatomical regions on imaging.” The resulting ROC curve and diagnostic performance metrics at selected cut-off values are presented in Fig. 3; Table 4. ROC curve for the WHOLE score Table 4Diagnostic accuracy statistics for WHOLE score cut points in terms of the need for whole body CT scanScale: WHOLE scoreCut pointSens (%)Spec (%)(+) LR(-) LRPPV (%)NPV (%)Accuracy>= 1100,0 (97.6–100.0)3,5 (1.6–6.6)1.04 (1.01–1.06)0.038,4 (37.9–39.0)100,0 (66.4–100.0)39.8 (35.0-44.7)>= 296,7 (92.5–98.9)35,8 (29.9–42.1)1.51 (1.37–1.66)0.09 (0.04–0.22)47,6 (45.2–50.0)94,8 (88.3–97.8)58.7 (53.8–63.6)>= 379,7 (72.5–85.8)72,4 (66.5–77.8)2.89 (2.33–3.59)0.28 (0.20–0.39)63,5 (58.4–68.4)85,6 (81.1–89.1)75.2 (70.7–79.3)AUC = 0.841 (0.802–0.881); p < 0.001 Diagnostic accuracy statistics for WHOLE score cut points in terms of the need for whole body CT scan AUC = 0.841 (0.802–0.881); p < 0.001 Discussion Limitations This study has several limitations. First, it is a single-center and retrospective study, and patients with incomplete medical records were excluded from the analysis. The high exclusion rate due to missing data may have introduced selection bias and influenced model estimates. However, given the ethical challenges inherent in conducting randomized controlled trials comparing WBCT and SCT in blunt abdominal trauma, the present study provides a reasonable methodological approach for data collection in this context. The study population consisted exclusively of patients who had already undergone WBCT; therefore, this retrospective cohort does not represent pediatric patients presenting to the emergency department with minor trauma. Since only patients who already underwent WBCT were included, the score is derived from a pre-selected high-risk population, which limits its clinical applicability to the broader pediatric trauma population. Importantly, this design implies that the WHOLE score is intended for use within the population of patients already being considered for WBCT by the treating physician — it is not designed to risk-stratify undifferentiated pediatric trauma presentations. Users of the score should be aware of this scope, and the findings should not be extrapolated to lower-acuity patients who would not ordinarily be candidates for WBCT. In addition, because patients who underwent SCT were not included, a direct comparison between WBCT and SCT could not be performed. Furthermore, outcome data for patients managed with selective CT or clinical observation were not available within the study dataset, precluding comparison of clinical outcomes between different imaging strategies. The study was not designed to compare the WHOLE score against selective imaging strategies or existing guideline-based approaches. Due to the inclusion criteria, patients who died before any imaging could be performed upon emergency department arrival, as well as those taken directly to surgery based on bedside ultrasonography findings without undergoing CT, were excluded from the study. Caution is therefore warranted when interpreting the results due to the absence of an independent external validation sample. The findings of this study require validation in an independent cohort. A potential limitation of a SCT strategy guided by a prediction model that incorporates laboratory parameters is the possibility of delayed imaging while awaiting laboratory results [17]. In hemodynamically stable patients who do not exhibit “high-risk” clinical variables according to our model and who have a reliable physical examination, deferring CT imaging until screening laboratory results are available may be considered appropriate [17]. In the present study, no distinction was made regarding the clinical significance of CT-detected findings, such as the need for surgical intervention or hospital admission. The study did not distinguish between clinically important injuries and less severe trauma-related findings detected on CT. Outcome measure is injury distribution in WBCT, rather than clinical need for CT. Accordingly, the outcome should not be interpreted as a direct indicator of true need for WBCT. All trauma-related findings identified on WBCT were classified as positive, with the exception of incidental findings unrelated to trauma. During the study period, imaging decisions were made at the discretion of the attending physicians. In addition, the WHOLE score assigns equal weight to all predictor variables, including physical examination findings, regardless of their severity. As a result, both minor and more severe findings contribute equally to the total score, which may lead to higher scores in patients with relatively low clinical risk. This is reflected in the distribution of scores and the proportion of patients discharged from the emergency department despite undergoing WBCT in our study. Future studies may consider incorporating weighted variables or severity-based scoring approaches to improve clinical discrimination. And lastly, calibration of the model was not assessed; therefore, agreement between predicted probabilities and observed outcomes remains uncertain. Interpretation Based on the diagnostic performance statistics of the WHOLE score, patients with a score of 0–1 were found to be at low risk for multi-region injury and may represent candidates in whom WBCT can be safely withheld, pending clinical judgment. Lower WHOLE scores were associated with a lower probability of multi-region injury, whereas higher scores were associated with increased risk. However, threshold-based clinical use of the WHOLE score should be considered preliminary and requires prospective external validation before implementation in decision-making. The score should be understood as a risk-stratification aid, not as a standalone decision rule; a high WHOLE score does not independently justify performing WBCT, and the final imaging decision must always incorporate the clinician’s overall assessment of the patient. It should also be noted that the real-world impact of implementing the WHOLE score prospectively remains uncertain and may influence imaging utilization in different ways depending on clinical context. According to the results of our study, when a cut-off value of 2 points was applied in the WHOLE scoring system, the sensitivity was 96.7% (CI 92.5–98.9) and the NPV was 94.8% (CI 88.3–97.8) suggesting that the WHOLE score may serve as a reliable screening tool to support emergency physicians in safely excluding the need for WBCT in low-risk pediatric trauma patients. It is critical to emphasize that the WHOLE score is designed as a rule-out tool, not a rule-in tool. A low score (0–1 points) may provide supportive evidence for withholding WBCT in patients otherwise considered for whole-body imaging; however, a high score should not be interpreted as an independent mandate for WBCT. Clinical judgment, hemodynamic status, mechanism of injury, and institutional protocols must always guide the final imaging decision. Clinicians should be aware that applying the WHOLE score outside its intended context — particularly in settings where the baseline threshold for WBCT is already low — could paradoxically contribute to increased CT utilization. The score is intended to reduce, not increase, radiation exposure in pediatric trauma patients. The age and sex distribution of patients in our cohort was comparable to that reported in other studies focusing on pediatric trauma populations [18–20]. In addition, the age-specific distribution of the most common mechanisms of injury observed in our study was consistent with previously published literature [21] supporting the external validity of our cohort and suggesting that the findings are generalizable to typical pediatric trauma populations encountered in emergency departments. A substantial number of pediatric patients continue to undergo WBCT, largely driven by clinicians’ concern about missing potential trauma-related sequelae, even though only a small proportion of these patients ultimately require surgical intervention [22]. In previous studies comparing SCT with WBCT, mortality or the detection of injuries requiring surgical intervention has most commonly been used as the primary outcome. In contrast, in the present study, we selected the identification of injuries involving more than one anatomical region on WBCT as the primary outcome. This approach was chosen because, in the context of increasing concerns about medical malpractice, many clinicians aim to identify all traumatic injuries in trauma patients, including those that are not immediately life-threatening or do not require operative management. While this approach captures the overall burden of injury, it does not distinguish between clinically important injuries requiring intervention and minor findings that may not alter patient management. In a large population-based study involving approximately 680,000 individuals who underwent CT imaging, an increased risk was demonstrated for several solid tumors—including those of the gastrointestinal organs, melanoma, soft tissue, female genital system, urinary system, brain, and thyroid—as well as for leukemia, myelodysplastic syndromes, and certain other lymphoid malignancies [4]. This increased risk has been shown to be dependent on radiation dose and the cumulative number of CT examinations, with the youngest patients representing the highest risk group. Consequently, routine WBCT should be avoided in pediatric trauma patients [5]. Reducing the rate of unnecessary CT examinations is essential to minimize radiation exposure among children [23]. These findings reinforce that pediatric CT exposure is associated with increased lifetime malignancy risk signals, particularly in younger patients and with cumulative exposure. Therefore, imaging decisions in pediatric trauma must carefully balance the risk of missed injuries against the potential long-term harms of radiation exposure. Timely and accurate identification of injuries is critical for the effective management of life-threatening conditions in trauma patients [24]. Several studies have reported that the use of emergent WBCT, particularly in patients with severe trauma, may improve survival by detecting occult injuries that might otherwise be missed with SCT imaging [25]. Consequently, the utilization of WBCT has increased substantially and has become the primary imaging modality in many centers for patients presenting with blunt trauma [26]. However, in a study conducted by Traynor et al., the high proportion of normal findings and the low rates of subsequent interventions following head, cervical spine, and abdominal CT examinations were interpreted as potential indicators of overutilization of this imaging modality [23]. In the present study, only patients who underwent WBCT were included, and because the decision to perform WBCT was made by the treating clinician, this cohort can be assumed to represent a higher-risk trauma population compared with the general trauma population. Despite this, no pathological findings were identified in any anatomical region in 33.2% (n = 135) of WBCT examinations performed in this group. In addition to computed tomography-based strategies, bedside ultrasonography—particularly focused assessment with sonography for trauma (FAST)—plays an important role in the evaluation of pediatric trauma patients [27, 28]. In hemodynamically stable children, a stepwise imaging approach incorporating clinical assessment and ultrasonography may allow for safe management without immediate CT in selected cases [16, 27]. Several studies have demonstrated that a positive FAST examination in stable patients can often be managed non-operatively, thereby reducing unnecessary radiation exposure [28]. Incorporating such stepwise approaches into clinical decision-making may further decrease reliance on WBCT in pediatric trauma care. Considerable variability exists among trauma centers in the utilization of WBCT for pediatric trauma patients, despite the absence of differences in patient outcomes [3]. Children treated in adult trauma centers are more likely to undergo WBCT, thereby increasing their exposure to ionizing radiation without a corresponding improvement in clinical outcomes [3]. Although multiple studies have demonstrated reduced mortality among injured adults undergoing WBCT, it remains uncertain whether pediatric patients derive a similar benefit [4]. In some studies comparing SCT and WBCT, WBCT has been associated with lower mortality rates and improved outcomes [11]; however, other studies have failed to demonstrate a significant advantage of WBCT over SCT [29]. Although several studies have developed algorithms to guide the use of abdominal CT in blunt abdominal trauma, we did not identify any scoring system in the existing literature specifically designed to guide WBCT utilization in pediatric patients [17, 30]. Consequently, the establishment of standardized guidelines aimed at minimizing unnecessary WBCT utilization across trauma centers is warranted [3]. Mechanism of injury alone cannot be considered a sufficient justification for the routine use of WBCT in pediatric patients with blunt trauma. Similarly, the use of CT imaging in injured children who do not exhibit physiological instability or identifiable anatomical abnormalities appears unwarranted [5, 31]. To date, no single clinical variable has been identified that can reliably serve as the sole indication for WBCT in pediatric polytrauma patients [32]. In their study examining vital signs in trauma patients, Chalari et al. reported higher injury rates among patients presenting with abnormal vital signs, including hypotension, tachycardia, and hypoxia [33]. Consistent with these findings, our study demonstrated that patients who had abnormal vital signs at presentation—specifically hypotension, tachycardia, or hypoxia—were more likely to have injuries involving multiple anatomical regions on WBCT. While higher transaminase thresholds have been proposed in previous studies, particularly in the context of isolated liver injury, the threshold used in our study (AST ≥ 50) was determined based on ROC analysis for predicting multi-region injury. We did not perform additional analyses using alternative cut-off values, which may influence the balance between sensitivity and specificity and could be explored in future studies [34]. The present study should be interpreted in the context of existing pediatric imaging strategies (e.g. AAP, PECARN, NICE) which generally advocate selective and region-specific CT use rather than routine WBCT. Our findings support a multivariable, risk-based decision making approach by demonstrating that a combination of readily available clinical, physiological, and laboratory parameters—rather than any single factor—can aid emergency physicians in identifying pediatric trauma patients who are at higher risk for multi-region injuries and are more likely to benefit from WBCT. Conclusions The WHOLE score offers a practical clinical decision support tool for emergency physicians managing pediatric trauma patients in the acute care setting. Using readily available information at initial assessment, including vital signs, physical examination findings, and basic laboratory parameters, the WHOLE score represents a potential clinical decision support tool that may assist emergency physicians in risk stratifying pediatric trauma patients for WBCT utilization. However, its real-world impact on imaging utilization remains uncertain and will depend substantially on the clinical context in which the score is applied. In settings where WBCT is already used liberally, the score has the greatest potential to reduce unnecessary imaging; in settings where CT utilization is already appropriately conservative, its incremental benefit may be limited. Clinicians and institutions should carefully consider local practice patterns when evaluating the applicability of the score. In the emergency department, where time-sensitive decisions must balance diagnostic accuracy with radiation exposure, the WHOLE score has the potential to support more consistent, risk-based imaging strategies. External and prospective validation is required before routine implementation in emergency care practice. Acknowledgements Not applicable. Abbreviations Whole-body computed tomography Receiver operating characteristics Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis Selective computed tomography Computed tomography Negative predictive value Positive predictive value Author contributions Conceptualization: [GPG], [EÇ], [AŞ]; Methodology: [GPG], [ÇY], [EÇ]; Data acquisition: [GPG], [ÇY], [AŞ]; Formal analysis and investigation: [GPG], [EÇ], [AŞ], [ÇY]; Writing - original draft preparation: [GPG], [ÇY], [EÇ]; Writing - review and editing: [GPG], [ÇY], [AŞ], [EÇ]; Statistical Expertise: [GPG], [AŞ]; Supervision: [ÇY], [GPG] Funding None. Data availability The datasets generated and/or analysed during the current study are available in the Elsevier’s Mendeley Data repository, 10.17632/x4kh298ndx.1. Declarations Ethics approval and consent to participate The study protocol was approved by the Ankara Bilkent City Hospital Clinical Research Ethics Committee No. 2 (approval date: 23.11.2022; decision number: E2-22-2872), and the study was conducted in accordance with the Declaration of Helsinki. As this was a retrospective chart review study involving analysis of existing medical records without any intervention or direct patient contact, the requirement for individual informed consent to participate was waived by the ethics committee. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated and/or analysed during the current study are available in the Elsevier’s Mendeley Data repository, 10.17632/x4kh298ndx.1.