Nonaccidental trauma in pediatric patients: evidence-based screening criteria for ophthalmologic examination

Ip, Samantha S.; Zafar, Sidra; Liu, T. Y. Alvin; Srikumaran, Divya; Repka, Michael X.; Goldstein, Mitchell; Woreta, Fasika A.

doi:10.1016/j.jaapos.2020.03.012

Cited by 4 publications

(5 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Subdural hematoma is the most frequent intracranial hemorrhage in AHT and has been highly correlated with RH in prior studies. 4,5,[24][25][26] Several authors Given that the results of 23% to 84% of retinal examinations conducted for abuse evaluation are positive for hemorrhages 3,5 and that retinal examinations should be performed within 48 hours of presentation before hemorrhages begin to resolve, 31 this technology could also offer timely decision support about who urgently needs a direct fundoscopic examination. In the near term, deep learning-assisted RH detection would not replace the fundoscopic examination, which remains the criterion standard of evidence for medical and legal evaluation.…”

Section: Discussionmentioning

confidence: 99%

“…When risk prediction from our deep learning model was added as a feature in our combined light GBM model, the presence of subdural hematoma contributed minimally. Subdural hematoma is the most frequent intracranial hemorrhage in AHT and has been highly correlated with RH in prior studies . Several authors have made an evidence-based argument for forgoing dilated retinal examinations in children without intracranial hemorrhage, even if physical abuse is suspected, which raises the question of why deep learning should be applied to this problem.…”

Section: Discussionmentioning

confidence: 99%

“…Subdural hematoma is the most frequent intracranial hemorrhage in AHT and has been highly correlated with RH in prior studies . Several authors have made an evidence-based argument for forgoing dilated retinal examinations in children without intracranial hemorrhage, even if physical abuse is suspected, which raises the question of why deep learning should be applied to this problem. One answer is statistical: our deep learning model in its current form would have fewer false-positive results compared with using intracranial hemorrhage for triage (positive predictive value of 68.6% vs 53.9%) as well as fewer false-negative results (negative predictive value of 87.1% vs 82.6%) (eTable 2 in Supplement 1).…”

Section: Discussionmentioning

confidence: 99%

“…This subspecialty is not readily available in many communities. Furthermore, the examination is uncomfortable and can require sedation, and it temporarily nullifies pupillary response as an indication of neurologic status . Thus, although head CTs are obtained routinely, dilated fundoscopic examinations are reserved for those patients with the highest likelihood of abuse.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Development of a Deep Learning Model for Retinal Hemorrhage Detection on Head Computed Tomography in Young Children

Gunturkun,

Bakir-Batu,

Siddiqui

et al. 2023

JAMA Netw Open

View full text Add to dashboard Cite

ImportanceAbusive head trauma (AHT) in children is often missed in medical encounters, and retinal hemorrhage (RH) is considered strong evidence for AHT. Although head computed tomography (CT) is obtained routinely, all but exceptionally large RHs are undetectable on CT images in children.ObjectiveTo examine whether deep learning–based image analysis can detect RH on pediatric head CT.Design, Setting, and ParticipantsThis diagnostic study included 301 patients diagnosed with AHT who underwent head CT and dilated fundoscopic examinations at a quaternary care children’s hospital. The study assessed a deep learning model using axial slices from 218 segmented globes with RH and 384 globes without RH between May 1, 2007, and March 31, 2021. Two additional light gradient boosting machine (GBM) models were assessed: one that used demographic characteristics and common brain findings in AHT and another that combined the deep learning model’s risk prediction plus the same demographic characteristics and brain findings.Main Outcomes and MeasuresSensitivity (recall), specificity, precision, accuracy, F1 score, and area under the curve (AUC) for each model predicting the presence or absence of RH in globes were assessed. Globe regions that influenced the deep learning model predictions were visualized in saliency maps. The contributions of demographic and standard CT features were assessed by Shapley additive explanation.ResultsThe final study population included 301 patients (187 [62.1%] male; median [range] age, 4.6 [0.1-35.8] months). A total of 120 patients (39.9%) had RH on fundoscopic examinations. The deep learning model performed as follows: sensitivity, 79.6%; specificity, 79.2%; positive predictive value (precision), 68.6%; negative predictive value, 87.1%; accuracy, 79.3%; F1 score, 73.7%; and AUC, 0.83 (95% CI, 0.75-0.91). The AUCs were 0.80 (95% CI, 0.69-0.91) for the general light GBM model and 0.86 (95% CI, 0.79-0.93) for the combined light GBM model. Sensitivities of all models were similar, whereas the specificities of the deep learning and combined light GBM models were higher than those of the light GBM model.Conclusions and RelevanceThe findings of this diagnostic study indicate that a deep learning–based image analysis of globes on pediatric head CTs can predict the presence of RH. After prospective external validation, a deep learning model incorporated into CT image analysis software could calibrate clinical suspicion for AHT and provide decision support for which patients urgently need fundoscopic examinations.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Development of a Deep Learning Model for Retinal Hemorrhage Detection on Head Computed Tomography in Young Children

Gunturkun,

Bakir-Batu,

Siddiqui

et al. 2023

JAMA Netw Open

View full text Add to dashboard Cite

show abstract

“…Specific indications for ophthalmologic examination to look for retinal hemorrhages are included in Figure 6 (box G). [28][29][30][31][32] The provider is also given the reminder to inform the state's child protection services (CPS) agency if they have not already done so.…”

Section: Disposition and Treatmentmentioning

confidence: 99%