Nonaccidental Trauma and Child Abuse

Value Proposition

Nonaccidental trauma most commonly includes head trauma, spine trauma, and skeletal trauma but can also include visceral injuries of the abdomen. These injuries occur typically because of severe angular acceleration/deceleration, compression forces, and blunt impact forces if there is contact with a solid object. The forces needed to cause these injuries are very high and abnormal for routine accidental childhood injuries.

More child abuse deaths occur from head injuries than other types of injuries, and those that survive go on to have serious neurological sequela..3 Non-accidental head trauma is defined as an injury to the skull or intracranial contents of an infant or young child (< 5 years old) due to inflicted blunt impact and/or violent shaking.1

Detecting non-accidental head trauma is often difficult due to nonspecific symptoms and especially difficult in non-verbal patients whose care-takers do not provide a clear injury history. It has been reported that 31% of non-accidental head trauma is not recognized at initial presentation, 28% are reinjured prior to accurate diagnosis, 19% were seen by a medical provider within a month prior to death, 13% misinterpretation caused delayed diagnosis, and 4 in 5 deaths in unrecognized non-accidental head trauma group may have been prevented.

An algorithm meeting this use case will make it easier to detect non-accidental head trauma considering the factors differentiating it from accidental trauma and other abnormalities which may look similar to a radiologist and emergency room clinician not trained in pediatric subspecialties. In addition, combining clinical data with imaging findings may also flag cases with nonspecific clinical findings and potentially help with earlier identification and avoiding misinterpretation. These automated measurements will also improve consistency between operators and provide the radiologist and clinicians a tool for timely and reliable management and optimize resource allocation.

Narrative(s)

A caregiver and patient present after an injury or with nonspecific neurological symptoms in a community hospital setting. The algorithm evaluates Brain/Spine MRI, provides a probability of non-accidental trauma, accidental trauma, or abnormalities related to other conditions, or “unknown”. The radiologist is informed to review/confirm the results of this information at the time of interpretation to aid in final clinical management.

Workflow Description

An image(s) is obtained from the MRI scanner are sent to PACS/Viewer and the AI engine. The image is analyzed by the engine. The system segments the region of concern and analyzes in comparison with a model trained from known cases of each class. A message is sent to PACS from the engine with the classification information. If the location of the lesion is also identified, location information can also be sent to PACS to highlight the region the engine identified.

Consideration for future work would be to provide a combined probability score derived from the assessment of multiple modalities and clinical data. Therefore, this system may perform best if x-ray, CT, and MR images are sent to individual AI Models designed for analyzing each modality to add more variables to improve accuracy of diagnosis. In addition, extracting EMR data containing social/family history, demographic information, and past medical/surgical history as well the pertinent information on physical exam may also be incorporated into the assessment to strengthen the model. Other novel techniques currently in the pipeline including “Black Bone” MRI may also provide some utility to speed up diagnoses and reduce exposure to radiation (not require additional Head CT to confirm fracture).

Inputs

 DICOM Study

Primary Outputs

 Parenchymal Injury/Lesion

Collections

Spine Injury

Ocular Injury

Skull Fracture

Secondary Outputs

Nonaccidental Trauma Probability Score

1. Parks, SE, et al. Pediatric Abusive Head Trauma: Recommended Definitions for Public Health Surveillance and Research. Atlanta (GA): Centers for disease Control and Prevention; 2012

2. Paul K. Kleinman. Diagnostic Imaging of Child Abuse. 3rd Edition. Cambridge Press.

3. Jenny C, Hymel KP, Ritzen A, Reinert SE, Hay TC. Analysis of Missed Cases of Abusive Head Trauma. JAMA. 1999;281(7):621–626. doi:10.1001/jama.281.7.621

4. Kralik SF, Supakul N, Wu IC, et al. Black bone MRI with 3D reconstruction for the detection of skull fractures in children with suspected abusive head trauma. Neuroradiology. 2019;61(1):81‐87. doi:10.1007/s00234-018-2127-9

No known related public datasets at this time,  please alert us if you know of any.