Search Results

The EyeBOX is intended to measure and analyze eye movements as an aid in the diagnosis of concussion, also known as mild traumatic brain injury (mTBI), within one week of head injury in patients 5 through 67 years of age in conjunction with a standard neurological assessment of concussion.

A negative EyeBOX classification may correspond to eye movement that is consistent with a lack of concussion.

A positive EyeBOX classification corresponds to eye movement that may be present in both patients with or without concussion.

Oculogica's EyeBOX is an eye-tracking device with custom software. The device is comprised of a host PC with integrated touchscreen interface for the operator, eye tracking camera, LCD stimulus screen and head-stabilizing rest (chin rest and forehead rest) for the patient, and data processing algorithm. The data processing algorithm detects subtle changes in eye movements resulting from concussion. The eye tracking task takes about 4 minutes to complete and involves watching a video move around the perimeter of an LCD monitor positioned in front of the patient while a high speed near-infrared (IR) camera records gaze positions 500 times per second. The post-processed data are analyzed automatically to produce one or more outcome measures.

The provided text describes the Oculogica EyeBOX, Model OCL 02, a device intended to aid in the diagnosis of concussion (mild traumatic brain injury, mTBI). Since this is a 510(k) submission for a modified device (OCL 02) that is deemed substantially equivalent to a previously cleared device (OCL 01), the document focuses on demonstrating that the changes do not raise new questions of safety or effectiveness. Therefore, detailed information about the original clinical study that proved the device met its acceptance criteria is not explicitly repeated in this 510(k) summary; rather, it refers back to the data from the predicate device (DEN170091).

Based on the information provided, here's what can be extracted and inferred:

1. A table of acceptance criteria and the reported device performance

The provided 510(k) summary for OCL 02 does not explicitly state acceptance criteria or device performance metrics for this specific submission because it's a modification focusing on substantial equivalence. It asserts that the EyeBOX algorithm which processes the eye tracking data and outputs the BOX score is not changed. This implies that the performance established for the predicate device (OCL 01 under DEN170091) is maintained.

To fully answer this, one would typically need access to the original DEN170091 submission. However, we can infer the type of performance metrics that would have been evaluated: sensitivity and specificity for concussion diagnosis. The "Indications for Use" statement gives a clue:

• "A negative EyeBOX classification may correspond to eye movement that is consistent with a lack of concussion." (Implies high negative predictive value/sensitivity for ruling out concussion)
• "A positive EyeBOX classification corresponds to eye movement that may be present in both patients with or without concussion." (Implies that a positive result needs to be interpreted in conjunction with other clinical assessments and not as a definitive diagnosis, suggesting emphasis might be on sensitivity rather than high specificity as a standalone diagnostic.)

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

The 510(k) for OCL 02 does not provide details about the test set sample size or data provenance, as it refers to the predicate device's data. For the predicate device (OCL 01, DEN170091), a clinical study would have been conducted. This information is not present in the provided text.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)

This information is also not present in the provided text for the OCL 02 submission, as it relies on the predicate device's data, which is not detailed here. For concussion diagnosis, the ground truth would typically be established by a consensus of neurologists, sports medicine physicians, or other specialists experienced in concussion diagnosis, likely based on a combination of clinical assessment (e.g., SCAT5, BESS, neurocognitive testing), imaging (if applicable for other purposes), and follow-up.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

This information is not present in the provided text. Adjudication methods would have been part of the clinical study design for the predicate device to establish the ground truth for concussion diagnosis.

5. If a multi reader multi case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance

The EyeBOX is described as an "aid in the diagnosis," not a standalone diagnostic that replaces human assessment. The text states it is used "in conjunction with a standard neurological assessment of concussion." This implies that it is meant to assist clinicians. However, the provided document does not contain information regarding an MRMC comparative effectiveness study or the effect size of human reader improvement with AI assistance. Such a study might have been part of the predicate device's clinical evidence, but it's not detailed here.

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

The device provides a "BOX score" and a "classification" (positive or negative). The statement "A negative EyeBOX classification may correspond to eye movement that is consistent with a lack of concussion. A positive EyeBOX classification corresponds to eye movement that may be present in both patients with or without concussion." implies the algorithm generates a classification on its own. However, the overriding indication for use is "as an aid in the diagnosis... in conjunction with a standard neurological assessment." This strongly suggests that a standalone, algorithm-only diagnosis is not the intended use or claim, and therefore, a standalone performance study as a definitive diagnosis without human-in-the-loop for the final diagnostic decision would be inconsistent with the stated indication. The algorithm produces a result, but that result is an aid to a human clinician making the final diagnosis.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)

The type of ground truth used for the predicate EyeBOX (OCL 01) would almost certainly have been expert clinical consensus of concussion diagnosis, based on standard neurological assessment, symptom questionnaires, and potentially neurocognitive testing. Concussion diagnosis is primarily clinical, so pathology is not a typical ground truth for this condition, and while outcomes data is important, the initial diagnostic ground truth typically relies on expert assessment at the time of diagnosis. This is inferred, as it is not explicitly stated in the provided document.

8. The sample size for the training set

The provided 510(k) summary for OCL 02 focuses on the current device and its substantial equivalence to its predicate. It explicitly states that "The EyeBOX algorithm which processes the eye tracking data and outputs the BOX score is not changed." This suggests the algorithm was developed and trained prior to the OCL 01 submission. The sample size for the training set for the original algorithm development is not provided in this document.

9. How the ground truth for the training set was established

Similar to the test set, the ground truth for the training set would have been established through expert clinical consensus based on comprehensive neurological assessments. This information is not provided in this document but is inferred based on standard practices for clinical AI/ML device development in this field.

Ask a specific question about this device