Overjet Caries Assist (OCA) is a radiological, automated, concurrent-read, computer-assisted detection (CADe) software intended to aid in the detection and segmentation of caries on bitewing and periapical radiographs. The device provides additional information for the dentist to use in their diagnosis of a tooth surface suspected of being carious. The device is not intended as a replacement for a complete dentist's review or their clinical judgment that takes into account other relevant information from the image, patient history, or actual in vivo clinical assessment.

Device Description

OCA is a software-only device which operates in three layers: a Network Layer, a Presentation Layer, and a Decision Layer. Images are pulled in from a clinic/dental office, and the Machine Learning model creates predictions in the Decision Layer and results are pushed to the dashboard, which are in the Presentation Layer.

The machine learning system with the Decision Layer processes bitewing and periapical radiographs and annotates suspected carious lesions. It is comprised of four modules:

Image Preprocessor Module
Tooth Number Assignment Module
Caries Module
Post Processing

AI/ML Overview

This document describes the Overjet Caries Assist (OCA) device, a computer-assisted detection (CADe) software intended to aid dentists in the detection and segmentation of caries on bitewing and periapical radiographs.

Here's an analysis of the acceptance criteria and the study that proves the device meets them:

1. Table of Acceptance Criteria and Reported Device Performance:

The document doesn't explicitly state numerical acceptance criteria for sensitivity or specificity in a "table" format as initial goals. However, it does state a performance objective for the clinical reader improvement study: "Increase in dentist's sensitivity of greater than 15%". The other metrics are presented as reported performance from standalone and clinical evaluation studies.

Metric	Acceptance Criteria (if stated)	Reported Device Performance	Comments
Standalone Performance			Bitewing Images (n=1,293)
Overall Sensitivity	Not explicitly stated	76.6% (73.8%, 79.4%)	Based on surfaces (27,920)
Primary Caries Sensitivity	Not explicitly stated	79.9% (77.1%, 82.7%)
Secondary Caries Sensitivity	Not explicitly stated	60.9% (53.5%, 68.2%)
Enamel Caries Sensitivity	Not explicitly stated	74.4% (70.4%, 78.3%)
Dentin Caries Sensitivity	Not explicitly stated	79.5% (75.8%, 83.2%)
Overall Specificity	Not explicitly stated	99.1% (98.9%, 99.2%)
Primary Caries Dice Score	Not explicitly stated	0.77 (0.76, 0.78)	Pixel-level metric for true positives
Secondary Caries Dice Score	Not explicitly stated	0.73 (0.70, 0.75)	Pixel-level metric for true positives
Enamel Caries Dice Score	Not explicitly stated	0.76 (0.75, 0.77)	Pixel-level metric for true positives
Dentin Caries Dice Score	Not explicitly stated	0.77 (0.76, 0.79)	Pixel-level metric for true positives
			Periapical Images (n=1,314)
Overall Sensitivity	Not explicitly stated	79.4% (76.1%, 82.8%)	Based on surfaces (16,254)
Primary Caries Sensitivity	Not explicitly stated	79.8% (76.0%, 83.7%)
Secondary Caries Sensitivity	Not explicitly stated	77.9% (71.4%, 84.5%)
Enamel Caries Sensitivity	Not explicitly stated	67.9% (60.7%, 75.1%)
Dentin Caries Sensitivity	Not explicitly stated	84.9% (81.3%, 88.4%)
Overall Specificity	Not explicitly stated	99.4% (99.2%, 99.5%)
Primary Caries Dice Score	Not explicitly stated	0.79 (0.78, 0.81)	Pixel-level metric for true positives
Secondary Caries Dice Score	Not explicitly stated	0.79 (0.77, 0.82)	Pixel-level metric for true positives
Enamel Caries Dice Score	Not explicitly stated	0.75 (0.73, 0.77)	Pixel-level metric for true positives
Dentin Caries Dice Score	Not explicitly stated	0.81 (0.80, 0.82)	Pixel-level metric for true positives
Clinical Evaluation (Reader Improvement)			Bitewing Images (n=330)
Increase in reader sensitivity (overall)	> 15%	78.5% (assisted) vs. 64.6% (unassisted)	Increase = 13.9%. This falls slightly below the stated >15% criterion, though the document concludes it demonstrates a "clear benefit". The text states "overall reader sensitivity improved from 64.6% (56.4%, 72.1%) to 78.5% (72.6%, 83.6%) unassisted vs assisted", if calculated as a direct percentage difference (78.5-64.6 = 13.9), it is slightly below 15%. If interpreted as (assisted/unassisted)-1 * 100 ((78.5/64.6)-1)*100 = 21.5%, then it meets the criterion. The framing in the document implies the latter.
Overall reader specificity (decrease)	Not explicitly stated (implied minimal decrease is acceptable)	98.6% (assisted) vs. 99.0% (unassisted)	Decrease of 0.4%
Overall wAFROC AUC (increase)	Not explicitly stated	0.785 (assisted) vs. 0.729 (unassisted)	Increase of 0.055, statistically significant (p<0.001)
Reader Dice Score (overall)	Not explicitly stated	0.72 (assisted) vs. 0.66 (unassisted)	Average across caries types. Mean increased from 0.67 to 0.76 (Primary), 0.65 to 0.67 (Secondary), 0.65 to 0.74 (Enamel), 0.67 to 0.74 (Dentin).
			Periapical Images (n=330)
Increase in reader sensitivity (overall)	> 15%	79.0% (assisted) vs. 65.6% (unassisted)	Increase = 13.4%. Similar to bitewing, falls slightly below the stated >15% criterion as a direct percentage difference. If interpreted as ((79.0/65.6)-1)*100 = 20.4%, then it meets the criterion.
Overall reader specificity (decrease)	Not explicitly stated (implied minimal decrease is acceptable)	97.6% (assisted) vs. 98.0% (unassisted)	Decrease of 0.4%
Overall wAFROC AUC (increase)	Not explicitly stated	0.848 (assisted) vs. 0.799 (unassisted)	Increase of 0.050, statistically significant (p<0.001)
Reader Dice Score (overall)	Not explicitly stated	0.77 (assisted) vs. 0.71 (unassisted)	Average across caries types. Mean increased from 0.73 to 0.80 (Primary), 0.69 to 0.74 (Secondary), 0.64 to 0.73 (Enamel), 0.76 to 0.81 (Dentin).

2. Sample Size for Test Set and Data Provenance:

Standalone Testing:
- Sample Size: 1,293 Bitewing images (27,920 surfaces) and 1,314 Periapical images (16,254 surfaces).
- Data Provenance: Not explicitly stated regarding country of origin. The images were from "the following sensor manufacturers: Carestream, Dexis, e2v, Gendex, Hamamatsu, Jazz Imaging, ScanX, Schick, Soredex Digora." The context of US licensed dentists for ground truth suggests the data is likely from the US or a similar dental practice environment. The document specifies "Digital files of bitewing and periapical radiographs whose longer edge is greater than 500 pixel resolution."
- Retrospective/Prospective: Not explicitly stated, but the description of collecting and annotating existing images suggests it was a retrospective study.
Clinical Evaluation (Reader Improvement Study):
- Sample Size: 330 bitewing images (94 containing caries / 236 without caries) and 330 periapical images (94 containing caries / 236 without caries).
- Data Provenance: Not explicitly stated regarding country of origin, but the "US licensed dentists" implies the study was conducted within the US.
- Retrospective/Prospective: Not explicitly stated, but similar to the standalone testing, the use of a predefined image set for readers suggests retrospective.

3. Number of Experts and Qualifications for Ground Truth:

Standalone Testing: Ground truth established by consensus of three US licensed dentists. Non-consensus labels were adjudicated by an oral radiologist. No specific years of experience are listed, but "US licensed dentists" and "oral radiologist" imply professional qualifications relevant to dental imaging interpretation.
Clinical Evaluation (Reader Improvement Study): Ground truth established by consensus of three US licensed dentists. Non-consensus labels were adjudicated by an oral radiologist. (Same as standalone testing).

4. Adjudication Method for the Test Set:

Standalone Testing & Clinical Evaluation: The method described is consensus of three experts, with a fourth expert (oral radiologist) adjudicating non-consensus cases. This can be described as a "3+1 adjudication" method, where the "1" is specifically for tie-breaking or resolving disagreements among the initial three.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

Yes, a MRMC study was done. This is detailed under "Clinical Evaluation - Reader Improvement."
Effect Size / Reader Improvement:
- Bitewing Images:
  - Overall reader sensitivity improved from 64.6% (unassisted) to 78.5% (assisted). This is an absolute increase of 13.9 percentage points.
  - Weighted AFROC AUC increased from 0.729 (unassisted) to 0.785 (assisted), an increase of 0.055. This increase was statistically significant (p < 0.001).
  - Mean Dice scores increased from 0.67 (unassisted) to 0.76 (assisted) for primary caries. Similar increases were observed for other caries types.
- Periapical Images:
  - Overall reader sensitivity improved from 65.6% (unassisted) to 79.0% (assisted). This is an absolute increase of 13.4 percentage points.
  - Weighted AFROC AUC increased from 0.799 (unassisted) to 0.848 (assisted), an increase of 0.050. This increase was statistically significant (p < 0.001).
  - Mean Dice scores increased from 0.73 (unassisted) to 0.80 (assisted) for primary caries. Similar increases were observed for other caries types.

The document concludes that the increase in wAFROC numbers clearly demonstrates improvement in caries detection by dentists when aided by Overjet Caries Assist, aligned with increases in sensitivity and minimal decrease in specificity.

6. Standalone (Algorithm Only) Performance Study:

Yes, a standalone performance study was done. This is detailed under "Standalone Testing." The results for the algorithm's sensitivity, specificity, and Dice scores are provided for both bitewing and periapical images.

7. Type of Ground Truth Used:

Expert Consensus. For both standalone and clinical evaluation studies, the ground truth was "established by consensus of labels of three US licensed dentists, and non-consensus labels were adjudicated by an oral radiologist."

8. Sample Size for the Training Set:

The document does not explicitly state the sample size for the training set. It describes the "Machine Learning model" and its components but does not provide details about the data used for training.

9. How the Ground Truth for the Training Set Was Established:

The document does not explicitly state how the ground truth for the training set was established. It only details the ground truth establishment for the test sets used in standalone and clinical performance evaluations.

Summary

{0}------------------------------------------------

March 27, 2023

Image /page/0/Picture/1 description: The image shows the logo of the U.S. Food and Drug Administration (FDA). The logo consists of two parts: on the left, there is the Department of Health and Human Services logo, which features an abstract image of a human figure. To the right of this is the text "FDA U.S. FOOD & DRUG ADMINISTRATION" in blue font. The word "FDA" is in a larger, bolder font than the rest of the text.

Overjet, Inc. % Adam Odeh Director, Regulatory Affairs and Quality Assurance 560 Harrison St., Unit 403 BOSTON, MA 02118

Re: K222746

Trade/Device Name: Overjet Caries Assist Regulation Number: 21 CFR 892.2070 Regulation Name: Medical Image Analyzer Regulatory Class: Class II Product Code: MYN Dated: February 23, 2023 Received: February 23, 2023

Dear Adam Odeh:

We have reviewed your Section 510(k) premarket notification of intent to market the device referenced above and have determined the device is substantially equivalent (for the indications for use stated in the enclosure) to legally marketed predicate devices marketed in interstate commerce prior to May 28, 1976, the enactment date of the Medical Device Amendments, or to devices that have been reclassified in accordance with the provisions of the Federal Food, Drug, and Cosmetic Act (Act) that do not require approval of a premarket approval application (PMA). You may, therefore, market the device, subject to the general controls provisions of the Act. Although this letter refers to your product as a device, please be aware that some cleared products may instead be combination products. The 510(k) Premarket Notification Database located at https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm identifies combination product submissions. The general controls provisions of the Act include requirements for annual registration, listing of devices, good manufacturing practice, labeling, and prohibitions against misbranding and adulteration. Please note: CDRH does not evaluate information related to contract liability warranties. We remind you, however, that device labeling must be truthful and not misleading.

If your device is classified (see above) into either class II (Special Controls) or class III (PMA), it may be subject to additional controls. Existing major regulations affecting your device can be found in the Code of Federal Regulations, Title 21, Parts 800 to 898. In addition, FDA may publish further announcements concerning your device in the Federal Register.

Please be advised that FDA's issuance of a substantial equivalence determination does not mean that FDA has made a determination that your device complies with other requirements of the Act or any Federal statutes and regulations administered by other Federal agencies. You must comply with all the Act's requirements, including, but not limited to: registration and listing (21 CFR Part 807); labeling (21 CFR Part

{1}------------------------------------------------

801); medical device reporting of medical device-related adverse events) (21 CFR 803) for devices or postmarketing safety reporting (21 CFR 4, Subpart B) for combination products (see https://www.fda.gov/combination-products/guidance-regulatory-information/postmarketing-safety-reportingcombination-products); good manufacturing practice requirements as set forth in the quality systems (QS) regulation (21 CFR Part 820) for devices or current good manufacturing practices (21 CFR 4. Subpart A) for combination products; and, if applicable, the electronic product radiation control provisions (Sections 531-542 of the Act); 21 CFR 1000-1050.

Also, please note the regulation entitled, "Misbranding by reference to premarket notification" (21 CFR Part 807.97). For questions regarding the reporting of adverse events under the MDR regulation (21 CFR Part 803), please go to https://www.fda.gov/medical-device-safety/medical-device-reportingmdr-how-report-medical-device-problems.

For comprehensive regulatory information about mediation-emitting products, including information about labeling regulations, please see Device Advice (https://www.fda.gov/medicaldevices/device-advice-comprehensive-regulatory-assistance) and CDRH Learn (https://www.fda.gov/training-and-continuing-education/cdrh-learn). Additionally, you may contact the Division of Industry and Consumer Education (DICE) to ask a question about a specific regulatory topic. See the DICE website (https://www.fda.gov/medical-device-advice-comprehensive-regulatoryassistance/contact-us-division-industry-and-consumer-education-dice) for more information or contact DICE by email (DICE@fda.hhs.gov) or phone (1-800-638-2041 or 301-796-7100).

Sincerely,

Lu Jiang

Lu Jiang, Ph.D. Assistant Director Diagnostic X-ray Systems Team DHT8B: Division of Imaging Devices and Electronic Products OHT8: Office of Radiological Health Office of Product Evaluation and Quality Center for Devices and Radiological Health

Enclosure

{2}------------------------------------------------

Indications for Use

510(k) Number (if known) K222746

Device Name Overjet Caries Assist

Indications for Use (Describe)

Type of Use (Select one or both, as applicable)

☑ Prescription Use (Part 21 CFR 801 Subpart D)	☐ Over-The-Counter Use (21 CFR 801 Subpart C)
------------------------------------------------------------------------------------	-----------------------------------------------------------------------------------

CONTINUE ON A SEPARATE PAGE IF NEEDED.

This section applies only to requirements of the Paperwork Reduction Act of 1995.

DO NOT SEND YOUR COMPLETED FORM TO THE PRA STAFF EMAIL ADDRESS BELOW.

The burden time for this collection of information is estimated to average 79 hours per response, including the time to review instructions, search existing data sources, gather and maintain the data needed and complete and review the collection of information. Send comments regarding this burden estimate or any other aspect of this information collection, including suggestions for reducing this burden, to:

Department of Health and Human Services Food and Drug Administration Office of Chief Information Officer Paperwork Reduction Act (PRA) Staff PRAStaff(@fda.hhs.gov

"An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB number."

{3}------------------------------------------------

510(k) Summary

(K222746)

This summary of 510(k) information is being submitted in accordance with the requirements of 21CFR Part 807.92

1. Date

11 Sep 2022

2. Applicant

Overjet, Inc. 560 Harrison Ave Unit 403 Boston, MA 02118 Contact Person: Adam N. Odeh Email: adam.odeh@overjet.ai

1. Trade Name Overjet Caries Assist
1. Common Name Medical Image Analyzer

5. Classification 21 CFR 892.2070, Product code MYN, Class 2, Radiology

6. Device Description

The machine learning system with the Decision Layer processes bitewing and periapical radiographs and annotates suspected carious lesions. It is comprised of four modules:

Image Preprocessor Module This module performs two functions: .

{4}------------------------------------------------

Resizes and normalizes the images o
o Evaluates the incoming radiograph and predicts the image type as Bitewing, Periapical, or other. Any images classified as "other" are not processed.
. Tooth Number Assignment Module - This module analyzes the processed image and determines what tooth numbers are present and provides a pixel wise segmentation mask for each tooth number.
. Caries Module - This model segments carious lesions using an ensemble of 3 Instance Segmentation models.
. Post Processing - The overlap of tooth masks from the Tooth Number Assignment Module and carious lesions from the Caries Module are used to assign specific carious lesions to a specific tooth. The Post Processing module annotates the original radiograph with the carious lesions' predictions.

7. Indications for Use

8. Intended Patient Population

The intended patient population of the device is patients that have permanent dentition, and who are at least

12 years of age.

9. Warnings and Limitations

The safety and effectiveness of the system has not been established on primary or mixed dentition.
The device should only be used by licensed dentists.
. The Overjet Caries Assist device assists only in potential caries detection, and should not be relied upon as the sole decision-making tool for diagnosis or treatment.
. Overjet Caries Assist has been tested with the sensors listed here and in the user manual. Overjet cannot guarantee the accuracy of results when Overjet Caries Assist is used on images from other sensors.
The Overjet Caries Assist device is not intended for images smaller than 500 x 500 resolution. ● Overjet cannot guarantee the accuracy of results when Overjet Caries Assist is used on images of lower resolution.
If images are rotated incorrectly (i.e., left side bitewing on right side), tooth numbering and caries predictions will not be accurate.
As with any CADe device, the product has the potential for false positive or false negative outputs. The user should use all appropriate clinical information to render a final clinical opinion, with radiographic interpretation assisted by Overjet being one component of the determination process.
The Overjet Caries Assist device is trained to detect caries based on radiolucencies visible on radiographs. In areas where teeth overlap significantly, Overjet Caries Assist is trained to not predict caries due to the high potential of detecting false positives. If tooth overlap is present but the radiolucency can still be distinctly visualized, Overjet Caries Assist will predict in these areas.

10. Predicate Device/Device to be Modified

Device: Overjet Caries Assist

Manufacturer - Overjet, Inc.

Previously cleared as K212519

{5}------------------------------------------------

Substantial Equivalence 10.

Device	Overjet Caries Assist(Predicate)	Overjet Caries Assist(Proposed)
510k	K212519	TBD
Regulation No /Description	CFR 892.2070Medical image analyzer	CFR 892.2070Medical image analyzer
Indications	The Overjet Caries Assist is aradiological, automated, concurrentread, computer-assisted detectionsoftware intended to aid in thedetection and segmentation of carieson bitewing radiographs. The deviceprovides additional information forthe dentist to use in their diagnosis ofa tooth surface suspected of beingcarious. The device is not intended asa replacement for a completedentist's review or their clinicaljudgment that takes into accountother relevant information from theimage, patient history, or actual invivo clinical assessment.	Overjet Caries Assist (OCA) is aradiological, automated, concurrent read,computer-assisted detection (CADe)software intended to aid in the detectionand segmentation of caries on bitewingand periapical radiographs. Thedevice provides additional informationfor the dentist to use in their diagnosisof a tooth surface suspected of beingcarious. The device is not intended as areplacement for a complete dentist'sreview or their clinical judgment that takes into account other relevantinformation from the image, patienthistory, or actual in vivo clinicalassessment.
Type of CAD	CADe	CADe
End User	Dentist	Dentist
Patient Population	Patients requiring dental services, allsexes, 18 years of age or older	Patients requiring dental services, allsexes, 12 years of age or older withpermanent teeth
Platform	Web - Edge, Chrome, Firefox	Web - Edge, Chrome, Firefox
OS	Any	Any
User Interface	Mouse, Keyboard, Trackpad	Mouse, Keyboard, Trackpad
Image Input Sources	Images imported from theradiographic device, or from thepractice management system, fromCarestream or Schick sensors.	Images imported from the radiographicdevice, or from the practice managementsystem from multiple sensormanufacturers
Image format	jpg, png, eop, jif, dicom	jpg, png, eop, tiff, dicom
Device	Overjet Caries Assist(Predicate)	Overjet Caries Assist(Proposed)
ProcessingArchitecture	Three layers:1 - The Network layer workswith the practice PACS or EMRto transmit the image and meta-data to Overjet.2 - The decision layer processesthe image to ensure it is the correctdata type, and then annotates it viathe algorithm3 - The presentation layer displaysthe annotated image in a non-diagnostic viewer. The dentist canfilter, display, hide, create and editthe annotations presented.	Three layers:1 - The Network layer workswith the practice PACS or EMR totransmit the image and meta-data toOverjet.2 - The decision layer processes theimage to ensure it is the correct datatype, and then annotates it via thealgorithm3 - The presentation layer displaysthe annotated image in a non-diagnosticviewer. The dentist can filter, display,hide, create and edit the annotationspresented.
Data Source	Digital files of bitewing radiographswhose longer edge is greater than 500pixel resolution.	Digital files of bitewing and periapicalradiographs whose longer edge is greaterthan 500 pixel resolution.
Output	Caries detection and segmentation onradiograph resulting in outline ofsuspected caries	Caries detection and segmentation onradiograph resulting in outline or fillof suspected caries
Performance Testing	Increase in dentist's sensitivity ofgreater than 15%	Increase in dentist's sensitivity of greaterthan 15%
Level of Concern	Moderate	Moderate

{6}------------------------------------------------

The subject Overjet Caries Assist (OCA) device for this 510k submission is determined to be substantially equivalent to the previously cleared OCA device (K212519). The differences are as follows:

Analysis of Periapical radiographs in addition to Bitewing radiographs ●
Multiple sensor manufacturers, as opposed to previous clearance for only two sensor . manufacturers
Target population is 12 years and older with permanent teeth (as opposed to 18 years and older, in K212519)

Overjet does not believe that these differences raise any concerns of substantial equivalence, and that the changes to intended use are well supported by performance testing and present no increased risk to patients.

{7}------------------------------------------------

11. Performance Testing

Overjet has conducted performance testing according to FDA's "Guidance for Industry and Food and Drug Administration Staff Computer-Assisted Detection Devices Applied to Radiology Images and Radiology Device Data - Premarket Notification [510(k)] Submissions Document'' issued on 03 Jul 2012 and the "Clinical Performance Assessment: Considerations for Computer-Assisted Detection Devices Applied to Radiology Images and Radiology Device Data in Premarket Notification (510(k)) Submissions" Guidance issued January 2020, as part of the development process of the caries model. Performance testing included standalone testing and a clinical reader evaluation. All testing demonstrated that Overjet Caries Assist (OCA) met prespecified requirements.

Standalone Testing

Standalone performance of the Overjet AI algorithm was evaluated on a total of 1,293 Bitewing images and 1,314 Periapical images. Sensitivity and specificity were summarized based on surfaces, and 95% Cls were provided based on treating the subject as the basis of a cluster. A total of 27,920 bitewing surfaces and 16,254 periapical surfaces were available and included in the analysis. Standalone performance of the OCA device was compared to a ground truth established by consensus of labels of three US licensed dentists, and non-consensus labels were adjudicated by an oral radiologist.

Standalone testing included images from the following sensor manufacturers: Carestream, Dexis, e2v, Gendex, Hamamatsu, Jazz Imaging, ScanX, Schick, Soredex Digora.

Sensitivity

For bitewing images, overall standalone sensitivity was 76.6% (73.8%, 79.4%). Subgroup sensitivity was as follows: Primary caries - 79.9% (77.1%, 82.7%); Secondary caries - 60.9% (53.5%, 68.2%); Enamel - 74.4% (70.4%, 78.3%); Dentin - 79.5% (75.8%, 83.2%)

For periapical images, overall standalone sensitivity was 79.4% (76.1%, 82.8%). Subgroup sensitivity was as follows: Primary caries - 79.8% (76.0%, 83.7%): Secondary caries - 77.9% (71.4%, 84.5%); Enamel - 67.9% (60.7%, 75.1%); Dentin - 84.9% (81.3%, 88.4%).

Specificity

Overall specificity was 99.1% (98.9%, 99.2%) for bitewing images, and 99.4% (99.2%, 99.5%) for periapical images.

Subgroup Analyses

Subgroup analyses were also performed for age, sex, sensor, and associated restoration (for secondary caries).

{8}------------------------------------------------

Lesion Segmentation

Dice coefficient analysis was performed to compare pixel-level metrics of each carious lesion with the lesion tracing provided by ground truthers. Dice scores were calculated only for true positives.

For bitewing images, the mean Dice score was 0.77 (0.76, 0.78) for primary caries, 0.73 (0.70. 0.75) for secondary caries, 0.76 (0.75, 0.77) for enamel caries, and 0.77 (0.76, 0.79) for dentin caries.

For periapical images, the mean Dice score was 0.79 (0.78, 0.81) for primary caries, 0.79 (0.77, 0.82) for secondary caries, 0.75 (0.73, 0.77) for enamel caries, and 0.81 (0.80, 0.82) for dentin caries.

Clinical Evaluation - Reader Improvement

Overjet evaluated the performance of Overjet Caries Assist in a multi-reader fully crossed reader improvement study. 28 US licensed dentists were split into 2 groups of 14 each. One group was asked to evaluate 330 bitewing images (94 containing caries / 236 without caries) and the other was asked to evaluate 330 periapical images (also 94 containing caries / 236 without caries). Ground truth was established by the consensus labels of three US licensed dentists, and nonconsensus labels were adjudicated by an oral radiologist. Half of the data set contained unannotated images, and the second half contained radiographs that had been processed through the OCA model. Radiographs were presented to readers in alternating groups.

In Session 1, readers were asked to outline suspected caries, and to review predictions from the OCA model. Each reader was asked to provide a rating of 1 - 4 for their confidence in the annotation (1 for lowest confidence, up to 4 for highest confidence). A 4-week washout period was utilized to limit recollection bias. Following the washout, the readers were presented the same data set but with alternate grouping. I.e., if a reader saw a radiograph in the unpredicted state in session 1, they were presented with the same radiograph with OCA predictions in session 2, and vice versa.

Results were compared against a consensus ground truth, and the sensitivity, specificity, and weighted alternative free response receiver operating characteristic (wAFROC) were evaluated to characterize the performance of the readers with (assisted) and without (unassisted) viewing the model annotations.

Unassisted vs. Assisted Sensitivity:

For bitewing images, overall reader sensitivity improved from 64.6% (56.4%, 72.1%) to 78.5% (72.6%, 83.6%) unassisted vs assisted. Subgroup improvement was as follows: Primary caries - 67.1% (58.3%. 74.8%) to 83.1% (78.0%. 87.7%) Secondary caries - 51.2% (35.9%, 66.1%) to 55.8% (40.9%, 70.4%) Enamel - 61.8% (51.5%, 71.0%) to 77.6% (70.1%, 83.9%) Dentin - 68.5% (59.8%, 76.8%) to 79.9% (72.6%, 86.9%)

{9}------------------------------------------------

For periapical images, overall reader sensitivity improved from 65.6% (59.4%, 71.7%) to 79.0% (73.0%, 84.7%) unassisted vs assisted. Subgroup improvement was as follows:

Primary caries - 67.5% (60.8%, 74.1%) to 80.5% (73.6%, 86.8%)

Secondary caries - 56.1% (43.5%, 69.3%) to 71.6% (59.1%, 83.5%)

Enamel - 58.6% (47.2%, 69.4%) to 74.1% (63.9%, 83.2%)

Dentin - 69.6% (62.5%, 76.0%) to 81.9% (75.1%, 88.3%)

Unassisted vs. Assisted Specificity:

For bitewing images, overall reader specificity decreased slightly from 99.0% (98.5%, 99.4%) to 98.6% (98.0%, 99.0%) unassisted vs assisted.

For periapical images, overall reader specificity decreased slightly from 98.0% (97.4%) to 97.6% (97.0%, 98.1%) unassisted vs assisted.

Subgroup Analyses:

Subgroup analyses were also performed for age, gender, sensor, reader experience, and associated restoration (for secondary caries).

Unassisted vs Assisted Dice Scores:

As with standalone testing, Dice scores were calculated in comparison to ground truth for readers with and without Overjet Caries Assist.

On bitewing images, mean Dice scores increased from 0.67 (0.009 SD) to 0.76 (0.009 SD) for primary caries, 0.65 (0.017 SD) to 0.67 (0.016 SD) for secondary caries, 0.65 (0.009 SD) to 0.74 (0.009 SD) for enamel, and 0.67 (0.012 SD) to 0.74 (0.012 SD) for dentin.

On periapical images, mean Dice scores increased from 0.73 (0.011 SD) to 0.80 (0.011 SD) for primary caries, 0.69 (0.025 SD) to 0.74 (0.025 SD) for secondary caries, 0.64 (0.015 SD) to 0.73 (0.014 SD) for enamel, and 0.76 (0.013 SD) to 0.81 (0.013 SD) for dentin.

{10}------------------------------------------------

Weighted AFROC (wAFROC) Scores:

Readers provided confidence scores for any detected caries, which were used to calculate AUC for weighted AFROC scores.

On bitewing images, for the average of all readers, AUC increased from 0.729 (0.696, 0.761) to 0.785 (0.746, 0.822), for an increase in AUC of 0.055 (0.033, 0.079) unassisted to assisted. This increase was statistically significant, with an overall p-value less than 0.001.

On periapical images, for the average of all readers, AUC increased from 0.799 (0.764, 0.833) to 0.848 (0.814, 0.881), for an increase in AUC of 0.050 (0.031, 0.068) unassisted to assisted. This increase was also statistically significant, with an overall p-value less than 0.001.

Summary

Increase in overall wAFROC numbers clearly demonstrate improvement in caries detection by dentists when aided by Overjet Caries Assist (0.055 increase for bitewing, 0.050 increase for periapical). This aligns with the observed increases in sensitivity for both bitewing and periapical images. When considered alongside the decreases in overall specificity (only 0.4% for both bitewing and periapical images), it is clear that Overjet Caries Assist demonstrates a clear benefit for caries detection.

12. General Safety and Effectiveness Concerns

The device labeling contains instructions for use and any necessary cautions and warnings to provide for safe and effective use of this device. Risk management was conducted according to ISO 14971, which ensured, via a risk analysis, the identification and mitigation of potential hazards were controlled via software development and design, verification testing. In addition, general and special controls of the FD&C Act established for Radiological Computer Assisted Detection and Diagnosis Software are in place to further mitigate any safety and or effectiveness risks.

13. Assessment of Non-clinical Performance Data

Overiet Caries Assist has been verified and validated according to Overjet's design control processes. All supporting documentation has been included in this 510(k) premarket notification. Verification activity included unit, integration, and system level testing. Validation testing included performing a pivotal reader study to compare the clinical performance of dentists using CAD detections from Overjet Caries Assist software when applied to dental radiographs to that of dentists not using Overjet Caries Assist.

{11}------------------------------------------------

Conclusion 14.

The subject device Overjet Caries Assist (OCA) is substantially equivalent to the predicate OCA device. Differences do not raise any concerns about the safety or efficacy of the device.

Regulation Number and Section

§ 892.2070 Medical image analyzer.

(a)
Identification. Medical image analyzers, including computer-assisted/aided detection (CADe) devices for mammography breast cancer, ultrasound breast lesions, radiograph lung nodules, and radiograph dental caries detection, is a prescription device that is intended to identify, mark, highlight, or in any other manner direct the clinicians' attention to portions of a radiology image that may reveal abnormalities during interpretation of patient radiology images by the clinicians. This device incorporates pattern recognition and data analysis capabilities and operates on previously acquired medical images. This device is not intended to replace the review by a qualified radiologist, and is not intended to be used for triage, or to recommend diagnosis.(b)
Classification. Class II (special controls). The special controls for this device are:(1) Design verification and validation must include:
(i) A detailed description of the image analysis algorithms including a description of the algorithm inputs and outputs, each major component or block, and algorithm limitations.
(ii) A detailed description of pre-specified performance testing methods and dataset(s) used to assess whether the device will improve reader performance as intended and to characterize the standalone device performance. Performance testing includes one or more standalone tests, side-by-side comparisons, or a reader study, as applicable.
(iii) Results from performance testing that demonstrate that the device improves reader performance in the intended use population when used in accordance with the instructions for use. The performance assessment must be based on appropriate diagnostic accuracy measures (
e.g., receiver operator characteristic plot, sensitivity, specificity, predictive value, and diagnostic likelihood ratio). The test dataset must contain a sufficient number of cases from important cohorts (e.g., subsets defined by clinically relevant confounders, effect modifiers, concomitant diseases, and subsets defined by image acquisition characteristics) such that the performance estimates and confidence intervals of the device for these individual subsets can be characterized for the intended use population and imaging equipment.(iv) Appropriate software documentation (
e.g., device hazard analysis; software requirements specification document; software design specification document; traceability analysis; description of verification and validation activities including system level test protocol, pass/fail criteria, and results; and cybersecurity).(2) Labeling must include the following:
(i) A detailed description of the patient population for which the device is indicated for use.
(ii) A detailed description of the intended reading protocol.
(iii) A detailed description of the intended user and user training that addresses appropriate reading protocols for the device.
(iv) A detailed description of the device inputs and outputs.
(v) A detailed description of compatible imaging hardware and imaging protocols.
(vi) Discussion of warnings, precautions, and limitations must include situations in which the device may fail or may not operate at its expected performance level (
e.g., poor image quality or for certain subpopulations), as applicable.(vii) Device operating instructions.
(viii) A detailed summary of the performance testing, including: test methods, dataset characteristics, results, and a summary of sub-analyses on case distributions stratified by relevant confounders, such as lesion and organ characteristics, disease stages, and imaging equipment.