Lunit INSIGHT MMG

K211678 · Lunit, Inc. · QDQ · Nov 17, 2021 · Radiology

Device Facts

• Record ID: K211678
• Device Name: Lunit INSIGHT MMG
• Applicant: Lunit, Inc.
• Product Code: QDQ · Radiology
• Decision Date: Nov 17, 2021
• Decision: SESE
• Submission Type: Traditional
• Regulation: 21 CFR 892.2090
• Device Class: Class 2
• Attributes: AI/ML, Software as a Medical Device

Intended Use

Lunit INSIGHT MMG is a radiological Computer-Assisted Detection and Diagnosis (CADe/x) software device based on an artificial intelligence algorithm intended to aid in the detection, and characterization of suspicious areas for breast cancer on mammograms from compatible FFDM systems. As an adjunctive tool, the device is intended to be viewed by interpreting physicians after completing their initial read. It is not intended as a replacement for a complete physician's review or their clinical judgement that takes into account other relevant information from the image or patient history. Lunit INSIGHT MMG uses screening mammograms of the female population.

Device Story

Software-only CADe/x device; analyzes screening mammograms from FFDM systems. Inputs: DICOM mammography images. Processing: Deep learning AI algorithm trained on biopsy-proven databases of breast cancer, benign lesions, and normal tissue. Outputs: Visualized maps (heatmaps) of suspicious lesions; lesion-level abnormality scores (1-100); per-breast abnormality scores (maximum lesion score). Workflow: De-identifies images; processes in background; saves results to PACS/storage. Usage: Adjunctive tool for interpreting physicians after initial read. Clinical impact: Assists in detection/characterization of malignant lesions; reduces false positives/negatives; improves diagnostic accuracy. Benefits: Enhanced physician interpretation performance; potential for earlier breast cancer detection.

Clinical Evidence

Retrospective, multi-case (MRMC) reader study with 12 radiologists and 240 mammograms. Primary endpoint: ROC AUC improvement with CAD assistance. Results: ROC AUC increased from 0.754 (unaided) to 0.805 (aided) (difference 0.051, p=0.0001). Standalone performance study (2,412 mammograms) showed ROC AUC of 0.903 (95% CI: 0.889-0.917). Secondary endpoints (LCM LROC, LCM ROC AUC, recall rates) also showed improved performance with device assistance.

Technological Characteristics

Software-only radiological CADe/x device. Deep learning algorithm for lesion detection/characterization. Inputs: DICOM FFDM images. Outputs: Heatmaps, lesion scores, per-breast abnormality scores. Connectivity: Integrates with PACS/imaging equipment. Level of concern: Moderate.

Indications for Use

Indicated for female patients undergoing screening mammography to aid physicians in the detection and characterization of suspicious breast cancer lesions on FFDM images.

Regulatory Classification

Identification

A radiological computer-assisted detection and diagnostic software is an image processing device intended to aid in the detection, localization, and characterization of fracture, lesions, or other disease-specific findings on acquired medical images (e.g., radiography, magnetic resonance, computed tomography). The device detects, identifies, and characterizes findings based on features or information extracted from images, and provides information about the presence, location, and characteristics of the findings to the user. The analysis is intended to inform the primary diagnostic and patient management decisions that are made by the clinical user. The device is not intended as a replacement for a complete clinician's review or their clinical judgment that takes into account other relevant information from the image or patient history.

Special Controls

A radiological computer assisted detection and diagnosis software must comply with the following special controls: Design verification and validation must include: 1. i. A detailed description of the image analysis algorithm, including but not limited to a description of the algorithm inputs and outputs, each major component or block, how the algorithm and output affects or relates to clinical practice or patient care, and any algorithm limitations. ii. A detailed description of pre-specified performance testing protocols and dataset(s) used to assess whether the device will provide improved assisted-read detection and diagnostic performance as intended in the indicated user population(s), and to characterize the standalone device performance for labeling. Performance testing includes standalone test(s), side-by-side comparison(s), and/or a reader study, as applicable. iii. Results from standalone performance testing used to characterize the independent performance of the device separate from aided user performance. The performance assessment must be based on appropriate diagnostic accuracy measures (e.g., receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Devices with localization output must include localization accuracy testing as a component of standalone testing. The test dataset must be representative of the typical patient population with enrichment made only to ensure that the test dataset contain a sufficient number of cases from important cohorts (e.g., subsets defined by clinically relevant confounders, effect modifiers, concomitant disease, 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. Results from performance testing that demonstrate that the device provides improved assisted-read detection and/or diagnostic performance as intended in the indicated user population(s) when used in accordance with the instructions for use. The reader population must be comprised of the intended user population in terms of but not limited to clinical training, certification, and years of experience. The performance assessment must be based on appropriate diagnostic accuracy measures (e.g., receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Test datasets must meet the requirements described in 1(iii) above. v. Appropriate software documentation, including device hazard analysis, software requirements specification document, software design specification document, traceability analysis, system level test protocol, pass/fail criteria, testing results, and cybersecurity measures. 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 device instructions for use, including the intended reading protocol and how the user should interpret the device output. iii. A detailed description of the intended user, and any user training materials as programs that addresses appropriate reading protocols for the device to ensure that the end user is fully aware of how to interpret and apply the device output. iv. A detailed description of the device inputs and outputs. v. A detailed description of compatible imaging hardware and imaging protocols. vi. 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. 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 anatomical characteristics, patient demographics and medical history, user experience, and imaging equipment.

*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 algorithm, including a description of the algorithm inputs and outputs, each major component or block, how the algorithm and output affects or relates to clinical practice or patient care, and any algorithm limitations. (ii) A detailed description of pre-specified performance testing protocols and dataset(s) used to assess whether the device will provide improved assisted-read detection and diagnostic performance as intended in the indicated user population(s), and to characterize the standalone device performance for labeling. Performance testing includes standalone test(s), side-by-side comparison(s), and/or a reader study, as applicable. (iii) Results from standalone performance testing used to characterize the independent performance of the device separate from aided user performance. The performance assessment must be based on appropriate diagnostic accuracy measures ( *e.g.,* receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Devices with localization output must include localization accuracy testing as a component of standalone testing. The test dataset must be representative of the typical patient population with enrichment made only to ensure that the test dataset contains a sufficient number of cases from important cohorts (*e.g.,* subsets defined by clinically relevant confounders, effect modifiers, concomitant disease, 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) Results from performance testing that demonstrate that the device provides improved assisted-read detection and/or diagnostic performance as intended in the indicated user population(s) when used in accordance with the instructions for use. The reader population must be comprised of the intended user population in terms of clinical training, certification, and years of experience. The performance assessment must be based on appropriate diagnostic accuracy measures ( *e.g.,* receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Test datasets must meet the requirements described in paragraph (b)(1)(iii) of this section.(v) Appropriate software documentation, including device hazard analysis, software requirements specification document, software design specification document, traceability analysis, system level test protocol, pass/fail criteria, testing results, and cybersecurity measures. (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 device instructions for use, including the intended reading protocol and how the user should interpret the device output. (iii) A detailed description of the intended user, and any user training materials or programs that address appropriate reading protocols for the device, to ensure that the end user is fully aware of how to interpret and apply the device output. (iv) A detailed description of the device inputs and outputs. (v) A detailed description of compatible imaging hardware and imaging protocols. (vi) 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) 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 anatomical characteristics, patient demographics and medical history, user experience, and imaging equipment.

Predicate Devices

• Transpara™ (K192287)

Submission Summary (Full Text)

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

Image /page/0/Picture/0 description: The image contains the logo of the U.S. Food and Drug Administration (FDA). On the left is the Department of Health & Human Services USA logo. To the right of that is the FDA logo, which is a blue square with the letters "FDA" in white. To the right of the blue square is the text "U.S. FOOD & DRUG ADMINISTRATION" in blue.

Lunit Inc. % John Smith Partner Hogan Lovells US LLP 555 Thirteenth Street NW Washington, District of Columbia 20004

February 7, 2022

Re: K211678

Trade/Device Name: Lunit INSIGHT MMG Regulation Number: 21 CFR 892.2090 Regulation Name: Radiological computer assisted detection and diagnosis software Regulatory Class: Class II Product Code: QDQ

Dear John Smith:

The Food and Drug Administration (FDA) is sending this letter to notify you of an administrative change related to your previous substantial equivalence (SE) determination letter dated October 7, 2021. Specifically, FDA is updating this SE Letter due to a typographical error in the 510(k) summary as an administrative correction. Specifically the substantial equivalence table contains an additional column, which could trigger a misinterpretation for members of the public.

Please note that the 510(k) submission was not re-reviewed. For questions regarding this letter please contact Jessica Lamb, OHT7: Office of In Vitro Diagnostics and Radiological Health, 301-796-6167, Jessica.Lamb(@fda.hhs.gov.

Sincerely.

Jessica Lamb

Jessica Lamb Assistant Director Division of Radiological Health OHT7: Office of In Vitro Diagnostics and Radiological Health Office of Product Evaluation and Quality Center for Devices and Radiological Health

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

Image /page/1/Picture/0 description: The image contains the logo of the U.S. Food and Drug Administration (FDA). On the left is the Department of Health & Human Services logo. To the right of that is the FDA logo, which is a blue square with the letters "FDA" in white. To the right of the blue square is the text "U.S. FOOD & DRUG ADMINISTRATION" in blue.

Lunit Inc. % John J. Smith, M.D., J.D. Partner Hogan Lovells US LLP 555 Thirteenth Street NW WASHINGTON DC 20004

November 17, 2021

## Re: K211678

Trade/Device Name: Lunit INSIGHT MMG Regulation Number: 21 CFR 892.2090 Regulation Name: Radiological computer assisted detection and diagnosis software Regulatory Class: Class II Product Code: QDQ Dated: October 7, 2021 Received: October 7, 2021

Dear Dr. Smith:

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 801); medical device reporting of medical device-related adverse events) (21 CFR 803) for

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

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 ([e-mail]) or phone [phone] or 301-796-7100).

Sincerely.

Jessica Lamb

For

Thalia T. Mills, Ph.D. Director Division of Radiological Health OHT7: Office of In Vitro Diagnostics and Radiological Health Office of Product Evaluation and Quality Center for Devices and Radiological Health

Enclosure

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

#### DEPARTMENT OF HEALTH AND HUMAN SERVICES Food and Drug Administration Indications for Use

510(k) Number (if known) K211678

Device Name

Lunit INSIGHT MMG Indications for Use (Describe)

Lunit INSIGHT MMG is a radiological Computer-Assisted Detection and Diagnosis (CADe/x) software device based on an artificial intelligence algorithm intended to aid in the detection, and characterization of suspicious areas for breast cancer on mammograms from compatible FFDM systems. As an adjunctive tool, the device is intended to be viewed by interpreting physicians after completing their initial read.

It is not intended as a replacement for a complete physician's review or their clinical judgement that takes into account other relevant information from the image or patient history. Lunit INSIGHT MMG uses screening mammograms of the female population.

区 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 [e-mail]

"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

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

# 510(k) Summary

# Lunit INSIGHT MMG (K211678)

This 510(k) summary of safety and effectiveness information is prepared in accordance with the requirements of 21 CFR §807.92.

Date Prepared: November 15, 2021

## I. SUBMITTER

#### Manufacturer:

Lunit Inc. 15 Floor, 27 Teheran-ro 2-gil, Gangnam-gu, Seoul, Republic of Korea 06241 Phone [phone] Fax [phone]

## Contact Person:

Joohee Lee Regulatory Affairs Specialist

## II. DEVICE

Name of Device: Lunit INSIGHT MMG Common or Usual Name: Lunit INSIGHT MMG Classification Name: Radiological Computer Assisted Detection/Diagnosis Software For Suspicious Lesions For Cancer Classification Regulation: 21 CFR 892.2090 Requlatory Class: II Product Code: QDQ

## III. PREDICATE DEVICE

Predicate Device Manufacturer Name: Screenpoint Medical B.V. Device Trade Name: Transpara™ 510(k) Number: K192287

#### IV. DEVICE DESCRIPTION

Lunit INSIGHT MMG is a radiological Computer-Assisted Detection and Diagnosis (CADe/x) software for aiding interpreting physicians with the detection, and characterization of suspicious areas for breast cancer on mammograms from compatible FFDM (full-field digital mammography) systems. The software applies an artificial intelligence algorithm for recognition of suspicious lesions,

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

which are trained with large databases of biopsy proven examples of breast cancer, benign lesions and normal tissues.

Lunit INSIGHT MMG automatically analyzes the mammograms received from the client's image storage system (e.g., Picture Archiving and Communication System (PACS)) or other radiological imaging equipment. Following receipt of mammograms, the software device de-identifies the copies of images in DICOM format (.dcm) and then automatically analyzes each image and identifies and characterizes suspicious areas for breast cancer. The analysis result is converted into DICOM file and the result is saved within the designated storage location (e.g., PACS, x-ray system, etc.)

Lunit INSIGHT MMG processes mammograms and the output of the device can be viewed by interpreting physicians after completing their initial read. As an analysis result, the software device allows a visualization and quantitative estimation of the presence of a malignant lesion. The suspicious lesions are marked by a visualized map and an abnormality score, which reflects general likelihood of presence of malignancy, is presented for each lesion, as well as for each breast.

## V. INDICATIONS FOR USE

Lunit INSIGHT MMG is a radiological Computer-Assisted Detection and Diagnosis (CADe/x) software device based on an artificial intelligence algorithm intended to aid in the detection, and characterization of suspicious areas for breast cancer on mammograms from compatible FFDM systems. As an adjunctive tool, the device is intended to be viewed by interpreting physicians affer completing their initial read.

lt is not intended as a replacement for a complete physician's review or their clinical judgement that takes into account other relevant information from the image or patient history. Lunit INSIGHT MMG uses screening mammograms of the female population.

## VI. COMPARISON OF TECHNOLOGICAL CHARACTERSTICS WITH THE PREDICATE DEVICE

The Lunit INSIGHT MMG and the predicate Transpara have the similar indications where both devices are intended to assist a clinical user in the detection, and characterization suspicious areas for breast cancer on mammography images. Also, both devices have intended user, and patient populations, and use the same imaging modality. The differences in the indications are with respect to the clinical workflow where the predicate device is intended to be used for concurrent read of images while the Lunit INSIGHT MMG is intended to be viewed by interpreting physicians after completing their initial read. Neither device is intended as a replacement for the review of a clinician or their clinical judgement and the patient management decisions are not made solely on the basis of analysis of the device. The minor difference in the indications with respect to the clinical workflow does not raise different questions of safety or effectiveness.

Lunit INSIGHT MMG and predicate share similar technological characteristics. Both devices are radiological computer assisted detection and diagnostic software based on an artificial intelligence algorithm that detect, identify and characterize findings based on features or information extracted from images. Both devices provide information about the presence and location of the findings to the user. As both devices process radiological images using similar machine learning techniques to

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

highlight abnormalities, the same types of safety and effectiveness questions are raised between devices. Both devices are support tools that provide relevant clinical data as an additional resource to the physician and will not replace the clinical expertise and judgement of the clinical user.

Both devices provide users with a numeric score indicating the likelihood of a presence of a malignant lesion however, the scoring system is different. Both devices provide a lesion score indicating likelihood of presence of malignancy. Lunit INSIGHT MMG analyzes per-breast abnormality score that takes the maximum scores from two views (i.e. CC, MLO). However, the predicate device analyzes the overall likelihood that cancer is present in a mammogram and categorizes the exams on a scale of 1 to 10 with increasing likelihood of cancer. The minor differences do not raise any different questions for its performance or safety; both devices provide clinically useful information to physicians who are reviewing a mammogram.

A comparison of the technological characteristics with the predicate device is summarized in Table 1.

## VII. PERFORMANCE DATA

#### Non-clinical performance Data

The Lunit INSIGHT MMG is considered as a "moderate" level of concern, since a failure or a latent design flaw of the software device could result in Minor injury, either to a patient or to a user of the device. Based on results of verification and validation tests, Lunit INSIGHT MMG is effective in the detection of suspicious lesion for breast cancer at an appropriate safety level in mammograms. Verification tests which consisted of software unit testing and system testing were successfully completed to assure that the software application satisfied the software requirements. Validation testing determining standalone performance of the algorithms in Lunit INSIGHT MMG was also completed successfully. Validation and verification test results demonstrated that Lunit INSIGHT MMG is as safe and effective as the predicate device.

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

#### Standalone Performance Testing

Standalone Performance Study of the Lunit INSIGHT MMG assessed the performance of software algorithm for breast cancer detection in screening mammography.

Total of 2412 mammograms were collected using Hologic, GE Healthcare, and Siemens mammography equipment. The standalone performance of Lunit INSIGHT MMG for breast cancer detection in screening mammography is examined as comparing the reference standards and the analysis results of the device. The dataset used in the study is independent from the dataset used for development of the algorithm and the US pivotal reader study.

The primary endpoint of the standalone performance measured the Lunit INSIGHT MMG performance compared to radiologist performance alone. ROC AUC in the standalone performance analysis was 0.903 (95% Cl: 0.889-0.917) with statistical significance (p < 0.0001), which demonstrates improvement compared to the interpretation performance of radiologists when reading mammograms unaided. The secondary exploratory endpoints are two types of LROC AUC, sensitivity and specificity and these results are as follow:

#### [Standalone Analysis Results]

| ROC AUC [95% CI] | 0.903 [0.889, 0.917] |
|---------------------------|----------------------|
| | *p < 0.0001 |
| Type I LROC AUC [95% CI] | 0.781 [0.751, 0.812] |
| Type II LROC AUC [95% CI] | 0.792 [0.763, 0.822] |
| Sensitivity (%) [95% CI] | 85.74 [82.95, 88.53] |
| Specificity (%) [95% Cl] | 75.62 [73.64, 77.60] |

## Clinical Testing - Reader Study

Clinical Performance Assessment was conducted to evaluate effectiveness of Lunit INSIGHT MMG in the assistance of detection and diagnosis of breast cancer during screening mammography interpretation. A retrospective, multi-case (MRMC) study was conducted comparing the reading panel's interpretation performance with and without the assistance of the Lunit INSIGHT MMG during the screening mammography interpretation.

Total of 240 mammograms were collected using Hologic and GE Healthcare mammography equipment in the US. During the study, 12 MQSA qualified reading panelists performed their interpretation independently using a setting similar to a screening procedure in the U.S. Reading panelists interpret cases without CAD assistance first then re-interpret of the same cases with Test 1 reading results in assistance of the Lunit INSIGHT MMG. To minimize any potential bias, the reading order for each reading panelists was randomized.

The primary objective of the reader study is to determine that the diagnostic ability of radiologist with CAD assistance is superior to without CAD assistance. Radiologist performance was assessed with an inter-test comparison of measuring FBR (Forced Bl-RADS) ROC AUC between reader's interpretation and device-assisted interpretation for the detection of malignant lesions in the study. Success criteria of the study was defined as the p-value of the ROC AUC difference test is less than

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

the significance level of 5% and the lower bound of two-sided 95% Cl (Confidence Interval) of the AUC difference (Test 2-Test 1) is above 0. Secondary exploratory endpoints of the reader study assessed inter-test difference in LCM (level of confidence of malignancy) LROC AUC, LCM ROC AUC, and recall rate in cancer and non-cancer group (sensitivity and 1-specificity).

For the primary endpoint, the average inter-test difference in ROC AUC between Test 2 and Test 1 was 0.051 (95% Cl: 0.027-0.075) with statistical significance (P=0.0001), which indicates that physician's interpretation ability for breast cancer in mammograms is improved in Test 2 (with CAD) from Test 1 (without CAD).

## [Primary Endpoint Results of the Pivotal Reader Study]

| | Test1 | Test2 | Test2-Test1 | P-value |
|-----------------|-------------------------|-------------------------|-------------------------|---------|
| ROC AUC (95%CI) | 0.754<br>(0.702, 0.807) | 0.805<br>(0.759, 0.852) | 0.051<br>(0.027, 0.075) | 0.0001 |

All secondary exploratory endpoints have shown higher performance in assistance of the device, the inter-test differences for each assessment are as follow: LCM LROC 0.094 (95% C1: 0.056 - 0.132), LCM ROC AUC 0.052 (95% Cl: 0.026 - 0.079), recall rate in cancer group (sensitivity) 5.97(95% Cl: 2.48 - 9.46), and recall rate in non-cancer group (1-specificity) -1.46 (95% Cl: -3.41 - 0.05). Results of the primary and secondary analysis of the clinical testing demonstrate improvement for breast cancer detection and diagnosis using Lunit INSHGT MMG in mammograms.

Separated from the standalone performance test, an additional assessment of standalone algorithm performance was conducted with same cases from the reader study, using the Lunit INIGHT MMG without reader intervention as a sole means of reading mammograms compared to the software output. ROC AUC in the standalone performance analysis is 0.863 (95% Cl: 0.818 - 0.909) which exceeds the results of every reading panelist in Test 1 (CAD unaided).

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

Image /page/9/Figure/0 description: The image is a plot of three ROC curves, showing the performance of different models. The x-axis is the false positive rate (FPF), and the y-axis is the true positive rate (TPF). The red curve represents the Standalone Lunit MMG model, which has an AUC of 0.863. The green curve represents the Unaided model, which has an AUC of 0.754, and the blue curve represents the With Lunit MMG model, which has an AUC of 0.805.

[Plot Comparison of the standalone performance, CAD unaided and aided reader interpretation]

#### VIII. CONCLUSIONS

The non-clinical performance data and clinical data showed that Lunit INSIGHT MMG is as safe and effective as the predicate device. In particular, the effectiveness of the Lunit INSIGHT MMG in assistance of detection and diagnosis of breast cancer during screening mammography interpretation has been verified in the US population by confirming that the improvement for breast cancer detection and diagnosis before and after use of the Lunit INSIGHT MMG. In addition, the study results lead a positive expectation in an aspect of the clinical benefit that false positives and false negatives are reduced in the screening mammograms with the use of the Lunit INSIGHT MMG. In a conclusion, with all the clinical evidence achieved from the clinical study, the Lunit INSIGHT MMG can be considered to perform effectively in assistance of detection and diagnosis of breast cancer.

The Lunit INSIGHT MMG has the similar intended uses, indications, technological characteristics, and principles of operation as its predicate device. Both devices detect and characterize findings in radiological breast images and provide information about the presence, location, and characteristics of the findings to the user in a similar way. The minor differences in indications do not alter the intended of the device and do not affect its safety and effectiveness. In addition, the minor technological differences between the Lunit INSIGHT MMG and its predicate devices raise no new questions of safety or effectiveness. Performance data demonstrate that the Lunit INSIGHT MMG is as safe and effective as the predicate Transpara. Thus, the Lunit INSIGHT MMG is substantially equivalent.

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

| Technical<br>Characteristics | Subject Device<br>Lunit INSIGHT MMG | Predicate Device<br>Transpara™ (K192287) |
|------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Classification<br>Regulation | 21 CFR 892.2090<br>Radiological Computer Assisted<br>Detection and Diagnosis software | 21 CFR 892.2090<br>Radiological Computer Assisted<br>Detection and Diagnosis software |
| Medical Device<br>Classification | Class II | Class II |
| Product Code | QDQ | QDQ |
| Level of<br>Concern | Moderate | Moderate |
| Intended Use | A reading aid for physicians<br>interpreting screening FFDM<br>acquired with compatible<br>mammography systems, to identify<br>findings and assess their level of<br>suspicion. | A reading aid for physicians<br>interpreting screening FFDM<br>acquired with compatible<br>mammography systems, to identify<br>findings and assess their level of<br>suspicion. |
| Target patient<br>population | Women undergoing FFDM screening<br>mammography | Women undergoing FFDM screening<br>mammography |
| Target user<br>population | Physicians interpreting FFDM<br>screening mammograms | Physicians interpreting FFDM<br>screening mammograms |
| Design | Software-only device | Software-only device |
| Indication for<br>Use | Lunit INSIGHT MMG is a radiological<br>Computer-Assisted Detection and<br>Diagnosis (CADe/x) software device<br>based on an artificial intelligence<br>algorithm intended to aid in the<br>detection, localization, and | The ScreenPoint Transpara™ system is intended for use as a<br>concurrent reading aid for physicians<br>interpreting screening mammograms<br>from compatible FFDM system, to<br>identify regions suspicious for breast |
| Technical<br>Characteristics | Subject Device<br>Lunit INSIGHT MMG | Predicate Device<br>Transpara™ (K192287) |
| | characterization of suspicious areas<br>for breast cancer on mammograms<br>from compatible FFDM systems. As<br>an adjunctive tool, the device is<br>intended to be viewed by interpreting<br>physicians after completing their<br>initial read. It is not intended as a<br>replacement for a complete<br>physician's review or their clinical<br>judgement that takes into account<br>other relevant information from the<br>image or patient history. The Lunit<br>INSIGHT MMG uses screening<br>mammograms of the female<br>population. | cancer and assess their likelihood of<br>malignancy. Output of the device<br>includes marks placed on suspicious<br>soft tissue lesions and suspicious<br>calcifications; region-based scores,<br>displayed upon the physician's query,<br>indicating the likelihood that cancer is<br>present in specific regions; and an<br>overall score indicating the likelihood<br>that cancer is present on the<br>mammogram. Patient management<br>decisions should not be made solely<br>on the basis of analysis by<br>Transpara™™. |
| Device output<br>in case of<br>positive<br>detection | Output of the device includes marks<br>with visualized map such as heatmap<br>placed on suspicious lesions for<br>breast cancer; lesion score,<br>indicating the likelihood of presence<br>of the malignancy in specific regions;<br>and an abnormality score indicating<br>the likelihood of the presence of<br>malignancy per-breast in the<br>mammogram. The per-breast<br>abnormality score takes maximum<br>score from two views (i.e., CC, MLO)<br>of a breast which is the maximum<br>lesion score in the breast. | Output of the device includes marks<br>in outlines placed on suspicious soft<br>tissue lesions and suspicious<br>calcifications; region-based scores,<br>displayed upon the physician's query,<br>indicating the likelihood that cancer is<br>present in specific regions; and an<br>overall score indicating the likelihood<br>that cancer is present on the<br>mammogram. |
| Score | Finding level: | Finding level: |
| Technical<br>Characteristics | Subject Device<br>Lunit INSIGHT MMG | Predicate Device<br>Transpara™ (K192287) |
| | Score 1-100 indicating likelihood of<br>presence of malignancy (from low<br>suspicion to high suspicion). | Continuous score 1-100 indicating<br>the level of suspicion of malignancy<br>(from low suspicion to high<br>suspicion). |
| | Breast level:<br>Score 1-100 indicating the level of<br>suspicious of malignancy.<br>The maximum scores from two view<br>(i.e., CC, MLO) of a breast which is<br>the maximum lesion score in the<br>breast | Breast level: None |
| | Exam level: None | Exam level: 10-point scale score<br>indicative of higher frequency of<br>cancer positive |
| Performance | Reader study:<br>• 240 cases<br>• 12 radiologists<br>cf. Reading time was not measured<br>in this study. | Reader study:<br>• 240 cases<br>• 14 radiologists<br>Reading time:<br>• 146 seconds (unaided<br>session) |
| | | • 149 seconds (with<br>Transpara™) |
| | ROC AUC:<br>radiologists AUC (unaided) = 0.754<br>radiologists AUC (aided) = 0.805<br>Standalone Performance Test:<br>• 2412 cases | AUC:<br>• radiologists AUC (unaided) =<br>0.866<br>• radiologists AUC (aided) =<br>0.886 |
| Technical<br>Characteristics | Subject Device<br>Lunit INSIGHT MMG | Predicate Device<br>Transpara™ (K192287) |
| | ROC AUC = 0.903 | standalone AUC = 0.887 |
| Image Source<br>Modality | FFDM | FFDM |
| Fundamental<br>scientific<br>technology | In Lunit INSIGHT MMG, a range of<br>medical image processing and<br>machine learning techniques are<br>implemented. The system includes<br>'deep learning' algorithm applied to<br>images for recognition of suspicious<br>lesions for breast cancer. The<br>machine learning components are<br>trained to detect suspicious lesions<br>for breast cancer with large<br>databases of biopsy-proven cases of<br>breast cancer, benign lesions and<br>normal tissues. | In Transpara, a range of medical<br>image processing and machine<br>learning techniques are implemented.<br>The system includes 'deep learning'<br>algorithm applied to images for<br>recognition of suspicious lesions for<br>breast cancer. The machine learning<br>components are trained to detect<br>suspicious lesions for breast cancer<br>with large databases of biopsy-<br>proven cases of breast cancer,<br>benign lesions and normal tissues. |

## Table 1 – Comparison Between Subject and Predicate Devices

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

{12}------------------------------------------------

{13}------------------------------------------------