Search Results

AliveCor's Corvair ECG analysis system assists the healthcare professional (HCP) in measuring and interpreting resting diagnostic ECGs for rhythm and morphological information by providing an initial automated interpretation. The interpretation by the analysis program may then be confirmed, edited, or deleted by the HCP. The analysis program is intended for use in the general population ranging from healthy subjects to patients with cardiac abnormalities. Corvair is intended for use by healthcare professionals, or trained personnel in healthcare facilities (e.g. the doctor's office or hospital) and in acute settings.

Corvair analyses should be used only as an adjunct to clinical history, symptoms, and the results of other non-invasive and/or invasive tests. Corvair analyses are considered unconfirmed and must be reviewed by a qualified physician. The provisional automated ECG analysis should not be used for clinical action if it has not been reviewed by a qualified healthcare professional capable of independently interpreting the ECG signal.

Corvair is Software as a Medical Device (SaMD) intended for use by healthcare professionals to analyze a diagnostic-bandwidth ECG. Corvair analyzes a 10-second ECG and provides rhythm analysis, morphological analysis, and ECG interval estimation. Corvair provides 35 separate determinations with 14 rhythm and 21 morphology determinations. Rhythm determinations include Normal Sinus Rhythm, Atrial fibrillation, Atrial flutter, Paced Rhythm, Junctional Rhythm, and Bigeminy, with the modifiers of 1* Degree AV Block, Higher Degree AV Block (including 2nd and 3rd degree AV blocks), Sinus Arrhythmia, Marked Sinus Arrhythmia, Marked Bradycardia, Sinus Tachycardia, and PVCs. Morphology determinations include Intraventricular block (RBBB, LBBB, and Other Intraventricular Block), Hypertrophy (LVH, and RVH), Atrial Enlargement (LAE and RAE), Acute Myocardial Infarction (Anterior MI, Inferior MI, Lateral MI), Old/Previous Myocardial Infarction (Anterior Old MI, Inferior Old MI, Lateral Old MI), Ischemia (Anterior, Inferior, Lateral), Prolonged QT, Paced ECG, Other Morphological Defects (Early Repolarization, Wolff-Parkinson-White Syndrome (WPW)), and Normal or Otherwise Normal. Rhythm and morphology determinations are overlapping. i.e., an ECG could receive multiple rhythm and morphology determinations (e.g., Sinus Rhythm, Acute MI). The device also provides global ECG measurements (PR, ORS, OT, OTcB, OTcF, and Heart Rate). No beat-level analysis is provided by the device. Corvair may fail to detect or misidentify conduction system pacing and demand pacing. Corvair does not detect sinus pause. While Corvair provides PR interval estimation and does detect WPW, it does not have a separate determination of abnormally short PR intervals.

This SaMD provides these capabilities in the form of an Application Program Interface (API) library. Any software or device ("target device") can incorporate the Corvair API library into its device software to provide users with resting ECG analytics. The input ECG is provided by the target device to Corvair, to which the various Corvair algorithms are applied, and outputs generated accordingly. Corvair has a C++ interface and a distributed binary (library), which is used by the target device to statically link to Corvair. Viewing of Corvair's ECG analysis is handled by the target device.

Corvair is intended to be used with standard diagnostic-bandwidth, resting ECG recordings collected using 'wet' Ag/AgCl electrodes with conductive gel/paste. Corvair only requires 4 ECG leads for analysis, specifically, either Leads {I, II, V2, and V4}, or Leads {I, II, V1, and V4}. Compatible devices include resting ECGs from GE Medical Systems® (e.g., K081437, MAC 1600, K110266, MAC 5500, K173830, MAC VU360, etc.), and AliveCor's Impala (K232035). Regardless of the lead configuration, Corvair provides the same set of rhythm, morphological, and interval determinations. Corvair has two modes of operation, Symptomatic Mode, which is used when the pre-test probability for a specific rhythm or morphology is high, and Asymptomatic Mode, which optimizes the PPV, by optimizing the specificity, to detect the various rhythms and morphologies. The target device can choose which lead set and which mode of determinations to utilize based on the target clinical application and the patient's clinical presentation.

Corvair utilizes several deep neural networks (DNNs) for its analysis. These DNNs were trained on a dataset of approximately 1 million 12-Lead ECGs acquired from about 400K clinical patients at the Emory University Hospital over several decades between 1985 and 2010. Each ECG has a physician overread confirmed diagnosis with multiple diagnostic codes. The dataset had a 52%/48% ratio of ECGs from male and female patients, respectively. The average age of the patient was 61.3 ± 16. The dataset included 56% white, 33% African American, 2.2% Asian, 9% other races/ethnicities.

Here's a breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:

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

The document states that Corvair was evaluated against a large set of ECGs and compared its analysis output against a known reference using standard ECG performance metrics. These outputs were evaluated against clinically relevant acceptance criteria. However, the specific numerical acceptance criteria for sensitivity, specificity, PPV, and error margins for interval outputs are not explicitly stated in the provided text. The text only mentions that acceptable performance was demonstrated.

Table of Acceptance Criteria and Reported Device Performance (Summary based on text):

2. Sample size used for the test set and the data provenance

• Test Set Size: The document states that Corvair was evaluated "against a large set of ECGs" and mentions "additional large validation datasets" created from sites independent of the training data. For PR, QRS, QT interval estimation, the Common Standards for Quantitative Electrocardiography Standard Database (CSEDB) and an AliveCor proprietary dataset developed from ECGs collected in a clinical study at the Mayo Clinic's Genetic Heart Rhythm Clinic were used. For Heart Rate and QTcF validation, the AliveCor proprietary dataset was also used. The exact numerical sample sizes for these test sets are not explicitly provided.
• Data Provenance:
  • Training Data: Approximately 1 million 12-Lead ECGs acquired from about 400,000 clinical patients at the Emory University Hospital over several decades (1985-2010).
  • Test Data:
    • CSEDB (Common Standards for Quantitative Electrocardiography Standard Database) - an established public database.
    • AliveCor proprietary dataset from clinical study at Mayo Clinic's Genetic Heart Rhythm Clinic.
    • Additional large validation datasets from sites independent of the training data.
  • Retrospective/Prospective: The Emory University Hospital data (1985-2010), used for training, is retrospective. The Mayo Clinic data used for the proprietary dataset likely has a prospective component if collected specifically for this study, but the text doesn't explicitly state its collection method. The "additional large validation datasets" are not detailed regarding their collection method.
  • Country of Origin: Emory University Hospital and Mayo Clinic are in the United States.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

The document mentions that the training dataset ECGs "each has a physician overread confirmed diagnosis with multiple diagnostic codes." For the test sets, the ground truth is against a "known reference." While it refers to "physician overread confirmed diagnosis" for the training set, it does not explicitly state the number or specific qualifications of experts who established the ground truth for the test set. For CSEDB, the ground truth is part of the established database, which typically involves expert consensus. For the AliveCor proprietary dataset, the text implies a clinical study setting, but details on ground truth establishment by experts are missing.

4. Adjudication method for the test set

The document does not explicitly state the adjudication method used for establishing the ground truth of the test set (e.g., 2+1, 3+1, none). It refers to "physician overread confirmed diagnosis" for the training data and "known reference" for the test data, implying an established ground truth, but the process of its establishment is not detailed.

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

The document explicitly states: "No clinical testing was required or conducted to support a determination of substantial equivalence." This indicates that an MRMC comparative effectiveness study was not performed for this submission. The device is intended to assist healthcare professionals, providing an "initial automated interpretation" that "may then be confirmed, edited, or deleted by the HCP," but its impact on human performance was not part of this specific submission's evidence.

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

Yes, a standalone performance evaluation was done. The nonclinical performance testing sections describe evaluating Corvair's analysis output against a known reference using standard ECG performance metrics (sensitivity, specificity, PPV, mean error, standard deviation of error, mean absolute error). This directly assesses the algorithm's performance without a human in the loop. The device provides its analysis as an API library, which integrates into other software, implying a standalone analysis capability.

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

The ground truth for the training set was based on "physician overread confirmed diagnosis with multiple diagnostic codes." For the test sets, it was against a "known reference." For the CSEDB, it's generally accepted expert consensus reference data. For the proprietary dataset from Mayo Clinic, it likely involves clinical diagnoses and expert review, but the specific process (e.g., expert consensus vs. single physician overread) is not detailed.

8. The sample size for the training set

The training set consisted of approximately 1 million 12-Lead ECGs acquired from about 400,000 clinical patients.

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

The ground truth for the training set was established through "a physician overread confirmed diagnosis with multiple diagnostic codes" for each ECG. This indicates that medical professionals reviewed and confirmed the diagnoses, which were then used as the labels for training the deep neural networks.

Ask a specific question about this device

Impala is intended to record, store, and transfer a 12-lead resting electrocardiogram (ECG). Impala acquires four standard diagnostic-bandwidth leads (Leads I, II, V2, V4, or Leads I, II, V1, V4). The device derives four standard diagnosticbandwidth, Lead-III and unipolar limb leads aVR, aVF and aVL. The device also synthesizes Leads V1 or V2, V3, V5, V6, which are similar to but not identical to the same leads of a standard diagnostic 12-lead. The 4 synthesized chest leads are not intended for diagnostic use and may fail to show important findings limited to those leads. This device is not a substitute for a diagnostic 12-lead ECG and is contraindicated for use in ruling out any condition (including but not limited to certain ischemia/infarcts, Brugada syndrome) for which the diagnosis may be solely dependent on the synthesized leads.

The device also provides ECG measurements and ECG analysis (rhythm and morphological interpretation) using the acquired leads. The automated ECG analysis results are provisional and should not be used for clinical action if it has not been reviewed by a qualified physician capable of interpreting the ECG signal in the context of the patient's condition. The automated analysis may then be confirmed, edited, or qualified physician. ECG analysis should be used only as an adjunct to clinical history, symptoms, and the results of other non-invasive tests.

Impala is intended for use with patients aged 18 years and older. Impala is intended for use by healthcare professionals, or trained personnel in healthcare facilities (e.g. the doctor's office or hospital) and in acute settings.

Impala is a portable 12-Lead resting electrocardiograph (ECG) device that acquires 4 standard diagnostic-bandwidth ECG leads from a patient (I, II, V1 or V2, V4), derives 4 standard diagnostic-bandwidth leads (aVL, aVR, aVF, III), and using software generates the remaining leads (V2 or V1, V3, V5, V6) to create a 12-Lead ECG recording. The device can be used by healthcare professionals (HCPs) to record a reduced lead, diagnostic-bandwidth, resting ECG, where traditional 10 electrode, 12 lead ECG recorders are not practical to administer due to size, time, or need for specialized clinicians to administer. Examples may include physician offices, and remote and field locations.

The Impala hardware consists of the Impala ECG Module that connects to the Patient Lead Wire. The Patient Lead Wire is a single cable that includes five snap-on electrodes. Impala also consists of a mobile software application, the Impala App that executes on a mobile computing platform (MCP), such as an Apple iPhone smartphone. To record an ECG, the user positions standard off-the-shelf (OTS) ECG gel electrodes on the patient and snaps on the connectors in the Patient Lead Wire on to the electrodes. Impala allows for two options for which set of leads are acquired:

1. Lead Set 1: Leads {I, II, V2, and V4}, with electrodes on RA, LA, LL, V2, V4; and
2. Lead Set 2: Leads {I, II, V1, and V4} with electrodes on RA, LA, LL, V1, and V4.

Impala incorporates AliveCor's Corvair (K231010), a software as a medical device, to provide rhythm analysis, morphological analysis, and ECG measurements. Corvair has two modes of operation, Symptomatic Mode, which is used when the pre-test probability for a specific rhythm is high, and Asymptomatic Mode, which optimizes the PPV, by optimizing the specificity, to detect the various rhythms and morphologies. The user can choose which Impala Lead Set and which Corvair mode of determinations to utilize based on the target clinical application.

The Impala device, an electrocardiograph, underwent nonclinical testing to demonstrate its performance and substantial equivalence to a predicate device.

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

The document does not provide a specific table of acceptance criteria with corresponding performance metrics for the synthesized leads. However, it implicitly states an acceptance criterion related to the similarity of synthesized leads to standard 12-lead ECGs and the device's intended use.

| Acceptance Criteria (Implicit) | Reported Device Performance |
| :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Synthesized leads are similar to corresponding standard 12-leads, and their display is not likely to confuse human interpreters, while acknowledging that significant differences may exist and these leads are not for diagnostic use to rule out certain conditions. | Results suggest that the synthesized leads tend to be similar to the corresponding 12-leads, so the display of these leads is not likely to confuse human interpreter. However, significant differences exist in many cases. Thus, the 4 synthesized chest leads are not intended for diagnostic use and may fail to show important findings limited to those leads. Further, this device is contraindicated for the diagnosis of any condition (including but not limited to certain ischemia/infarcts, Brugada syndrome) for which the diagnosis may be solely dependent on the synthesized leads. |

2. Sample size used for the test set and the data provenance:

• Sample Size: The test set for the performance of the synthesized leads utilized data from a proprietary AliveCor database described as consisting of 3,000 12-Lead ECGs. Additionally, two public databases (Physionet PTB-XL and the Common Standards for Quantitative Electrocardiography Standard Database (CSEDB) as referenced in IEC 60601-2-25) were used, though their specific sample sizes for this test are not explicitly stated.
• Data Provenance:
  • Proprietary Dataset: The 3,000 ECGs were acquired from clinical patients at the Mayo Clinic over several decades between 1985 and 2010.
  • Public Databases: Physionet PTB-XL and CSEDB. The origin countries are not specified for the public databases but are generally global or multi-center for such resources.
• Retrospective or Prospective: The proprietary dataset from the Mayo Clinic (1985-2010) is clearly retrospective. The nature of the public databases would also be retrospective.

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):

The document mentions that each ECG in the proprietary dataset "has a physician overread confirmed diagnosis with multiple diagnostic codes." It does not specify the number of physicians or their specific qualifications (e.g., years of experience as cardiologists/ECG interpreters) for establishing the ground truth for the test set ECGs.

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

The term "physician overread confirmed diagnosis" suggests that a physician reviewed the ECGs and assigned diagnoses. However, the exact adjudication method (e.g., single physician, consensus of multiple physicians, tie-breaking by a senior expert) is not specified.

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:

No MRMC comparative effectiveness study involving human readers with and without AI assistance was reported. The testing focused on the bench performance of the synthesized leads and their similarity to standard leads, rather than a human-in-the-loop diagnostic accuracy study. The device's automated ECG analysis (Corvair) is explicitly stated to be "provisional and should not be used for clinical action if it has not been reviewed by a qualified physician." This implies a human-in-the-loop is always required for clinical action, but no study comparing human performance with/without AI assistance was described.

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

Yes, a standalone performance evaluation of the algorithm for lead synthesis was performed. The "Bench performance testing" section specifically describes the evaluation of "the 4 synthesized chest leads generated by Impala from the 4 recorded leads in comparison with the corresponding chest leads of a standard 10-electrode resting 12-lead ECG." This is an algorithm-only evaluation of the lead synthesis component.

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

The ground truth for evaluating the synthesized leads was the corresponding chest leads of a standard 10-electrode resting 12-lead ECG. For the underlying diagnoses of the ECGs in the proprietary dataset, the ground truth was "physician overread confirmed diagnosis with multiple diagnostic codes."

8. The sample size for the training set:

The machine learning model for lead synthesis was trained on a dataset of approximately 110,000 12-Lead ECGs.

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

The training dataset of 110,000 ECGs was "acquired from clinical patients at the Mayo Clinic over several decades between 1985 and 2010. Each ECG has a physician overread confirmed diagnosis with multiple diagnostic codes." This indicates that the ground truth for the training set's clinical characteristics was established by physician overread and confirmed diagnoses. For the lead synthesis part of the training, the direct 12-lead ECG itself would serve as the ground truth against which the synthesized leads were learned.

Ask a specific question about this device

KardiaMobile 6L is intended to record, store and two-channel electrocardiogram (ECG) rhythms. In single channel mode, KardiaMobile 6L can record Lead-I. In two channel mode, KardiaMobile 6L can record Lead-I and Lead-II simultaneously and derive Lead-III and unipolar limb leads aVR, aVF and aVL. KardiaMobile 6L also displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAI platform including detecting the presence of normal sinus rhythm, atrial fibrillation, bradycardia, and others. KardiaMobile 6L is intended for use by healthcare professionals, patients with known or suspected heart conditions and health conscious individuals. The device has not been tested and is not intended for pediatric use.

KardiaMobile 6L is a trans-telephonic (transmission by smartphone) electrocardiogram (ECG) event recorder that records, stores, transfers, and analyzes single-channel or two channel ECG rhythm recordings. KardiaMobile 6L provides output of one or six ECG leads, including Lead I, Lead II, Lead III, aVL, aVR and aVF. The device utilizes the computing power of iOS-based or Android-based devices (referred to as "Mobile Computing Platforms" (MCP) within this 510(k)) to record and analyze ECG signals. KardiaMobile 6L consists of KardiaMobile 6L Hardware (portable small wireless hardware with electrodes) and the Kardia Core app. which is installed on an MCP (i.e., iOS or Android devices). KardiaMobile 6L Hardware uses Bluetooth to transmit the ECG signal from the electrodes to the Kardia Core app on the MCP, which then displays the recorded ECG on the MCP's screen. The device is intended to be used by patients with known or suspected heart conditions and health conscious individuals as well as by healthcare professionals (HCPs) who want to remotely monitor their patient's heart health. The device is available for Over-the-Counter (OTC) as well prescription use. The Kardia Core app provides the complete ECG recording and analysis workflow, from acquisition of the signal from the KardiaMobile 6L hardware, to the display of the ECG and analysis results, to printing of the ECG rhythm strip. The app utilizes KardiaAI (K181823, K201985) to provide ECG analysis, which includes the determinations of Normal Sinus Rhythm, Atrial Fibrillation, Bradycardia, Tachycardia, or Unclassified to OTC users and additionally, Sinus Rhythm determinations (Sinus Rhythm with Wide ORS, Sinus Rhythm with Supraventricular Ectopy, and Sinus Rhythm with Premature Ventricular Contractions) for prescription-use only users.

The document provided is a 510(k) summary for the AliveCor KardiaMobile 6L device. It describes the device, its intended use, and its substantial equivalence to a previously cleared predicate device.

Based on the provided information, the current submission (K220350) is for a minor modification (software reorganization and API addition) to an existing device (K210753, KardiaMobile 6L). Therefore, the study described here is primarily focused on demonstrating that these software changes do not adversely affect the device's safety or effectiveness as previously established.

Here's a breakdown of the requested information based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance:

The document explicitly states that the subject device (K220350) is being compared to its predicate (K210753). The performance criteria are implicitly met by demonstrating that the changes do not raise new questions of safety or effectiveness, as stated in the "Nonclinical Testing Summary" and "Conclusions" sections.

2. Sample size used for the test set and the data provenance:

The document states: "No clinical testing was required for the minor changes made to the software." This implies that the current submission (K220350) did not involve a new clinical test set with human subjects to prove performance, but rather relied on nonclinical testing, specifically software verification activities.

Therefore, information on the sample size for a test set (e.g., number of ECGs, number of patients) or data provenance (country of origin, retrospective/prospective) directly related to a new clinical performance study for K220350 is not provided as it was explicitly deemed not required. The performance is being established by comparing to the predicate and ensuring the software changes did not introduce new risks. The previous predicate devices (K210753, K183319) would have had their own clinical data.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

Not applicable, as no new clinical testing was performed for this specific submission (K220350). The performance claims are based on the substantial equivalence to the predicate, K210753, which would have established its own ground truth using experts.

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

Not applicable, as no new clinical testing was performed for this specific submission (K220350).

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:

Not applicable. The document states "No clinical testing was required for the minor changes made to the software." This submission is not an MRMC study and does not evaluate AI assistance for human readers. It focuses on software reorganization and API addition.

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

The device description indicates that the Kardia Core app utilizes KardiaAI (K181823, K201985) to provide ECG analysis, detecting conditions like Normal Sinus Rhythm, Atrial Fibrillation, Bradycardia, Tachycardia, or Unclassified. This implies a standalone algorithmic analysis component. However, the current submission (K220350) primarily focuses on the software reorganization and API addition and explicitly states no new clinical testing was required. The performance of these AI determinations would have been established in the previous predicate clearances (K181823 and K201985), not in this particular submission.

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

Not directly addressed for this specific submission (K220350), as no new clinical data was collected. For the underlying KardiaAI platform (K181823, K201985), the ground truth for ECG analysis typically involves expert consensus of cardiologists reviewing the ECGs.

8. The sample size for the training set:

The document does not provide details of the training set size, as this submission is not about training a new AI model but rather a software modification to an already cleared device that utilizes existing AI algorithms (KardiaAI K181823, K201985). The training data for those AI algorithms would have been described in their respective 510(k) submissions.

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

Not directly addressed for this specific submission (K220350), as it pertains to the training of the underlying KardiaAI algorithms (K181823, K201985) rather than the software reorganization. For such algorithms, ground truth for training is typically established by expert adjudication/consensus of ECG interpretations by qualified medical professionals (e.g., cardiologists).

Ask a specific question about this device

AliveCor QT Service analyses 30 seconds of a previously acquired electrocardiogram (ECG) from AliveCor designed 6-Lead ambulatory ECG devices analyzed as normal sinus rhythm for QT interval measurements.

AliveCor OT Service is intended for use in a professional medical facility, such as a hospital, clinic, or doctor's office by a qualified health care professional, including trained ECG technician.

AliveCor QT Service is indicated for use on adult patients (older than 18 years). The device has not been tested for and is not intended for pediatric use. The service is not intended for use in life supporting, or sustaining systems, or continuous ECG monitors, or cardiac alarm devices, or OTC use only devices.

AliveCor QT Service is a cloud-based Software as a Medical Device (SaMD) that is used to measure the QT and heart-rate corrected QT ("QTc") interval measurements from electrocardiograms (ECG) recorded from adult patients (older than 18 years) using AliveCordesigned 6-Lead ambulatory ECG devices. AliveCor QT Service provides QT and QTc interval measurements only on ECGs analyzed as Normal Sinus Rhythm by KardiaAI (K201985).

AliveCor QT Service is a prescription (Rx) use only device intended for use by qualified healthcare professionals, including trained ECG technicians. Healthcare professionals can access AliveCor OT Service over the internet from patient management and ECG storage Medical Device Data Systems (MDDS), or from other prescription-use only medical data software devices. These devices provide a previously recorded ECG from an AliveCor-designed 6-Lead ambulatory ECG device, such as the KardiaMobile 6L (K210753), in a compatible format using the AliveCor QT Service's REST-based application program interface (API). REST or REpresentational State Transfer is a software methodology that defines rules for creating web services to access resources over Hypertext Transfer Protocol (HTTP). AliveCor QT Service responds to the analysis request with the following measurements:

• QT interval, measured from the first 30 seconds of the ECG
• Heart-rate corrected QT (QTc) interval based on the Bazett's formula and the Fridericia's ● formula

AliveCor QT Service utilizes various internal algorithms including deep neural networks (DNN) to analyze an ECG to compute the QT interval. AliveCor QT Service also includes algorithms to compute the RR-interval, which is used to provide both the Bazett's and Fridericia corrected OTc intervals. These algorithms were trained and validated on datasets with ECGs from patients representative of the device's intended use. The training included more than 750K ECGs from over 250K patients. The training dataset included more than 200K ECGs with approximately 49% data from females, approximately 70% from subjects who were reported as whites, 3% nonwhites, and 27% from those who did not report their race The validation was conducted on two datasets. The first dataset included more than 34K ECGs with approximately 51% data from females, approximately 80% from subjects who were reported as whites, 3.6% non-whites, and 1 7% from those who did not report their race. The second dataset included 226 ECGs with approximately 60% data from females, approximately 87% from subjects who were reported as whites, 3.2% non-whites, and 10% from those who did not report their race.

Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided FDA 510(k) summary for the AliveCor QT Service:

1. Table of Acceptance Criteria and Reported Device Performance

Note: The FDA 510(k) summary provides a high-level summary of the performance testing. While the acceptance criteria are stated, the precise numerical results for the mean and standard deviation of the interval differences are not explicitly detailed in this public document, beyond the assertion that the device "performs to its specifications" and achieves "substantially equivalent performance."

2. Sample Sizes Used for the Test Set and Data Provenance

• Test Set (Validation Datasets):
  • First Dataset: More than 34,000 ECGs
    • Provenance: Not explicitly stated (e.g., country of origin). Appears to be retrospective, as it's described as a "validation was conducted on two datasets."
    • Demographics: Approximately 51% females, ~80% whites, 3.6% non-whites, and 17% did not report their race.
  • Second Dataset: 226 ECGs
    • Provenance: Not explicitly stated (e.g., country of origin). Appears to be retrospective.
    • Demographics: Approximately 60% females, ~87% whites, 3.2% non-whites, and 10% did not report their race.
• Data Provenance: The document does not specify the country of origin for the validation datasets, only that they contain ECGs from "patients representative of the device's intended use." The testing was "nonclinical" and involved "standards-based and AliveCor proprietary ECG databases." It implies retrospective data collection.

3. Number of Experts Used to Establish Ground Truth for the Test Set and Their Qualifications

The document mentions "expert annotated reference interval from a supine, diagnostic bandwidth, 12-lead ECG" as the ground truth. However, it does not specify the following:

• The exact number of experts used.
• The specific qualifications of those experts (e.g., "radiologist with 10 years of experience" - for ECGs, this would typically involve cardiologists or electrophysiologists).
• The process for their annotation (e.g., independent reviews, consensus).

4. Adjudication Method for the Test Set

The document does not specify an explicit adjudication method (e.g., 2+1, 3+1, none) for the expert annotations used to establish ground truth for the test set. It simply states "expert annotated reference interval."

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

• Was an MRMC study done? No. The document explicitly states: "No clinical testing was required to establish substantial equivalence." This indicates that the study focused on device performance against a predefined ground truth in a non-clinical setting, rather than a human-in-the-loop comparative effectiveness study for human readers.

6. Standalone (Algorithm Only) Performance

• Was a standalone performance study done? Yes. The entire nonclinical testing section describes the performance of the "AliveCor QT Service" (the algorithm) independently against expert-annotated ground truth. The acceptance criteria relate directly to the algorithm's output (mean and standard deviation of interval differences).

7. Type of Ground Truth Used

• Type of Ground Truth: The ground truth used was expert consensus / expert annotated reference interval. Specifically, it states "expert annotated reference interval from a supine, diagnostic bandwidth, 12-lead ECG."

8. Sample Size for the Training Set

• Training Set Sample Size: The training included more than 750,000 ECGs from over 250,000 patients.
  • Demographics: More than 200,000 ECGs within this set had demographic data: ~49% females, ~70% whites, 3% non-whites, and 27% did not report their race.

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

The document states that the algorithms (including deep neural networks) were "trained and validated on datasets with ECGs from patients representative of the device's intended use." However, it does not explicitly detail how the ground truth for the training set was established. It is common practice for such large datasets to be assembled from institutional databases or clinical trials where ECGs might have pre-existing annotations or be reviewed by a panel of experts, but the specific methodology for establishing ground truth for the training set is not provided in this summary.

Ask a specific question about this device

The KardiaMobile Card System is intended to record, store and transfer single-channel electrocardiogram (ECG) rhythms. The KardiaMobile Card System also displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAl platform including detecting the presence of normal sinus rhythm, atrial fibrillation, bradycardia, and others. The KardiaMobile Card System is intended for use by healthcare professionals, patients with known or suspected heart conditions and health-conscious individuals. The device has not been tested and is not intended for pediatric use.

The AliveCor KardiaMobile Card System is a single-channel ambulatory electrocardiogram (ECG) device that is intended to record, store, transfer, display, and analyze single-channel ECG rhythms in an ambulatory setting. The device utilizes the computing power of Apple iOS- and Google Android-based smartphones to obtain and analyze single-channel ECGs. These smartphones are termed Mobile Computing Platforms (MCPs). The device consists of the hardware (that has the electrodes), and the Kardia phone app (installed on an MCP). The same software is implemented in the iOS and Android MCP. In either configuration, the same hardware is used to sense the ECG. The KardiaMobile Card Hardware transmits the ECG signal from the electrode to the Kardia phone app on the MCP to be analyzed and presented to the user. All ECGs are synced with the user's account.

The provided text, K211668, is an FDA 510(k) clearance letter for the AliveCor KardiaMobile Card System. It outlines the device description, indications for use, and a comparison to a predicate device. However, it does not include detailed information regarding clinical study data to demonstrate the device meets acceptance criteria for the KardiaAI platform's specific ECG analysis functionalities (detecting normal sinus rhythm, atrial fibrillation, bradycardia, and others).

The "PERFORMANCE DATA" section primarily describes bench testing related to the hardware modifications (change in data transmission method from ultrasonic acoustics to BLE, and change in hardware material). It lists various IEC and ISO standards for ECG acquisition, transmission, biocompatibility, electrical safety, and electromagnetic compatibility.

Therefore, I cannot provide a complete answer to your request based on the provided text, as it lacks the specific clinical study data (acceptance criteria, reported performance, sample sizes, expert ground truth establishment, MRMC studies, standalone performance, training set details) related to the diagnostic performance of the KardiaAI platform itself.

The document states: "No modifications were made to the Kardia app software's clinical functionalities with respect to ECG acquisition, display, and analysis because of this change since its prior clearance under K191406, K182396 and K201985." This implies that the clinical performance evaluation of the KardiaAI platform was conducted during the clearance of the predicate devices (K191406, K182396, K201985), and this 510(k) focuses on the substantial equivalence of the new hardware (KardiaMobile Card) to those previously cleared systems.

To answer your question fully, information from the 510(k) submissions for K191406, K182396, and/or K201985 would be required, as they would contain the detailed clinical performance studies for the KardiaAI platform.

Based only on the provided K211668 document, I can infer the following about what is not present:

• No acceptance criteria for diagnostic performance: The document does not state specific accuracy, sensitivity, or specificity thresholds for detecting cardiac rhythms.
• No reported device performance for diagnostic accuracy: There are no tables or summaries of clinical performance metrics for the KardiaAI algorithms.
• No sample sizes for a clinical test set (for diagnostic performance): The document only mentions bench testing for hardware.
• No details on data provenance (for diagnostic performance): No information on origin, retrospective/prospective nature of a clinical test set.
• No information on experts for ground truth (for diagnostic performance): No details on number, qualifications, or adjudication methods for ECG interpretations.
• No MRMC study details: The document does not describe any human reader studies.
• No standalone (algorithm-only) performance data (for diagnostic performance): The focus is on hardware changes, not re-evaluation of the AI algorithms.
• No information on training set size or ground truth establishment for the AI algorithms.

In summary, the provided document K211668 does not contain the information requested about the acceptance criteria and the study that proves the device meets diagnostic performance acceptance criteria for the KardiaAI platform. It only addresses the substantial equivalence of the hardware changes for the KardiaMobile Card.

Ask a specific question about this device

KardiaMobile 6L is intended to record, store and two-channel electrocardiogram (ECG) rhythms. In single channel mode, KardiaMobile 6L can record Lead-I. In two channel mode, KardiaMobile 6L can record Lead-I and Lead-II simultaneously and derive Lead-III and unipolar limb leads aVE. KardiaMobile 6L also displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAI platform including detecting the presence of normal sinus rhythm, atrial fibrillation, bradycardia, and others. KardiaMobile 6L is intended for use by healthcare professionals, patients with known or suspected heart conditions and health conscious individuals. The device has not been tested and is not intended for pediatric use.

KardiaMobile 6L is a trans-telephonic (transmission by smartphone) electrocardiogram (ECG) event recorder that records, stores, transfers, and analyzes single-channel or two channel ECG rhythm recordings. KardiaMobile 6L provides output of one or six ECG leads, including Lead I, Lead II, Lead III, aVL, aVR and aVF. The device utilizes the computing power of iOS-based or Android-based devices (referred to as "Mobile Computing Platforms" (MCP) within this submission) to obtain and analyze ECG signals. KardiaMobile 6L consists of KardiaMobile 6L Hardware (portable small wireless hardware with electrodes) and the Kardia App, which is installed on an MCP (i.e., iOS or Android devices). KardiaMobile 6L hardware uses Bluetooth to transmit the ECG signal from the electrodes to the Kardia App on the MCP. KardiaMobile 6L displays thus record their ECG and additionally provides ECG analysis using the KardiaAI (K181823, K201985) ECG analysis suite, which includes the determinations of Atrial Fibrillation, Normal Sinus Rhythm, Bradycardia, Tachycardia, or Unclassified. The device is intended to be used by patients with known or suspected heart conditions and health conscious individuals as well as by healthcare professionals (HCPs) who want to remotely monitor their patient's heart health. Patients can forward the recorded ECG to their HCP, who can review the ECG for rhythm and to measure the QT interval. The device is available for Over-the-Counter (OTC) purchase and for purchase with a prescription. The Kardia App is also comes in an alternate variant, called the KardiaStation App that is exclusively used within hospitals and clinics; this app is identical to the Kardia App with the exception of incorporating patient administration workflows.

Here's a summary of the acceptance criteria and study details based on the provided text, focusing on the expanded use case of QT interval measurement for healthcare professionals:

1. Table of Acceptance Criteria and Reported Device Performance

The provided document specifically addresses the expanded use case for healthcare professionals to measure the QT interval. It compares the KardiaMobile 6L's ability to measure QT intervals against a gold standard 12-lead diagnostic ECG device.

2. Sample Size Used for the Test Set and Data Provenance

• Sample Size: 313 subjects.
• Data Provenance: Not explicitly stated regarding country of origin, but the study design suggests prospective data collection as ECG recordings were "taken comparison" and included "both healthy volunteers and patients suspected of long QT syndrome or other genetic heart disease."

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications

• The document states that a "core lab used to provide precise QT measurements for Thorough QT studies was utilized to measure the QT and heart-rate corrected QTc in each ECGs."
• Number of Experts: Not specified.
• Qualifications of Experts: Implied expertise in QT measurement for Thorough QT studies, indicating highly specialized cardiac electrophysiology expertise.

4. Adjudication Method for the Test Set

• The document mentions that the core lab measured QT and QTc "in a randomized and blinded order, using standard measurement techniques." This suggests a robust, independent assessment, but a specific "adjudication method" (e.g., 2+1, 3+1) is not explicitly stated.

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

• No, a MRMC comparative effectiveness study was not explicitly described for the expanded QT interval measurement use. The study focused on the equivalence of QT measurements from the device compared to a gold standard, not on the improvement of human readers with AI assistance. The "human readers" in this context were the core lab experts establishing ground truth, not users of the device whose performance was being evaluated for improvement with AI.

6. Standalone (Algorithm Only) Performance

• No, an explicit standalone (algorithm only without human-in-the-loop performance) study for QT interval measurement was not described. The study validated that "a healthcare professional can review and measure the QT interval reliably using an ECG recorded using KardiaMobile 6L." While the device records the ECG, the measurement itself is performed by a human expert/core lab. The device's role here is data acquisition for human interpretation.
• It's important to note that the KardiaAI platform (K181823, K201985) does run algorithms for detecting Atrial Fibrillation, Normal Sinus Rhythm, Bradycardia, Tachycardia, or Unclassified rhythms, and these would have had standalone performance studies in their original clearances. However, for the expanded QT interval measurement claim, the study focuses on the reliability of the ECG recording for human measurement.

7. Type of Ground Truth Used

• Expert Consensus / Highly Specialized Core Lab Measurement: The QT and heart-rate corrected QTc measurements were performed by a "core lab used to provide precise QT measurements for Thorough QT studies." This represents a high-level, specialized form of expert consensus and measurement standard.

8. Sample Size for the Training Set

• Not provided. The document focuses on the clinical validation study for the QT interval measurement claim, which is a test set evaluation. Information about the training set for any underlying algorithms is not included in this 510(k) summary.

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

• Not provided. As the training set size is not mentioned, neither is the method for establishing its ground truth.

Ask a specific question about this device

KardiaAI is a software analysis library intended to assess ambulatory electrocardiogram (ECG) rhythms from adult subjects (when prescribed or used under the care of a physician). The device supports analyzing data recorded in compatible formats from any ambulatory ECG devices such as event recorders, or other similar devices. The library is intended to be integrated into other device software. The library is not intended for use in life supporting, or sustaining systems, or ECG monitors, or cardiac alarm, or OTC use only devices.

The KardiaAI library provides the following capabilities:

• Filtering ECG noise,
• Reporting heart rate measurement from ECGs,
• Detecting noisy ECGs.
• Reporting ECG rhythm analysis for the presence of sinus rhythm, atrial fibrillation, bradycardia, for ECGs detected as sinus rhythm, detecting normal sinus rhythm with with wide QRS, sinus rhythm with premature ventricular contractions (PVC), and sinus rhythm with supraventricular ectopy;
• Detecting QRS complexes in an ECG.
• For ECGs detected as sinus rhythm, classifying individual beats as a PVC or non-PVC beat, and
• Generating an average beat from an ECG

The device is not intended for use in patients who have pacemakers, ICDs, or other implanted electronic devices.

KardiaAI is a software library that implements various ECG processing and analysis algorithms. This Software as a Medical Device (SaMD) computes various physiologic parameters from an ECG and provides these capabilities in the form of an Application Program Interface (API) library. AliveCor-designed ECG devices ("target device") incorporate the API library into their device software to enable algorithmic analysis of ECGs to provide analytical capabilities. KardiaAI provides ECG processing functions, including ECG noise filtering and detection of noisy ECGs. It performs rhythm analysis on ECGs, specifically detecting atrial fibrillation, bradycardia, tachycardia and sinus rhythm, which can be further classified as normal sinus rhythm, sinus rhythm with wide QRS, sinus rhythm with premature ventricular contractions (PVCs), and sinus rhythm with supraventricular ectopy. It further provides beat-level annotations, including beat-level ORS locations, and, for sinus rhythm ECGs, PVC/not-PVC annotations. It also provides an average beat ECG representation, and the R-R interval tachogram. Recording and viewing of ECGs and the results of the KardiaAI analyses are to be provided by other AliveCor FDA-cleared devices (i.e., the target devices) into which the API library is incorporated, such as AliveCor's Triangle System (K183319) and KardiaMobile System (K182396).

The provided text describes the KardiaAI, a software analysis library intended to assess ambulatory electrocardiogram (ECG) rhythms. The information regarding acceptance criteria and the study proving the device meets these criteria is fragmented across different sections.

Here's an organized breakdown of the requested information based on the provided document:

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

The document states that "All analysis outputs were found to meet their performance specifications" and "it was found that the subject device demonstrated equivalent performance to the predicate device." However, specific numerical acceptance criteria (e.g., sensitivity, specificity, accuracy thresholds) and their corresponding reported device performance values are not explicitly detailed in the provided text. The table below represents the types of performance claimed to be met, but the precise numerical targets and outcomes are absent.

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

• Sample Size for Test Set: Not explicitly stated. The document mentions an "AliveCor proprietary ECG database" and "databases from the ANSVAAMI EC57" were used for algorithm performance testing. No specific number of ECGs or patients is given for either database.
• Data Provenance:
  • AliveCor proprietary ECG database: No information on country of origin.
  • ANSVAAMI EC57 databases: No information on country of origin.
  • Retrospective or Prospective: Not specified.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

This information is not provided in the document. The method for establishing ground truth is mentioned as "AliveCor proprietary ECG database" and "databases from the ANSVAAMI EC57", but details on expert involvement and qualifications are missing.

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

This information is not provided in the document.

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

A Multi-Reader Multi-Case (MRMC) comparative effectiveness study focused on human readers improving with AI assistance was not described in the provided text. The document refers to "comparative testing" between the subject device and the predicate device's algorithm performance, but this is a comparison of algorithms, not human readers with and without AI assistance.

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

Yes, a standalone algorithm-only performance study was conducted. The "Nonclinical Testing Summary" states: "Specifically, algorithm performance testing was assessed using an AliveCor proprietary ECG database. Additional comparative testing was also performed on databases from the ANSVAAMI EC57. All analysis outputs were found to meet their performance specifications." This indicates testing of the algorithm's performance independent of human intervention.

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

The document does not explicitly state the specific type of ground truth used. It mentions using an "AliveCor proprietary ECG database" and "databases from the ANSVAAMI EC57" for algorithm performance testing. This implies that these databases contained pre-established "ground truth" annotations for the ECGs, but the method by which that ground truth was established (e.g., expert interpretation, comparison to other diagnostic tests) is not detailed.

8. The sample size for the training set

The document does not provide the sample size for the training set. It only mentions the databases used for "algorithm performance testing," which typically refers to evaluation on a test set, distinct from a training set.

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

The document does not provide information on how the ground truth for the training set was established, nor does it explicitly mention details about a training set.

Ask a specific question about this device

The KardiaMobile System is intended to record, store and transfer single-channel electrocardiogram (ECG) rhythms. The KardiaMobile System also displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAl platform including detecting the presence of normal sinus rhythm, atrial fibrillation, bradycardia, and others. The KardiaMobile System is intended for use by healthcare professionals, patients with known or suspected heart conditions and health conscious individuals. The device has not been tested and is not intended for pediatric use.

The KardiaMobile System is a trans-telephonic (transmission by telephone) ECG (electrocardiogram) event recorder that records, stores and transfers single-channel electrocardiogram rhythms. The device utilizes the computing power of Apple iOS- and Google Android-based smartphones to obtain and analyze single-channel ECG. These smartphones are termed Mobile Computing Platforms (MCPs). The device consists of the hardware (that has the electrodes), and the Kardia phone app (installed on an MCP). The same software is implemented in the iOS and Android MCP. In either configuration, the same hardware is used to sense the ECG. The KardiaMobile Hardware transmits the ECG signal from the electrode to the Kardia phone app on the MCP to be analyzed and presented to the user. All ECGs are synced with the user's account.

The provided text describes the acceptance criteria and the study conducted for the KardiaMobile System, primarily focusing on proving that the device meets special controls for Electrocardiograph Software for Over-the-Counter Use, especially after the removal of a "clinician overread" function.

Here's a breakdown of the requested information:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not provide a direct table of numerical "acceptance criteria" (e.g., minimum sensitivity/specificity percentages) and corresponding "reported device performance" values for the AI algorithm suite (KardiaAI SaMD). Instead, it states that the performance characteristics (sensitivity and specificity) were "tested to meet the system requirements" against ANSI/AAMI EC57:2012 databases and AliveCor's proprietary databases.

However, the "Special Control" table implicitly functions as acceptance criteria for different aspects of the device's performance and the "Summary of Conformance" column indicates the reported performance/compliance.

Acceptance Criteria (Implicit from Special Controls) and Reported Device Performance (Summary of Conformance):

2. Sample Sizes Used for the Test Set and Data Provenance

• Sample Size for Algorithmic Performance (KardiaAI SaMD): The document states that the KardiaAI SaMD (K181823), leveraged by the KardiaMobile System for ECG analysis, was validated using "ANSI/AAMI EC57:2012 databases and AliveCor's proprietary databases." It also notes that these "validation datasets are representative of the patient population of the proposed device."
  • Specific sample sizes are NOT provided for these databases.
  • Data Provenance: Not explicitly stated regarding country of origin. The use of "ANSI/AAMI EC57:2012 databases" suggests a standardized, likely diverse, source, while "AliveCor's proprietary databases" could be from various global or specific regions. The document does not specify if the data was retrospective or prospective for the algorithmic validation, but typically such databases are compiled retrospectively.
• Sample Size for Human Factors and Usability Testing: Not explicitly stated, but it refers to "representative users."

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

• For Algorithmic Performance (KardiaAI SaMD): The document states that the detection algorithm performance was validated using a "previously adjudicated data set" (Special Control 3) and mentions that for the primary predicate, the "overread unlock" mechanism involved review by a "board-certified cardiologist." However, it does not explicitly state the number or specific qualifications (e.g., years of experience) of experts used to establish the ground truth for the test sets used for the KardiaAI algorithm validation. Adjudicated data implies expert review, but details are absent.

4. Adjudication Method for the Test Set

• For Algorithmic Performance (KardiaAI SaMD): The document mentions "previously adjudicated data set." No specific adjudication method (e.g., 2+1, 3+1) is detailed.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done

• No MRMC comparative effectiveness study, comparing human readers with AI vs. without AI assistance, is mentioned. The focus of the changes and testing described is on the device's standalone performance and human factors/usability for over-the-counter use after removing the "overread" requirement.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

• Yes. The document explicitly states: "The KardiaMobile System leverages the KardiaAI SaMD (K181823) for ECG analysis. KardiaAI algorithm suite ECG detection algorithm outputs of Atrial Fibrillation, Normal, Bradycardia, Tachycardia, and Noise as well as the heart rate calculations were tested to meet the system requirements for sensitivity and specificity." This indicates that the core AI algorithm's performance was evaluated independently (without human-in-the-loop for its direct analytical output).

7. The Type of Ground Truth Used

• For Algorithmic Performance (KardiaAI SaMD): "Previously adjudicated data set." This typically implies expert consensus (e.g., cardiologists reviewing ECGs). It does not mention pathology or outcomes data as the ground truth directly for the AI algorithm's performance.

8. The Sample Size for the Training Set

• Not specified. The document focuses on the validation/test sets (ANSI/AAMI EC57:2012 and AliveCor's proprietary databases) for the KardiaAI algorithm. Information regarding the training set's size is not provided.

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

• Not specified. As the training set size itself is not mentioned, neither is the method for establishing its ground truth.

Ask a specific question about this device

The Triangle System is intended to record, store and two-channel electrocardiogram (ECG) rhythms. In single channel mode, the Triangle System can record Lead-I. In two channel mode, the Triangle System can record Lead-I and Lead-II simultaneously and derive Lead-II and unipolar limb leads aVR, aVF and aVL. The Triangle System also displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAl platform including detecting the presence of normal sinus rhythm, atral fibrillation, bradycardia, and others. The Triangle System is intended for use by healthcare professionals, patients with known or suspected heart conditions and health conscious individuals. The device has not been tested and is not intended for pediatric use.

The Triangle System is a trans-telephonic (transmission by smartphone) electrocardiogram (ECG) event recorder that records, stores, transfers, and analyzes single-channel or two-channel ECG rhythm recordings. The Triangle System provides output of one or six ECG leads by recording Lead-I and Lead-II simultaneously and derive Lead-III and unipolar limb leads aVR, aVF, and aVL for identifying cardiac arrhythmias. The device utilizes the computing power of iOS-based or Android-based smartphones (referred to as "Mobile Computing Platforms" (MCP) within this submission) to obtain and analyze ECG signals. The device consists of the Triangle Hardware (portable small wireless hardware with electrodes) and the Kardia app, which is installed on an MCP (i.e., iPhone or Android phone). The Triangle System is a new addition to AliveCor's ambulatory ECG devices that are intended to be used by lay users to record their ECG and obtain ECG analysis of Atrial Fibrillation. Normal Sinus Rhythm, Bradycardia, Tachycardia, or unclassified file. The Triangle System hardware transmits the ECG signal from the electrodes to the Kardia phone app on the MCP to be analyzed and presented to the user. All ECGs are synced with the user's account. The device is available for Over-the-Counter (OTC) purchase and for purchase with clinician prescription.

Here's an analysis of the provided text, outlining the acceptance criteria and the study that proves the device meets those criteria:

1. Table of Acceptance Criteria and Reported Device Performance:

The document doesn't explicitly state "acceptance criteria" in a numerical or target format for the qualitative and quantitative analyses. Instead, it describes the results of the assessments that demonstrate equivalence.

2. Sample Size for the Test Set and Data Provenance:

• Sample Size: 44 subjects
• Data Provenance: The text states, "Overall, 44 subjects participated in the study, that included nearly equal numbers of healthy volunteers and arrhythmia patients." It does not explicitly state the country of origin or if the study was prospective or retrospective, but the clinical study setup (simultaneous ECG recordings) strongly suggests a prospective design. The subjects appear to be a mix of healthy individuals and those with arrhythmias.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications:

• Number of Experts: Two
• Qualifications of Experts: "Two board-certified electrophysiologists."

4. Adjudication Method for the Test Set:

• Adjudication Method: The text states, "two board-certified electrophysiologists compared 6-lead ECG rhythm strips acquired from the Triangle device and the corresponding leads from the reference standard 12-lead ECG device for diagnostic equivalence." It then concludes, "All paired recordings (100%, n=44 subjects), were deemed equivalent for assessing cardiac arrhythmias by both electrophysiologists." This implies that both experts had to agree for a recording to be deemed equivalent. There is no mention of a tie-breaking third expert, suggesting unanimosity was required for the reported 100% equivalence.

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

• Was an MRMC study done? No. The study described is a comparison of the device's output to a gold standard, not a comparison of human readers' performance with and without AI assistance. While the device utilizes the KardiaAI platform, this specific clinical study focused on the hardware's ability to record equivalent ECG signals, not the AI's impact on human interpretation.

6. Standalone (Algorithm Only Without Human-in-the-Loop Performance) Study:

• Was a standalone study done? Yes, to a degree outlined in the "Validation of KardiaAI integration" under non-clinical testing. However, the specific clinical study described does involve human experts (electrophysiologists) for the qualitative assessment, indicating a human-in-the-loop component for validation of the output. The quantitative assessment is purely algorithmic comparison (cross-correlation and RMS error). The statement, "The Triangle System also displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAI platform including detecting the presence of normal sinus rhythm, atrial fibrillation, bradycardia, and others," suggests the AI provides its own analysis which would be a standalone capability, but the clinical study discussed here is primarily about the signal equivalence of the hardware. The KardiaAI itself would have had standalone performance studies, but those are not detailed here for this K183319 submission.

7. Type of Ground Truth Used:

• Type of Ground Truth: The ground truth for the clinical study was established using a gold standard 12-lead ECG device (for both qualitative and quantitative comparison of ECG signals) and the expert consensus of two board-certified electrophysiologists (for qualitative diagnostic equivalence).

8. Sample Size for the Training Set:

• The document does not provide information regarding the sample size for the training set. The clinical study described involves performance validation of the Triangle System hardware, not training of the KardiaAI algorithms.

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

• The document does not provide information on how the ground truth for the training set of the KardiaAI platform was established.

Ask a specific question about this device

The KardiaMobile System is intended to record, store and transfer single-channel electrocardiogram (ECG) rhythms. The KardiaMobile System also displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAl platform including detecting the presence of normal sinus rhythm, atrial fibrillation, bradycardia, and others (when prescribed or used under the care of a healthcare professional). The KardiaMobile System is intended for use by healthcare professionals, patients with known or suspected heart conditions and health conscious individuals. The device has not been tested and is not intended for pediatric use.

The KardiaMobile System is a trans-telephonic (transmission by telephone) ECG (electrocardiogram) event recorder that records, stores and transfers single-channel electrocardiogram rhythms. The device utilizes the computing power of Apple iOS- and Google Android-based smartphones to obtain and analyze single-channel ECG. These smartphones are termed Mobile Computing Platforms (MCPs). The device consists of the hardware (that has the electrodes), and the Kardia phone app (installed on an MCP). The same software is implemented in the iOS and Android MCP. In either configuration, the same hardware is used to sense the ECG. The KardiaMobile Hardware transmits the ECG signal from the electrode to the Kardia phone app on the MCP to be analyzed and presented to the user. All ECGs are synced with the user's account.

The provided documents describe the KardiaMobile System and its substantial equivalence to a predicate device. However, the specific acceptance criteria for the device's performance related to its AI algorithms (atrial fibrillation, normal sinus rhythm, tachycardia, bradycardia, and noise detection) and the detailed study that proves these criteria are met are NOT explicitly detailed within the provided text.

The document states:

• "The KardiaMobile System... displays ECG rhythms and output of ECG analysis from AliveCor's KardiaAI platform including detecting the presence of normal sinus rhythm, atrial fibrillation, bradycardia, and others..."
• "Available Algorithms: Atrial Fibrillation, Noise Algorithm, Normal Sinus Rhythm, Tachycardia, Bradycardia (implements the same algorithms of the KardiaAI reference device, K181823)"
• "All necessary performance testing was conducted on the KardiaMobile System to support a determination of substantial equivalence to the predicate device. This testing included the following: - validation of KardiaAI integration"

While it confirms that "validation of KardiaAI integration" was part of the testing, it does not provide the specific acceptance criteria (e.g., sensitivity, specificity, accuracy targets for each rhythm detection) or the details of the study (sample size, ground truth establishment, expert qualifications, etc.) for the AI algorithms themselves.

Therefore, for aspects related to the performance of the AI algorithms, the requested information cannot be fully extracted from the provided text. The document focuses more on the substantial equivalence of the overall system (hardware and software integration) to a predicate device, rather than a detailed performance study for the AI algorithms against specific statistical targets.

However, based on the information provided for the overall system's substantial equivalence:

1. Table of acceptance criteria and the reported device performance:

The document does not provide a table with specific quantitative acceptance criteria (e.g., sensitivity, specificity, accuracy percentages) for the rhythm detection algorithms. Instead, the acceptance is based on the system demonstrating substantial equivalence to the predicate device (AliveCor Heart Monitor K142672) and the reference device's (KardiaAI K181823) algorithms, and meeting established specifications through nonclinical testing.

Reported Device Performance (General):

• Validation of KardiaAI integration: Performed.
• Verification of the device's specification: Performed.
• Testing to software level of concern requirements: Performed.
• Compliance with standards: ISO 10993-1:2009, IEC 60601-1:2012, IEC 60601-1-2:2007, IEC 60601-1-11:2015, IEC 60601-2-47:2012.
• Conclusion: "The collective results of the performance testing demonstrate that the KardiaMobile System meets the established specifications and complies with the aforementioned standards." and "The evaluation and testing results showed that differences between the subject and predicate device do not raise different questions of safety or effectiveness."

The following information applies to the overall system's validation and substantial equivalence as described, but not specifically to the detailed performance of the AI algorithms against quantifiable targets, which is not provided in the text.

2. Sample size used for the test set and the data provenance:

• Sample Size: Not specified in the provided documents. The text mentions "validation of KardiaAI integration" and "verification of the device's specification" but does not detail the size of the dataset used for these tests.
• Data Provenance: Not specified. It is not mentioned if the data was retrospective or prospective, or the country of origin.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• Not specified in the provided documents. The text does not detail the process of establishing ground truth for any test sets related to the AI algorithms.

4. Adjudication method for the test set:

• Not specified.

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:

• A multi-reader multi-case (MRMC) comparative effectiveness study focusing on human readers improving with AI assistance is not described in the provided documents. The submission focuses on the substantial equivalence of the device, including its AI algorithms, rather than a comparative effectiveness study with human readers.

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

• The document implies that the algorithms were evaluated independently as part of the "KardiaAI integration validation" and reference to "KardiaAI K181823". However, the specific details of such a standalone study (e.g., metrics, test set, ground truth) are not provided for the algorithms themselves. The overall device is described as having "output of ECG analysis from AliveCor's KardiaAI platform," suggesting that the algorithms perform analysis independently before being displayed to the user.

7. The type of ground truth used:

• Not specified in the provided documents.

8. The sample size for the training set:

• Not specified in the provided documents.

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

• Not specified in the provided documents.

Ask a specific question about this device