FLUOBEAM LX and FLUOBEAM LM are intended to provide real-time near infrared (NIR) fluorescence imaging of tissue during surgical procedures. Upon intravenous administration and use of an ICG consistent with its approved labeling, the FLUOBEAM LX and FLUOBEAM LM are indicated for use in capturing and viewing fluorescent images for the visualization of vessels, blood flow and tissue perfusion before, during and after organ transplant, plastic, micro- and reconstructive surgeries.

The FLUOBEAM LX and FLUOBEAM LM can also be used to assist in the imaging of parathyroid glands and can be used as an adjunctive method to assist in the location of parathyroid glands due to the auto-fluorescence of this tissue.

Use of the FLUOBEAM LX and FLUOBEAM LM devices are intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care. The system is not to be used to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues.

Upon interstitial administration and use of ICG consistent with its approved labeling, the FLUOBEAM LX and FLUOBEAM LM are used to perform intraoperative fluorescence imaging and visualization of the lymphatic system, including lymphatic vessels and lymph nodes.

Upon administration and use of pafolacianine consistent with its approved labeling, the FLUOBEAM LX and FLUOBEAM LM are used to perform intraoperative fluorescence imaging of tissues that have taken up the drug.

FLUOBEAM LX and FLUOBEAM LM are imaging systems intended to provide real-time near infrared (NIR) fluorescence imaging of tissue during surgical procedures.

Class 1 infrared laser light is used to excite the fluorescent tissues of parathyroid glands or the ICG or the pafolacianine and illuminate the regions of a patient's body to be observed. A camera inside the optical head captures the fluorescent image that is used to visualize the parathyroid glands or assess the blood vessels and related tissue perfusion. FLUOBEAM LX and FLUOBEAM LM consist of the following components: a hardware part with a camera unit (optical head) linked by a specific cable to a control box and a software part with FLUOSOFT LX or FLUOSOFT LM imaging software. The optical head contains a video camera and light sources (laser and LEDs) and is used by hand. The control box receives the video signal of the fluorescent image from the optical head, it digitizes it and sends it to a computer that outputs it on a display screen and/or records it. Adjustments of the fluorescent image are possible either by the optical head or via the FLUOSOFT LX imaging software and the FLUOSOFT LM imaging software on the computer.

The subject devices FLUOBEAM LX and FLUOBEAM LM have therefore exactly the same principle of operation of the predicate device. Only aesthetic aspects are different between FLUOBEAM LX and FLUOBEAM LM.

This Traditional 510(k) premarket notification of the FLUOBEAM LX and FLUOBEAM LM is to expand the indication for use statement to include the usage in the lymphatic system with the use of an ICG consistent with its approved label.

This Traditional 510(k) premarket notification of the FLUOBEAM LX and FLUOBEAM LM is to expand the indication for use statement to include the additional cleared infrared dye, pafolacianine, for use with infrared imaging.

The provided FDA 510(k) clearance letter and summary for the FLUOBEAM LX and LM Imaging Systems describe the device and its intended use, as well as the types of studies conducted to demonstrate substantial equivalence to predicate devices. However, the document does not contain a specific table of acceptance criteria nor detailed results of a study designed to explicitly "prove the device meets the acceptance criteria" in terms of clinical performance metrics. Instead, the focus is on demonstrating comparable performance to the predicate and reference devices through bench testing and, by extension, substantial equivalence for the expanded indications.

Based on the information provided, here's a breakdown of the requested elements:

1. Table of Acceptance Criteria and Reported Device Performance

The submission does not explicitly provide a table of acceptance criteria with corresponding reported device performance values in the style typically found in a clinical study report with quantitative metrics like sensitivity, specificity, accuracy, or a specific statistical threshold for performance.

Instead, the document states:

• "The results of these performance evaluations demonstrated that the FLUOBEAM LX and FLUOBEAM LM met the acceptance criteria defined in the product specification, functioned as intended, and performed comparably to the predicate device."
• "The FLUOBEAM LX and FLUOBEAM LM was able to visualize similar concentration samples of ICG compared to the reference device, both by analysis of the image contrast (SNR) and by observation of the images."
• "The FLUOBEAM LX and FLUOBEAM LM was able to visualize lower concentration samples of pafolacianine compared to the reference device, both by analysis of the image contrast (SNR) and by observation of the images."

This indicates that the "acceptance criteria" were qualitative (e.g., "capable of imaging ICG at different concentrations") and comparative (performing "similarly" or "better" than a reference device in terms of visualization capabilities and image contrast/SNR).

Without explicit quantitative acceptance criteria in the document, a direct table cannot be constructed. The reported performance is primarily descriptive and comparative to other devices.

2. Sample Size for the Test Set and Data Provenance

The document describes bench testing for the performance evaluation.

• Sample Size for Test Set: Not explicitly stated as a number of "cases" or "patients." The testing involved "different concentrations" of ICG and pafolacianine in in vitro settings. This implies multiple samples at varying concentrations were used for the bench tests.
• Data Provenance: In vitro bench testing. No mention of human or animal data for the performance evaluation described. Therefore, there is no country of origin or retrospective/prospective designation relevant to clinical data.

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

For the bench testing described, the "ground truth" would be the known concentrations of ICG and pafolacianine and the objective measurement of image contrast (SNR). This type of ground truth does not typically involve human expert interpretation in the same way clinical imaging studies do.

Therefore, no experts were used to establish ground truth in the context of the in vitro performance tests.

4. Adjudication Method for the Test Set

Given that the performance data presented is from in vitro bench testing involving objective measurements of image contrast (SNR) and visual observation of images by the testers/engineers, an adjudication method for a test set (e.g., 2+1, 3+1 by clinical experts) is not applicable or mentioned.

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

No MRMC comparative effectiveness study is mentioned in the provided document. The performance evaluation focuses on the standalone device's capability and its comparison to predicate/reference devices through bench testing, not on human reader performance with or without AI assistance.

6. Standalone Performance (Algorithm Only without Human-in-the-Loop Performance)

The performance evaluation described is implicitly a standalone (algorithm only) assessment. The bench tests evaluated the system's ability to image ICG and pafolacianine at different concentrations and analyze image contrast (SNR) without human interpretation as part of the core performance metric. While "observation of the images" by humans is mentioned, the primary performance measure (SNR) and the ability to visualize specific concentrations are inherent properties of the imaging system itself.

7. Type of Ground Truth Used

For the performance evaluation described:

• Type of Ground Truth: The ground truth used was known concentrations of ICG and pafolacianine for the in vitro imaging tests. The assessment involved comparing the device's ability to visualize these known concentrations and analyzing the Signal-to-Noise Ratio (SNR).

8. Sample Size for the Training Set

The document describes the FLUOBEAM LX and LM as imaging systems, which typically do not involve machine learning algorithms that require a "training set" in the traditional sense. The software updates mentioned likely relate to operational software, not an AI model trained on data.

Therefore, a "training set" is not applicable or mentioned in the context of this device and its clearance.

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

As noted above, a "training set" in the context of machine learning is not applicable for this device as described.

Ask a specific question about this device

The SnapshotGLO is a handheld imaging tool that allows clinicians diagnosing and treating skin wounds, at the point of care, to (i) View and digitally record images of a wound, (ii) Measure and digitally record the size of a wound, and (iii) View and digitally record images of fluorescence emitted from a wound when exposed to an excitation light. The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >104 CFU per gram as compared to examination of clinical signs and symptoms alone. The SnapshotGLO device should not be used to rule-out the presence of bacteria in a wound. The SnapshotGLO does not diagnose or treat skin wounds.

SnapshotGLO is a medical device that operates like a camera. It is a point-of-care, wound imaging device. This device is a non-contact imaging device wherein the wound images are captured from a height of ~12 cm using 395 nm LEDs and a white LED to produce a resultant fluorescence image that aids in visualising the bacteria on the wound and a colour-based "RGB" image or clinical image. Resultant images are viewed on the 7-inch touchscreen display. SnapshotGLO is based on autofluorescence imaging technology and uses native fluorescence of bacteria to determine the presence of bacterial bioburden and displays a two-dimensional, colour-coded highlight of bioburden presence on the wounds. SnapshotGLO is a handheld imaging tool that allows clinicians diagnosing, monitoring and treating skin wounds at the point of care with the help of the following features: . View and digitally record images of a wound, . Measure and digitally record the size of a wound, and View and digitally record images of fluorescence emitted from a wound when exposed to an ● excitation light SnapshotGLO consists of: - SnapshotGLO device - Medical grade adapter - User Manual - Quick Start Guide ● SnapshotGLO is intended for wound care applications as an adjunct tool that uses autofluorescence to detect tissues or structures. The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing high bacteria bioburden compared to clinical symptoms alone. SnapshotGLO should not be used to rule-out the presence of bacteria in a wound. This device is not intended to provide a diagnosis.

The provided text is a 510(k) summary for the SnapshotGLO (KB100) device, aiming to demonstrate its substantial equivalence to the MolecuLightDX as a predicate device. While it details the device's function and provides some study information, it does not explicitly state "acceptance criteria" in a tabulated format alongside "reported device performance." Instead, it discusses the outcomes of non-clinical and clinical studies that support the device's equivalence and performance.

Based on the information provided, here's an attempt to structure the response according to your request, extracting the closest equivalents to "acceptance criteria" and "reported performance" from the study descriptions.

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly define quantitative acceptance criteria for the clinical study. However, the non-clinical tests imply an acceptance criterion of "comparable performance" or "substantially equivalent performance" to the predicate device. For the clinical study, the acceptance was based on showing "improved accuracy" compared to the predicate device when used with clinical signs and symptoms.

2. Sample Size and Data Provenance

• Test Set Sample Size:
  • Clinical Study: 40 patients.
  • Non-clinical Studies: Not explicitly stated, but conducted on "culture plates" and "wound dimensions."
• Data Provenance: The document does not specify the country of origin for the clinical study. The clinical study was described as retrospective.

3. Number of Experts and Qualifications for Ground Truth

• Number of Experts: Two clinical evaluators were involved in the retrospective clinical study.
• Qualifications: Not explicitly stated in the document.

4. Adjudication Method for the Test Set

The document states "This blinded assessment" for the clinical study, indicating that the evaluators were blinded to some information, but it does not describe a specific adjudication method (e.g., 2+1, 3+1 consensus).

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

• Was an MRMC study done? A clinical study comparing the SnapshotGLO and the predicate device was done, involving two clinical evaluators. While it's a comparative study with multiple readers, the format described does not fully align with a typical MRMC study designed to assess reader improvement with AI assistance. It rather compares the device's performance (with CSS) against the predicate device's performance (with CSS) validated against culture.
• Effect Size of Human Readers Improvement with AI vs. Without AI Assistance: The study compared SnapshotGLO + CSS versus MolecuLightDX + CSS versus CSS alone (implicit, as the basis for comparison), all validated against culture results. It demonstrated:
  • SnapshotGLO + CSS accuracy: 75-82.5%
  • MolecuLightDX + CSS accuracy: 52.5-65%
  • The phrase "increase the likelihood that clinicians can identify wounds containing bacterial loads >10^4 CFU per gram as compared to examination of clinical signs and symptoms alone" (from the Indications for Use) suggests that the device, when combined with CSS, improves performance over CSS alone. The specific "effect size" of improvement of human readers with AI vs. without AI assistance (meaning AI as an added tool for human readers) is not directly quantified as a comparative value in terms of reader gain. The comparison shown is between two different devices (both of which are imaging tools that provide additional information to clinicians) when used with CSS, against culture results.

6. Standalone (Algorithm Only) Performance

The document does not report on standalone (algorithm only without human-in-the-loop performance). The indications for use consistently state that the fluorescence image is to be used "in combination with clinical signs and symptoms."

7. Type of Ground Truth Used

• Clinical Study: The ground truth for identifying wounds with bacterial loads >10^4 CFU per gram was established using culture results.
• Non-clinical Studies: The ground truth for bacterial fluorescence detection was based on bacterial presence in in vitro culture plates. For wound dimensions, it was likely based on known or carefully measured dimensions.

8. Sample Size for the Training Set

The document is a 510(k) summary for a medical device and does not provide information regarding the training set sample size as it primarily focuses on the device's performance for regulatory submission. This device description points to an "autofluorescence imaging technology" for directly visualizing bacterial compounds, rather than a machine learning algorithm that requires a training set. If there is an AI component for image processing or interpretation not explicitly detailed, the training set information is not included in this document.

9. How Ground Truth for Training Set Was Established

As no training set is discussed concerning an AI/ML algorithm, no information is provided on how its ground truth was established. The device utilizes physical principles of autofluorescence.

Ask a specific question about this device

FLUOBEAM® LX and FLUOBEAM® LX Red are intended to provide real-time near infrared (NIR) fluorescence imaging of tissue during surgical procedures. Upon intravenous administration and use of an ICG consistent with its approved labeling, the FLUOBEAM® LX and FLUOBEAM® LX Red are indicated for use in capturing and viewing fluorescent images for the visualization of vessels, blood flow and tissue perfusion before, during and after organ transplant, plastic, micro- and reconstructive surgeries.

The FLUOBEAM® LX and FLUOBEAM® LX Red can also be used to assist in the imaging of parathyroid glands and can be used as an adjunctive method to assist in the location of parathyroid glands due to the auto-fluorescence of this tissue.

Use of the FLUOBEAM® LX and FLUOBEAM® LX Red devices are intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care. The system is not to be used to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues.

FLUOBEAM® LX and FLUOBEAM® LX Red are imaging systems intended to provide realtime near infrared (NIR) fluorescence imaging of tissue during surgical procedures. The FLUOBEAM® LX and FLUOBEAM® LX Red are indicated for use in capturing and viewing fluorescent images for the visualization of vessels, blood flow and tissue perfusion before, during and after organ transplant, plastic, micro- and reconstructive surgeries.

FLUOBEAM® LX and FLUOBEAM® LX Red can also be used to assist in the imaging of parathyroid glands and can be used as an adjunctive method to assist in the location of parathyroid glands due to the auto-fluorescence of this tissue. Use of the FLUOBEAM® LX and FLUOBEAM® LX Red devices are intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care. The system is not to be used to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues.

FLUOBEAM® LX and FLUOBEAM® LX Red enable surgeons to observe fluorescent images of parathyroid glands, blood vessels and related tissue perfusion. Fluorescence can be observed as a result of natural fluorescence of parathyroid glands or as a result of a fluorescent product, indocyanine green (ICG), injected intravenously into patients before the surgery allowing the perfusion assessment.

Class 1 infrared laser light is used to excite the fluorescent tissues of parathyroid glands or the ICG and illuminate the regions of a patient's body to be observed. A camera inside the optical head captures the fluorescent image that is used to visualize the parathyroid glands or assess the blood vessels and related tissue perfusion. FLUOBEAM® LX and FLUOBEAM® LX Red consist of the following components: a hardware part with a camera unit (optical head) linked by a specific cable to a control box and a software part with FLUOSOFT™ LX or FLUOSOFT™ LX Red imaging software. The optical head contains a video camera and light sources (laser and LEDs) and is used by hand. The control box receives the video signal of the fluorescent image from the optical head, it digitizes it and sends it to a computer that outputs it on a display screen and/or records it. Adjustments of the fluorescent image are possible either by the optical head or via the FLUOSOFT™ LX imaging software and the FLUOSOFT™ LX Red imaging software on the computer.

This special 510(k) premarket notification is intended to re-frame the indications for use statement to be consistent with specific indications of legally marketed ICG products, eliminating the need for the co-packaging with ICG.

The provided document is a 510(k) Premarket Notification from the FDA, specifically concerning a "Special 510(k)" for the FLUOBEAM® LX and FLUOBEAM® LX Red Imaging Systems.

Crucially, this document states: "No performance data are needed to support the modified indications for use. As noted above, there are no technological changes associated with the proposed labeling changes. Additionally, no new surgical procedures or tissue types are being referenced in the modified indications for use."

This means the submission is not presenting new performance studies or data to demonstrate the device meets acceptance criteria. Instead, it's leveraging the substantial equivalence to previously cleared devices (K190891 and K230898) and a reference device (K223020) because the changes are limited to refining the "Indications for Use" statement to align with existing ICG product labeling, thereby eliminating the need for co-packaging with ICG.

Therefore, for the information requested in your prompt, based solely on the provided text, we cannot fill in most of the table or answer many of the questions as no new performance study data is presented.

Here's what can be extracted from the document:

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

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

• Sample size: Not applicable, as no new performance study data is presented. The submission relies on substantial equivalence.
• Data provenance: Not applicable.

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 performance study data is presented.

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

• Not applicable, as no new performance study data is presented.

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. This device is an imaging system, not an AI/CADe/CADx device that assists human readers in interpretation or diagnosis. It aids in visualization during surgery. No MRMC study is mentioned.

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

• Not applicable. This is not an algorithm-only device; it's an imaging system where a human observes the generated images.

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

• Not explicitly stated for new data, as none was required. For the original clearance of the predicate devices, ground truth would likely have involved direct visual confirmation during surgery, correlation with anatomical knowledge, and potentially histopathological confirmation where applicable (e.g., for parathyroid tissue). The document reiterates that the device "is not to be used to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues," implying that standard clinical and pathological evaluations remain the definitive ground truth for such aspects.

8. The sample size for the training set:

• Not applicable, as no new performance study data for a machine learning model is presented.

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

• Not applicable, as no new performance study data for a machine learning model is presented.

In summary: This 510(k) submission primarily focuses on a labeling change for an existing device, asserting that no new performance data is needed because "there are no technological changes associated with the proposed labeling changes. Additionally, no new surgical procedures or tissue types are being referenced in the modified indications for use." Therefore, the detailed performance study information requested is not present in this specific FDA clearance document.

Ask a specific question about this device

FLUOBEAM® LX Red is intended to provide real-time near infrared (NIR) fluorescence imaging of tissue during surgical procedures. The FLUOBEAM® LX Red is indicated for use in capturing fluorescent images for the visual assessment of blood flow in adults as an adjunctive method for the evaluation of tissue perfusion, perfused organs, and related tissue-transfer circulation in tissue and free flaps used in plastic, micro- and reconstructive and organ transplant surgeries.

The FLUOBEAM® LX Red can also be used to assist in the imaging of parathyroid glands and can be used as an adjunctive method to assist in the location of parathyroid glands due to the auto-fluorescence of this tissue. Use of the FLUOBEAM® LX Red device is intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care. The system is not to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues.

FLUOBEAM® LX Red is an imaging system intended to provide real-time near infrared (NIR) fluorescence imaging of tissue during surgical procedures. The FLUOBEAM® LX Red is indicated for use in capturing and viewing fluorescent images for the visual assessment of blood flow in adults as an adjunctive method for the evaluation of tissue perfusion, perfused organs, and related tissue-transfer circulation in tissue and free flaps used in plastic, micro- and reconstructive and organ transplant surgeries.

FLUOBEAM® LX Red can also be used to assist in the imaging of parathyroid glands and can be used as an adjunctive method to assist in the location of parathyroid glands due to the autofluorescence of this tissue. Use of the FLUOBEAM® LX Red device is intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care. The system is not to be used to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues.

FLUOBEAM® LX Red enables surgeons to observe fluorescent images of parathyroid glands, blood vessels and related tissue perfusion. Fluorescence can be observed thanks to natural fluorescence of parathyroid glands or thanks to a fluorescent product, indocyanine green (ICG), injected intravenously into patients before the surgery allowing the perfusion assessment.

Class 1 infrared laser light is used to excite the fluorescent tissues of parathyroid glands or the ICG and illuminate the regions of a patient's body to be observed. A camera inside the optical head captures the fluorescent image that is used to visualize the parathyroid glands or assess the blood vessels and related tissue perfusion. FLUOBEAM® LX Red consists of the following components: a hardware part with a camera unit (optical head) linked by a specific cable to a control box and a software part with FLUOSOFT™ LX Red imaging software. The optical head contains a video camera and light sources (laser and LEDs) and is used by hand. The control box receives the video signal of the fluorescent image from the optical head, it digitizes it and sends it to a computer that outputs it on a display screen and/or records it. Adjustments of the fluorescent image are possible either by the optical head or via the FLUOSOFT™ LX Red imaging software on the computer.

The modified device FLUOBEAM® LX Red has therefore exactly the same principle of operation of the predicate device. The modified device FLUOBEAM® LX Red is a device modification of the FLUOBEAM® LX device. Compared to the predicate device FLUOBEAM® LX, change of wavelength and detection range is being implemented for modified device FLUOBEAM® LX Red to improve performance in terms of sensitivity for the location of parathyroid glands due to the auto-fluorescence of this tissue while maintaining sensitivity to indocyanine green (ICG).

The FLUOBEAM® LX Red is an imaging system designed to provide real-time near-infrared (NIR) fluorescence imaging of tissue during surgical procedures. It is indicated for visual assessment of blood flow using indocyanine green (ICG) as an adjunctive method, and for assisting in the location of parathyroid glands due to their autofluorescence.

Here’s a breakdown of the acceptance criteria and the study information based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance:

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

• Test Set Sample Size: 5 patients for the clinical tests on ICG fluorescence imaging.
• Data Provenance: Not explicitly stated, but given it's for a European company (FLUOPTICS SAS, France), the data is likely from a European country. The text does not specify if it was retrospective or prospective, but clinical tests are generally prospective.

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 general indication statement mentions that the device is intended to "assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care," suggesting that expert assessment and histology are the ground truth, but doesn't specify the number or qualifications of experts involved in the study itself.

4. Adjudication method 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:

• No, a multi-reader multi-case (MRMC) comparative effectiveness study involving human readers and AI assistance was not mentioned. This device is an imaging system, not an AI-driven algorithm with a human-in-the-loop component.

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

• This device is an imaging system, not solely an algorithm. Its performance is intrinsically linked to the physical hardware (camera, light source) and software for image acquisition and display. Therefore, a "standalone algorithm only" performance study, as typically understood for AI, is not applicable in this context. The document describes bench tests for image quality and sensitivity, and clinical tests for its imaging capabilities.

7. The type of ground truth used:

• For ICG fluorescence imaging for blood flow assessment: The ground truth is implied by comparison to the predicate device in clinical settings, and by the "visual assessment of blood flow." Ultimately, the standard of care for confirming tissue perfusion and viability would involve clinical observation, surgical assessment, and potentially other diagnostic tools.
• For parathyroid gland location: The ground truth is implied to be "visually identified gland/tissues" by experienced visual assessment and "biopsy with conventional histopathological confirmation per standard of care."

8. The sample size for the training set:

• This information is not provided as the document does not mention any machine learning or AI models with distinct training sets. The device's performance is established through engineering design, bench testing, and limited clinical validation.

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

• This information is not applicable as the document does not mention a training set in the context of machine learning or AI.

Ask a specific question about this device

The MolecuLight i:X is a handheld imaging tool that allows clinicians diagnosing and treating skin wounds, at the point of care, to

(i) View and digitally record images of a wound,

(ii) Measure and digitally record the size of a wound, and

(iii) View and digitally record images of fluorescence emitted from a wound when exposed to an excitation light.

The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >10^4 CFU per gram) as compared to examination of clinical signs and symptoms alone. The MolecuLight i:X device should not be used to rule-out the presence of bacteria in a wound.

The MolecuLight i:X does not diagnose or treat skin wounds.

The MolecuLight i:X Imaging Device is a handheld medical imaging device comprised of a high-resolution color LCD display and touch-sensitive screen with integrated optical and microelectronic components. MolecuLight i:X uses its patented technology to enable real-time standard digital imaging and fluorescence (FL) imaging in wounds and surrounding healthy skin of patients as well as wound area measurements.

The MolecuLight i:X is a handheld imaging tool that helps clinicians identify wounds containing elevated bacterial loads. The provided text outlines the acceptance criteria and the study that supports the device's claims.

1. Table of Acceptance Criteria and Reported Device Performance

The provided documentation does not explicitly state "acceptance criteria" as a set of predefined thresholds that the device had to meet for clearance. Instead, it presents performance metrics from a clinical study, which implicitly serve as evidence for addressing the added labeling claims. The key performance metrics are related to the device's ability to guide wound sampling for detecting bacterial burden.

Derived Acceptance Criteria (based on the device's claims and study outcomes) and Reported Device Performance:

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

• Sample Size for Clinical Test Set: Data from a post hoc retrospective analysis of 78 patients were analyzed. The analysis specifically looked at "all wounds that had two samples obtained in the study."
• Data Provenance: The document does not explicitly state the country of origin. The study was a retrospective analysis.

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

The document does not specify the number or qualifications of experts used to establish the ground truth for the clinical test set. The ground truth for bacterial load was established through "wound sampling" implying microbiological culture results (CFU/g).

4. Adjudication Method for the Test Set

The document does not describe any specific adjudication method for the clinical test set. Given that the ground truth appears to be based on microbiological culture results from wound samples, an adjudication process by experts, as might be seen for image interpretation consensus, would not be directly applicable for this type of ground truth.

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

A Multi-Reader Multi-Case (MRMC) comparative effectiveness study, where human readers interpret images with and without AI assistance, was not explicitly described in the provided text. The clinical performance testing focused on comparing the effectiveness of FL-guided wound sampling versus Standard of Care (SoC) guided sampling to detect bacterial burden, rather than a direct comparison of human reader performance with and without AI assistance in image interpretation. The device's role is presented as a tool to guide sampling, not as an AI for image interpretation by clinicians.

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

The provided information focuses on the device's clinical utility in guiding sampling with a clinician in the loop. While the device itself processes images to show fluorescence, its performance is evaluated in the context of improving a clinical procedure (wound sampling), not as a standalone diagnostic algorithm for automated interpretation. The non-clinical testing mentions a "custom algorithm" to detect red fluorescence, which implies standalone analytical capability, but this is distinct from clinical diagnostic performance.

7. The Type of Ground Truth Used

• Clinical Performance Testing: The ground truth for bacterial load was established by microbiological culture results from wound samples, quantified as Colony Forming Units (CFU) per gram (CFU/g). Specifically, "bacterial loads >10^4 CFU per gram" were the key threshold. Pathogen identification was also a part of the ground truth.
• Non-Clinical Testing: The ground truth for demonstrating red fluorescence production for specific bacterial species was based on sub-culturing on Porphyrin Test Agar (PTA) and direct observation of fluorescence using the MolecuLight i:X.

8. The Sample Size for the Training Set

The document primarily discusses a retrospective analysis for clinical validation and in-vitro testing. It does not mention a separate training set size for any specific AI/algorithm development, as the device's core functionality appears to be based on known autofluorescence properties of bacteria rather than a deep learning model trained on large datasets for complex pattern recognition. The study mentioned ([K191371](https://510k.innolitics.com/search/K191371)) was used to support subsequent claims, implying this may have provided data for initial claims.

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

As no explicit training set for a machine learning algorithm is detailed, the method for establishing ground truth for such a set is not provided. The non-clinical testing described involves culturing specific bacterial species to verify their autofluorescent properties, which would serve as a 'ground truth' for the physical principle the device leverages.

Ask a specific question about this device

The MolecuLightDX is a handheld imaging tool that allows clinicians diagnosing and treating skin wounds, at the point of care, to

• (i) View and digitally record images of a wound,
• (ii) Measure and digitally record the size of a wound, and
• (iii) View and digitally record images of fluorescence emitted from a wound when exposed to an excitation light.

The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >104 CFU per gram as compared to examination of clinical signs and symptoms alone. The MolecuLightDX device should not be used to rule-out the presence of bacteria in a wound.

The MolecuLightDX does not diagnose or treat skin wounds.

The MolecuLightDX Imaging Device is a handheld medical imaging device comprised of a high-resolution color AMOLED display and touch-sensitive screen with integrated optical and microelectronic components. MolecuLightDX uses its patented technology to enable real-time standard digital imaging and fluorescence imaging in wounds and surrounding healthy skin of patients as well as wound area measurements.

This document is a 510(k) summary for the MolecuLightDX device. It describes non-clinical testing performed on the device but does not contain information about clinical studies or acceptance criteria for identifying bacterial loads >10^4 CFU per gram. The summary focuses on established equivalence to a predicate device (MolecuLight i:X K191371) based on similar technological characteristics and compliance with electrical safety and software validation standards, rather than direct performance metrics against clinical acceptance criteria for bacterial load detection.

Therefore, the requested information regarding acceptance criteria and a study proving the device meets those criteria, specifically concerning the identification of bacterial loads, cannot be fully provided from the given text.

However, based on the information provided, here's what can be extracted:

1. Table of Acceptance Criteria and Reported Device Performance:

The document primarily discusses compliance with safety and technical standards rather than specific clinical performance acceptance criteria for bacterial load detection. The key performance claim related to its indication for use is: "The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >10^4 CFU per gram as compared to examination of clinical signs and symptoms alone."

While this is an indication of efficacy, the document does not specify quantitative acceptance criteria (e.g., sensitivity, specificity, or AUC thresholds) for this "increased likelihood" or present a study showing the MolecuLightDX device itself met such criteria. Instead, it refers to its predecessor's demonstrated capability.

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

Not provided in this document for the MolecuLightDX's bacterial load identification capability. The non-clinical testing mentioned includes:

• Standards Compliance Testing
• Software Verification and Validation
• System Verification and Validation
• Accuracy and Inter/Intra Reader Variability Testing of Wound Measurement Function
• Packaging and Transport validation

These tests do not involve clinical test sets for bacterial load identification. The effectiveness claim for bacterial load identification references earlier work/evidence from the predicate device, implicitly assuming the newer device maintains this capability due to similar technology.

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

Not provided in this document.

4. Adjudication method for the test set:

Not provided in this 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:

Not provided in this document. The device is a "handheld imaging tool" and not an AI-driven diagnostic algorithm in the typical sense that would necessitate an MRMC comparative effectiveness study for "AI vs without AI assistance". Its primary function is to enable visualization of fluorescence which clinicians then interpret.

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

Not applicable. The device is explicitly described as a "handheld imaging tool that allows clinicians... to view and digitally record images... The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds..." This indicates a human-in-the-loop system.

7. The type of ground truth used:

Not explicitly stated for the MolecuLightDX in this document regarding bacterial load. However, the indication for use refers to "bacterial loads >10^4 CFU per gram," which implies quantitative microbiology (CFU per gram) as the ground truth method in the referenced studies from which the claim of "increased likelihood" originated (likely from the predicate device's clearance).

8. The sample size for the training set:

Not applicable, as this document does not describe a machine learning algorithm requiring a training set for the bacterial load identification claim. The device is an imaging tool.

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

Not applicable.

Summary based on available information:

The provided document (a 510(k) summary) focuses on demonstrating substantial equivalence to a predicate device (MolecuLight i:X K191371) through non-clinical testing, regulatory compliance, and comparison of technological characteristics. It does not detail a specific clinical study for MolecuLightDX to establish performance against acceptance criteria for identifying bacterial loads. Instead, it leverages the known efficacy claim from its predicate device for its indication of use, implying that the MolecuLightDX, being technologically similar, shares this established capability. The "increased likelihood" claim for identifying bacterial loads >10^4 CFU per gram is presented as a previously established benefit from the predicate device's performance.

Ask a specific question about this device

The MolecuLight i:X is a handheld imaging tool that allows clinicians diagnosing and treating skin wounds, at the point of care, to
(i) View and digitally record images of a wound.
(ii) Measure and digitally record the size of a wound, and
(iii) View and digitally record images of fluorescence emitted from a wound when exposed to an excitation light.

The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >104 CFU per grams as compared to examination of clinical signs and symptoms alone. The MolecuLight i:X device should not be used to rule-out the presence of bacteria in a wound.

The MolecuLight i:X does not diagnose or treat skin wounds.

The MolecuLight i:X Imaging Device is a handheld medical imaging device comprised of a high-resolution color LCD display and touch-sensitive screen with integrated optical and microelectronic components. MolecuLight i:X uses its patented technology to enable real-time standard digital imaging and fluorescence imaging in wounds and surrounding healthy skin of patients as well as wound area measurements.

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 MolecuLight i:X:

The MolecuLight i:X device is an autofluorescence detection device for general surgery and dermatological use. The 510(k) summary details a modification to the device's labeling to clarify the relationship between cyan fluorescence and the increased likelihood of Pseudomonas aeruginosa bacterial loads. The study's purpose is to demonstrate that this additional labeling statement does not raise new questions of safety or efficacy and is supported by existing clinical data.

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria are implicitly tied to the performance metrics of the device in identifying bacterial loads, specifically Pseudomonas aeruginosa and Total Bacterial Load (TBL), as established in the original K191371 clearance. The current submission's goal is to demonstrate that the expanded labeling around cyan fluorescence's association with P. aeruginosa is supported by the data and does not alter the previous safety and effectiveness profile.

The performance is reported in terms of sensitivity, specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV), and Likelihood Ratio for different fluorescence signatures (Cyan, Red, Red or Cyan) in detecting bacterial loads.

Note: The document states "The additional labeling statement does not raise different questions of safety or efficacy. Retrospective analysis has demonstrated the safety and effectiveness of MolecuLight i:X with regards to the additional labeling statement." This implies that the acceptance criteria are met if the performance metrics continue to support the device's utility in identifying bacterial loads and the new labeling is consistent with the observed data.

2. Sample Size and Data Provenance

• Sample Size for Test Set: Data from 350 patients were retrospectively analyzed.
• Data Provenance: The document does not explicitly state the country of origin but refers to "retrospective analysis" of existing clinical study data. Given MolecuLight Inc. is based in Toronto, Canada, and the clinical study was "reported in support of K191371", it is likely the data was collected in either Canada or the US, or potentially a multi-site international study. The study was retrospective.

3. Number of Experts and their Qualifications

The document does not specify the number of experts used to establish the ground truth for the test set, nor does it explicitly state their qualifications. The interpretation of "fluorescence image, when used in combination with clinical signs and symptoms" suggests that the ground truth was established by clinicians based on the convergence of factors, potentially including microbiological culture results (as indicated by CFU/g measurements).

4. Adjudication Method for the Test Set

The document does not describe any adjudication method for the test set. Given that the analysis is "retrospective analysis" of existing clinical study data, the ground truth was likely established as part of the original study design.

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

A Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not performed for this 510(k) submission. This submission is a "Real-World Data" (retrospective analysis) supporting a labeling change for a previously cleared device, not a new device clearance or a comparative effectiveness study with human readers. The document states: "The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >10^4 CFU per grams as compared to examination of clinical signs and symptoms alone." This sentence refers to a finding from the original K191371 clearance, not a new MRMC study in this submission.

6. Standalone (Algorithm Only) Performance

This device, the MolecuLight i:X, is an imaging tool used by clinicians to view and record images of fluorescence. It is not an AI algorithm that generates a diagnosis or interpretation autonomously. Therefore, a standalone (algorithm only) performance study was not applicable or performed in the context of this 510(k). The device provides visual information (fluorescence images) that clinicians interpret in conjunction with clinical signs and symptoms.

7. Type of Ground Truth Used

The ground truth for the test set appears to be microbiological culture results (bacterial loads measured in CFU/g) combined with clinical assessment. The specific phrases "identify wounds containing bacterial loads >10^4 CFU per grams" and "Pseudomonas aeruginosa at Species Specific Levels ≥ 10^4 CFU/q" clearly indicate that quantitative bacterial culture was the definitive ground truth for bacterial presence and load.

8. Sample Size for the Training Set

The document does not describe a training set in the context of an AI/algorithm. The "study" described is a retrospective analysis of previously collected clinical data to support a labeling claim for a medical device. If there was any machine learning involved (which is not directly implied for this device's function as an imaging tool), that would have been part of the original K191371 submission and is not detailed here.

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

As no training set (in the context of an AI/ML algorithm) is described, this question is not applicable based on the provided document. The device's function described (capturing and displaying fluorescence) implies it is a viewing tool, not an AI-powered diagnostic algorithm requiring a training phase for its output beyond the initial development of its optical and imaging capabilities.

Ask a specific question about this device

The MolecuLight i:X is a handheld imaging tool that allows clinicians diagnosing and treating skin wounds, at the point of care, to

• View and digitally record images of a wound, (i)
• Measure and digitally record the size of a wound, and (ii)
• View and digitally record images of fluorescence emitted from a wound when exposed to an excitation light. (iii)

The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >104 CFU per gram as compared to examination of clinical signs and symptoms alone. The MolecuLight i:X device should not be used to rule-out the presence of bacteria in a wound.

The MolecuLight i:X does not diagnose or treat skin wounds.

The MolecuLight i:X Imaging Device is a handheld medical imaging device comprised of a high-resolution color LCD display and touch-sensitive screen with integrated optical and microelectronic components. Moleculight i:X uses its patented technology to enable real-time standard digital imaging and fluorescence imaging in wounds and surrounding healthy skin of patients as well as wound area measurements.

Here's a breakdown of the acceptance criteria and the study that proves the MolecuLight i:X device meets them, based on the provided FDA 510(k) summary:

1. Table of Acceptance Criteria and Reported Device Performance

The core clinical claim for the MolecuLight i:X related to bacterial identification is: "The fluorescence image, when used in combination with clinical signs and symptoms, has been shown to increase the likelihood that clinicians can identify wounds containing bacterial loads >10^4 CFU per gram as compared to examination of clinical signs and symptoms alone."

While the document doesn't explicitly state quantitative acceptance criteria in the "we will achieve X performance" format, the demonstrated performance serves as the evidence for meeting their stated claim. The crucial part of the performance is the improvement in identifying wounds with relevant bacterial loads.

Acceptance Criteria (Implied by Clinical Claim and Study Results)

^1 The document states: "The use of the MolecuLight i:X in combination with clinical signs and symptoms resulted in a <10% increase as compared to CSS alone in the rate in wounds that were incorrectly identified as having bacterial load >10^4 CFU/g, whose resulting bacterial load determined by conventional microbiological analysis was <10^4 CFU/q." This is a specific claim about the false positive rate.

2. Sample Size and Data Provenance

• Sample Size for Test Set: n = 350 wound sites (referred to as "All patients")
  • Microbiology Positive (>10^4 CFU/g): n = 287
  • Microbiology Negative (<10^4 CFU/g): n = 63
• Data Provenance: The document does not specify the country of origin. It indicates it was a "clinical study," implying prospective data collection for this purpose.

3. Number of Experts and Qualifications for Ground Truth

• The document mentions "clinicians" evaluating clinical signs and symptoms (CSS) and then interpreting MolecuLight i:X fluorescence images. However, it does not specify the number or qualifications of these clinicians related to establishing the clinical ground truth for the test set.
• The primary ground truth for bacterial load (>10^4 CFU/g) was "quantitative microbiological analysis," which is a lab-based, objective method, not dependent on expert consensus.

4. Adjudication Method for the Test Set

• The document does not describe an adjudication method for the clinical evaluation of CSS or the interpretation of MolecuLight i:X images. It compares these clinical assessments directly against the quantitative microbiological analysis. This suggests that individual clinicians' interpretations were the data points, rather than a consensus interpretation.

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

• A formal MRMC study is not explicitly mentioned as having been performed. The study evaluates the likelihood that clinicians can identify wounds, comparing performance with and without the device. While it involves multiple clinicians, it's presented as a direct comparison of the combined CSS+iX approach versus CSS alone, rather than a statistical comparison of reader performance improvement.
• Effect Size:
  • The sensitivity for identifying high bacterial loads increased from 15.33% (CSS alone) to 60.98% (CSS + iX). This is a substantial increase, making the device significantly more likely to flag relevant wounds.
  • The specificity decreased from 93.65% (CSS alone) to 84.13% (CSS + iX). This indicates a trade-off where more wounds were incorrectly identified as having high bacterial loads with the device, but the report explicitly states this increase in false positives was "<10%".

6. Standalone Performance (Algorithm Only)

• Not applicable/Not done. The MolecuLight i:X is described as a "handheld imaging tool" used by "clinicians." Its function is to provide an image that clinicians interpret. It is not an AI/algorithm that provides a standalone diagnosis or classification. The "performance" reported is the combination of the device's output and human interpretation ("clinician's interpretation of the MolecuLight i:X fluorescence image and CSS").

7. Type of Ground Truth Used

• The primary ground truth for validation of the bacterial load identification claim was quantitative microbiological analysis (i.e., bacterial culture results defining >10^4 CFU per gram). This is an objective and laboratory-confirmed ground truth, considered a strong reference standard for bacterial burden.

8. Sample Size for the Training Set

• The document does not specify a separate training set or its size. This is typical for a medical device that provides direct imaging for human interpretation, rather than a machine learning algorithm that is "trained" on data. The clinical study described here functions as the pivotal performance validation.

9. How Ground Truth for the Training Set was Established

• As no separate training set or AI/ML training is indicated, this point is not applicable. The ground truth for the clinical study was established by quantitative microbiological analysis of wound samples.

Ask a specific question about this device

Fluobeam® LX is intended to provide real-time near infrared (NIR) fluorescence imaging of tissue during surgical procedures. The Fluobeam® LX is indicated for use in capturing and viewing fluorescent images for the visual assessment of blood flow in adults as an adjunctive method for the evaluation of tissue perfusion, perfused organs, and related tissue-transfer circulation in tissue and free flaps used in plastic, micro- and reconstructive and organ transplant surgeries.

The Fluobeam® LX can also be used to assist in the imaging of parathyroid glands and can be used as an adjunctive method to assist in the location of parathyroid glands due to the auto-fluorescence of this tissue.

Use of the Fluobeam® LX device is intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care. The system is not to be used to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues.

Fluobeam® LX is an imaging system intended to provide real-time near infrared (NIR) fluorescence imaging of tissue during surgical procedures. This device is indicated for use in capturing and viewing fluorescent images for the visual assessment of blood flow in adults as an adjunctive method for the evaluation of tissue perfusion, perfused organs, and related tissuetransfer circulation in tissue and free flaps used in plastic, micro- and reconstructive and organ transplant surgeries.

Fluobeam® LX can also be used to assist in the imaging of parathyroid glands and can be used as an adjunctive method to assist in the location of parathyroid glands due to the autofluorescence of this tissue. Use of the Fluobeam® LX device is intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care. The system is not to be used to confirm the absence of parathyroid tissue or glands and is only to be used to assist in locating visually identified gland/tissues.

Fluobeam® LX enables surgeons to observe fluorescent images of parathyroid glands, blood vessels and related tissue perfusion. Fluorescence can be observed thanks to natural fluorescence of parathyroid glands or thanks to a fluorescent product, indocyanine green (ICG), injected intravenously into patients before the surgery allowing the perfusion assessment.

Class 1 infrared laser light is used to excite the fluorescent tissues of parathyroid glands or the ICG and illuminate the regions of a patient's body to be observed. A camera inside the optical head captures the fluorescent image that is used to visualize the parathyroid glands or assess the blood vessels and related tissue perfusion. Fluobeam® LX consists of the following components: a hardware part with a camera unit (optical head) linked by a specific cable to a control box and a software part with Fluosoft™ LX imaging software. The optical head contains a video camera and light sources (laser and LEDs) and is used by hand. The control box receives the video signal of the fluorescent image from the optical head, it digitizes it and sends it to a computer that outputs it on a display screen and/or records it. Adjustments of the fluorescent image are possible either by the optical head or via the Fluosoft™ LX imaging software on the computer.

The subject device Fluobeam® LX has therefore exactly the same principle of operation of the predicate device. Fluobeam® LX is only a second generation than the predicate Fluobeam® device.

The provided text describes the Fluobeam LX device and its substantial equivalence to a predicate device, but it does not contain specific acceptance criteria or an explicit study proving performance against those criteria in the format requested.

Instead, it refers to:

• Bench testing to support a determination of substantial equivalence, covering homogeneity of excitation illumination, live image quality (spatial resolution and acquisition frame rate), and fluorescence sensitivity.
• Clinical data acquired by independent surgeons in Europe to confirm bench test results.
• Performance and safety testing according to international standards (IEC 60601-1, IEC 60601-1-2, IEC 60825-1).

The document states: "The results of these performance evaluations demonstrated that the Fluobeam® LX met the acceptance criteria defined in the product specification, functioned as intended, and performed comparably to the predicate device." However, the specific acceptance criteria and the detailed results of the studies are not provided in this document.

Therefore, I cannot populate the table or answer all of your questions directly from the provided text.

Here is what I can infer or state based on the given information:

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

2. Sample sized 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 specified.
• Data provenance: Clinical data was acquired by independent surgeons in Europe. It's not explicitly stated if it was retrospective or prospective, but the phrasing "acquired by independent surgeons" suggests a form of prospective or concurrent data collection during clinical use.

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)

• Number of experts: "Independent surgeons" were involved, but the specific number is not provided.
• Qualifications of experts: They are described as "independent surgeons." No further details on their specific qualifications or experience are given.

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

• Not specified. The text only mentions that surgeons "accepted to share the images with Fluoptics to compare the two devices."

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, a multi-reader multi-case (MRMC) comparative effectiveness study focusing on human reader improvement with AI assistance was not described. The device is for real-time fluorescence imaging and is stated to assist, not replace, experienced visual assessment. The comparison was primarily between the new device (Fluobeam LX) and its predicate device (Fluobeam 800 Clinic).

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

• The device is an imaging system intended to "assist, not replace, experienced visual assessment." It creates images for human interpretation, so a standalone "algorithm only" performance study in the sense of an automated diagnostic AI would not be applicable or described in this context.

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

• For the parathyroid gland location, the device is an "adjunctive method to assist in the location of parathyroid glands" and its use "is intended to assist, not replace, experienced visual assessment, and biopsy with conventional histopathological confirmation per standard of care." This implies that the standard of care, including histopathological confirmation (pathology), would serve as the ultimate ground truth for actual parathyroid tissue, but the device assists in locating visually identified glands.

8. The sample size for the training set

• Not specified. This document only mentions "clinical data were also acquired by independent surgeons" (implying a test or validation set) and does not refer to a separate training set for an AI/algorithm in the way modern machine learning devices often do. The Fluobeam LX is an imaging system, and its performance evaluation focused on its imaging capabilities and equivalence to a predicate device, rather than training a predictive model.

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

• Not applicable/Not specified, as a training set for an AI model is not described in this document.

Ask a specific question about this device

The AiBiomed Parathyroid Detection System (PTeye) is an adjunctive tool intended to aid in the identification of parathyroid tissue by confirming parathyroid tissue already visually located by the surgeon.

The Parathyroid Detection (PTeye) System aids surgeons in differentiating parathyroid tissue during common surgical procedures. The handheld probe assembly includes one glass fiber-optic element that emits non-ionizing radiation at 785 nm in the near IR range (NIR) and one fiber optic detector element that collects and transmits the fluorescence emitted by the tissue to a photo detector. The Parathyroid Detection (PTeye) System consists of the following components: 1. A console that includes: An LED display that indicates if the laser is on. A display for visual feedback. A speaker for auditory feedback. 2. A handheld fiber-optic probe assembly that interfaces into the console unit using two unique connectors. One connector plugs into the laser output and the other plugs into the photo detector input (fluorescent signal). 3. A foot pedal attached by a cable to the rear of the unit, used to control power to the laser and initiate data collection. 4. An external power supply and power cord that plugs into the power supply. Tissue detection is based on the ratio of the fluorescent response of parathyroid to thyroid tissue; with the assumption that the fluorescence of thyroid tissue is much lower than parathyroid. During surgery, five thyroid data points are collected by touching thyroid tissue with the probe assembly. The system calculates a baseline median for the thyroid tissue based on those points. The baseline value establishes a reference point for distinguishing parathyroid tissue, which produces a higher level of fluorescence. Once the baseline thyroid value has been calculated, a new operational screen will display to support continuous parathyroid search mode. To operate in this mode, the foot switch must be pressed in order to activate the laser. When laser emission is taking place, the LASER ON LED at the front of the system will illuminate. Responses indicating parathyroid tissue are communicated to the user through a bar graph, a detection percentage, a detection ratio, and audio feedback.

Here's a breakdown of the acceptance criteria and study details for the Parathyroid Detection (Model PTeye) System, based on the provided text:

Acceptance Criteria and Device Performance for PTeye System

Study Details:

1. Sample sizes used for the test set and the data provenance:

   • Test Set (Effectiveness Data): 81 patients (40 from Site A (Vanderbilt), 41 from Site B (Ohio State University Medical Center)).
   • Additional exclusion: 2 patients with secondary hyperparathyroidism (r-SHPT) were excluded from the final effectiveness analysis due to known limitations of the device in these cases, resulting in data from 79 patients for core effectiveness analysis.
   • Total tissues measured: 181 PGs and 546 non-PG tissues (194 thyroid, 116 fat, 119 neck muscle, and 117 trachea).
   • Data Provenance: Prospective, multi-center clinical study conducted in the USA (Vanderbilt University Medical Center, Nashville, TN; Ohio State University Medical Center).

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

   • Experts: 4 surgeons (high-volume endocrine surgeons).
   • Qualifications: Perform >25 thyroid surgeries per year and >15 parathyroid surgeries per year.
     • Surgeon #1: 16 years of experience as a practicing endocrine surgeon.
     • Surgeon #2: 4 years of experience as a practicing endocrine surgeon.
     • Surgeon #3: 16 years of experience as a practicing endocrine surgeon.
     • Surgeon #4: 4 years of experience as a practicing endocrine surgeon.

3. Adjudication method for the test set:

   • Ground truth was established by a combination of:
     • Histological examination: For excised tissues (frozen section or Hematoxylin-Eosin stained). 68 PGs (37.6% of all PG measurements) were confirmed with histology.
     • Expert surgeon's visual assessment: For in-situ tissues not excised, the surgeon's visual identification was used, but only those with "high" or "moderate" confidence were included. "Low confidence" assessments were excluded unless histological validation was obtained. The study was single-blinded, meaning surgeons were blinded to the PTeye device output during their initial visual assessment.

4. 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:

   • Yes, a comparative effectiveness study was done comparing the PTeye's performance to human visual assessment. However, it was not an MRMC study designed to show how much human readers improve with AI assistance (human-in-the-loop). Instead, it compared the PTeye's standalone performance to the surgeons' standalone visual assessment.
   • Comparison of PTeye vs. Surgeon's Visual Determination (validated by histology for excised PGs, excluding r-SHPT patients):
     • Overall PG detection rate (validated by histology):
       • Surgeons: 93.3% (56/60)
       • PTeye: 93.3% (56/60)
     • Overall non-PG tissue identification (excised, assumed as PG, validated by histology):
       • Surgeons: 0.0% (0/5) correctly identified as non-PG.
       • PTeye: 80.0% (4/5) correctly identified as non-PG.
   • (Effect size of human improvement with AI vs. without AI assistance): This specific "human-in-the-loop" effect size (synergy or improvement) was not reported or evaluated in this study, as the surgeons were blinded to the PTeye's output for their initial assessment. The study focused on the PTeye's standalone accuracy compared to visual assessment.

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

   • Yes, the primary effectiveness endpoint measures the standalone performance of the PTeye system in identifying parathyroid glands. The overall accuracy, sensitivity, specificity, PPV, and NPV figures (e.g., 96.3% accuracy) represent the PTeye's performance independently of the surgeon's real-time interpretation of its output (as the surgeon was blinded). The study also directly compared this standalone PTeye performance against the surgeons' standalone visual assessment.

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

   • A combined ground truth was used:
     • Histopathology/Pathology: For excised tissues (frozen section or Hematoxylin-Eosin stained tissue section analysis by pathologists). This was considered the "gold standard" when available.
     • Expert Surgeon Consensus/Opinion: For in-situ tissues not excised, the visual assessment of expert surgeons, expressed with "high" or "moderate" confidence, served as the ground truth.

7. The sample size for the training set:

   • The document does not explicitly state the sample size for a training set. The study described is a clinical study to support the safety and effectiveness of the device, likely using a pre-trained model. If the device's internal model was developed using a training set, that information is not provided in this regulatory document.

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

   • As the training set sample size is not explicitly provided, the method for establishing its ground truth is also not described in this document.

Ask a specific question about this device