Search Results

HYPER AiR is an image processing function intended to be used by radiologists and nuclear medicine physicians to reduce noise and improve contrast of fluorodeoxyglucose (FDG) PET images.

HYPER AiR is a software-only device. HYPER AiR is an image reconstruction technique which incorporates pre-trained neural networks in the iteration reconstruction process to control image noise and contrast. It is intended to be implemented on previously cleared PET/CT devices uMI 550 (K193241) and uMI 780 (K172143). HYPER AiR serves as an alternative to the existing image reconstruction algorithm that are available on the predicate devices.

The provided text describes the 510(k) summary for the HYPER AiR device, a software-only image processing function for FDG PET images. Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the information provided:

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria are implicitly defined by the performance tests and clinical image evaluation described. The device's performance is reported in terms of improvement over the conventional OSEM (Ordered Subset Expectation Maximization) algorithm.

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

• Test Set Sample Size: The exact number of cases or images in the test set for the clinical image evaluation is not specified. It only states "The clinical image evaluation was performed by comparing HYPER AiR with OSEM."
• Data Provenance: The raw datasets used for evaluation were "obtained on UH's uMI 780 and uMI 550," which are devices from United Imaging Healthcare. The country of origin of this data is not explicitly stated, but given the sponsor's location (Shanghai, China), it can be inferred that the data likely originated from China. The data was retrospective as it involved applying two different reconstruction algorithms (HYPER AiR and OSEM) to the identical raw datasets already obtained.

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

• Number of Experts: "Each image was read by three board-certified nuclear medicine physicians."
• Qualifications of Experts: "board-certified nuclear medicine physicians." No specific years of experience are mentioned.

4. Adjudication Method for the Test Set

The adjudication method is not explicitly stated. It says "Each image was read by three board-certified nuclear medicine physicians who provided an assessment of image contrast, image noise and image quality." It does not describe how discrepancies among the three readers were resolved or if a consensus mechanism was used.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done, and Effect Size

• MRMC Study: A comparative evaluation was done with human readers comparing HYPER AiR reconstructed images to OSEM reconstructed images. This is akin to an MRMC study if the multiple readers evaluated the same cases under both conditions.
• Effect Size: The document states that "HYPER AiR produces images with better image contrast and lower image noise than OSEM while the image quality was sufficient for clinical diagnosis." However, a quantitative effect size (e.g., statistical significance of improvement, specific metrics like AUC difference, or reader confidence scores) is not provided in this summary. It's a qualitative statement of improvement. The study focuses on the standalone performance of the image processing rather than human readers improving with AI assistance vs without, although improved image quality implies potential for human improvement.

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

Yes, standalone performance was evaluated through "Engineering bench testing" where the "evaluation and analysis used the identical raw datasets obtained on UH's uMI 780 and uMI 550, and then applies both HYPER AiR and OSEM to do image reconstruction. The resultant images were then compared for: Performance on noise reduction, Performance on image contrast, Performance on contrast to noise ratio." The aim was to show HYPER AiR's intrinsic ability to improve image characteristics compared to OSEM.

7. The Type of Ground Truth Used

The ground truth used for the evaluation was expert consensus/reader assessment by three board-certified nuclear medicine physicians for the clinical image evaluation. For the non-clinical bench testing, the "ground truth" was essentially the quantitative improvement in objective image metrics (noise reduction, contrast, CNR) based on the algorithm's output compared to OSEM. This is not a "true" clinical ground truth like pathology, but rather a technical performance measure.

8. The Sample Size for the Training Set

The document does not specify the sample size used for the training set of the neural networks integrated into HYPER AiR. It only mentions "pre-trained neural networks."

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. It merely states that HYPER AiR "incorporates pre-trained neural networks in the iteration reconstruction process to control image noise and contrast."

Ask a specific question about this device

HYPER Focus can be used to correct respiratory motion in PET images. Relative to non - corrected images, HYPER Focus can reduce respiratory motion effects and thus improve the measurement accuracy of SUV and lesion volume.

HYPER Focus is a software-only device. It is intended to be implemented on previously cleared PET/CT devices uMI 550 (K193241) and uMI 780 (K172143). HYPER Focus serves as an additional function for uMI 550 and uMI 780 to carry the respiratory correction. It uses the similar respiratory motion correction technique, non-rigid image registration, as the predicate device.

The provided text describes the regulatory clearance of a medical device called "HYPER Focus" (K210418), a software-only device designed to correct respiratory motion in PET images.

Based on the information provided, here's a breakdown of the acceptance criteria and the study that proves the device meets them:

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

The document does not explicitly state a table of quantifiable acceptance criteria with corresponding performance metrics like a typical clinical study report would. Instead, the acceptance criteria are implicitly tied to the device's ability to achieve "substantial equivalence" to a predicate device (GE Q.Freeze software, K113408) in terms of its ability to reduce respiratory motion effects and improve the accuracy of SUV and lesion volume.

The reported device performance is described qualitatively as:

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: The document does not specify a numerical sample size for the test set. It mentions "identical raw datasets obtained from UIH's uMI 780 (K172143) and uMI 550 (K193241)." This suggests that existing datasets were used, but the quantity of these datasets or individual patient cases is not provided.
• Data Provenance: The data was obtained from UIH's uMI 780 and uMI 550 devices. Given that the company, Shanghai United Imaging Healthcare Co., Ltd., is based in China, it is highly probable that the data originated from China. The document does not explicitly state whether the data was retrospective or prospective, but given they are "identical raw datasets obtained" and "existing data" for bench testing, it strongly implies retrospective use of previously acquired patient data.

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

This information is not provided in the document. The document describes "engineering bench testing" and "performance verification" using "identical raw datasets," which suggests a technical analysis rather than an expert-read clinical study to establish ground truth for the test set.

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

This information is not provided in the document. The study described appears to be a technical bench test comparing reconstructed images with and without motion correction, rather than a reader study requiring adjudication.

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, an MRMC comparative effectiveness study involving human readers is not described in this document. The study focuses on the device's quantitative performance (SUV and lesion volume accuracy) and its ability to reduce motion effects in comparison to non-corrected images, and on demonstrating substantial equivalence to a predicate device.

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

Yes, a standalone performance evaluation was done. The "Performance Verification" section states: "Engineering bench testing was performed to support substantial equivalence and product performance claims. The evaluation and analysis used the identical raw datasets obtained from UIH's uMI 780 (K172143) and uMI 550 (K193241), and then respectively performed image reconstruction with/without HYPER Focus." This indicates that the algorithm's performance was assessed independently of human interpretation.

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

The document primarily refers to "bench testing" and "analysis" of quantitative metrics like SUV (Standardized Uptake Value) and sphere/lesion volume accuracy. The "ground truth" for this type of test is typically based on:

• Known physical properties of phantoms: For sphere volume and potentially SUV accuracy, phantom studies with known dimensions and activity concentrations are commonly used. While not explicitly stated, "bench test" often implies phantom studies.
• Comparison to "ideal" or "reference" motion-corrected images: The document states a comparison "in comparison with no motion correction." This implies an assessment against a baseline reference, where the ground truth is the improved accuracy obtained by the algorithm. For motion correction, perfect motion-free images are the ideal ground truth, which is often approximated or modeled.
• The document implies that the "ground truth" for proving efficacy is the demonstrated improvement in SUV and lesion volume accuracy and reduction of motion effects when HYPER Focus is applied, compared to images without motion correction.

The document does not suggest the use of expert consensus, pathology, or outcomes data as a ground truth for this particular submission, which is focused on validating the technical performance of motion correction software for PET images in the context of substantial equivalence.

8. The sample size for the training set

The document does not provide any information about the sample size used for the training set of the HYPER Focus algorithm.

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

The document does not provide any information about how the ground truth for the training set was established. It focuses solely on the performance verification (testing) of the final algorithm.

Ask a specific question about this device

HYPER DLR is an image processing function intended to be used by radiologists and nuclear medicine physicians to reduce noise of the fluorodeoxyglucose (FDG) PET images.

HYPER DLR is a software-only device. HYPER DLR is intended to be implemented on previously cleared PET/CT devices uMI 550 (K182237) and uMI 780 (K172143). HYPER DLR serves as an alternative to the existing image smoothing options that are available on the predicate devices. HYPER DLR is an image post-processing technique which uses a pre-trained neural network to predict low noise PET image from high noise PET image. After training, the network could extract the noise component from the image, thus reducing the image noise.

Here's a breakdown of the acceptance criteria and the study information for the HYPER DLR device, based on the provided FDA 510(k) summary:

Acceptance Criteria and Reported Device Performance

The document describes the device performance in qualitative terms and through common quantitative metrics, but does not provide a specific table of numerical acceptance criteria with corresponding performance values for clinical image evaluation. The acceptance for clinical evaluation is described as:

For non-clinical (bench) testing, the document states: "Bench test showed overall image quality improvement based on the commonly used quantitative metrics. HYPER DLR can significantly improve SNR and CNR while preserving image consistency." The specific acceptance values for "significant improvement" are not detailed in this summary. The quantitative metrics evaluated include:

• Peak signal to noise ratio
• Structural similarity index
• Pearson correlation coefficient
• Signal to noise ratio (SNR)
• Contrast to noise ratio (CNR)
• Normalized root mean square error
• Bland-Altman plot of body & brain VOI SUVmean values

Study Details

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

   • Test Set Sample Size: Not explicitly stated as a number of cases or images. The document mentions "clinical image evaluation were performed for typical clinical scan times of uMI 550 and uMI 780 systems." This implies a set of clinical images, but the exact count is not provided.
   • Data Provenance: Not explicitly stated. The manufacturer is Shanghai United Imaging Healthcare Co., Ltd. in China, so it's plausible the data originates from studies conducted there or affiliated sites. The document does not specify if the data was retrospective or prospective.

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

   • Number of Experts: Not explicitly stated as a specific number. The document mentions "Each image was read by board-certified nuclear medicine physicians." It implies multiple physicians but doesn't specify how many.
   • Qualifications: Board-certified nuclear medicine physicians.

3. Adjudication method for the test set:

   • The document does not explicitly describe an adjudication method for disagreements among the physicians. It states that physicians "provided an assessment," implying individual assessments that were then analyzed.

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:

   • A multi-reader, multi-case study was done for clinical image evaluation, comparing HYPER DLR images with Gaussian filtered images. However, this study appears to be a standalone reader study comparing two different image processing methods, not an AI-assisted vs. non-AI-assisted human reader study. Therefore, no effect size for human reader improvement with AI assistance is reported because the study design was different. The physicians were evaluating images generated by the AI algorithm (HYPER DLR) versus images generated by conventional post-smoothing (Gaussian filtering).

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

   • Yes, the "Bench test" section describes standalone algorithm performance evaluation using quantitative metrics like SNR, CNR, etc. This solely assesses the algorithm's output characteristics without human interpretation.

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

   • For the clinical image evaluation, the ground truth was based on the "assessment of both image noise and overall image quality" by board-certified nuclear medicine physicians. This could be interpreted as a form of expert consensus or expert opinion on image characteristics rather than a definitive "true positive/negative" ground truth for disease detection, as the device's purpose is noise reduction and image quality improvement.
   • For the bench testing, the ground truth for quantitative metrics would typically be derived from the inherent properties of the phantoms/datasets used for those measurements (e.g., known signal levels, known noise levels).

7. The sample size for the training set:

   • Not explicitly stated. The document mentions that HYPER DLR "uses a pre-trained neural network," but the size of the dataset used for this training is not disclosed in this summary.

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

   • Not explicitly stated in this summary. For deep learning noise reduction, the training process often involves pairs of 'noisy' and 'clean' images, where the 'clean' image serves as the ground truth for the noise reduction task. These 'clean' images might be generated from higher dose acquisitions or simulated to represent noise-free data.

Ask a specific question about this device