Search Results

This device is indicated to acquire and display cross-sectional volumes of the whole body, to include the head. The Aquilion Exceed LB has the capability to provide volume sets. These volume sets can be used to perform specialized studies, using indicated software/hardware, by a trained and qualified physician.

AiCE (Advanced Intelligent Clear-IQ Engine) is a noise reduction algorithm that improves image quality and reduces image noise by employing Deep Convolutional Network methods for abdomen, pelvis, lung, cardiac, extremities, head and inner ear applications.

Aquilion Exceed LB (TSX-202A/3) V10.9 with AiCE-i (Advanced intelligent Clear-IQ Engineintegrated) is a whole body multi-slice helical CT scanner, consisting of a gantry, couch and a console used for data processing and display. This device captures cross sectional volume data sets used to perform specialized studies, using indicated software, by a trained and qualified physician. This system is based upon the technology and materials of previously marketed Canon CT systems.

The provided document is a 510(k) summary for the Canon Medical Systems Corporation's Aquilion Exceed LB (TSX-202A/3) V10.9 with AiCE-i. It discusses the device's substantial equivalence to a predicate device, focusing on an expanded clinical use for the AiCE (Advanced Intelligent Clear-IQ Engine) feature, specifically to include cardiac applications.

Here's an analysis of the acceptance criteria and the study proving the device meets them, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The document primarily focuses on demonstrating that the subject device performs similarly to or better than the predicate device. The acceptance criteria are implicitly tied to maintaining or improving upon the predicate's performance, particularly in image quality and the diagnostic utility of AiCE-reconstructed images.

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

• Sample Size for Test Set: The document does not specify exact sample sizes for the phantom studies (e.g., number of scans, repetitions). For the human review of cardiac images, it states "Representative cardiac diagnostic images," but does not provide a numerical count.
• Data Provenance: The document implies the studies were conducted internally by Canon Medical Systems, given the context of a 510(k) submission. No specific country of origin is mentioned for the data, but the submitter is based in Japan (Canon Medical Systems Corporation, Shimoishigami Otawara-Shi, Tochigi-ken, Japan). The studies are inherently "prospective" in the context of validating the device for its submission.

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

• Number of Experts: For the evaluation of cardiac diagnostic images, only one expert is explicitly mentioned: "reviewed by an American Board Certified Radiologist."
• Qualifications of Experts: The expert was an "American Board Certified Radiologist." No specific tenure or experience level is mentioned beyond being Board Certified.

4. Adjudication Method for the Test Set

• For the phantom studies, no human adjudication is mentioned, as the results are based on objective image quality metrics.
• For the cardiac diagnostic image review, with only one radiologist reviewing, there was no adjudication method described (e.g., no 2+1 or 3+1 consensus process). The assessment was a single expert's confirmation.

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

• No, an MRMC comparative effectiveness study was not explicitly done or described. The document focuses on technical bench testing (phantom studies) and a single expert's review of "representative cardiac diagnostic images" to confirm diagnostic quality. There is no mention of human readers improving with or without AI assistance, or any comparative analysis of human performance. The AiCE itself is a noise reduction algorithm, not a diagnostic AI intended to assist human interpretation directly in terms of detection or diagnosis.

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

• Yes, in essence, standalone performance (algorithm only) was assessed for the AiCE feature. The "Image Quality Evaluation" using phantoms measured the intrinsic performance of the AiCE algorithm (and the overall system) in terms of various image quality metrics (CNR, CT Number Accuracy, etc.). These are objective measurements of the algorithm's output quality, independent of human interpretation.
• The "Dual Energy" phantom studies also represent a form of standalone performance assessment for those features of the system.

7. The Type of Ground Truth Used

• Physiological Phantom Measurements: For the bulk of the image quality and dual-energy assessments, the ground truth was established by known physical properties of phantoms. These phantoms are designed with specific, measurable characteristics (e.g., known electron densities, accurate CT numbers, specific contrast elements) that serve as the "true" values against which the device's measurements are compared.
• Expert Consensus (single expert): For the "representative cardiac diagnostic images," the ground truth regarding "diagnostic quality" relied on the opinion/assessment of a single American Board Certified Radiologist. While not a "consensus" in the multi-reader sense, this expert's judgment served as the ground truth for diagnostic utility in this specific clinical application.

8. The Sample Size for the Training Set

• The document does not provide information on the sample size used for the training set for the AiCE Deep Convolutional Network. Information on training data is typically proprietary and not included in 510(k) summaries unless specifically relevant to substantial equivalence arguments.

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

• The document does not provide information on how the ground truth was established for the training set of the AiCE Deep Convolutional Network. This information is usually part of the internal development and validation of the AI algorithm itself, which predates the specific 510(k) submission for this device modification.

Ask a specific question about this device

This device is indicated to acquire and display cross-sectional volumes of the whole body, to include the head. The Aquilion Exceed LB has the capability to provide volume sets. These volume sets can be used to perform specialized studies, using indicated software/hardware, by a trained and qualified physician.

AiCE (Advanced Intelligent Clear-IQ Engine) is a noise reduction algorithm that improves image quality and reduces image noise by employing Deep Convolutional Network methods for abdomen, pelvis, lung, extremities, head and inner ear applications.

Aquilion Exceed LB (TSX-202A/3) V10.6 with AiCE-i (Advanced intelligent Clear-IQ Engineintegrated) is a whole body multi-slice helical CT scanner, consisting of a gantry, couch and a console used for data processing and display. This device captures cross sectional volume data sets used to perform specialized studies, using indicated software, by a trained and qualified physician. This system is based upon the technology and materials of previously marketed Canon CT systems.

In addition, the subject device incorporates the latest reconstruction technology, AiCE-i (Advanced intelligent Clear-IQ Engine - integrated), intended to reduce image noise and improve image quality by utilizing Deep Convolutional Network methods. These methods can more fully explore the statistical properties of the signal and noise. By learning to differentiate structure from noise, the algorithm produces fast, high quality CT reconstruction. The AiCE algorithm has not been modified or retrained since the previous clearance in the predicate, K192832.

The provided text is a 510(k) summary for the Aquilion Exceed LB (TSX-202A/3) V10.6 with AiCE-i, a Computed Tomography (CT) system. The document focuses on demonstrating substantial equivalence to a predicate device (Aquilion Prime SP with AiCE-i, K192832) primarily through bench testing and phantom studies, rather than a clinical multi-reader, multi-case (MRMC) study with human subjects. Therefore, many of the requested criteria related to clinical study design and human reader performance are not directly addressed in this submission.

Here's a breakdown based on the provided information:

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

The acceptance criteria are implicitly defined by demonstrating "substantially equivalent or improved performance relative to the predicate device" in various image quality metrics and dose reduction, based on phantom studies. Quantitative targets are stated for some metrics.

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

The testing described is primarily bench testing utilizing phantoms. Therefore, the "sample size" refers to the number of phantom acquisitions and measurements, not patient data. The provenance is internal to the manufacturer's testing facility, as it's not a clinical study. The data is prospective in the sense that the tests were specifically conducted for this submission.

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

Not applicable for this submission. The ground truth for phantom studies is established by the physical properties of the phantom and known input parameters. No human experts were used for ground truth establishment as it was not a clinical reading study.

4. Adjudication method for the test set

Not applicable. Since the measurements are quantitative from phantom studies, human adjudication is not part of the process.

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 comparative effectiveness study involving human readers was not reported in this 510(k) summary. The submission explicitly states: "Representative clinical images were not necessary to demonstrate substantial equivalence of the subject device." The focus was on engineering performance demonstrated through phantom studies.

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

Yes, the performance of the AiCE algorithm (as integrated into the CT system, referred to as "AiCE-i") was evaluated in a standalone manner using phantom studies. The results in the table above demonstrate the algorithm's impact on image quality metrics and noise reduction. The AiCE algorithm itself was not modified or retrained since its previous clearance (K192832), indicating its performance characteristics are presumed stable.

7. The type of ground truth used

The ground truth used for these performance tests was phantom-based. This includes:

• Physical properties of the phantoms (e.g., known material densities, lesion sizes, contrast values).
• Quantitative measurements derived from the phantom scans (e.g., comparing reconstructed values to known phantom values).
• Comparison to gold standards for image quality metrics (e.g., ideal MTF, NPS curves).

8. The sample size for the training set

The document states that "The AiCE algorithm has not been modified or retrained since the previous clearance in the predicate, K192832." This means the training of the AiCE algorithm itself was done prior to the submission for the predicate device. The actual sample size for the training data is not provided in this specific 510(k) summary (K203042). It would have been part of the K192832 submission.

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

The document does not detail how the ground truth for the training data of the AiCE algorithm was established, as the algorithm itself was not retrained for this submission. This information would typically be found in the original submission for the AiCE algorithm (K192832). Generally, for deep learning algorithms in medical imaging, ground truth for training data is established through a combination of expert consensus, high-quality reference scans, or other validated methods, but this is not specified here.

Ask a specific question about this device