Search Results

The device is a diagnostic imaging system that combines Positron Emission Tomography (PET) and X-ray Computed Tomography (CT) systems. The CT component produces cross-sectional images of the body by computer reconstruction of X-ray transmission data. The PET component images the distribution of PET radiopharmaceuticals in the patient body. The PET component utilizes CT images for attenuation correction and anatomical reference in the fused PET and CT images.

This device is to be used by a trained health care professional to gather metabolic and functional information from the distribution of the radiopharmaceutical in the body for the assessment of metabolic and physiologic functions. This information can assist in the evaluation, detection, localization, diagnosis, staging, follow-up, therapeutic planning and therapeutic outcome assessment of (but not limited to) oncological, cardiovascular, neurological diseases and disorders. Additionally, this device can be operated independently as a whole body multi-slice CT scanner.

AiCE-i for PET is intended to improve image quality and reduce image noise for FDG whole body data by employing deep learning artificial neural network methods which can explore the statistical properties of the signal and noise of an input PET image. The AiCE algorithm can be applied to improve image quality and denoising of PET images.

Cartesion Prime (PCD-1000A) V10.8 system combines a high-end CT and a high-throughput PET designed to acquire CT, PET and fusion images. The high-end CT system is a multi-slice helical CT scanner with a gantry aperture of 780 mm and a maximum scan field of view (FOV) of 700 mm. The high-throughput PET system has a digital PET detector utilizing SiPM sensors with temporal resolution of

Here's a breakdown of the acceptance criteria and the study information for the Canon Medical Systems Corporation's Cartesion Prime (PCD-1000A/3) V10.8 with AiCE-i for PET, based on the provided FDA 510(k) summary:

1. Table of Acceptance Criteria and Reported Device Performance

The document primarily describes a series of tests conducted to demonstrate the improvement provided by AiCE-i for PET rather than explicit "acceptance criteria" with defined numerical thresholds. However, we can infer the performance metrics and the reported outcomes from the "Testing" section.

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

• NEMA NU 2-2018 & Phantom Studies (Artifact Check, Quantification Accuracy, Preservation of Quantification): These studies used phantoms (e.g., IEC Body phantom, NEMA NU 2-2018 phantom). The number of phantoms is not specified.
• Clinical Data Check & PSNR Measurements:
  • Test Set Size: 5 patients for visual inspection by physicians.
  • Data Provenance: Clinical data was used. For PSNR measurements, "clinical data of long scan duration that is not used in the DCNN training process" was utilized. The country of origin is not specified but is likely internal data from Canon Medical Systems based on the context of the submission. The studies appear to be retrospective as they involve existing clinical data for evaluation.

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

• Number of Experts: Three (3) physicians.
• Qualifications: "at least 20 years of experience in nuclear medicine".

4. Adjudication Method for the Test Set

The document states: "All three physicians determined that all five AiCE-i for PET images were of diagnostic quality. Overall image quality, image sharpness and image noise were determined to be either improved or significantly improved in AiCE-i for PET images when compared to Gaussian images, with the exception of image noise, where one physician determined that the noise in AiCE-i for PET images is about the same as Gaussian images."

This suggests a consensus or majority opinion approach. While not explicitly stated as "2+1" or "3+1", the fact that "all three physicians determined" diagnostic quality implies a unanimous decision for that criterion. For image quality aspects (sharpness, noise), it acknowledges a single dissenting opinion but emphasizes the overall improvement. So, we can infer a form of consensus-based adjudication, with individual expert opinions recorded.

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

• Was it done? A limited MRMC-style evaluation was done with 3 physicians reviewing 5 cases. However, it was focused on assessing image quality improvements and diagnostic utility, not specifically on comparative effectiveness with vs. without AI assistance in terms of diagnostic accuracy or reader performance metrics.
• Effect size of human reader improvement: The document does not report a specific effect size in how much human readers improve with AI vs. without AI assistance in terms of diagnostic performance metrics (e.g., AUC, sensitivity, specificity). The physicians assessed image quality and diagnostic utility, concluding improvement and diagnostic quality, but not a quantifiable improvement in diagnostic accuracy compared to a baseline without AI.

6. Standalone (i.e., algorithm only without human-in-the-loop performance) Study

Yes, standalone performance was evaluated through various bench tests and phantom studies:

• NEMA NU 2-2018 Image Quality (CRC, BGV, Lung Residual Error measured on phantoms)
• AiCE-i for PET Phantom Artifact Check (visual inspection of phantom images)
• AiCE-i for PET Quantification Accuracy (phantom study)
• AiCE-i for PET Preservation of Quantification (phantom study)
• AiCE-i for PET PSNR Measurements (using clinical data, comparing algorithm output to long scan duration images)
• Statements like "AiCE significantly improved Signal to Noise Ratio, improved quantification at the same noise, reduced the count rates while preserving noise" demonstrate standalone algorithmic performance.

7. Type of Ground Truth Used

• Phantom Studies: The ground truth is the known physical properties and activity distribution within the phantoms.
• Clinical Data (Visual Inspection & PSNR):
  • For the visual inspection by physicians, the ground truth for diagnostic quality and image characteristics was expert consensus/opinion.
  • For PSNR measurements, the ground truth was considered the "long scan duration image," which represents a reference image with higher signal and lower noise due to extended acquisition time, against which the processed images were compared for similarity. This acts as a proxy for an ideal image.

8. Sample Size for the Training Set

The document mentions that "clinical data of long scan duration that is not used in the DCNN training process" was used for PSNR measurements. However, the sample size for the training set itself is not specified in the provided text.

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

The document does not explicitly state how the ground truth for the training set was established. It only implies that a "Deep Learning Artificial Neural Network" (DCNN) was trained. Typically, for such AI systems in medical imaging, training data ground truth is established through:
* Expert annotations/labels: Radiologists or nuclear medicine physicians marking regions of interest, identifying pathologies, or rating image quality.
* Higher quality reference scans: Using longer acquisition times or different imaging modalities as a "gold standard" for what an ideal image should look like for noise reduction and image enhancement tasks (as was partially done for the test set PSNR comparison).

Without further information, the specific method for training ground truth establishment remains undisclosed in this document.

Ask a specific question about this device