(42 days)
AutoSPECT® produces images, which depict the three-dimensional distribution of radiopharmaceutical tracers in a patient. This software is intended to provide enhancements to gamma camera emission image processing by automating previously manual image processing functions, providing manual and automated motion correction, providing enhanced reconstruction algorithms that include resolution recovery, scatter correction, noise compensation, and attenuation correction via application of a transmission dataset.
AutoSPECT® is a software application that produces images, which depict the three-dimensional distribution of radiopharmaceutical tracers in a patient via automatic or manual processing. One or more cardiac SPECT, gated SPECT, or MCD cardiac data sets may be processed automatically using AutoSPECT®. Additionally, one or more non-cardiac SPECT, or MCD data sets may be processed manually. AutoSPECT® contains basic data-processing algorithms that have been cleared previously; in addition to enhanced data reconstruction algorithms that include scatter correction, resolution recovery, map-based attenuation correction, and OSEM (Astonish SPECT) reconstruction.
The AutoSPECT software option may be used on images from a gamma camera system that are DICOM 3.0 compatible. The following data sets may be used:
Cardiac, brain, or other (bone, liver, etc.) SPECT datasets .
Gated SPECT datasets .
Vantage SPECT datasets
SPECT-CT datasets ●
Total Body SPECT datasets .
MCD and MCD-AC datasets .
AutoSPECT® provides the user three options for automatically processing cardiac datasets: AutoAll, ' > Auto Recon, and Auto Reorient. Each option is described in greater detail in the Software Description section, Section 4.
AutoSPECT® also allows the user to process non-cardiac SPECT and MCD datasets. In this case, the operator manually positions the reconstruction limit lines to reconstruct transverse data sets. If necessary, the data set can be reoriented manually by positioning the azimuth, elevation, and twist lines to the desired locations.
In addition, the capability of processing groups of SPECT data sets in a batch mode fashion is provided. Once the operator has selected the datasets and determined the processing method, AutoSPECT® processes the first dataset, followed by all remaining datasets without further interaction from the user.
AutoSPECT® application runs on Microsoft Windows XP Professional environment. The minimum hardware requirements are listed:
. Intel x86/Pentium class processor > 1 GHz ;
Graphics capability must meet or exceed 1280x1024 pixels with 32 bit pixel depth; .
30 GB of disk space (minimum); .
512 MB of DRAM (minimum); .
. 10/100 BaseTX Ethernet interface;
Port capable of supporting a dongle; .
. CD drive- capable of reading and writing;
. 56 Kbps modem (minimum)
Here's a breakdown of the acceptance criteria and study details for the AutoSPECT® device, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The submission claims to expand marketing relative to half-count imaging using Astonish-AC. The acceptance criteria revolve around achieving equivalent or improved diagnostic accuracy, image quality, and interpretive certainty compared to a baseline method (full-time back projection).
| Acceptance Criteria | Reported Device Performance |
|---|---|
| For Astonish (half count density) vs. full-time back projection: | |
| Equivalent diagnostic accuracy (sensitivity, specificity, normalcy) | Achieved: Equivalent diagnostic accuracy (equivalent sensitivity, specificity, and normalcy) |
| Better image quality for perfusion imaging | Achieved: Better image quality for perfusion imaging |
| Improved equivalent interpretive certainty | Achieved: Improved equivalent interpretive certainty |
| For Astonish-AC (Astonish + Attenuation Correction) vs. full-time back projection: | |
| Improved diagnostic accuracy (specificity, normalcy) | Achieved: Improved diagnostic accuracy (improved specificity and normalcy) |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size: 297 patient images.
- Data Provenance: The data was acquired using Philips' imaging systems and AutoSPECT® (with Astonish & Astonish AC). It was a "multi-center evaluation" and used "previously scanned images," indicating it was retrospective and likely from various centers where Philips' equipment was used. The country of origin is not explicitly stated, but Philips Medical Systems (Cleveland), Inc. is the submitting entity.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications
The document does not explicitly state the number of experts or their specific qualifications (e.g., "radiologist with 10 years of experience") used to establish ground truth. It vaguely refers to "multi-center evaluations" and conclusions about "diagnostic accuracy" and "interpretive certainty," implying expert interpretation was involved, but details are missing.
4. Adjudication Method for the Test Set
The document does not specify an adjudication method (e.g., 2+1, 3+1, none) for the test set.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- Was an MRMC study done? The text describes "multi-center evaluations" that compared different processing techniques and their impact on "diagnostic accuracy" and "interpretive certainty." This strongly suggests a comparative effectiveness study, likely involving multiple readers to assess the different image sets. However, it does not explicitly label it an "MRMC study."
- Effect Size of human readers' improvement with AI vs. without AI assistance: The study compared different reconstruction techniques (Astonish, Astonish-AC, and full-time back projection) on images, not directly the improvement of human readers with AI assistance versus without AI assistance. The tools (Astonish, Astonish-AC) are enhancements to the imaging process, which then impact the diagnostic accuracy and interpretive certainty. The benefit is reported for the technique, not explicitly as a reader-AI collaboration enhancement.
6. Standalone (Algorithm Only Without Human-in-the-Loop Performance) Study
The study evaluates the output of the AutoSPECT® software's reconstruction algorithms (Astonish and Astonish-AC) in terms of diagnostic accuracy, image quality, and interpretive certainty. Since these are measures that require human interpretation to determine against clinical outcomes or established diagnoses, the study is not a standalone (algorithm only) performance assessment. It assesses the impact of the algorithm's output on human interpretation.
7. Type of Ground Truth Used
The ground truth used is implied to be expert consensus or established clinical diagnoses/outcomes to determine "diagnostic accuracy," "sensitivity," "specificity," and "normalcy." The document does not specify how this ground truth was definitively established (e.g., based on pathology confirmation for all cases, or long-term follow-up for outcomes).
8. Sample Size for the Training Set
The document does not provide any information about the sample size used for the training set. The study detailed is an evaluation of existing reconstruction techniques and their expanded claims.
9. How the Ground Truth for the Training Set Was Established
Since no information is provided about a training set, the method for establishing its ground truth is not mentioned.
§ 892.1200 Emission computed tomography system.
(a)
Identification. An emission computed tomography system is a device intended to detect the location and distribution of gamma ray- and positron-emitting radionuclides in the body and produce cross-sectional images through computer reconstruction of the data. This generic type of device may include signal analysis and display equipment, patient and equipment supports, radionuclide anatomical markers, component parts, and accessories.(b)
Classification. Class II.