Search Results

This computed tomography system is intended to generate and process cross-sectional images of patients by computer reconstruction of x-ray transmission data.

The images delivered by the system can be used by a trained staff as an aid in diagnosis and treatment as well as for diagnostic and therapeutic interventions.

This CT system can be used for low dose lung cancer screening in high risk populations*.

• As defined by professional medical societies. Please refer to clinical literature, including the results of the National Lung Screening Trial (N Engl J Med 2011; 365:395-409) and subsequent literature, for further information.

The subject device NAEOTOM Alpha with software version SOMARIS/10 syngo CT VB10 is a Computed Tomography X-ray system which features two continuously rotating tube-detector systems, denominated as A- and B-systems respectively (dual source CT scanner system). The detectors' function is based on photon-counting technology. The NAEOTOM Alpha with SOMARIS/10 syngo CT VB10 produces CT images in DICOM format, which can be used by trained staff for postprocessing applications commercially distributed by Siemens and other vendors as an aid in diagnosis and treatment as well as for diagnostic and therapeutic interventions. The computer system delivered with the CT scanner is able to run optional post-processing applications.

The platform software for the NAEOTOM Alpha is syngo CT VB10 (SOMARIS/10 syngo CT VB10). It is a command-based program used for patient management, data management, X-ray scan control, image reconstruction, and image archive/evaluation.

The software platform provides plugin software interfaces that allow for the use of specific commercially available post-processing software algorithms in an unmodified form from the cleared stand-alone post-processing version.

The provided text describes the Siemens NAEOTOM Alpha CT Scanner System with software version SOMARIS/10 syngo CT VB10. While it details extensive non-clinical testing and verification/validation activities, it does not include acceptance criteria for specific AI/software performance metrics (e.g., sensitivity, specificity, accuracy) nor does it describe a comparative clinical study (MRMC or standalone AI performance) with human readers or clinical outcomes. The submission focuses on demonstrating substantial equivalence to predicate devices through engineering verification and validation of new and modified features, rather than a clinical performance study of an AI-powered diagnostic aid.

Therefore, many of the requested details regarding AI performance acceptance criteria and a study proving device performance in a clinical AI context are not present in the provided document. The document primarily describes the general engineering and regulatory testing performed for a CT system and its software updates.

However, I can extract information related to the non-clinical testing performed to support the modifications, as well as the types of ground truth used where applicable.

Here's an attempt to answer the questions based only on the provided text, highlighting where information is absent:

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

The document defines "acceptance criteria" generally as part of the system validation tests (workflow and user manual tests, legal and regulatory tests) and system verification tests (system integration, functionality verification, image quality evaluation). For specific new/modified features, the acceptance criteria are generally qualitative or comparative relative to the predicate device, demonstrating comparability or improvement.

2. No new artifacts introduced.
3. Equivalent image quality in quantitative standard physics phantom-based measurements (noise, homogeneity, high-contrast resolution, slice thickness, CNR) compared to non-corrected standard reconstruction.
4. Equivalent image quality in quantitative and qualitative phantom-based measurements with respect to metal objects.
5. Algorithm can be successfully applied to phantom data from motion phantom. | "If misalignment artefacts are identified in non-corrected standard ECG-gated reconstructed sequence or spiral images, the feature 'Cardiac Stack Artefact Correction' (SAC, marketing name: ZeeFree) enables optional stack artefact corrected images, which reduce the number of alignment artefacts."
   "The SAC reconstruction does not introduce new artefacts, which were previously not present in the non-corrected standard reconstruction."
   "The SAC reconstruction does realize equivalent image quality in quantitative standard physics phantom-based measurements (ACR, Gammex phantom) in terms of noise, homogeneity, high-contrast resolution, slice thickness and CNR compared to a non-corrected standard reconstruction."
   "The SAC reconstruction does realize equivalent image quality in quantitative and qualitative phantom-based measurements with respect to metal objects compared to a non-corrected standard reconstruction."
   "The SAC algorithm can be successfully applied to phantom data if derived from a suitable motion phantom demonstrating its correct technical function on the tested device." |
   | myNeedle Guide (with myNeedle Detection) | 1. Accuracy of automatic needle detection algorithm.
6. Reduction of necessary user interactions for navigating to a needle-oriented view. | "It has been shown that the algorithm was able to consistently detect needle-tips over a wide variety of scans in 90.76% of cases."
   "Further, the results of this bench test clearly shows that the auto needle detection functionality reduces the number of interactions steps needed to generate a needle-aligned view in the CT Intervention SW. Zero user interactions are required and a needle-aligned view is displayed right away after a new scan, if auto needle detection is switched on in the SW configuration." |
   | Quantum Spectral Imaging | 1. T3D reconstructions in Quantumpeak mode possible with sharpest available kernels.
7. Quantumpeak scan mode allows monoenergetic images from 40 to 190 keV.
8. Monoenergetic reconstructions free of artifacts.
9. Measured CT values precisely match reference values.
10. Accuracy of monoenergetic reconstructions in iodine and calcium inserts comparable or better than secondary predicate. | "The results showed that: with T3D reconstructions from Quantumpeak scan modes, high resolution images with sharp kernel up to Br98 are obtained. The resolution is comparable to other Highresolution scan modes of the NAEOTOM Alpha."
    "Monoenergetic reconstructions from Quantumpeak scan modes are free of artifacts. Measured CT values precisely match the reference values."
    "The accuracy of monoenergetic reconstructions in iodine and calcium inserts at the NAEOTOM Alpha is comparable or better than on the secondary predicate device SOMATOM Force." |
    | Quantum HD Cardiac | Substantial equivalence in image quality (UHR and standard resolution spectral images) compared to single source spectral capable 120x0.2mm UHR scan mode. | "Based on the results it can be concluded that substantial equivalence in image quality is achieved by the images derived from the spectral capable cardiac acquisition mode 96x0.2mm for both, the high-resolution UHR and the standard resolution spectral image cases, compared to the single source spectral capable 120x0.2mm UHR scan mode." |
    | HD FoV (High Definition Field of View) | HU accuracy in extended field of view region. | "In the phantom study, an HU value accuracy of about +/- 40 HU was achieved with skin-line accuracy of about +/- 3 mm."
    "HD FoV enables the reconstruction of images while significantly improving the visualization of anatomy in the regions outside the scan field of view of 50 cm." |

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: Not specified in terms of patient counts for clinical validation, as this document focuses on non-clinical (phantom and bench) testing. For the "myNeedle Guide" feature, it states "a wide variety of scans" were used, and the success rate was "90.76% of cases" but doesn't quantify the number of cases. Phantom studies are mentioned for other features without specific numbers of scans/readings.
• Data Provenance: The testing described is non-clinical (phantom, bench, and system integration/verification testing). There is no mention of human patient data or its country of origin. The manufacturing site is Siemens Healthcare GmbH in Forchheim, Germany, implying the testing likely occurred there.
• Retrospective or Prospective: Not applicable as it's non-clinical testing.

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: Not specified.
• Qualifications of Experts: Not specified.
• Role of Experts: For the non-clinical tests, "ground truth" generally refers to the known physical properties of the phantoms or the expected performance based on engineering specifications. Human experts are mentioned as trained staff who would use the device, but not as part of a formal ground truth establishment process for the performance studies presented.

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

• Adjudication Method: Not applicable, as no multi-reader human-based test set or clinical study is described. The assessment of performance is based on measurements from phantoms and internal engineering verification.

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

• MRMC Study: No MRMC comparative effectiveness study is mentioned. The document focuses on demonstrating that new/modified features of the CT system are comparable to or improve upon the predicate device through non-clinical testing. The device is a CT system, not an AI-powered diagnostic assist that would typically be evaluated with MRMC. The "myNeedle Detection" feature is a software algorithm within the CT system to aid in procedures, but its evaluation is described as a bench test of its detection accuracy and reduction of user interaction steps, not an MRMC study.

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

• Standalone Performance: The performance data provided is for the CT system and its integrated software features, including algorithms like ZeeFree and myNeedle Detection. The "algorithm only" performance is implicitly covered in the bench testing of these features (e.g., myNeedle Detection achieving 90.76% detection accuracy). However, this is not presented as a "standalone AI" product in the sense of a distinct AI diagnostic algorithm being submitted for clearance. It's an integrated feature of the CT system.

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

• Type of Ground Truth: For the non-clinical tests described, the ground truth is primarily based on:
  • Known physical properties of phantoms: Used for evaluating image quality metrics (noise, homogeneity, resolution, CNR, HU accuracy, etc.) and quantitative measurements.
  • Engineering specifications and expected functional behavior: For features like FAST 3D Camera accuracy and Multi-Purpose Table movement.
  • Reference values: For monoenergetic reconstructions in Quantum Spectral Imaging.
  • Manual verification/observation: For testing user interaction steps in myNeedle Guide.

8. The sample size for the training set

• Training Set Sample Size: Not specified. The document does not describe the development or training of AI models. It focuses on the verification and validation of specific software features within the CT system.

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

• Training Set Ground Truth Establishment: Not specified, as no training set or AI model development is described in this document.

Ask a specific question about this device