Search Results

SpotLight /SpotLight Duo (with DLIR option) is intended to produce cross-sectional images of the body by computer reconstruction of x-ray transmission projection data taken at different angles. The system has the capability to image cardiovascular and thoracic anatomies, in a single rotation. The system may acquire data using Axial, Cine, and Cardiac scan techniques from patients of all ages (DLR is limited for patient use above the age of 2 years). These images may be obtained either with or without contrast. This device may include signal analysis and display equipment, patient and equipment supports, components and accessories.

This device may include data and image processing to produce images in a variety of trans-axial and reformatted planes.

The system is indicated for X-ray Computed Tomography imaging of cardiovascular and thoracic anatomies that fit in the scan field of view. The device output is useful for diagnosis of disease or abnormality and for planning of therapy procedures.

The SpotLight / SpotLight Duo (with DLIR option) is a multi-slice (192 detector rows), dual tube CT scanner consisting of a gantry, patient table, operator console, power distribution unit (PDU) and interconnecting cables. The system includes image acquisition hardware, image acquisition and reconstruction software and software for operator interface and image handling. The Deep Learning Image Reconstruction (DLIR) algorithm is a deep learning technology-based software sub-system that is integrated into the image reconstruction software. As in other CT scanners, a scanned subject is irradiated by X rays and a detector array measures attenuation data of X rays that have been attenuated by the subject from multiple view angles. This is achieved by rotation of the radiation source and the detector about the subject while acquiring the attenuation data. A computer is used to reconstruct cross sectional images of the subject from the attenuation data.

Here's a summary of the acceptance criteria and the study details for the SpotLight/SpotLight Duo (with DLIR option) CT system, based on the provided FDA 510(k) summary:

1. Acceptance Criteria and Reported Device Performance

The document doesn't explicitly list specific quantitative acceptance criteria in a table format. However, it indicates that the DLIR (Deep Learning Image Reconstruction) algorithm was evaluated to demonstrate non-inferiority to the predicate device's ASiR-CV noise reduction algorithm in terms of image quality. The performance tests focused on several image quality parameters:

2. Sample Size and Data Provenance for the Test Set

• Sample Size (Test Set): Not explicitly stated as a number of cases, but implied to be a collection of "clinical cases of different anatomies, using different types of scans, from patients with a wide range of BMIs and clinical features."
• Data Provenance: Retrospective clinical data acquired by SpotLight / CardioGraphe scanners. Collected from "multiple clinical sites with at least 50% of the cases performed in the US."

3. Number of Experts and Qualifications for Ground Truth

• Number of Experts: Five (5) certified CT readers.
• Qualifications of Experts: 3 radiologists and 2 cardiologists. "4 out 5 are US board certified." Specific years of experience are not mentioned.

4. Adjudication Method

• Adjudication Method: The study was a "retrospective blinded image evaluation." Each exam was reviewed using both standard (ASiR-CV) and alternative (DLIR) methods. The data was "coded to avoid readers' bias." This suggests a comparative reading where readers likely assessed both image sets for the same patient without knowing which was which, and potentially without direct consensus discussions as a formal adjudication step, but rather an independent assessment that collectively informed the non-inferiority conclusion. No explicit "2+1" or "3+1" adjudication method is described.

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

• Was an MRMC study done? Yes, a form of multi-reader evaluation was performed, comparing DLIR to ASiR-CV on clinical cases by five expert readers.
• Effect Size (human readers improve with AI vs. without AI assistance): The document states that "DLIR was found to provide diagnostic image quality that is not inferior to ASiR." It does not provide a specific quantitative effect size or a measure of improvement for human readers with AI (DLIR) vs. without AI (ASiR-CV, which is also an algorithm, a noise reduction one). The focus was on non-inferiority rather than an enhancement measurement for the readers themselves.

6. Standalone Performance Study

• Was a standalone study done? Yes, "DLIR bench tests were performed by applying DLIR and ASIR on phantoms." This represents a standalone evaluation of the algorithm's performance on objective image quality metrics using phantoms.

7. Type of Ground Truth Used

• For Bench Tests: Physical phantoms (water phantoms, CATPHAN, QA phantom) with known properties were used.
• For Clinical Evaluation: "Diagnostic image quality that is not inferior to ASiR" as determined by the expert readers serves as the ground truth/comparison metric. This is effectively expert consensus/opinion on diagnostic quality. There's no mention of pathology or outcomes data being used to establish ground truth for the clinical cases.

8. Sample Size for the Training Set

• The document does not specify the sample size used for training the Deep Learning Image Reconstruction (DLIR) algorithm.

9. How Ground Truth for the Training Set Was Established

• The document does not explicitly describe how ground truth for the training set was established. It only mentions that DLIR is a "deep learning technology-based software sub-system." For deep learning reconstruction, training typically involves pairs of noisy (or lower-dose) input images and corresponding high-quality (or higher-dose/reference standard) images, often reconstructed with a traditional full-dose, high-quality algorithm (not explicitly stated here, but common practice).

Ask a specific question about this device