Search Results

producing panoramic X-ray images for diagnostic examination (teeth), jaws and oral structures;
producing radiographs of maxillofacial region and parts of the skull for cephalometric examination, if equipped with CEPH arm;
producing radiographs of hands and wrists for carpus examination, if equipped with CEPH arm;
producing tomographic images of the oral and maxillofacial region, for diagnostic examination of dentition (teeth), jaws, oral structures and some cranial bones, if equipped with CBCT option.
From a clinical point of view, X-MIND trium can be applied for the following medical indications:
Generic dentistry
Dental implantology
Dental surgery
Maxillo-facial surgery
Cephalometric analysis
Carpus radiology

Device Description

X-MIND trium is a digital panoramic, cephalometric and tomographic extra-oral X-ray system. The system consists of the X-MIND trium SCANNER and the OPERATOR'S WORKSTATION. The scanner includes a control panel, X-ray generator, sliding body, U-arm, detector sliding group, CBCT detector, PAN detector, TLD CASSETTE (optional), patient support, F group, and column. The workstation runs ACTEON IMAGING SUITE (AIS) + 2D diagnostic analysis and AIS 3D app for CBCT dataset display software. The device can be configured as PAN only, PAN / CBCT, PAN / CEPH, PAN / CBCT / CEPH (PAN Full), PAN / CBCT TLD version, and PAN / CBCT / CEPH TLD version. It can be wall mounted or self-standing. Main changes include the addition of a new accessory, the X-MIND trium TLD cassette, including a new model of CBCT sensor that can be used for both panoramic and CBCT examination, and the introduction of CEPH and PAN sensors with a wider admitted range of power supply value. A new movement of the cassette has been added to modify the SID, allowing for special exams for children with reduced SID. A new AUXCBCT board has been designed to support the new movement and cassette. A new functionality, the Low Dose functionality, allows for CBCT exams using Low Dose protocols as an alternative to Standard Dose protocols, ensuring a radiation dose saving of about 50%.

AI/ML Overview

Here's an analysis of the acceptance criteria and study details for the X-MIND trium device, based on the provided document:

Acceptance Criteria and Device Performance

The document describes several changes to the X-MIND trium device, and the acceptance criteria and performance are tied to these specific changes, particularly related to the new PAN-CBCT detector and the Low Dose (LD) functionality.

1. Table of Acceptance Criteria and Reported Device Performance

Device Change / Characteristic	Acceptance Criteria	Reported Device Performance
New PAN Detector (1)	Same technology. DQE values for new detector are in the range ±0.1 on the whole frequency range (1 – 4 LP/mm).	PASS (DQE values are within range, same technology)
New CEPH Detector (2)	Same technology. DQE values for new detector are in the range ±0.1 on the whole frequency range (1 – 4 LP/mm).	PASS (DQE values are within range, same technology)
New PAN-CBCT Detector (3)	Image Quality vs. Dose Trade-off (System Indicator - SI_DAP): Same or superior trade-off between dose and image quality (i.e. System Indicator (SI) in CBCT) $S_{DAP} ≥ S_{DAP ; predicate}$ where $S_{DAP} = \frac{CNR}{DAP \cdot (\frac{1}{2 \cdot V_{50%}})^2}$ (CNR = contrast to noise ratio, DAP = dose area product, V50% = resolution indicator at 50% modulation). Voxel Noise: Voxel Noise [a.u.] with TLD cassette including new PAN-CBCT sensor (VNTLD) ≤ Voxel Noise [a.u.] predicate (VN). Noise Power Spectrum (NPS): NPS0D [a.u.²] with TLD cassette including new PAN-CBCT sensor ≤ NPS0D [a.u.²] predicate. X-ray Field to Image Receptor Correspondence (IEC 60601-2-63): along each of the two axes of the IMAGE RECEPTION AREA, the edges of the X-RAY FIELD shall not exceed the corresponding edges of the EFFECTIVE IMAGE RECEPTION AREA by more than 2 cm or 3 % of the indicated FOCAL SPOT TO IMAGE RECEPTOR DISTANCE, whichever the larger; the sum of the discrepancies on both axes shall not exceed 3 cm or 4 % of the indicated FOCAL SPOT TO IMAGE RECEPTOR DISTANCE, whichever the larger. X-ray Field to Image Receptor Correspondence (21 CFR 1020.31 f) 4)): field does not exceed each dimension of the image receptor by more than 2 percent of the SID; align the center of the x-ray field with the center of the image receptor to within 2 percent of the SID.	Image Quality vs. Dose Trade-off (SI_DAP): - Woman Medium Standard quality FOV 80X80-80X90: SDAP TLD > SDAP pred (PASS) - Woman Medium Standard quality FOV 110X80: SDAP TLD > SDAP pred (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Adult: SDAP TLD > SDAP pred (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Child: SDAP TLD > SDAP pred (PASS) Voxel Noise: - Woman Medium Standard quality FOV 80X90: VNTLD < VN (PASS) - Woman Medium Standard quality FOV 110X80: VNTLD < VN (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Adult: VNTLD < VN (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Child: VNTLD < VN (PASS) Noise Power Spectrum (NPS): - Woman Medium Standard quality FOV 80X90: NPS0D TLD < NPS0D (PASS) - Woman Medium Standard quality FOV 110X80: NPS0D TLD < NPS0D (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Adult: NPS0D TLD < NPS0D (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Child: NPS0D TLD < NPS0D (PASS) X-ray Field to Image Receptor Correspondence (IEC 60601-2-63): - SID Adult edge ≤ 20mm (PASS) - SID Child edge ≤ 20mm (PASS) - Sum of discrepancies: < 30mm (PASS) - Along scanning axis: ≤ 1mm (PASS) - Perpendicular to scanning axis: =0 (PASS) X-ray Field to Image Receptor Correspondence (21 CFR 1020.31 f) 4)): - PAN: ≤ 2% (PASS) - PAN reduced SID: ≤ 2% (PASS) - CBCT: ≤ 2% (PASS) - CBCT reduced SID: ≤ 2% (PASS) - Center alignment (PAN and CEPH only): PASS
New Child Special Exam (4)	Usability (IEC 60601-1-6): All comments are positive. Biocompatibility (ISO 10993-5: 2009 & ISO 10993-10: 2010 for new tool): Cytotoxicity: No cytotoxic. Hypersensitivity: Not sensitizing.	Usability: PASS (all comments positive). Biocompatibility: PASS (no cytotoxicity, not sensitizing, same material and process as predicate)
True Low Dose (TLD) / Reduced SID (4)	Image Quality (PAN mode): Absence of artifacts. Equal high contrast resolution (Both 3.1 lp/mm) and equal low contrast resolution (Both 4 holes visible). Declared DAP (PAN mode): Declared DAP PAN 38.77 < 52. (image CHILD NEDIUM SID) < predicate. Image Quality vs. Dose Trade-off (CBCT mode - SI_DAP): Same or superior trade-off between dose and image quality (i.e. System Indicator (SI) in CBCT) $S_{DAP} ≥ S_{DAP ; predicate}$.	Image Quality (PAN mode): Both verified PASS (Absence of artifacts, 3.1 lp/mm high contrast resolution, 4 holes visible low contrast resolution). Declared DAP (PAN mode): PASS (38.77 < predicate for child). Image Quality vs. Dose Trade-off (CBCT mode - SI_DAP): - Woman Medium Standard quality FOV 80X80-80X90: SDAP TLD > SDAP pred (PASS) - Woman Medium Standard quality FOV 110X80: TLD > pred (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Adult: SDAP TLD = > SDAP pred (PASS) - Child medium Standard quality FOV 80X80-80X90 SID Child: SDAP TLD > SDAP pred (PASS)
Detectability Study (3 & 7)	Same pathological or clinically relevant conditions identifiable both in the STD protocol scan and in the TLD protocol one (for TLD cassette / reduced SID) AND Same pathological or clinically relevant conditions identifiable both in the STD protocol scan and in the LD protocol one (for Low Dose functionality).	TLD cassette / Reduced SID: PASS. All pathological or clinically relevant conditions identified with STD protocol scan have been identified with the LD protocol one. Note Pos. 45 "worn out dental neck" not identified nor in standard mode nor in TLD configuration (this implies both performed equally, neither identified this specific condition). Low Dose Functionality (LD): PASS. All pathological or clinically relevant conditions have been identified both in the STD protocol scan and in the LD protocol one.
Low Dose Functionality (7)	Image Quality vs. Dose Trade-off (SI_DAP): Same or superior trade-off between dose and image quality. $S_{DAP low dose} ≥ S_{DAP standard protocol}$. Voxel Noise (VN): $\Delta%VN \le 10%$, where $\Delta%VN = \frac{VNld - VNs}{100 + VNsd}$ (VNld = Voxel Noise [a.u.] low dose protocol, VNsd = Voxel Noise [a.u.] standard protocol). Noise Power Spectrum (NPS): $\Delta%NPS \le 30%$, where $\Delta%NPS = \frac{NPSld - NPSsd}{100 + NPSsd}$ (NPSld = NPS0D [a.u.²] low dose protocol, NPSsd = NPS0D [a.u.²] standard dose protocol).	Image Quality vs. Dose Trade-off (SI_DAP): - Man: SDAP low dose > SDAP standard protocol (PASS) - Woman: SDAP low dose > SDAP standard protocol (PASS) - Child: SDAP low dose > SDAP standard protocol (PASS) Voxel Noise: - Man: $\Delta VN% < 10%$ (PASS) - Woman: $\Delta VN% < 10%$ (PASS) - Child: $\Delta VN% < 10%$ (PASS) Noise Power Spectrum (NPS): - Man: $\Delta NPS < 30%$ (PASS) - Woman: $\Delta NPS < 30%$ (PASS) - Child: $\Delta NPS < 30%$ (PASS)

2. Sample Size and Data Provenance

The document mentions "bench tests" and a "clinical detectability analysis" using "radiographs of cadavers".

Test Set Sample Size: For the detectability analysis, the "radiographs of cadavers" constitute the test set for evaluating clinical performance. The number of cadavers scanned is not specified in the provided text. The number of specific conditions tested is also not explicitly stated, beyond the general statement of "All pathological or clinically relevant conditions".
Data Provenance: The data appears to be prospective as it involves new bench tests and a detectability study specifically designed to evaluate the changes in the device. The country of origin of the data is not specified, but the manufacturer is based in Italy.

3. Number of Experts and their Qualifications

Number of Experts: For the detectability analysis, a "set of intended users" were used. The exact number is not specified.
Qualifications of Experts: The experts included "radiologist, dentist, oral surgeon." Specific years of experience are not mentioned.

4. Adjudication Method

The document implies a consensus-based evaluation for the detectability study, stating that experts "evaluated the visibility of anatomical structures and features of interests in the different scanning conditions." However, a formal adjudication method (e.g., 2+1, 3+1, none) is not explicitly described. It states that "All pathological or clinically relevant conditions identified with STD protocol scan have been identified with the LD protocol one," suggesting an agreement on findings rather than a formal disagreement resolution process between multiple readers.

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

A comparative effectiveness study in the sense of comparing human reader performance with and without AI assistance is not explicitly described or quantified.
The detectability study compares the device's performance (standard vs. low dose protocols) by assessing whether experts can identify pathological conditions in both, rather than assessing the uplift in human reader performance using AI. The changes described are enhancements to the imaging system itself (new detector, low dose protocols, reduced SID), not an AI-assisted diagnostic tool for readers.

6. Standalone Performance Study

Yes, standalone performance studies were conducted for various technical aspects of the device. The document details extensive "bench testing" to evaluate parameters like:
- DQE values for new PAN and CEPH detectors.
- Trade-off between dose and image quality (SI_DAP) for the new PAN-CBCT detector and Low Dose functionality.
- Voxel Noise and Noise Power Spectrum.
- X-ray field to image receptor correspondence.
- Mechanical stability (Self-standing version).
These tests were performed on the algorithm/device itself without human-in-the-loop intervention for the quantitative measurements, demonstrating the technical performance of the device's imaging capabilities.

7. Type of Ground Truth Used

For the detectability analysis: The ground truth appears to be based on the identification of "pathological or clinically relevant conditions" by the experts in the standard protocol scan. The low dose scan was then compared against this baseline to ensure similar conditions could be identified. This suggests a form of expert consensus or established clinical findings from the standard protocol as the reference.
For the bench tests: The ground truth for technical parameters (DQE, CNR, DAP, Voxel Noise, NPS, field correspondence) is based on physical phantom measurements and engineering standards/specifications (e.g., IEC standards, physical dimensions, signal characteristics).

8. Sample Size for the Training Set

The document describes changes to a medical imaging device (X-ray system), not an AI algorithm that requires a "training set" in the context of machine learning. Therefore, a training set sample size is not applicable in the conventional sense. The device's operation is based on established physics and engineering principles, and its performance is validated through bench tests and clinical detectability studies.

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

As explained above, the concept of a "training set" in this context is not applicable for an X-ray imaging system. The system's functionalities are developed based on engineering design, and its performance is validated against physical and clinical criteria, not by training on a dataset with pre-established ground truth.

Ask a Question

Ask a specific question about this device

Page 1 of 1