Search Results

The Auto Lung Nodule Detection is computer-aided detection software to identify and mark regions in relation to suspected pulmonary nodules from 10 to 30 mm in size. It is designed to aid the physician to review the PA chest radiographs of adults as a second reader and be used as part of S-Station, which is operation software installed on Samsung Digital X-ray Imaging systems. Auto Lung Nodule Detection cannot be used on the patients who have lung lesions other than abnormal nodules.

Auto Lung Nodule Detection is a Computer-Aided Detection (CADe) device that is designed to perform CAD processing in Chest X-ray images for indication of locations for high nodule probability, which has an effective detection sizes from 10 mm to 30 mm.

Auto Lung Nodule Detection receives images acquired with SAMSUNG Digital X-ray Imaging Systems as an input and identifies suspected nodules, and then sends information of suspected nodules to the visualization part of S-Station, which is installed on all kinds of SAMSUNG Digital X-ray Imaging Systems, to generate output images with circular marks. The CAD performed images, are displayed on the screen by S-Station without defeat of original images and used as a second reader only after the initial read is completed.

Here's a breakdown of the acceptance criteria and study details for the Auto Lung Nodule Detection device, based on the provided FDA 510(k) Premarket Notification:

1. Table of Acceptance Criteria and Reported Device Performance

The document describes the clinical study's objective as demonstrating that the device improves human readers' nodule detection performance. While specific numerical acceptance criteria (e.g., minimum sensitivity, maximum FPPI) are not explicitly called out in a "table of acceptance criteria," the clinical performance testing section clearly states that the results have demonstrated that all readers' nodule detection performances using the proposed device have increased with statistical significance. This implicitly defines the acceptance criteria: a statistically significant improvement in nodule detection performance metrics.

2. Sample Size and Data Provenance

• Test Set Sample Size: Not explicitly stated. The document mentions "a test dataset containing both normal and diseased images."
• Data Provenance: Not explicitly stated (e.g., country of origin). The study is described as a "clinical evaluation," implying real patient data. It is not specified if the data was retrospective or prospective.

3. Number of Experts and Qualifications for Ground Truth

• Number of Experts: Not explicitly stated.
• Qualifications of Experts: Not explicitly stated. However, given the context of a CADe device for radiological use, these would typically be board-certified radiologists or pulmonary specialists.

4. Adjudication Method for the Test Set

The document mentions that "Readers were asked to mark their region of nodule suspicion on the images while also providing confidence scores on each decision they have made." It then states, "After independent reading, readers were allowed to adjust their confidence scores after reviewing the ALND's detection results." This describes a workflow, but does not explicitly describe the ground truth adjudication method (e.g., consensus reading by multiple experts, pathology confirmation). The term "independent reading" before ALND review suggests individual reads prior to any potential adjudication or consensus if it occurred.

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

Yes, an MRMC comparative effectiveness study was conducted.

• Effect Size of Improvement: The document states, "The results have demonstrated that all readers' nodule detection performances using the proposed device have increased with statistical significance." While specific numerical effect sizes (e.g., percentage increase in sensitivity, reduction in FPPI, or change in JAFROC FOM) are not provided, the key finding is the statistical significance of the improvement when radiologists used ALND as an assistant tool.

6. Standalone (Algorithm Only) Performance

The document does not explicitly report standalone performance of the algorithm without human-in-the-loop. The clinical evaluation focuses on the human-in-the-loop performance (with ALND assistance vs. without).

7. Type of Ground Truth Used

The type of ground truth used is not explicitly stated. It can be inferred that it involved expert consensus or a gold standard based on follow-up, but the document does not specify if it was pathology, follow-up imaging, or expert consensus.

8. Sample Size for the Training Set

The sample size for the training set is not provided in this document.

9. How Ground Truth for the Training Set was Established

The method for establishing ground truth for the training set is not provided in this document.

Ask a specific question about this device

The GC70 Digital X-ray Imaging System is intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

The GU60A & GU60A-65 Digital X-ray Imaging Systems are intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

GF50 Digital X-ray Imaging System is intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

The GF50A Digital X-ray Imaging System is intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

The GR40CW Digital X-ray Imaging System is intended for use in general projection radiographic applications wherever conventional screen-film systems or CR systems may be used. This device is not intended for mammographic applications.

The GM85 Digital Mobile X-ray imaging System is intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

GC70, GU60&GU60A-65, GF50, GF50A, GR40CW and GM85 are used to capture images by transmitting X-ray to a patient's body. The X-ray passing through a patient's body is sent to the detector and then converted into electrical signals. These signals go through the process of amplification and digital data conversion in the signal process on the S-station, which is the Operation Software (OS) of Samsung Digital Diagnostic X-ray System, and save in DICOM file, a standard for medical imaging. The captured images are tuned up by an Image Post-processing Engine (IPE) which is exclusively installed in S-station, and sent to the Picture Archiving & Communication System (PACS) sever for reading images.

An IPE operates, from the input image, the roles of a region-of-interest extraction, tonescale mapping, noise reduction and texture restoration. The IPE employing an advanced noise reduction algorithm is shown that the image quality of PA radiograph for average adult chest, exposed at the condition of 50% lower dose at Entrance Skin Exposure (ESE) in comparison with the condition of the conventional noise reduction algorithm, is substantially equivalent. It was cleared with K172229 that using the IPE is able to reduce dose of 50% for chest PA of average adult in GC85A.

This submission is purposed to get 510(k) clearance for expanding the scope of the claim, cleared with K172229, from GC85A to the proposed devices, Samsung x-ray systems.

The IPE, a software which has no relation with imaging chain, is applied to the proposed devices and it is evaluated that images acquired by a various imaging chain are substantially equivalent to GC85A in a non-clinical evaluation.

The proposed devices with the IPE employing an advanced noise reduction algorithm are able to reduce dose of 50% for chest PA of average adult for marketing purpose.

This claim is based on a limited study of an anthropomorphic phantom that simulates the x-ray properties of an average size adult, and on a small clinical study at one facility. Only routine PA chest radiography was studied, and results for larger-size adults (body mass index) greater than 30 was not studied to statistical significance. The pediatric patients was not studied and the clinical site is responsible for determining whether the particular radiographic imaging needs are not impacted by such x-ray dose reduction.

The provided text is a 510(k) Premarket Notification summary for several Samsung Digital X-ray Imaging Systems (GC70, GU60A&GU60A-65, GF50, GF50A, GR40CW, and GM85). The core of the submission is to expand the scope of a previously cleared Image Post-processing Engine (IPE) with an advanced noise reduction algorithm (from K172229 for GC85A) to these additional devices, claiming that the IPE can achieve a 50% dose reduction for chest PA of average adults while maintaining image quality.

However, the document explicitly states: "This submission does not required clinical data." This means that a clinical study proving the device meets specific acceptance criteria based on human reader performance with or without AI (MRMC study) was not conducted or submitted as part of this premarket notification. The evaluation relies heavily on non-clinical data and the substantial equivalence to a predicate device that previously demonstrated the dose reduction claim for a specific device (GC85A).

Given this, I cannot provide a table of acceptance criteria based on a clinical study or details about a MRMC study, as none were required or performed for this specific submission's scope expansion.

Here's what can be extracted and inferred from the document regarding the device and its testing, tailored to the questions where information is available:

Acceptance Criteria and Device Performance (Based on Non-Clinical Data)

Since no clinical study was required for this submission, the "acceptance criteria" and "reported device performance" are based on non-clinical evaluations comparing the image quality of the proposed devices with the IPE to the predicate device (GC85A) and to images produced with conventional settings. The primary claim revolves around the ability to maintain image quality at a 50% reduced dose.

Table of Acceptance Criteria and Reported Device Performance (Non-Clinical)

Study Details (Based on Available Information)

1. Sample sized used for the test set and the data provenance:

   • Test Set (Non-clinical): The document mentions "a limited study of an anthropomorphic phantom that simulates the x-ray properties of an average size adult" and "a small clinical study at one facility." However, the quantitative assessment for this specific submission's purpose (expanding IPE scope) was non-clinical: "The quantitative assessment of image quality was conducted with the images of TOR CDR radiography phantom and TO20 contrast detail phantom taken at a various exposure condition with different Samsung X-ray imaging systems." No specific sample size (number of phantom images) is provided.
   • Data Provenance: The anthropomorphic phantom study and "small clinical study" mentioned for the initial GC85A clearance (K172229) are noted as having limitations: "Only routine PA chest radiography was studied, and results for larger-size adults (body mass index) greater than 30 was not studied to statistical significance. The pediatric patients was not studied..." The origin country of the data is not explicitly stated for these studies, but Samsung is a Republic of Korea company. The non-clinical test set for this submission involved "various Samsung X-ray imaging systems" which would likely be internally generated. The studies mentioned (phantom and small clinical) for the original IPE clearance (K172229) were likely a mix of retrospective/prospective, but this detail is not provided for those underlying studies, nor are they claimed as part of this current submission's direct evidence, only referenced as a basis.

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

   • For the non-clinical quantitative assessment in this submission (MTF, CNR, visibility of phantom images), human experts were not used to establish quantitative ground truth. These are objective measures.
   • For the original clearance of the IPE (K172229) where the "small clinical study" and "substantially equivalent" image quality claim was made with 50% dose reduction based on human perception, no details about expert readers, their number, or qualifications are provided in this summary.

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

   • Not applicable for the non-clinical quantitative assessment of phantom images.
   • Not provided for the underlying "small clinical study" that was part of the original IPE clearance.

4. 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:

   • No. The document explicitly states: "This submission does not required clinical data." Therefore, no MRMC comparative effectiveness study was performed or submitted for this specific 510(k) notification. The evaluation relies on substantial equivalence to the predicate and non-clinical data.

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

   • The IPE is a software component (algorithm) that processes images. The "quantitative assessment of image quality was conducted with the images of TOR CDR radiography phantom and TO20 contrast detail phantom," which constitutes a standalone (algorithm only) performance evaluation based on objective image quality metrics (MTF, CNR, visibility). This confirms the algorithm's ability to process images in a way that, quantitatively, maintains image quality.

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

   • For the non-clinical quantitative assessment (of image quality parameters like MTF, CNR, visibility), the "ground truth" is derived from physical phantoms with known properties and objective measurements. It is not based on expert consensus, pathology, or outcomes data.
   • The claim of "substantially equivalent" image quality at 50% dose reduction (from K172229) implicitly relies on human perception from a "small clinical study," but the ground truth establishment method for that is not detailed here.

7. The sample size for the training set:

   • Not mentioned in the document. The document describes the "proposed devices" (SAMSUNG X-ray systems) and the "IPE" (Image Post-processing Engine) as software. It asserts that the IPE "operates, from the input image, the roles of a region-of-interest extraction, tonescale mapping, noise reduction and texture restoration." It also notes the IPE "employing an advanced noise reduction algorithm." While algorithms require training data, the submission does not delve into the development or training of the IPE.

8. How the ground truth for the training set was established:

   • Not mentioned in the document, as details about the IPE's development and training process are outside the scope of this 510(k) summary, which focuses on demonstrating substantial equivalence for the application of an already cleared IPE to new devices.

Ask a specific question about this device

The GU60A & GU60A-65 Digital X-ray Imaging Systems are intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

The GU60A & GU60A-65 digital X-ray imaging systems are to be used to take and store image for diagnosis of patients. It consists of HVG(High voltage generator), U-arm positioner, Detector, X-ray tube, Collimator, AEC(Auto Exposure Control), DAP(Dose Area Product), CIB(Control Interface Box), Remote controller, Grid, Barcode scanner and Autostitching stand.

These systems are used to capture images by transmitting X-ray to a patient's body. The X-ray passing through a patient's body is sent to the detector and then converted into electrical signals. These signals go through the process of amplification and digital data conversion in the signal process device being sent to the S-Station (Operation Software) and saved in DICOM file, a standard for medical imaging. The captured images are sent to the Picture Archiving & Communication System (PACS) server, and can be used for reading images.

The GU60A & GU60A-65 digital X-ray imaging systems is stationary, and it was previously cleared under K151685. The software features cleared with K171119, SimGrid, BSI (Bone Suppression Image) and TLE (Tube & Line Enhancement), is added to the predicate x-ray system (K151685).

The software features called as SimGrid, BSI (Bone Suppression Image) and TLE (Tube & Line Enhancement) is a post-image processing software option which provides companion images to assist diagnosis in addition to the images obtained from normal diagnosis protocol.

The SimGrid software option is able to compensate the contrast loss due to scatter radiations, primarily acquisitions without a physical anti-scatter grid.

The BSI software option suppresses bone anatomy and the TLE software option enhances visibility of tube and catheter features in a companion image that is delivered in addition to the original diagnostic image.

These software features are designed to be exclusively installed in S-station, SAMSUNG digital X-ray operation software, while it uses the radiological image as an input and do not depend on how the image is acquired or which radiology device is used.

The provided document does not contain details about specific acceptance criteria, a study proving the device meets those criteria, or reported device performance metrics in the format requested. The document is a 510(k) premarket notification summary for an X-ray imaging system, focusing on demonstrating substantial equivalence to predicate devices rather than providing detailed performance study results against specific acceptance criteria.

Therefore, many of the requested fields cannot be filled from the provided text.

Here's an analysis of what can be extracted:

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

• Acceptance Criteria: Not explicitly stated in the provided text. The document refers to conformance with various standards (ES 60601-1, IEC 60601-1-2, IEC 60601-1-3, IEC 60601-2-28, IEC 60601-2-54, ISO14971, 21CFR1020.30, 21CFR1020.31, IEC 62220-1) and a guidance document for wireless technology. It also states that all test results were "satisfying the standards." However, specific numerical acceptance criteria for performance (e.g., sensitivity, specificity, accuracy) for the software features are not detailed.
• Reported Device Performance: No specific performance metrics (e.g., sensitivity, specificity, accuracy) are reported for the device, or its software features (SimGrid, BSI, TLE). It only mentions that MTF and DQE measurements "show no difference in non-clinical testing data... from the predicate device."

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

• Not specified. The document states "The application of SimGrid, TLE, S-DAP and BSI, cleared with K171119, to the proposed device GU60A & GU60A-65 does not require clinical data." This implies no clinical test set was used for the substantial equivalence demonstration of these specific software features. Non-clinical testing data (MTF and DQE) likely used phantoms, but the sample size or provenance of that data is not provided.

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)

• Not applicable as no clinical data or specific test set adjudicated by experts is described in the document.

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

• Not applicable as no clinical data or specific test set with expert adjudication is described in the document.

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

• No MRMC comparative effectiveness study is mentioned. The document explicitly states that clinical data was not required for the software features in question.

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

• The document implies that the software features (SimGrid, BSI, TLE) provide "companion images to assist diagnosis." This suggests an assistive role rather than a standalone diagnostic role, but no specific performance metrics for this assistance are provided. The fact that clinical data was not required for these features further supports that a standalone performance study was not conducted for this submission.

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

• Not applicable, as no clinical studies with established ground truth are described for the performance of the features. For non-clinical MTF and DQE measurements, the ground truth would be determined by the phantom properties and measurement protocols per IEC 62220-1.

8. The sample size for the training set

• Not specified. The document describes the device as a hardware system with added post-processing software features. There is no mention of an "AI" or "machine learning" algorithm in the modern sense that would require a distinct training set. The software features (SimGrid, BSI, TLE) are described as "post-image processing software option" rather than adaptive learning algorithms.

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

• Not applicable, as a training set for an AI/ML algorithm is not described.

Ask a specific question about this device

The GC85A Digital X-ray Imaging System is intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

The GC85A digital X-ray imaging system is used to capture images by transmitting X-ray to a patient's body. The X-ray passing through a patient's body is sent to the detector and then converted into electrical signals. These signals go through the process of amplification and digital data conversion in the signal process device before being sent to the S-Station (Operation Software) and saved in DICOM file, a standard for medical imaging. The captured images are sent to the Picture Archiving & Communication System (PACS) server, and can be used for reading images.

The Image Post-processing Engine is exclusively installed in S-station, which is a Samsung Digital X-ray Operation Software for Samsung Digital X-ray System. It has an image processing algorithm to improve an acquired image and previously cleared with K160997.

The proposed Image Post-processing Engine is upgraded with employing an advanced noise reduction algorithm to improve image quality. The proposed Engine is shown of a post-processed image as substantially equivalent as the image by the predicate Image Post-processing Engine at a certain low dose level.

The acceptance criteria and the study proving the device meets them are detailed below, based on the provided text.

1. Table of Acceptance Criteria and Reported Device Performance

The device is an upgraded Image Post-processing Engine for the GC85A Digital X-ray Imaging System, featuring an advanced noise reduction algorithm. The primary acceptance criterion is achieving substantially equivalent image quality at a 50% reduced radiation dose compared to the predicate device.

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

• Sample Size:
  • Clinical Study: 78 in-vivo data sets of chest PA images.
  • Phantom Study: Not explicitly stated as a number, but "a semi-anatomical chest phantom at a various radiation dose" and "anthropomorphic chest phantom images" were used.
• Data Provenance: The clinical study was conducted at "one facility" and involved "in-vivo data sets." This implies the data is prospective clinical data collected specifically for this study. The country of origin is not explicitly mentioned but the submitter is from the Republic of Korea.

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

• Number of Experts: Three professional radiologists.
• Qualifications: "Three experienced radiologists." Specific years of experience are not provided, but they are described as "professional" and "experienced."

4. Adjudication Method for the Test Set

• The text mentions that "Anthropomorphic chest phantom images were scored by Bureau of Radiological Health (BRH) method and inter-observer agreement was calculated."
• For the clinical images, "Seven anatomical landmarks were evaluated for image quality assessment by three readers."
• While it states that inter-observer agreement was calculated for phantom images and multiple readers evaluated clinical images, it does not explicitly describe a specific adjudication method like 2+1 or 3+1 for resolving discrepancies in reader interpretations for either phantom or clinical studies. It implies a consensus or agreement was sought, but the formal process isn't detailed.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size of How Much Human Readers Improve with AI vs. Without AI Assistance

• A MRMC study comparing human readers with AI vs. without AI assistance was not conducted. The study evaluated whether images processed by the new engine at reduced dose were substantially equivalent to those from the predicate engine, evaluating the device's standalone performance in producing equivalent diagnostic images, rather than its assistive role to humans.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

• Yes, a standalone evaluation of the algorithm's output was critically performed. The primary goal was to demonstrate that the images generated by the new Image Post-processing Engine at 50% dose reduction are diagnostically equivalent to those from the predicate device at full dose. This means the algorithm's ability to process images to an acceptable diagnostic quality was the focus, rather than its interaction with a human reader. SNR and CNR measurements and the radiologists' evaluations of image quality were direct assessments of the algorithm's output.

7. The Type of Ground Truth Used

• The ground truth was established through expert consensus/reader interpretation (three experienced radiologists) during the evaluation of both anthropomorphic phantom images (scored by BRH method) and clinical images (evaluation of seven anatomical landmarks and 18 cases of featured lung lesions). There is no mention of pathology or outcomes data being used as ground truth.

8. The Sample Size for the Training Set

• The document does not explicitly state the sample size for the training set used to develop or train the advanced noise reduction algorithm in the Image Post-processing Engine.

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

• The document does not provide information on how the ground truth for the training set was established. It focuses primarily on the testing and validation of the device's performance.

Ask a specific question about this device

The GM85 Digital X-ray imaging System is intended for use in generating radiographic images of human anatomy by a qualified/trained doctor or technician. This device is not intended for mammographic applications.

The GM85 Digital Mobile X-ray imaging System is the equipment that captures images by transmitting X-ray to a patient's body. The X-ray passing through a patient's body is sent to the detector and then converted into electrical signals. These signals go through the process of amplification and digital data conversion in the signal process device before being sent to the S-Station (Operation Software) and saved in DICOM file, a standard for medical imaging. The captured images are sent to the Picture Archiving & Communication System (PACS) server, and can be used for reading images.

The Bone Suppression Image (BSI) is a post-image processing software option which provides companion images to assist diagnosis in addition to the images obtained from normal diagnosis protocol. The BSI software option suppresses bone anatomy to enhance visualization of chest pathology in a companion image that is delivered in addition to the original diagnostic image.

The BSI option is designed to be exclusively installed in S-station, SAMSUNG digital Xray operation software, while it uses the radiological image as an input and do not depend on how the image is acquired or which radiology device is used.

The provided text describes a 510(k) premarket notification for the GM85 Digital X-ray imaging System, specifically focusing on the integration of the Bone Suppression Image (BSI) feature. However, the document explicitly states that the application of BSI to the proposed device does not require clinical data and therefore, clinical studies to establish acceptance criteria or device performance are not detailed within this document.

The document primarily focuses on demonstrating substantial equivalence to predicate devices based on non-clinical data and technical characteristics, rather than presenting a performance study with acceptance criteria.

Therefore, many of the requested details about acceptance criteria and the study that proves the device meets them cannot be extracted from this particular submission.

Here's what can be inferred and what is explicitly stated as not applicable or not provided:

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

• Acceptance Criteria: Not explicitly stated in terms of clinical performance metrics (e.g., sensitivity, specificity, AUC). The acceptance criteria for this submission appear to be focused on substantial equivalence to predicate devices through technical comparisons and non-clinical testing.
• Reported Device Performance: No clinical performance metrics are reported because clinical data was deemed not required and thus not provided. The performance is assessed based on the device having the same detectors and therefore showing "no difference in non-clinical testing data such as MTF and DQE measurements from the predicate device #1 and #2."

2. Sample size used for the test set and the data provenance

• Sample Size: Not applicable/not provided. No clinical test set.
• Data Provenance: Not applicable/not provided. No clinical test data.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

• Number of Experts: Not applicable/not provided. No clinical test set.
• Qualifications of Experts: Not applicable/not provided.

4. Adjudication method for the test set

• Adjudication method: Not applicable/not provided. No clinical test set.

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. The document explicitly states "The application of BSI... to the proposed device GM85 does not require clinical data." Therefore, an MRMC study was not performed for this submission.
• Effect Size: Not applicable/not provided.

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

• Standalone Performance: Not explicitly detailed as a performance study for the BSI feature itself. The BSI software option "provides companion images to assist diagnosis," implying it's an aid, not a standalone diagnostic tool without a human. However, its verification and validation were done through a "Software System Test Case" (non-clinical).

7. The type of ground truth used

• Type of Ground Truth: Not applicable for a clinical performance study. For the BSI software's non-clinical verification and validation, the ground truth would be against expected software output based on its design specifications, rather than clinical outcomes or expert consensus on patient data.

8. The sample size for the training set

• Sample Size for Training Set: Not applicable/not provided. The document focuses on performance relative to predicate devices and non-clinical validation. Details regarding the training of the BSI algorithm itself are not included in this submission.

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

• Ground Truth for Training Set: Not applicable/not provided as training set details are not included.

Summary of what is provided regarding acceptance and study:

The acceptance for this 510(k) submission is based on demonstrating substantial equivalence to predicate devices. This is achieved through:

• Technical Characteristics Comparison: Showing that the proposed device (GM85 with BSI) is identical or equivalent to Predicate Device #1 (GM85 without BSI, K162278) in its core components (HVG, X-ray Tube Assembly, Detectors) and shares the BSI feature with Predicate Device #2 (GC70 with BSI, K163115).
• Safety, EMC, and Performance Data (Non-clinical):
  • Electrical, mechanical, environmental safety, and performance testing according to standards (ES 60601-1, IEC 60601-1-2, IEC 60601-1-3, IEC 60601-2-28, IEC 60601-2-54, ISO14971, 21CFR1020.30, 21CFR1020.31).
  • EMC testing according to IEC 60601-1-2.
  • Wireless function testing according to guidance on "Radio frequency Wireless Technology in Medical Devices."
  • All these tests "were satisfying the standards."
• Non-clinical Data for Imaging Performance:
  • Conformance to FDA "Guidance for the Submission of 510(k)'s for Solid-State X-ray Imaging Devices," including MTF and DQE measurements as tested by IEC 62220-1.
  • The proposed device uses the same detectors as the predicate devices, "so it shows no difference in non-clinical testing data such as MTF and DQE measurements from the predicate device #1 and #2."
• Software Verification and Validation: "BSI installed in S-Station of GM85 has been verified and validated through GM85 Software System Test Case." (This is a non-clinical, internal software validation).

Conclusion from the document: "The non-clinical demonstrates that the proposed device is as safe, as effective, and performs as well as the legally marketed predicate devices."

The critical point for this document is that the FDA determined, and the submission reflects, that for the integration of the already cleared BSI software into the GM85 mobile X-ray system, no new clinical data was required to demonstrate substantial equivalence.

Ask a specific question about this device