Algorithm Accuracy Study: 2,478 WSIs (GT 450 DX) and 424 WSIs (GT 180 DX) for pen marks
>1 (qualified experts)
Indications for Use
Aperio iQC DX Software is an artificial intelligence-based software intended to be used as an aid in the identification of artifacts in scanned whole slide images (WSIs) of hematoxylin and eosin (H&E) and immunohistochemistry (IHC) stained, formalin-fixed paraffin embedded (FFPE) tissue that should undergo further evaluation for image quality prior to diagnostic review of the images by pathologists. Aperio iQC DX Software identifies the following artifacts: - Digital artifacts: out of focus, image striping, missing and clipped tissue - Slide preparation artifact: air bubbles - Other artifact: pen marks Aperio iQC DX Software outputs the following: - Binary 'artifact detected' or 'no artifact detected' for air bubbles, image striping, missing and clipped tissue, out of focus, and pen marks, - Overlays highlighting the location of air bubbles and out of focus artifacts only, and - Bounding boxes for missing and clipped tissue artifacts. Aperio iQC DX Software is intended to be used in conjunction with the complete in-house laboratory image quality control workflow. It is the responsibility of a qualified pathologist to employ appropriate procedures and safeguards to assure the validity of the interpretation of images when using the Aperio iQC DX Software. Aperio iQC DX Software is not intended to be used for diagnosis, prognosis, or prediction of disease.
Device Story
Software processes scanned whole slide images (WSIs) from Aperio GT 450 DX and GT 180 DX scanners; utilizes four deep learning convolutional neural network (CNN) models to detect digital artifacts (out of focus, image striping, missing/clipped tissue), slide preparation artifacts (air bubbles), and pen marks. Operates in a parallel workflow; automatically analyzes images from dedicated storage; outputs binary detection status, overlays for air bubbles/out-of-focus, and bounding boxes for tissue issues. Results displayed via browser-based dashboard for laboratory personnel to accept/reject images. Augments quality control; final diagnostic suitability determination remains with pathologist. Benefits include standardized, automated artifact detection to ensure image quality before diagnostic review.
Clinical Evidence
Bench testing only. Performance evaluated using independent datasets of H&E and IHC WSIs from Aperio GT 450 DX and GT 180 DX scanners. Accuracy study (n=4,752 for GT 450; n=2,368 for GT 180) demonstrated sensitivity and specificity ≥90% and ≥85% (lower bound 95% CI) for single and multiple artifact detection. Localization accuracy for missing/clipped tissue achieved 90.74% of WSIs with IoU ≥70%. Precision testing (repeatability/reproducibility) showed 100% correct call rate across all artifact categories under variable conditions.
Technological Characteristics
Software-only device; utilizes deep learning convolutional neural networks (CNNs). Inputs: WSI files from Aperio GT 450 DX/GT 180 DX scanners. Outputs: Binary detection, bounding boxes, and overlays. System requirements: Linux Ubuntu 24.04, 64GB RAM, Intel Core i7/AMD EPYC CPU. Connectivity: Networked via customer image storage. Locked algorithm; no continuous learning. No calibration required.
Indications for Use
Indicated for use as an aid in identifying artifacts in scanned whole slide images (WSIs) of H&E and IHC stained FFPE tissue to support laboratory image quality control workflows prior to pathologist diagnostic review. Not for diagnosis, prognosis, or prediction of disease.
Regulatory Classification
Identification
The whole slide imaging system is an automated digital slide creation, viewing, and management system intended as an aid to the pathologist to review and interpret digital images of surgical pathology slides. The system generates digital images that would otherwise be appropriate for manual visualization by conventional light microscopy.
Special Controls
A whole slide imaging system must comply with the following special controls: (1) Premarket notification submissions must include the following information: (i) The indications for use must specify the tissue specimen that is intended to be used with the whole slide imaging system and the components of the system. (ii) A detailed description of the device and bench testing results at the component level, including for the following, as appropriate: (A) Slide feeder; (B) Light source; (C) Imaging optics: (D)Mechanical scanner movement; (E) Digital imaging sensor; (F) Image processing software; (G)Image composition techniques; (H)Image file formats; (I) Image review manipulation software; (J) Computer environment; (K)Display system. (iii)Detailed bench testing and results at the system level, including for the following, as appropriate: (A)Color reproducibility; (B) Spatial resolution; (C) Focusing test; (D) Whole slide tissue coverage; (E) Stitching error: (F) Turnaround time. (iv) Detailed information demonstrating the performance characteristics of the device, including, as appropriate: (A)Precision to evaluate intra-system and inter-system precision using a comprehensive set of clinical specimens with defined, clinically relevant histologic features from various organ systems and diseases. Multiple whole slide imaging systems, multiple sites, and multiple readers must be included. (B) Reproducibility data to evaluate inter-site variability using a comprehensive set of clinical specimens with defined, clinically relevant histologic features from various organ systems and diseases. Multiple whole slide imaging systems, multiple sites, and multiple readers must be included. (C) Data from a clinical study to demonstrate that viewing, reviewing, and diagnosing digital images of surgical pathology slides prepared from tissue slides using the whole slide imaging system is non-inferior to using an optical microscope. The study should evaluate the difference in major discordance rates between manual digital (MD) and manual optical (MO) modalities when compared to the reference (e.g., main sign-out diagnosis). (D) A detailed human factors engineering process must be used to evaluate the whole slide imaging system user interface(s). (2) Labeling compliant with 21 CFR 809.10(b) must include the following: The intended use statement must include the information described in paragraph (i) (1)(i) of this section, as applicable, and a statement that reads, "It is the responsibility of a qualified pathologist to employ appropriate procedures and safeguards to assure the validity of the interpretation of images obtained using this device." (ii) A description of the technical studies and the summary of results, including those that relate to paragraph (1)(ii) and (1)(iii) of this section, as appropriate. (iii) A description of the performance studies and the summary of results, including those that relate to paragraph (1)(iv) of this section, as appropriate. (iv) A limiting statement that specifies that pathologists should exercise professional judgment in each clinical situation and examine the glass slides by conventional microscopy if there is doubt about the ability to accurately render an interpretation using this device alone.
*Classification.* Class II (special controls). The special controls for this device are:(1) Premarket notification submissions must include the following information:
(i) The indications for use must specify the tissue specimen that is intended to be used with the whole slide imaging system and the components of the system.
(ii) A detailed description of the device and bench testing results at the component level, including for the following, as appropriate:
(A) Slide feeder;
(B) Light source;
(C) Imaging optics;
(D) Mechanical scanner movement;
(E) Digital imaging sensor;
(F) Image processing software;
(G) Image composition techniques;
(H) Image file formats;
(I) Image review manipulation software;
(J) Computer environment; and
(K) Display system.
(iii) Detailed bench testing and results at the system level, including for the following, as appropriate:
(A) Color reproducibility;
(B) Spatial resolution;
(C) Focusing test;
(D) Whole slide tissue coverage;
(E) Stitching error; and
(F) Turnaround time.
(iv) Detailed information demonstrating the performance characteristics of the device, including, as appropriate:
(A) Precision to evaluate intra-system and inter-system precision using a comprehensive set of clinical specimens with defined, clinically relevant histologic features from various organ systems and diseases. Multiple whole slide imaging systems, multiple sites, and multiple readers must be included.
(B) Reproducibility data to evaluate inter-site variability using a comprehensive set of clinical specimens with defined, clinically relevant histologic features from various organ systems and diseases. Multiple whole slide imaging systems, multiple sites, and multiple readers must be included.
(C) Data from a clinical study to demonstrate that viewing, reviewing, and diagnosing digital images of surgical pathology slides prepared from tissue slides using the whole slide imaging system is non-inferior to using an optical microscope. The study should evaluate the difference in major discordance rates between manual digital (MD) and manual optical (MO) modalities when compared to the reference (
*e.g.,* main sign-out diagnosis).(D) A detailed human factor engineering process must be used to evaluate the whole slide imaging system user interface(s).
(2) Labeling compliant with 21 CFR 809.10(b) must include the following:
(i) The intended use statement must include the information described in paragraph (b)(1)(i) of this section, as applicable, and a statement that reads, “It is the responsibility of a qualified pathologist to employ appropriate procedures and safeguards to assure the validity of the interpretation of images obtained using this device.”
(ii) A description of the technical studies and the summary of results, including those that relate to paragraphs (b)(1)(ii) and (iii) of this section, as appropriate.
(iii) A description of the performance studies and the summary of results, including those that relate to paragraph (b)(1)(iv) of this section, as appropriate.
(iv) A limiting statement that specifies that pathologists should exercise professional judgment in each clinical situation and examine the glass slides by conventional microscopy if there is doubt about the ability to accurately render an interpretation using this device alone.
{0}
[LOGO]
FDA
U.S. FOOD & DRUG
ADMINISTRATION
### 510(k) SUBSTANTIAL EQUIVALENCE DETERMINATION
DECISION SUMMARY
### I Background Information:
A 510(k) Number
K253561
B Applicant
Leica Biosystems Imaging, Inc.
C Proprietary and Established Names
Aperio iQC DX Software
D Regulatory Information
| Product Code(s) | Classification | Regulation Section | Panel |
| --- | --- | --- | --- |
| SIX | Class II | 21 CFR 864.3700 - Whole Slide Imaging System | PA - Pathology |
### II Submission/Device Overview:
A Purpose for Submission:
New Device
B Type of Test:
Software only device
### III Intended Use/Indications for Use:
A Intended Use(s):
See Indications for Use below.
B Indication(s) for Use:
Aperio iQC DX Software is an artificial intelligence-based software intended to be used as an aid in the identification of artifacts in scanned whole slide images (WSIs) of hematoxylin and eosin
Food and Drug Administration
10903 New Hampshire Avenue
Silver Spring, MD 20993-0002
www.fda.gov
{1}
(H&E) and immunohistochemistry (IHC) stained, formalin-fixed paraffin embedded (FFPE) tissue that should undergo further evaluation for image quality prior to diagnostic review of the images by pathologists. Aperio iQC DX Software identifies the following artifacts:
- Digital artifacts: out of focus, image striping, missing and clipped tissue
- Slide preparation artifact: air bubbles
- Other artifact: pen marks
Aperio iQC DX Software outputs the following:
- Binary 'artifact detected' or 'no artifact detected' for air bubbles, image striping, missing and clipped tissue, out of focus, and pen marks,
- Overlays highlighting the location of air bubbles and out of focus artifacts only, and
- Bounding boxes for missing and clipped tissue artifacts.
Aperio iQC DX Software is intended to be used in conjunction with the complete in-house laboratory image quality control workflow. It is the responsibility of a qualified pathologist to employ appropriate procedures and safeguards to assure the validity of the interpretation of images when using the Aperio iQC DX Software. Aperio iQC DX Software is not intended to be used for diagnosis, prognosis, or prediction of disease.
### C Special Conditions for Use Statement(s):
Rx – For Prescription Use Only
### IV Device/System Characteristics:
### A Device Description:
Aperio iQC DX Software (Model:23iQCDXUS), version 1.0, is an Artificial Intelligence (AI) based software algorithm derived from deep learning convolutional neural networks. The Aperio iQC DX Software algorithm is locked; it is not a continuous learning (continual machine learning model) algorithm.
Aperio iQC DX Software utilizes scanned whole slide images (WSIs) as inputs from Aperio GT 450 DX and GT 180 DX scanners to identify the digital artifacts using four AI models which were developed and trained on WSIs of H&E and DAB (3,3'-diaminobenzidine) chromogen based IHC stained tissue slides originating from FFPE tissue sections. Aperio iQC DX Software generates two types of visualizations of the detected artifact regions as follows: a bounding box is provided for Missing and Clipped Tissue and overlays for Out of Focus and Air Bubble artifacts. Aperio iQC DX Software identifies the following artifacts: missing and clipped tissue, out of focus, image striping; slide preparation artifact: air bubbles, and pen marks on WSIs. It is used to augment the initial quality review performed by laboratory personnel prior to making the WSI available to the pathologist for diagnostic review. The final determination of whether the WSIs are sufficient for diagnostic review remains the responsibility of the pathologist and occurs outside of Aperio iQC DX Software.
K253561 - Page 2 of 14
{2}
Aperio iQC DX Software outputs the following:
- Binary ‘artifact detected’ or ‘no artifact detected’ for air bubbles, image striping, missing and clipped tissue, out of focus, and pen marks
- Overlays highlighting the location of air bubbles and out of focus artifacts only, and
- Bounding boxes for missing and clipped tissue artifacts
These outputs are displayed to the user in the Aperio iQC DX Dashboard (a browser-based user interface used only for non-diagnostic purposes).
Aperio iQC DX Software is operated as follows:
1. Copies of WSIs are sent to the customer’s dedicated image storage by the connected WSI system.
2. Once available, Aperio iQC DX Software automatically executes AI algorithms on the copies of the WSIs from the customer’s dedicated image storage.
3. For WSI, Aperio iQC DX Software outputs ‘artifact detected’ or ‘no artifact detected’ for each artifact. In addition, whenever out of focus or air bubble artifacts are detected, overlays indicating the presence of the artifact are available. For missing and clipped tissue, a bounding box is provided.
4. All findings are displayed to the user in the Aperio iQC DX Dashboard where the user (e.g., laboratory technicians) will perform image quality review and ‘accept’ or ‘reject’ the image and include comments for any rejection.
The intended use interoperable components of the Aperio iQC DX Software device are identified in Table 1 below.
Table 1: Intended Use Components
| Component |
| --- |
| Aperio GT 450 DX Scanner (cleared under K253554) |
| Aperio GT 180 DX Scanner (cleared under K253554) |
| Aperio SAM DX Software (cleared under K253554) |
Computer Environment/System Requirements for the Aperio iQC DX Software device are identified in Table 2 below.
Table 2: Customer Environment/System Requirements
| Workstation Component | Specification |
| --- | --- |
| Computer System | RAM: 64 GB per scanner CPU: Intel Core i7 or higher, or AMD EPYC 9004 and 8004 Series |
| Web Browser | Google Chrome: 126 or later Microsoft Edge: 126 or later Firefox: 127 or later |
| Operating System | Linux Ubuntu 24.04 LTS or higher |
K253561 - Page 3 of 14
{3}
| Workstation Component | Specification |
| --- | --- |
| Network | Internet access At least 100 Mbps upload speed |
### B Instrument Description Information:
1. Instrument Name:
Aperio iQC DX Software
2. Specimen Identification:
Aperio iQC DX Software uses an image of the barcode on the slide for identification.
3. Specimen Sampling and Handling:
Specimen sampling and handling are performed upstream and independent of the use of the subject device. Specimen sampling includes biopsy or resection specimens which are processed using histology techniques. The FFPE tissue section is H&E or IHC stained. Digital images are then obtained from these glass slides using the intended use scanners specified in Table 1 above.
4. Calibration:
Aperio iQC DX Software is a software only device that does not require calibration.
5. Quality Control:
After evaluation and identification of artifacts in the scanned WSIs of the intended use slides by the Aperio iQC DX Software, any WSI with artifacts is reviewed by the appropriate laboratory personnel, and the subsequent corrective action depends on the type of artifact identified. For digital artifacts (e.g., out of focus, image striping, missing and clipped tissue), the affected slide(s) may be rescanned as needed; however, for histological artifacts (e.g., air bubbles), the physical glass slide(s) must first be reprocessed or otherwise corrected before rescanning.
### V Substantial Equivalence Information:
A Predicate Device Name(s):
Aperio GT 450 DX System
B Predicate 510(k) Number(s):
K253554
K253561 - Page 4 of 14
{4}
# C Comparison with Predicate(s):
| Device & Predicate Device(s): | K253561 Subject Device | K253554 (Predicate Device) |
| --- | --- | --- |
| Device Trade Name | Aperio iQC DX Software | Aperio GT 450 DX |
| **General Device Characteristic Similarities** | | |
| **Intended Use/Indications For Use** | Aperio iQC DX Software is an artificial intelligence-based software intended to be used as an aid in the identification of artifacts in scanned whole slide images (WSIs) of hematoxylin and eosin (H&E) and immunohistochemistry (IHC) stained, formalin-fixed paraffin embedded (FFPE) tissue that should undergo further evaluation for image quality prior to diagnostic review of the images by pathologists. Aperio iQC DX Software identifies the following artifacts: • Digital artifacts: out of focus, image striping, missing and clipped tissue • Slide preparation artifact: air bubbles • Other artifact: pen marks Aperio iQC DX Software outputs the following: • Binary 'artifact detected' or 'no artifact detected' for air bubbles, image striping, missing and clipped tissue, out of focus, and pen marks, • Overlays highlighting the location of air bubbles and out of focus artifacts only, and • Bounding boxes for missing and clipped tissue artifacts. Aperio iQC DX Software is intended to be used in conjunction with the complete in-house laboratory image quality control workflow. It is the responsibility of a qualified pathologist to employ appropriate procedures and safeguards to assure the validity of the interpretation of images when using the Aperio iQC DX Software. Aperio iQC DX Software is not intended to be used for diagnosis, prognosis, or prediction of disease. | The Aperio GT 450 DX is an automated digital slide creation and viewing system. The Aperio GT 450 DX is intended for in vitro diagnostic use as an aid to the pathologist to review and interpret digital images of surgical pathology slides prepared from formalin-fixed paraffin embedded (FFPE) tissue. The Aperio GT 450 DX is for creation and viewing of digital images of scanned glass slides that would otherwise be appropriate for manual visualization by conventional light microscopy. The Aperio GT 450 DX is not intended for use with frozen section, cytology, or non-FFPE hematopathology specimens. It is the responsibility of a qualified pathologist to employ appropriate procedures and safeguards to assure the validity of the interpretation of images obtained using the Aperio GT 450 DX. |
| Specimen Type | Formalin-Fixed Paraffin- Embedded (FFPE) specimen | Same |
| Tissue Finding | Automatic | Same |
K253561 - Page 5 of 14
{5}
| Image Storage | Images stored in the end-user- provided image storage on the local network. | Same |
| --- | --- | --- |
| **General Device Characteristic Differences** | | |
| Device Type | Artificial intelligence-based software application based on convolutional neural networks (CNNs) | Traditional image analysis software feature (Tissue Finder) |
| Auto QC during scanning | No | Yes |
| Viewer | N/A; images are viewed within the Aperio iQC DX dashboard (not for diagnostic use) | Aperio WebViewer DX |
| Compatible Display (Monitor) | Off-the-Shelf Monitor since this is not for diagnostic use. | Aperio WebViewer DX is intended for use with specific cleared intended use interoperable displays. |
| Image Manipulation Functions | Panning/Zooming (non- continuous) | Aperio WebViewer DX permits continuous panning/zooming, ability to compare images, annotations, digital bookmarks, measurements (distance), gamma function |
| Operational Mode | Image QC; not to be used to diagnose disease | Diagnostic device |
| Hardware Inputs | Aperio GT 450 DX and Aperio GT 180 DX • Bounding Box • Macro image • Scan data • Scanner configurations (rack position, rescan capability, scanner identifiers) • Hash files | Tissue finder receives the following from the scanner: • Config Values • Macro Image • Dirt Image |
| Output | • Binary artifact detected or no artifact detected • Overlays (air bubbles, out of focus) • Bounding Box (missing and clipped tissue) | Bounding Box indicating where the scan area (Tissue Finder) and WSI |
### VI Standards/Guidance Documents Referenced:
1. Guidance for Industry and Food and Drug Administration Staff: Applying Human Factors and Usability Engineering to Medical Devices (February 2016).
2. Guidance for Industry and Food and Drug Administration Staff: Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions (June 2025).
3. Guidance for Food and Drug Administration Staff: Content of Premarket Submissions for Device Software Functions (June 2023).
### VII Aperio iQC DX Software Model Development
Aperio iQC DX Software development was performed on training, tuning, and test datasets. Each dataset contained slides from unique patients ensuring that training, tuning and test datasets
K253561 - Page 6 of 14
{6}
do not have any WSIs in common. WSIs were labeled as artifact present or 'none'. These datasets were completely independent from the validation dataset.
Table 3: Data Split for Training, Tuning and Test Sets, Algorithm Development
| Training Dataset | Tuning Dataset | Test Dataset |
| --- | --- | --- |
| De-identified slides were procured from external vendors or prepared at an internal site and scanned with Aperio GT 450 scanners. Number of slide images: 12,278 (5158 H&E WSIs, 7127 IHC WSIs) | De-identified slides were procured from external vendors or prepared at an internal site and scanned with Aperio GT 450 scanners. Number of slide images: 2,450 (1455 H&E WSIs, 995 IHC WSIs) | De-identified slides were procured from external vendors or prepared at an internal site and scanned with Aperio GT 450 scanners. Number of slide images: 4,752 (1828 H&E WSIs, 2924 IHC WSIs) |
Demographics and geographic considerations were not observed as the regional variation in tissue morphology was not expected to influence artifact presence or detection.
### VIII Performance Characteristics (if/when applicable):
Performance testing was conducted with the Aperio iQC DX Software which included: algorithm accuracy study, algorithm localization accuracy study and precision studies.
#### 1. Algorithm Accuracy Study
The algorithm accuracy study assessed the performance of the device to detect and localize digital and histological artifacts, including missing/clipped tissue, image striping, out of focus areas, air bubbles, and pen marks when single and multiple artifacts were present. The study included a diverse dataset of WSIs across multiple intended use tissue types, staining methods (H&E and IHC) and both the intended use scanners (Aperio GT 450 DX and GT 180 DX). This dataset was independent of the training dataset.
The evaluation was conducted by comparing Aperio iQC DX Software's results against the ground truth annotations determined by qualified experts.
Device accuracy was tested for single and multiple artifacts. For single artifacts, WSIs were created using the Aperio GT 450 DX scanner and the Aperio GT 180 DX scanner.
The Aperio GT 450 DX scanner dataset consisted of the following: Missing and Clipped Tissue - 2568 WSIs (1284 positive, 1284 negative), Image Striping - 2222 WSIs (1167 positive, 1055 negative), Out of Focus - 924 WSIs (221 positive, 703 negative), Air Bubbles - 2134 (883 positive, 1251 negative), and Pen Marks - 2478 WSIs (1376 positive, 1102 negative).
Study results are provided in the tables below.
K253561 - Page 7 of 14
{7}
Table 4: Performance data of Aperio iQC DX using Aperio GT 450 DX scanner by each artifact independently
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 93.73 | (92.73, 94.60) | 94.78 | (93.42, 95.87) | 92.68 | (91.12, 93.98) |
| Image Striping | 99.82 | (99.54, 99.93) | 99.74 | (99.24, 99.91) | 99.91 | (99.47, 99.98) |
| Out of Focus | 97.29 | (96.04, 98.16) | 93.67 | (89.65, 96.19) | 98.44 | (97.22, 99.12) |
| Air Bubble | 95.60 | (94.64, 96.39) | 93.20 | (91.35, 94.68) | 97.28 | (96.22, 98.05) |
| Pen Mark | 99.19 | (98.75, 99.47) | 99.13 | (98.48, 99.50) | 99.27 | (98.57, 99.63) |
Table 5: Performance of Aperio iQC DX Software using Aperio GT 450 DX Scanner and H&E-stained slides by evaluating each artifact independently
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 94.76 | (93.05, 96.06) | 95.80 | (93.47, 97.33) | 93.71 | (91.00, 95.64) |
| Image Striping | 100 | (99.42,100.00) | 100 | (99.07,100.00) | 100 | (98.50,100.00) |
| Out of Focus | 96.62 | (94.41,97.98) | 93 | (86.25,96.57) | 97.77 | (95.47,98.92) |
| Air Bubble | 98.40 | (97.16,99.10) | 99.61 | (97.82,99.93) | 97.69 | (95.80,98.74) |
| Pen Mark | 99.14 | (98.31,99.56) | 99.06 | (97.82,99.60) | 99.26 | (97.85,99.75) |
Table 6: Performance of Aperio iQC DX Software using Aperio GT 450 DX Scanner and IHC-stained slides by evaluating each artifact independently
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 93.22 | (91.93, 94.31) | 94.27 | (92.50, 95.64) | 92.16 | (90.17, 93.78) |
| Image Striping | 99.74 | (99.34,99.90) | 99.60 | (98.84,99.86) | 99.88 | (99.31,99.98) |
| Out of Focus | 97.84 | (96.18,98.79) | 94.21 | (88.53,97.17) | 98.97 | (97.38,99.60) |
| Air Bubble | 94.26 | (92.94,95.35) | 90.61 | (88.08,92.65) | 97.07 | (95.68,98.02) |
| Pen Mark | 99.22 | (98.64,99.55) | 99.17 | (98.30,99.60) | 99.28 | (98.33,99.69) |
The Aperio GT 180 DX scanner dataset consisted of the following: Missing and Clipped Tissue - 576 WSIs (288 positive, 288 negative), Image Striping - 439 WSIs (158 positive, 281 negative), Out of Focus - 486 WSIs (243 positive, 243 negative), Air Bubbles - 645 (190 positive, 455 negative), and Pen Marks - 424 WSIs (142 positive, 282 negative). Study results are provided in the tables below.
K253561 - Page 8 of 14
{8}
Table 7: Validation performance of Aperio iQC DX in Aperio GT 180 DX scanner by evaluating each artifact independently
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 93.40 | (91.07, 95.15) | 94.10 | (90.75, 96.28) | 92.71 | (89.11, 95.18) |
| Image Striping | 99.77 | (98.72, 99.96) | 99.37 | (96.51, 99.89) | 100 | (98.65, 100) |
| Out of Focus | 93.62 | (91.09, 95.47) | 92.18 | (88.11, 94.94) | 95.06 | (91.57, 97.15) |
| Air Bubble | 96.43 | (94.70, 97.61) | 93.16 | (88.65, 95.96) | 97.80 | (97.51, 99.55) |
| Pen Mark | 99.06 | (97.60, 99.63) | 99.30 | (96.13, 99.88) | 98.94 | (96.93, 99.64) |
Table 8: Performance of Aperio iQC DX Software using Aperio GT 180 DX scanner and H&E-stained slides by evaluating each artifact independently
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 94.27 | (90.03,96.77) | 95.83 | (89.77,98.37) | 92.71 | (85.71,96.42) |
| Image Striping | 100 | (97.37,100.00) | 100 | (92.29,100.00) | 100 | (96.15,100.00) |
| Out of Focus | 96.10 | (92.76,97.93) | 97.89 | (93.98,99.28) | 93.26 | (86.07,96.87) |
| Air Bubble | 97.51 | (94.31,98.93) | 95.65 | (87.98,98.51) | 98.48 | (94.64,99.58) |
| Pen Mark | 99.31 | (96.18,99.88) | 100 | (92.59,100.00) | 98.96 | (94.34,99.82) |
Table 9: Performance of Aperio iQC DX Software using Aperio GT 180 DX Scanner and IHC-stained slides by evaluating each artifact independently
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 92.97 | (89.96,95.12) | 93.23 | (88.76,96.00) | 92.71 | (88.14,95.61) |
| Image Striping | 99.66 | (98.11,99.94) | 99.11 | (95.12,99.84) | 100 | (97.97,100.00) |
| Out of Focus | 91.37 | (87.28,94.23) | 84.16 | (75.81,90.01) | 96.10 | (91.76,98.20) |
| Air Bubble | 95.95 | (93.69,97.42) | 91.74 | (85.46,95.45) | 97.52 | (95.19,98.74) |
| Pen Mark | 98.93 | (96.90,99.64) | 98.94 | (94.22,99.81) | 98.92 | (96.16,99.70) |
To evaluate multiple artifacts, WSIs were created using the GT 450 DX scanner to demonstrate performance for the detection of each artifact when multiple artifacts are present. Two data sets were used for multiple artifact testing. The first dataset was used for evaluating performance of air bubbles, image striping, out of focus, and pen marks and consisted of: 522 WSIs positive for multiple artifacts (241 doubles, 106 triples, 175
K253561 - Page 9 of 14
{9}
quadruple). The second dataset was used for evaluating performance of missing and clipped tissue and consisted of: 505 WSIs positive for multiple artifacts (172 doubles, 132 triples, 138 quadruple, 63 quintuple). Total positive and negative counts for multiple artifacts were as follows: Missing and Clipped tissue (305 positive, 200 negative), Image Striping (440 positive, 82 negative), Out of Focus (368 positive, 154 negative), Air Bubbles (308 positive, 214 negative), and Pen Marks (384 positive, 138 negative).
Table 10: Overall performance of Aperio iQC DX Software using Aperio GT 450 DX Scanner for multiple artifacts
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 93.27 | (90.74, 95.14) | 93.11 | (89.70, 95.45) | 93.50 | (89.20, 96.16) |
| Image Striping | 99.43 | (98.33, 99.81) | 99.32 | (98.02, 99.77) | 100 | (95.52, 100) |
| Out of Focus | 95.79 | (93.71, 97.20) | 96.47 | (94.05, 97.93) | 94.16 | (89.27, 96.90) |
| Air Bubble | 95.40 | (93.25, 96.90) | 92.53 | (89.04, 94.97) | 99.53 | (97.40, 99.92) |
| Pen Mark | 97.70 | (96.02, 98.68) | 99.48 | (98.12, 99.86) | 92.75 | (87.17, 96.01) |
Table 11: Performance of Aperio iQC DX Software using Aperio GT 450 DX scanner and H&E-stained slides for evaluating multiple artifacts
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 94.67 | (90.91, 96.92) | 97.14 | (92.88, 98.88) | 90.59 | (82.51, 95.15) |
| Image Striping | 99.22 | (97.19, 99.78) | 99.06 | (96.64, 99.74) | 100 | (91.62, 100.00) |
| Out of Focus | 95.69 | (92.44, 97.57) | 95.60 | (91.57, 97.76) | 95.89 | (88.60, 98.59) |
| Air Bubble | 96.86 | (93.93, 98.40) | 95.68 | (91.35, 97.89) | 98.92 | (94.16, 99.81) |
| Pen Mark | 96.08 | (92.93, 97.86) | 100 | (97.78, 100.00) | 88.37 | (79.90, 93.56) |
Table 12: Performance of Aperio iQC DX Software using Aperio GT 450 DX scanner and IHC-stained slides for evaluating multiple artifact
| | Accuracy (%) | 95% CI Accuracy (%) | Sensitivity (%) | 95% CI Sensitivity (%) | Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- | --- | --- |
| Missing and Clipped Tissue | 92.14 | (88.39, 94.75) | 89.70 | (84.12, 93.47) | 95.65 | (90.22, 98.13) |
| Image Striping | 99.63 | (97.91, 99.93) | 99.56 | (97.55, 99.92) | 100 | (91.24, 100.00) |
| Out of Focus | 95.88 | (92.77, 97.68) | 97.31 | (93.86, 98.85) | 92.59 | (84.77, 96.56) |
| Air Bubble | 94.01 | (90.49, 96.28) | 89.04 | (82.94, 93.14) | 100 | (96.92, 100.00) |
| Pen Mark | 99.25 | (97.31, 99.79) | 99.07 | (96.67, 99.74) | 100 | (93.12, 100.00) |
K253561 - Page 10 of 14
{10}
The results of the accuracy study demonstrate that overall performance of Aperio iQC DX Software achieved sensitivity and specificity of ≥90% and ≥85% for the lower bound of corresponding 95% CI for single and multiple artifact detection, for the analysis of the combined H&E and IHC dataset.
## 2. Localization Accuracy
To evaluate the performance of the device in localization of Missing and Clipped Tissue, the overlap of Aperio iQC DX Software's predicted bounding box and the ground truth bounding box was calculated using intersection over union (IoU) ratio. The slide level average of sensitivity and specificity were reported. IoU was calculated for each WSI, then averaged across the dataset and reported.
To evaluate the performance in Out of Focus and Air Bubble localization, the pixel-level sensitivity and specificity for each WSI were computed by comparing the predicted artifact regions against the ground truth artifact regions. The ground truth was established through annotations performed by qualified expert annotators. For Missing and Clipped Tissue, annotators drew bounding boxes encompassing all visible tissue in the macro image, and the ground truth label was determined by comparing the annotated bounding box against the scanner's scanned region. For Out of Focus and Air Bubble localization, annotators placed positive and negative non-overlapping bounding boxes on representative tissue regions with and without the respective artifacts. A second annotator independently reviewed all annotations for quality control purposes, with a third-party arbiter consulted in cases of discordance. Sensitivity and specificity were calculated for each WSI, then averaged across the dataset and reported.
The dataset for localization performance included 1,988 WSIs with missing and clipped tissue, 67 WSIs with out of focus artifacts and 60 WSIs with air bubble artifacts. Study results are shown in the Tables below.
The results of the localization accuracy study demonstrate that Aperio iQC DX Software achieved ≥90% of WSIs with IoU ≥70% for missing and clipped tissue artifact detection and achieved ≥90% in sensitivity and specificity for out of focus and air bubble artifact detection.
Table 13: Localization performance in Missing and Clipped Tissue detection
| | Average IoU | Number of WSIs achieve IoU ≥ 70% | Ratio of the WSIs achieve IoU ≥ 70% |
| --- | --- | --- | --- |
| Missing and Clipped Tissue | 88.41% | 1804 | 90.74% |
Table 14: Localization performance in Out of Focus and Air Bubble detection
| | Localization Sensitivity (%) | 95% CI Sensitivity (%) | Localization Specificity (%) | 95% CI Specificity (%) |
| --- | --- | --- | --- | --- |
| Out of Focus | 96.99 | (95.44%, 98.02%) | 99.72 | (99.02%, 99.92%) |
| Air Bubble | 91.29 | (88.75%, 93.30%) | 96.01 | (94.12%, 97.31%) |
K253561 - Page 11 of 14
{11}
### 3. Precision
The precision (repeatability and reproducibility) testing was conducted to evaluate the Aperio iQC DX Software using H&E and IHC WSIs generated from the Aperio GT 450 DX Scanner and Aperio GT 180 DX Scanner.
The dataset for H&E comprised of 63 whole slide images (WSIs) per scanner:
- 3 WSIs per artifact across five artifacts (positive cases)
- 3 WSIs containing multiple artifacts (positive cases)
- 3 WSIs that contain no artifacts (negative cases)
Each WSI was rotated in two different orientations generating three replicates per WSI. The dataset for IHC was comprised of 63 whole slide images (WSIs) per scanner:
- 3 WSIs per artifact across five artifacts (positive cases)
- 3 WSIs containing multiple artifacts (positive cases)
- 3 WSIs that contain no artifacts (negative cases)
Each WSI was rotated in two different orientations generating three replicates per WSI. Ground truth was established through annotations performed by a group of qualified annotators.
Evaluation Method: To assess precision under variable conditions, each WSI was subjected to two deterministic image orientations, generating three replicates per WSI, for a total of nine replicate evaluations per artifact category.
The validation data demonstrates that Aperio iQC DX Software achieves 100% for correct call for all types of cases. Results are provided in the tables below.
Table 15: Precision performance under variable conditions for H&E WSIs using GT 450 DX Scanner
| Type of Cases | Type of Artifacts | #WSIs | #Replicates /WSI | #Total Replicates | %Correct Call | 95%CI (%) * |
| --- | --- | --- | --- | --- | --- | --- |
| Positive Cases-Single Artifact | Missing and Clipped Tissue | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Image Striping | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Out of Focus | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Air Bubble | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Pen Mark | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Positive Cases-Multiple Artifacts | Multiple Artifacts | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Negative Cases | No Artifact | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Overall | | | | 63 | 100% | (94.3%, 100%) |
*95%CI was calculated by Wilson score method
K253561 - Page 12 of 14
{12}
Table 16: Precision performance under variable conditions for H&E WSIs using GT 180 DX Scanner
| Type of Cases | Type of Artifacts | #WSIs | #Replicates /WSI | #Total Replicates | %Correct Call | 95%CI (%) * |
| --- | --- | --- | --- | --- | --- | --- |
| Positive Cases-Single Artifact | Missing and Clipped Tissue | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Image Striping | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Out of Focus | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Air Bubble | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Pen Mark | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Positive Cases-Multiple Artifacts | Multiple Artifacts | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Negative Cases | No Artifact | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Overall | | | | 63 | 100% | (94.3%, 100%) |
*95%CI was calculated by Wilson score method
Table 17: Precision performance under variable conditions for IHC WSIs using GT 450 DX Scanner
| Type of Cases | Type of Artifacts | #WSIs | #Replicates /WSI | #Total Replicates | %Correct Call | 95%CI (%) * |
| --- | --- | --- | --- | --- | --- | --- |
| Positive Cases-Single Artifact | Missing and Clipped Tissue | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Image Striping | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Out of Focus | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Air Bubble | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Pen Mark | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Positive Cases-Multiple Artifacts | Multiple Artifacts | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Negative Cases | No Artifact | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Overall | | | | 63 | 100% | (94.3%, 100%) |
*95%CI was calculated by Wilson score method
K253561 - Page 13 of 14
{13}
Table 18: Precision performance under variable conditions for IHC WSIs using GT 180 DX Scanner
| Type of Cases | Type of Artifacts | #WSIs | #Replicates /WSI | #Total Replicates | %Correct Call | 95%CI (%) * |
| --- | --- | --- | --- | --- | --- | --- |
| Positive Cases-Single Artifact | Missing and Clipped Tissue | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Image Striping | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Out of Focus | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Air Bubble | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| | Pen Mark | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Positive Cases-Multiple Artifacts | Multiple Artifacts | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Negative Cases | No Artifact | 3 | 3 | 9 | 100% | (70.1%, 100%) |
| Overall | | | | 63 | 100% | (94.3%, 100%) |
*95%CI was calculated by Wilson score method
### IX Other Supportive Instrument Performance Characteristics Data:
Human Factors and Usability testing was completed, and documentation was provided in the submission as recommended in Guidance for Industry and Food and Drug Administration Staff: Applying Human Factors and Usability Engineering to Medical Devices (February 2016).
Software and cybersecurity documentation was reviewed and found to be acceptable.
### X Proposed Labeling:
The labeling supports the finding of substantial equivalence for Aperio iQC DX Software device.
### XI Conclusion:
The submitted information in this premarket notification is complete and supports a substantial equivalence decision.
K253561 - Page 14 of 14
Predicate graph will load when search results are available.
Embedding visualization will load when search results are available.
PDF viewer will load when search results are available.
Loading panels...
Select an item from Submissions
Click any panel, subpart, regulation, product code, or device to see details here.
Section Matches
Results will appear here.
Product Code Matches
Results will appear here.
Special Control Matches
Results will appear here.
Loading collections...
Loading
My Alerts
You will receive email notifications based on the filters and frequency you set for each alert.
Sort by:
Create Alert
Search Filters
Agent Token
Create a read-only bearer token for Claude, ChatGPT, or other agents that can call HTTP APIs.
Copy this now. It will not be shown again.
Connected apps
Apps you authorized through browser sign-in. Disconnecting revokes their access immediately.
Learn the FDA Browser
Two short videos show you everything — or skip straight to the written tutorial if you'd rather read. You can reopen this any time from the Tutorial button in the top bar.
Part 1 — Search, results, and everyday workflows 16 min
Part 2 — Embeddings: the galaxy map 3 min
1. Search: exact and fuzzy
Type a phrase like "coronary artery calcification" into the search box. You get two kinds of results. Exact results match the literal phrase — prefix searches work ("coronary artery calcificati") but suffix searches do not. Fuzzy results match on the meaning and intent of your phrase rather than the exact words, and are sorted by relevance score. Hover over the Exact or Fuzzy badge on any row to see exactly why it matched.
Use the checkboxes above the results to narrow: SaMD keeps only software-only devices, AI / ML keeps only devices with AI.
Exact vs. fuzzy search: what's the difference?
Exact matches on the literal phrase (prefix search works, suffix does not). Fuzzy matches on the meaning and intent of the phrase rather than the exact words. Hover over the badge on any row to see why it matched.
You search "coronary artery calcification" and want only software devices with AI. What two filters do you apply?
Narrow by SaMD (software-only devices), then narrow by AI/ML (devices with AI).
2. The results table
Scroll right in the results table. The intended use is extracted for you — no need to open the PDF. The device story gives a high-level snapshot of what the device does and how it's used. The AI Performance sub-table shows each output name, acceptance criteria, observed values, and development/test dataset descriptions — the same format Innolitics uses for regulatory strategy outputs, and the fastest high-level fingerprint of an AI device. It is AI-generated but has been very reliable in practice.
Where do you find a device's intended use without opening the PDF?
Scroll right in the search results table. The intended use column is extracted for you; no need to dig into the 510(k) summary PDF.
What does the AI Performance sub-table show, and why is it useful?
Output name, acceptance criteria, observed values, development dataset description, and test dataset description. It's the same format we use for regulatory strategy output and Fast 510(k) input, and the fastest high-level fingerprint of an AI device. AI-generated but reliable in practice.
3. Judging fuzzy relevance
Fuzzy results trail off in relevance as you scroll. Use three signals to decide how far down to go: the fuzzy badge explanations, the intended use column, and whether your target output (e.g., Cobb angle) still appears in the AI Performance sub-table. Once it stops appearing, you're past the relevant zone. A top hit with a low score (~0.4) and a stretched explanation is a hint the closest predicates are far away — the project may be headed for De Novo. Note the fuzzy search is a pattern match: it doesn't handle negation ("not") well, and hardware devices can appear — filter by SaMD/AI ML to cut them.
How do you judge how far down fuzzy search results to go?
Use the relevancy signals: the fuzzy badge explanations, the intended use column, and whether the target output (e.g., Cobb angle) still appears in the AI Performance sub-table. Once it stops appearing, results are trailing off in relevancy.
4. Device detail page: chat and citations
Click a device name to open its detail page: device facts on the left, a chat window on the right. Ask something like "Describe the training data". The answer carries little citation bubbles — click one to jump to the highlighted passage in the source PDF, so you can verify every AI answer against the document. There's also a Download PDF button for sharing.
How do you verify an AI chat answer on the device detail page?
Click the citation bubbles to jump to the relevant highlight in the source document.
Reading rule for every project: how many summaries do you read in full?
At least the three most relevant 510(k) or De Novo summaries, in full. After that, use targeted chat questions to confirm your memory quickly. The tool supports this professional habit — it doesn't replace it.
5. Side-by-side comparison
Select multiple rows in the results table (aim for under ~10), then open the PDF Viewer tab. Ask one question — it goes to all selected devices in parallel, each with citations. This is the fastest way to compare and contrast devices: training data, PCCP scope, how they handled adding new scanners, and so on.
What does the side-by-side PDF viewer mode do?
Select multiple devices, open the PDF viewer tab, and ask one question (e.g., "Describe the training data"). It queries all selected devices simultaneously with citations, so you can compare and contrast quickly.
6. Collections
With rows selected, go to the Collections tab and create a labeled collection (e.g., "Cobb Angle Project"). Reload that selection any time — before a client call, pull up the collection and ask questions across all of its devices at once.
How do you save a set of selected devices for later use?
Select the rows, go to the Collections tab, and create a labeled collection (e.g., "Cobb Angle Project"). You can reload the selection anytime and carry it into the PDF viewer and other tabs that support selections.
7. Product codes and the regulations tree
Click a product code in the results to jump to it in the regulations tree — identification text, sibling product codes, and devices you can open in a PDF viewer on the right. Click a regulation number to see its identification, special controls, and related product codes. You can also search by product code or regulation number at the top of the tree. Always read the special controls if any exist for your device — it broadens your search and sharpens pre-kickoff research.
What can you do from the regulations tree view?
Browse product codes and regulation numbers, read the identification text and special controls, browse sibling product codes, open device PDFs on the right, and search by product code or regulation number at the top of the tree.
8. Chart view
Click Show Chart and segment by regulation number (or product code) to see which regulations dominate your result set. Clicking a regulation takes you into the regulations tree. Great for spotting that most matches are, say, hardware laparoscopic devices — a cue to go back and filter.
How do you see which regulations dominate a search result set?
Click "Show Chart" and segment by Regulation Number. Clicking a regulation takes you to the regulations tree.
9. The predicate graph
Open the Predicates tab for a family-tree view of predicate relationships. Click a node to trace its parents and children; selections from search carry over pre-selected. Commonly predicated devices are worth reading — a lot of people predicated them for a reason. The visual lineage is also handy on client calls, e.g. to show how a predicate family evolved and justify why your predicate still holds.
In the predicate graph, why are commonly predicated devices worth reading?
A lot of people predicated them for a reason. Clicking a node traces parents and children, and selections from search carry over pre-selected.
10. Embeddings: the galaxy map
The Embeddings tab plots every matching document in a 2-D "galaxy map" where semantically similar devices cluster together. Hover or click clusters to explore, and let AI label the clusters for you. Embeddings beat product codes for grouping: two devices can carry different product codes (LLZ vs. QIH) yet do the same thing — the embedding captures the meaning of the intended use and device story. This is also exactly how retrieval-augmented generation (RAG) works under the hood, and it makes a great visual on client calls.
Try it yourself
Head to the search page and work through a few of these AI/ML fuzzy searches to build intuition: perivascular fat on CT · aortic valve calcification opportunistic screening on noncontrast CT · breast cancer prediction on digital pathology slides · autism detection · gestational age prediction · a hearing aid that can also detect a pulse · foundation model based analysis of ECG · large language models · penetration test. Watch how the relevance scores, intended use, and AI Performance tables tell you when results stop being meaningful.
