Pathology
Conditions
Brief summary
The primary objective of this study is to evaluate the safety and accuracy of the Hamamatsu WSI compared to those of the reference method (conventional light microscope (Glass)) under clinical use conditions as an aid for pathologists to view, review and diagnose digital images of surgical pathology slides. The primary endpoint is the indicator of major discordance in primary diagnosis between ground truth case diagnosis and case diagnosis by each modality, WSI and Glass, separately.
Interventions
DIAGNOSTIC_TESTWhole Slide Imaging
Scanning of a glass slide to create a digital image that can be viewed on a monitor
DIAGNOSTIC_TESTLight Microscopy
Use of traditional light microscopy per institutional standard practice
Sponsors
Hamamatsu Photonics K.K.
Study design
Allocation
RANDOMIZED
Intervention model
SEQUENTIAL
Primary purpose
DIAGNOSTIC
Masking
SINGLE (Outcomes Assessor)
Masking description
Adjudicators are blinded to the imaging modality.
Intervention model description
Cases are assessed using both glass slide light microscopy and NanoZoomer digital whole slide imaging in a random order
Eligibility
Sex/Gender
ALL
Healthy volunteers
No
Inclusion criteria
* Cases originating from and that were diagnosed at that local site * Cases are available in the site's archive * Cases are at least 1 year old since accessioning * Cases are selected because their primary diagnosis is consistent with the assigned target categories * Cases have a set of slides representative of the primary diagnosis for which it has been selected Slide selection for a given case must meet the following criteria: * Slide is obtained by surgical pathology and prepared from FFPE human tissue * Slides must be stained with H&E and accompanying special stains (histochemical and/or immunohistochemical) * All special stains slides (histochemical and/or immunohistochemical) where the slide and stain is used for diagnosis, not prognosis. * A chosen slide must demonstrate and be representative of the primary diagnosis; 1 slide selection may suffice for biopsy cases, * For resection cases, a minimum of 5 slides must be selected, which represent the primary diagnosis. If represented with less than 5 slides, additional slides (primary, secondary, or benign slides) from same case may be used to fulfill minimum number * Slide is intact, has correct size/thickness, good edges, undamaged coverslip, without pen markings that can't be removed, no air bubbles, tidy labels, and fulfills the quality checks per the general clinical practice
Exclusion criteria
* Case does not have relevant slides or if case information necessary for the study is missing * Case is still active (less than 1 year old) at the local site * Cases for which the control slides for immunohistochemistry and special stains are not available * Two cases from same individual * Gross-only cases that have no slides * Cases that are frozen section, cytology or hematology or immunofluorescence specimens only * Case where the only available set of slides have evidence only of secondary or no diagnoses and not the primary diagnosis for which the case is being screened. Slides for a given case will be excluded if they meet the following criterion: • Glass slide that is broken, has abnormal size/thickness, beveled edges, poor coverslip (cracks, waviness, scratches), is sticky, has many pen markings or dirt that cannot be removed, contains air bubbles and overhanging labels that can't be corrected, and if stain is severely faded.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Major Discordance Rate | 1 day | Indpendent reads by four Reading Pathologists (RPs) of both imaging modalitites (8 reads/case) were compared to the original diagnosis (ground truth or GT) by an independent adjudication process. This resulted in one of four adjudication outcomes for each read: Match (read = GT), Minor (minor discordance between read & GT), Major (major discordance between read & GT), or Deferred (read deferred by RP and excluded from the primary endpoint analysis). The outcome measure was the rate at which major discordances occurred for each modality. |
Countries
United States
Participant flow
Pre-assignment details
Cases were assessed using both glass slide light microscopy and digital whole slide imaging in a random sequence. Therefore, each case had 2 primary diagnoses in the study.
Participants by arm
| Arm | Count |
|---|---|
| All Cases All cases were assessed via Whole Slide Imaging using the Hamamatsu NanoZoomer S360MD Digital Slide Scanner System to assess pathological characteristics of scanned slides and traditional light microscopy using glass slides. The order in which each modality is used was randomly assigned. | 2,000 |
| Total | 2,000 |
Baseline characteristics
| Characteristic | All Cases | — |
|---|---|---|
| Age, Continuous | 59.6 Years STANDARD_DEVIATION 14.8 | — |
| Race and Ethnicity Not Collected | — | — Participants |
| Region of Enrollment United States | 2000 participants | — |
| Sex: Female, Male Female | 1046 Participants | — |
| Sex: Female, Male Male | 954 Participants | — |
Adverse events
| Event type | EG000 affected / at risk |
|---|---|
| deaths Total, all-cause mortality | 0 / 0 |
| other Total, other adverse events | 0 / 0 |
| serious Total, serious adverse events | 0 / 0 |
Outcome results
Primary
Major Discordance Rate
Indpendent reads by four Reading Pathologists (RPs) of both imaging modalitites (8 reads/case) were compared to the original diagnosis (ground truth or GT) by an independent adjudication process. This resulted in one of four adjudication outcomes for each read: Match (read = GT), Minor (minor discordance between read & GT), Major (major discordance between read & GT), or Deferred (read deferred by RP and excluded from the primary endpoint analysis). The outcome measure was the rate at which major discordances occurred for each modality.
Time frame: 1 day
Population: Two modalities (glass \& WSI) are read by 4 reading pathologists at each site for a total of 16,000 reads. Any reads deferred by a reading pathologist are excluded.
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| NanoZoomer Whole Slide Imaging | Major Discordance Rate | 3.5 Percentage of Major Discordances |
| Glass Slide Light Microscopy | Major Discordance Rate | 3.1 Percentage of Major Discordances |
95% CI: [-0.1, 1]Mixed Models Analysis