Alzheimer Disease
Conditions
Keywords
flortaucipir F 18, Positron Emission Tomography, 18F-AV-1451, Diagnostic imaging, Tau imaging
Brief summary
This study will evaluate the performance of physician readers trained to read flortaucipir-PET (positron emission tomography) scans.
Interventions
DRUGflortaucipir F 18
No study drug will be administered.
Sponsors
Avid Radiopharmaceuticals
Study design
Allocation
NA
Intervention model
SINGLE_GROUP
Primary purpose
DIAGNOSTIC
Masking
NONE
Masking description
PET scans were obtained in an open-label fashion.
Intervention model description
Physician PET scan readers are participants, blinded to demographic and clinical data from the source PET scans.
Eligibility
Sex/Gender
ALL
Healthy volunteers
No
Inclusion criteria
Scan Reader Criteria (5 total): * Board-certified in radiology or nuclear medicine * Professional experience interpreting PET scans * Naive to study protocol * No previous training or exposure to Avid Flortaucipir F 18 read methodology Scan Criteria: \- Previous enrollment in Study A05 confirmatory cohort (NCT02016560), or A16 (NCT02516046) Scan Study Population Criteria for FR01 (A05 confirmatory cohort): * Cognitively-impaired * mild cognitive impairment (MCI) or dementia with suspected neurodegenerative cause * mini-mental status exam (MMSE) score of 20-27, inclusive Scan Study Population Criteria for FR01 (Study A16): * Subjects at end of life (less than or equal to 6 months) * Imaged with flortaucipir F18 and came to autopsy
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | baseline scan | Sensitivity and specificity of 5 independent readers' interpretations of ante-mortem 18F-AV-1451 PET imaging for detection of a pattern of 18F-AV-1451 neocortical uptake that corresponds to neurofibrillary tangles (NFT) Score of B3 (Hyman et al., 2012; Montine et al., 2012). NFT B scores range from B0 (no NFTs in the brain) to B3 (widespread NFTs in the brain). Sensitivity and specificity are percentages that can range from 0 to 100%. The hypothesis tested was that, of the 5 independent imaging physicians, at least 3 will have the lower bounds of 2-sided 95% CIs ≥50%, for both sensitivity and specificity. |
| Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | baseline scan | Sensitivity and specificity of 5 independent readers' interpretations of ante-mortem 18F-AV-1451 PET imaging for detection of a pattern of 18F-AV-1451 neocortical uptake that corresponds to high levels of AD neuropathologic change (High ADNC) as defined by National Institute on Aging-Alzheimer's Association (NIA-AA) criteria (see Hyman et al. 2012). ADNC categories are None, Low, Intermediate and High, with High indicating the most severe level of AD-related pathology changes in the brain. The hypothesis tested was that, of the 5 independent imaging physicians, at least 3 will have the lower bounds of 2-sided 95% CIs ≥50%, for both sensitivity and specificity. |
| Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | baseline scan | As measured by Fleiss' Kappa across all scans read. Fleiss' kappa is a statistical measure for assessing the reliability of agreement between a fixed number of raters when assigning categorical ratings to a number of items or classifying items. Fleiss' kappa can range from 0 to 1 with 1 indicating perfect agreement between the readers. Scan results binarized as positive AD pattern versus negative AD pattern. Results are displayed as percentage of agreement for individual reader pairs, computed as follows: number of images for which reader had the same interpretation divided by the total number of images evaluated, multiplied by 100%. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | baseline scan | Reader agreement measured by Fleiss' Kappa across scans from non-autopsy cases from Study A05. Scan results binarized as positive for AD pattern versus negative for AD pattern. Results are displayed as percentage of agreement for individual reader pairs, computed as follows: number of images for which reader had the same interpretation divided by the total number of images evaluated, multiplied by 100%. |
| Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | baseline scan | Flortaucipir F 18 PET imaging will be classified by each reader as either neocortical uptake not consistent with AD (τAD-), neocortical uptake consistent with AD (τAD+), or neocortical uptake consistent with AD with uptake beyond the temporal/occipital regions (τAD++). NFT scoring is according to Hyman, et al 2012. Truth positive is a NFT B3 score. Truth negative is NFT \<B3 score. |
| Secondary Objective 4: Intra-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | baseline scan | Cohen's Kappa will be calculated for each of 5 readers to assess the intra-reader reliability of flortaucipir F 18 PET scan visual interpretation. The statistic takes into account the possibility of the agreement occurring by chance. Cohen's kappa values range from 0 to 1 with 1 representing perfect agreement. Results are displayed as percentage of agreement within a reader, computed as follows: number of images for which reader had the same interpretation at initial and second read divided by the total number of images evaluated twice by a reader, multiplied by 100%. |
| Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | baseline scan | Sensitivity and specificity of 5 independent readers' interpretations of ante-mortem 18F-AV-1451 PET imaging for detection of a pattern of 18F-AV-1451 neocortical uptake that corresponds to high levels of AD neuropathologic change (High ADNC) as defined by National Institute on Aging-Alzheimer's Association (NIA-AA) criteria (Hyman et al. 2012). Truth positive is a High ADNC score. Truth negative is No/Low/Intermediate ADNC score. |
| Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | baseline scan | Overall reader agreement as measured by Fleiss' Kappa statistic. Scan results binarized as τAD++ versus τAD+/τAD-. Results are displayed as percentage of agreement for individual reader pairs, computed as follows: number of images for which reader had the same interpretation divided by the total number of images evaluated, multiplied by 100%. |
Countries
United States
Participant flow
Recruitment details
Subjects were not recruited for this study. Study scans were selected from 2 previously completed imaging studies. Scans previously acquired from Study A16 (NCT02516046) and A05 (NCT02016560) were read by 5 independent, blinded to other study information.
Participants by arm
| Arm | Count |
|---|---|
| All Cases All eligible subject scans from contributing studies | 242 |
| Total | 242 |
Baseline characteristics
| Characteristic | All Cases |
|---|---|
| Age, Continuous Study A05 Non-autopsy Cases | 72.9 years STANDARD_DEVIATION 9.64 |
| Age, Continuous Study A16 Autopsy Cases | 81.6 years STANDARD_DEVIATION 9.91 |
| Ethnicity (NIH/OMB) Study A05 Non-autopsy Cases Hispanic or Latino | 9 Participants |
| Ethnicity (NIH/OMB) Study A05 Non-autopsy Cases Not Hispanic or Latino | 150 Participants |
| Ethnicity (NIH/OMB) Study A05 Non-autopsy Cases Unknown or Not Reported | 0 Participants |
| Ethnicity (NIH/OMB) Study A16 Autopsy Cases Hispanic or Latino | 4 Participants |
| Ethnicity (NIH/OMB) Study A16 Autopsy Cases Not Hispanic or Latino | 79 Participants |
| Ethnicity (NIH/OMB) Study A16 Autopsy Cases Unknown or Not Reported | 0 Participants |
| Race (NIH/OMB) Study A05 Non-autopsy Cases American Indian or Alaska Native | 0 Participants |
| Race (NIH/OMB) Study A05 Non-autopsy Cases Asian | 2 Participants |
| Race (NIH/OMB) Study A05 Non-autopsy Cases Black or African American | 3 Participants |
| Race (NIH/OMB) Study A05 Non-autopsy Cases More than one race | 0 Participants |
| Race (NIH/OMB) Study A05 Non-autopsy Cases Native Hawaiian or Other Pacific Islander | 0 Participants |
| Race (NIH/OMB) Study A05 Non-autopsy Cases Unknown or Not Reported | 0 Participants |
| Race (NIH/OMB) Study A05 Non-autopsy Cases White | 154 Participants |
| Race (NIH/OMB) Study A16 Autopsy Cases American Indian or Alaska Native | 0 Participants |
| Race (NIH/OMB) Study A16 Autopsy Cases Asian | 1 Participants |
| Race (NIH/OMB) Study A16 Autopsy Cases Black or African American | 1 Participants |
| Race (NIH/OMB) Study A16 Autopsy Cases More than one race | 0 Participants |
| Race (NIH/OMB) Study A16 Autopsy Cases Native Hawaiian or Other Pacific Islander | 0 Participants |
| Race (NIH/OMB) Study A16 Autopsy Cases Unknown or Not Reported | 0 Participants |
| Race (NIH/OMB) Study A16 Autopsy Cases White | 81 Participants |
| Sex: Female, Male Study A05 Non-autopsy Cases Female | 74 Participants |
| Sex: Female, Male Study A05 Non-autopsy Cases Male | 85 Participants |
| Sex: Female, Male Study A16 Autopsy Cases Female | 41 Participants |
| Sex: Female, Male Study A16 Autopsy Cases Male | 42 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
Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score)
Sensitivity and specificity of 5 independent readers' interpretations of ante-mortem 18F-AV-1451 PET imaging for detection of a pattern of 18F-AV-1451 neocortical uptake that corresponds to neurofibrillary tangles (NFT) Score of B3 (Hyman et al., 2012; Montine et al., 2012). NFT B scores range from B0 (no NFTs in the brain) to B3 (widespread NFTs in the brain). Sensitivity and specificity are percentages that can range from 0 to 100%. The hypothesis tested was that, of the 5 independent imaging physicians, at least 3 will have the lower bounds of 2-sided 95% CIs ≥50%, for both sensitivity and specificity.
Time frame: baseline scan
Population: Analysis included all autopsy subjects with an evaluable flortaucipir PET scan interpretation (n=82). One autopsy subject scan was rated as unevaluable. n=47 were truth positive for NFTs. n=35 were truth negative for NFTs.
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 2 | 89.4 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 4 | 93.6 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 3 | 87.2 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 5 | 89.4 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 1 | 91.5 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 5 | 77.1 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 1 | 77.1 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 2 | 91.4 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 3 | 85.7 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 1: Diagnostic Performance of Individual Readers (NFT Score) | Reader 4 | 62.9 percentage of cases correctly identified |
Primary
Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis)
Sensitivity and specificity of 5 independent readers' interpretations of ante-mortem 18F-AV-1451 PET imaging for detection of a pattern of 18F-AV-1451 neocortical uptake that corresponds to high levels of AD neuropathologic change (High ADNC) as defined by National Institute on Aging-Alzheimer's Association (NIA-AA) criteria (see Hyman et al. 2012). ADNC categories are None, Low, Intermediate and High, with High indicating the most severe level of AD-related pathology changes in the brain. The hypothesis tested was that, of the 5 independent imaging physicians, at least 3 will have the lower bounds of 2-sided 95% CIs ≥50%, for both sensitivity and specificity.
Time frame: baseline scan
Population: Analysis included all autopsy subjects with an evaluable flortaucipir PET scan interpretation (n=82). One autopsy subject scan was rated as unevaluable. n=41 were truth positive. n=41 were truth negative.
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 2 | 95.1 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 4 | 97.6 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 1 | 95.1 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 5 | 95.1 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 3 | 95.1 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 5 | 73.2 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 1 | 70.7 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 2 | 85.4 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 3 | 82.9 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Primary Objective 1 Analysis 2: Diagnostic Performance of Individual Readers (NIA-AA Autopsy Diagnosis) | Reader 4 | 58.5 percentage of cases correctly identified |
Primary
Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging
As measured by Fleiss' Kappa across all scans read. Fleiss' kappa is a statistical measure for assessing the reliability of agreement between a fixed number of raters when assigning categorical ratings to a number of items or classifying items. Fleiss' kappa can range from 0 to 1 with 1 indicating perfect agreement between the readers. Scan results binarized as positive AD pattern versus negative AD pattern. Results are displayed as percentage of agreement for individual reader pairs, computed as follows: number of images for which reader had the same interpretation divided by the total number of images evaluated, multiplied by 100%.
Time frame: baseline scan
Population: Analysis included all subjects who had a valid flortaucipir PET scan in the autopsy and non-autopsy groups (n=241)
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 2 v Reader 3 | 97.1 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 1 v Reader 2 | 95.0 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 1 v Reader 3 | 93.4 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 1 v Reader 4 | 94.6 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 1 v Reader 5 | 97.1 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 2 v Reader 4 | 91.3 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 2 v Reader 5 | 94.6 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 3 v Reader 4 | 91.3 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 3 v Reader 5 | 93.8 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Primary Objective 2: Inter-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 4 v Reader 5 | 95.0 percentage agreement of cases |
Comparison: Fleiss' kappa is a statistical measure for assessing the reliability of agreement between a fixed number of raters when assigning categorical ratings to a number of items or classifying items. The primary hypothesis to be tested was that the lower bound of two-sided 95% confidence interval of Fleiss' Kappa will be greater than or equal to 0.6.95% CI: [0.83, 0.91]
Secondary
Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs
Flortaucipir F 18 PET imaging will be classified by each reader as either neocortical uptake not consistent with AD (τAD-), neocortical uptake consistent with AD (τAD+), or neocortical uptake consistent with AD with uptake beyond the temporal/occipital regions (τAD++). NFT scoring is according to Hyman, et al 2012. Truth positive is a NFT B3 score. Truth negative is NFT \<B3 score.
Time frame: baseline scan
Population: Analysis included all autopsy subjects with an evaluable flortaucipir PET scan interpretation (n=82). One autopsy subject scan was rated as unevaluable. n=47 were truth positive for NFTs. n=35 were truth negative for NFTs.
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 2 | 80.9 percentage cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 4 | 89.4 percentage cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 3 | 76.6 percentage cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 5 | 83.0 percentage cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 1 | 87.2 percentage cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 5 | 94.3 percentage cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 1 | 88.6 percentage cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 2 | 100.0 percentage cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 3 | 94.3 percentage cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 1: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect B3 NFTs | Reader 4 | 82.9 percentage cases correctly identified |
Secondary
Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC
Sensitivity and specificity of 5 independent readers' interpretations of ante-mortem 18F-AV-1451 PET imaging for detection of a pattern of 18F-AV-1451 neocortical uptake that corresponds to high levels of AD neuropathologic change (High ADNC) as defined by National Institute on Aging-Alzheimer's Association (NIA-AA) criteria (Hyman et al. 2012). Truth positive is a High ADNC score. Truth negative is No/Low/Intermediate ADNC score.
Time frame: baseline scan
Population: Analysis included all autopsy subjects with an evaluable flortaucipir PET scan interpretation (n=82). One autopsy subject scan was rated as unevaluable. n=41 were truth positive. n=41 were truth negative.
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 2 | 85.4 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 4 | 92.7 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 3 | 85.4 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 5 | 87.8 percentage of cases correctly identified |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 1 | 92.7 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 5 | 87.8 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 1 | 82.9 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 2 | 92.7 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 3 | 92.7 percentage of cases correctly identified |
| Specificity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 1 Analysis 2: Diagnostic Performance of τAD++ Flortaucipir PET Images to Detect High ADNC | Reader 4 | 75.6 percentage of cases correctly identified |
Secondary
Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images
Overall reader agreement as measured by Fleiss' Kappa statistic. Scan results binarized as τAD++ versus τAD+/τAD-. Results are displayed as percentage of agreement for individual reader pairs, computed as follows: number of images for which reader had the same interpretation divided by the total number of images evaluated, multiplied by 100%.
Time frame: baseline scan
Population: Analysis included all subjects who had a valid flortaucipir PET scan in the autopsy and non-autopsy groups (n=241)
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 2 v Reader 4 | 90.5 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 1 v Reader 2 | 92.5 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 1 v Reader 3 | 90.0 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 1 v Reader 4 | 91.3 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 1 v Reader 5 | 95.0 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 2 v Reader 3 | 91.7 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 2 v Reader 5 | 94.2 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 3 v Reader 4 | 88.8 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 3 v Reader 5 | 92.5 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 2: Inter-reader Reliability of Reader Interpretation of τAD++ Flortaucipir PET Images | Reader 4 v Reader 5 | 92.9 percentage agreement of cases |
Comparison: Fleiss' kappa is a statistical measure for assessing the reliability of agreement between a fixed number of raters when assigning categorical ratings to a number of items or classifying items. The primary hypothesis to be tested was that the lower bound of two-sided 95% confidence interval of Fleiss' Kappa will be greater than or equal to 0.6.95% CI: [0.8, 0.88]
Secondary
Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use
Reader agreement measured by Fleiss' Kappa across scans from non-autopsy cases from Study A05. Scan results binarized as positive for AD pattern versus negative for AD pattern. Results are displayed as percentage of agreement for individual reader pairs, computed as follows: number of images for which reader had the same interpretation divided by the total number of images evaluated, multiplied by 100%.
Time frame: baseline scan
Population: Non-autopsy cases from Study A05
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 1 v Reader 3 | 93.1 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 1 v Reader 4 | 95.6 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 1 v Reader 2 | 97.5 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 1 v Reader 5 | 97.5 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 2 v Reader 3 | 93.1 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 2 v Reader 4 | 94.3 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 2 v Reader 5 | 96.2 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 3 v Reader 4 | 93.7 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 3 v Reader 5 | 93.1 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 3: Inter-reader Reliability of Flortaucipir PET Scan Interpretation in the Population of Intended Use | Reader 4 v Reader 5 | 96.9 percentage agreement of cases |
Comparison: Fleiss' kappa is a statistical measure for assessing the reliability of agreement between a fixed number of raters when assigning categorical ratings to a number of items or classifying items. The hypothesis to be tested was that the lower bound of two-sided 95% confidence interval of Fleiss' Kappa will be greater than or equal to 0.6.95% CI: [0.85, 0.95]
Secondary
Secondary Objective 4: Intra-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging
Cohen's Kappa will be calculated for each of 5 readers to assess the intra-reader reliability of flortaucipir F 18 PET scan visual interpretation. The statistic takes into account the possibility of the agreement occurring by chance. Cohen's kappa values range from 0 to 1 with 1 representing perfect agreement. Results are displayed as percentage of agreement within a reader, computed as follows: number of images for which reader had the same interpretation at initial and second read divided by the total number of images evaluated twice by a reader, multiplied by 100%.
Time frame: baseline scan
Population: 20 scans read twice to assess the intra-reader reliability
| Arm | Measure | Group | Value (NUMBER) |
|---|---|---|---|
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 4: Intra-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 1 | 95.0 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 4: Intra-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 2 | 85.0 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 4: Intra-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 3 | 80.0 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 4: Intra-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 4 | 95.0 percentage agreement of cases |
| Sensitivity of Flortaucipir vs Autopsy NFT Score | Secondary Objective 4: Intra-reader Reliability of Reader Interpretation of Flortaucipir-PET Imaging | Reader 5 | 90.0 percentage agreement of cases |
Comparison: Cohen's kappa statistic for Reader 195% CI: [0.69, 1]
Comparison: Cohen's kappa statistic for Reader 295% CI: [0.41, 1]
Comparison: Cohen's kappa statistic for Reader 395% CI: [0.24, 0.95]
Comparison: Cohen's kappa statistic for Reader 495% CI: [0.67, 1]
Comparison: Cohen's kappa statistic for Reader 595% CI: [0.52, 1]