Research

JAMA Neurology | Original Investigation

Clinical Utility of Tau Positron Emission Tomography

in the Diagnostic Workup of Patients With Cognitive Symptoms
Ruben Smith, MD, PhD; Douglas Hägerström, MD; Daria Pawlik, MD; Gregory Klein, PHD; Jonas Jögi, MD, PhD;
Tomas Ohlsson, PhD; Erik Stomrud, MD, PhD; Oskar Hansson, MD, PhD

Supplemental content
IMPORTANCE It is important to determine the added clinical value for tau positron emission
tomography (PET) in the diagnostic workup of patients with cognitive symptoms before
widespread implementation in clinical practice.

OBJECTIVE To prospectively study the added clinical value of PET detecting tau pathology in
Alzheimer disease (AD).

DESIGN, SETTING, AND PARTICIPANTS This prospective cohort study (Swedish BioFINDER-2
study) took place from May 2017 through September 2021. A total of 878 patients with
cognitive complaints were referred to secondary memory clinics in southern Sweden and
then recruited to the study. In total, 1269 consecutive participants were approached, but 391
did not meet inclusion criteria or did not complete the study.

EXPOSURES Participants underwent a baseline diagnostic workup, including clinical

examination, medical history, cognitive testing, blood and cerebrospinal fluid sampling,
magnetic resonance imaging of the brain, and a tau PET ([18F]RO948) scan.

MAIN OUTCOMES AND MEASURES The primary end points were change in diagnosis and
change in AD drug therapy or other drug treatment between the pre- and post-PET visits.
A secondary end point was the change in diagnostic certainty between the pre- and post-PET
visits.

RESULTS A total of 878 participants with a mean age of 71.0 (SD, 8.5) years (491 male [56%])
were included. The tau PET result led to a change in diagnoses in 66 participants (7.5%) and a
change in medication in 48 participants (5.5%). The study team found an association with
overall increased diagnostic certainty after tau PET in the whole data set (from 6.9 [SD, 2.3]
to 7.4 [SD, 2.4]; P < .001). The certainty was higher in participants with a pre-PET diagnosis of
AD (from 7.6 [SD, 1.7] to 8.2 [SD, 2.0]; P < .001) and increased even further in participants
with a tau PET positive result supporting an AD diagnosis (from 8.0 [SD, 1.4] to 9.0 [SD, 0.9];
P < .001). The association with tau PET results had the largest effect sizes in participants with
pathological amyloid-β (Aβ) status, whereas no significant change in diagnoses was seen in
participants with normal Aβ status. Author Affiliations: Clinical Memory
Research Unit, Department of Clinical
CONCLUSIONS AND RELEVANCE The study team reported a significant change in diagnoses and Sciences, Lund University, Malmö,
Sweden (Smith, Pawlik, Stomrud,
patient medication when tau PET was added to an already extensive diagnostic workup that
Hansson); Department of Neurology,
included cerebrospinal fluid AD biomarkers. Including tau PET was associated with a Skåne University Hospital, Lund,
significant increase in certainty of underlying etiology. The effect sizes for certainty of Sweden (Smith, Pawlik); Department
etiology and diagnosis were largest in the Aβ-positive group and the study team suggests that of Neurophysiology, Skåne University
Hospital, Lund, Sweden
clinical use of tau PET be limited to populations with biomarkers indicating Aβ positivity. (Hägerström); F. Hoffmann-La Roche
Ltd, Basel, Switzerland (Klein); Skåne
University Hospital, Department of
Clinical Physiology and Nuclear
Medicine, Lund, Sweden (Jögi);
Department of Radiation Physics,
Skåne University Hospital, Lund,
Sweden (Ohlsson); Memory Clinic,
Skåne University Hospital, Malmö,
Sweden (Hansson).
Corresponding Authors: Ruben
Smith, MD, PhD, Department of
Neurology (Ruben.Smith@med.lu.
se), and Oskar Hansson, MD, PhD,
Memory Clinic (Oskar.Hansson@
JAMA Neurol. 2023;80(7):749-756. doi:10.1001/jamaneurol.2023.1323 med.lu.se), Skåne University
Published online May 22, 2023. Hospital, SE-205 02 Malmö, Sweden.

(Reprinted) 749

Downloaded From: https://jamanetwork.com/ on 08/29/2023

 Research Original Investigation Clinical Utility of Tau Positron Emission Tomography in the Diagnostic Workup of Patients With Cognitive Symptoms
A
lzheimer disease (AD) is believed to be caused by the
accumulation of amyloid-β (Aβ) in the brain1-3 fol-
lowed by a gradual spread of tau pathology across the
brain as the clinical symptoms emerge.4-6 Aβ positron emis-
sion tomography (PET) for determining the presence of Aβ
pathological changes in the brain has been available for more
than a decade and there are now several tracers available for
clinicaluse(eg,[18F]florbetapir,[18F]flutemetamol,and[18F]flor-
betaben). A large prospective multicenter study from the US
(the IDEAS study)7 reported that the added information from
Aβ-PET led to a significant change in patient management and
diagnosis in patients where AD was among the considered dif-
ferential diagnoses. A recent retrospective study8 showed that
both Aβ-PET and tau PET increased diagnostic confidence and
changed diagnoses to a similar extent when added to a basic
clinical workup. There are now a number of tau PET tracers
available for research purposes.9 The most widely used tracer,
[18F]flortaucipir, has been shown to reliably detect tau as as-
sessed by neuropathology10-13 and has been approved by the
US Food and Drug Administration for use in the US as a diag-
nostic marker of neurofibrillary tau in AD.14 It is expected that
tau PET will also be available for clinical use in other parts of
the world within the coming years. The recently developed
second-generation tau tracer [18F]RO948, used in the current
study, has shown similar properties as flortaucipir.15 Before
widespread implementation of tau PET as a diagnostic method
in the clinic, it is vital to establish if the method shows an added
clinical value for diagnosis and treatment of patients with
memory complaints and to establish whether the method in-
creases the diagnostic confidence of treating clinicians.
The aim of the current study was to prospectively evalu-
ate the added clinical value of including visual read of tau PET
([18F]RO948 PET) in the diagnostic workup of AD. This was
evaluated in secondary memory clinics where the diagnostic
workup already included patient history, clinical examina-
tion, cognitive testing, magnetic resonance imaging (MRI) of
the brain, as well as relevant cerebrospinal fluid (CSF) mark-
ers (Aβ42, Aβ40, and pTau181). Specifically, we studied
whether tau PET led to change in diagnoses, change in treat-
ment, as well as in diagnostic certainty. As tau PET examina-
tions are costly, we have also assessed whether this bio-
marker is more informative in certain subpopulations.
Methods
Participants and Baseline Assessment
Between May 2017 and Sept 2021, 1269 patients referred for
cognitive or neurological symptoms were consecutively re-
cruited at the secondary memory clinic at Skåne University
Hospital, Malmö, Sweden; the secondary memory clinic at
Ängelholm hospital, Ängelholm, Sweden; and the secondary
neurology clinic, Skåne University Hospital, Lund, Sweden. The
study was part of the BioFINDER-2 study (NCT03174938).
Inclusion and exclusion criteria have been specified in detail
previously. 16 Patients included in this study had either
subjective cognitive decline (SCD) or objective reductions in
memory performance, either at mild cognitive impairment
750 JAMA Neurology July 2023 Volume 80, Number 7 (Reprinted)
Downloaded From: https://jamanetwork.com/ on 08/29/2023
Key Points
Question Does tau positron emission tomography (PET) provide
additional information on top of an extensive clinical workup in
participants with cognitive symptoms?
Findings In this cohort study of 878 patients referred to
secondary memory clinics in south Sweden, the study team found
that including tau PET in the diagnostic workup resulted in a
statistically significant change in diagnosis in 7.5% of the
participants and a significant change in medication in 5.5% of the
study population and also found a significant association of overall
increased diagnostic certainty with including tau PET.
Meaning Tau PET may have an added clinical value to increase
diagnostic certainty, especially in amyloid-β positive patients
where Alzheimer disease is a differential diagnosis.
(MCI) level or dementia level. A total of 391 participants were
excluded (eMethods in Supplement 1 for details) and 878
participants met the inclusion criteria and completed the study.
Participant characteristics are presented in the Table and the
eTable in Supplement 1. Written informed consent was
obtained from all participants prior to entering the study. The
study was approved by the regional review board for human
research ethics at Lund University. Strengthening the Reporting
of Observational Studies in Epidemiology (STROBE) reporting
guidelines were followed.
The study design is presented in Figure 1. Participants were
assessed with a baseline diagnostic workup, including clini-
cal examination, medical history, cognitive testing, CSF sam-
pling for biomarkers (mainly Aβ42, Aβ40, and pTau181), and
MRI imaging. Based on the baseline assessment, the treating
clinician was asked to fill out a report form stating (1) the most
likely diagnosis (etiology) underlying the cognitive symp-
toms, (2) how certain they were of this diagnosis on a scale from
0 (very uncertain) to 10 (very certain), (3) the cognitive status
(SCD/MCI/dementia) of the participant, (4) the certainty of the
cognitive status (scale 0 to 10), (5) if the patient was receiving
any medication to enhance cognitive function (for example,
acetylcholinesterase inhibitors or memantine) or antidepres-
sants, and (6) if any further investigations were planned for
establishing the diagnosis. Once the form was filled out, the
visual read of the [18F]RO948 PET was revealed together with
a template for interpretation (eAppendix in Supplement 1) and
the clinician was asked to fill out a follow-up report form, again
stating the most likely etiology, their certainty, as well as if they
planned changes in the medication regimen or patient man-
agement as a result of the [18F]RO948 PET information. The
full forms, translated into English, are available in the eAp-
pendix in Supplement 1. Outcomes were change in diagnosis
(from AD to non-AD or vice versa), change in medication to en-
hance cognitive function or antidepressant medication, and
change in diagnostic certainty.
Image Acquisition and Processing
Imaging acquisition details are provided in the eMethods in
Supplement 1. Rating was performed by 2 raters (D.H. and R.S.),
masked to clinical information of the patient and reaching a
jamaneurology.com
 Clinical Utility of Tau Positron Emission Tomography in the Diagnostic Workup of Patients With Cognitive Symptoms Original Investigation Research

Table. Participant Demographics

Characteristic AD Non-AD AD-SCD AD-MCI AD-dementia P value

No. 408 470 33 179 196 NA
Age, y, mean (SD) 72.7 (7.8)a 69.5 (8.9) 70.4 (8.4)b 72.2 (7.6) 73.5 (7.9) a,b

Sex
Female 192 195 11 79 102 .11
Male 216 275 22 100 94 .11
MMSE score, mean (SD) 24.4 (4.6)a 26.2 (4.2) 29.1 (1.1)c 26.7 (2.5)d 21.4 (4.4) a,c,d

Education, y, mean (SD) 12.6 (4.2) 12.3 (3.8) 14.4 (4.0)e 12.5 (4.3) 12.4 (4.2) e

a a
Aβ+, No./total No. (%) 386/394 (98) 167/454 (37) 32/32 (100) 169/172 (98) 185/190 (97)
f a g d a,g,d
Positive tau PET, No./total No. (%) 304/408 (75) 39/470 (8) 11/33 (33) 122/179 (68) 171/196 (87)
a c d a,c,d
Tau PET SUVR, temporal ROI, mean (SD) 1.77 (0.64) 1.19 (0.20) 1.28 (0.25) 1.59 (0.49) 2.02 (0.70)
c
Abbreviations: Aβ+, amyloid-β positive; AD, Alzheimer disease; MCI, mild P < .001 vs AD with MCI and AD with dementia.
cognitive impairment; MMSE, Mini-Mental State Examination; Non-AD, all d
P < .001 vs AD with dementia.
conditions other than AD pooled (eMethods in Supplement 1); NA, not e
P = .01 vs AD with MCI and AD with dementia.
applicable; PET, positron emission tomography; ROI, region of interest;
f
SCD, subjective cognitive decline; SUVR, standardized uptake value ratio. Based on visual read.
g
a
P < .001 vs non-AD. P < .001 vs AD with MCI and P < .001 AD with dementia.
b
P = .04 vs AD with dementia.

joint decision on which category to allot the image. In brief,
the categories used were (A) normal image; no discernible
[18F]RO948 retention, (B) retention of [18F]RO948 confined to
the temporal lobes, (C) more widespread retention of
[18F]RO948, reaching into the parietal, occipital, or frontal
lobes, and (D) inconclusive scan. A detailed description of the
visual read algorithm and example images are provided in the
eMethods and eFigure 1 in Supplement 1. For analyses using a
dichotomous negative/positive tau PET read, inconclusive
visual reads were considered negative (not having an AD-
typical pattern).
The visual read algorithm applied in this study was rela-
tively similar to the visual read algorithm used by Seibyl et al17
for assessing [18F]MK-6240 PET scans, but allowed a positive
visual read of the tau PET scan even with unilateral uptake in
the medial temporal lobe. The visual read algorithm used in
this study is also likely to be more sensitive to early temporal
accumulation of tau compared with published methods for
[18F]flortaucipir.8,10
Statistics
Statistical comparisons of baseline characteristics were as-
sessed using χ 2 test for categorical data (sex) or Mann-
Whitney U tests for continuous data. For comparison of pre-
and post-tau PET, paired Wilcoxon signed rank tests or the
McNemar χ2 tests with continuity correction were used. For
comparison between nonpaired groups, Wilcoxon rank sum
tests were performed. All statistical tests were 2-sided with a
significance level of .05. All analyses were performed using the
R programming language version 4.0.3 (The R Project).
Results
Participants
Patients (n = 878) from the 3 secondary clinics were consecu-
tively recruited into the BioFINDER-2 study. Participant de-
mographics are summarized in the Table. Based on the pre-
PET forms, 408 participants had an AD diagnosis and 470 had
non-AD diagnoses (details in the eMethods and the eTable in
Supplement 1). Patients with a pre-PET diagnosis of AD were
older and performed worse on Mini-Mental State Examina-
tion compared with the non-AD group. Of the participants with
a pre-PET diagnosis of AD, 304 of 408 (75%) had tau PET scans
that were read as positive (ie, read as early or late AD pattern).
Furthermore, 11 of 33 patients with SCD, where AD was sus-
pected as an underlying cause before PET (33%), had a positive
tau PET scan; 122 of 179 patients with MCI, where AD was sus-
pected as underlying cause before PET (68%), had a positive
tau PET scan; and 171 of 196 patients with dementia, where AD
was suspected as underlying cause before PET (87%), had a
positive tau PET scan. In the group with a non-AD diagnosis
before PET, 39 of 470 had positive tau PET scan (8%) (eFig-
ure 2 in Supplement 1).
Association With Tau PET on Diagnosis
and Diagnostic Certainty
The tau PET information was associated with a change in di-
agnoses in 66 out of 878 participants (7.5%). Diagnoses changed
from AD to non-AD in 47 participants (11.5% of participants with
a pre-PET diagnosis of AD) and from non-AD to AD in 19 par-
ticipants (4.0% of participants with a non-AD pre-PET diag-
nosis) (McNemar test, P < .001; Figure 2A). Changes in diag-
noses were found both in patients with a negative tau PET
visual read (47 changed diagnosis, total n = 535; McNemar test,
P < .001), as well as in patients within the group with a posi-
tive visual read (19 changed diagnosis, total n = 343; McNemar
test, P = .001). Significant changes in diagnoses were seen in
participants with dementia (23 changed diagnosis, total n = 318;
McNemar test, P = .04) and MCI (33 changed diagnosis, total
n = 419; McNemar test, P = .01), whereas the number of
changed diagnoses in the SCD group did not reach statistical
significance (10 changed diagnosis, total n = 141; McNemar test,
P = .75; eFigure 3 in Supplement 1).

jamaneurology.com (Reprinted) JAMA Neurology July 2023 Volume 80, Number 7 751

Downloaded From: https://jamanetwork.com/ on 08/29/2023

 Research Original Investigation Clinical Utility of Tau Positron Emission Tomography in the Diagnostic Workup of Patients With Cognitive Symptoms
Figure 1. Study Design
Enrollment into Swedish BioFINDER-2 study
Clinical workup including clinical examination,
cognitive testing, blood and CSF sampling,
and structural MRI
Pre-PET form completed
Visual read
Post-PET form completed
Patients with memory complaints were recruited to the study after a
comprehensive clinical workup, including clinical examination, cognitive testing,
structural imaging using magnetic resonance imaging (MRI), and
blood/cerebrospinal fluid (CSF) sampling. The clinician filled out the
pre–positron emission tomography (PET) form, indicating diagnosis, suspected
underlying etiology, certainty of etiology, and treatment. The clinician then
received the outcome of the visual read and filled out the post-PET form.
Eight hundred forty-eight of the participants (97%) had in-
formation on their Aβ status determined using the CSF Aβ42/40
ratio. Therefore, the study team studied the added value of tau
PET in Aβ-positive (n = 553) and Aβ-negative (n = 295) sub-
populations. In the Aβ-positive group, the tau PET result led
to a significant change in diagnoses in 55 participants (9.9%)
(38 changed from pre-PET AD to a post-PET non-AD diagno-
sis and 17 from a pre-PET non-AD to a post-PET AD diagnosis;
McNemar test, P = .01). The visual read results of the Aβ-
positive group are presented in eFigure 4 in Supplement 1. The
tau PET result was not associated with a change in diagnoses
in the Aβ-negative group (6 participants changed from pre-
PET AD to post-PET non-AD and 1 from a pre-PET non-AD di-
agnosis to post-PET AD; McNemar test, P = .13).
In the total population, the overall certainty of diagnosis
increased from 6.9 (SD, 2.3) to 7.4 (SD, 2.4) (P < .001) on a scale
from 0 (very uncertain) to 10 (complete certainty). The cer-
tainty was higher in the group having a pre-PET diagnosis of
AD (n = 408; Figure 3A and 3B), increasing from 7.6 (SD, 1.7)
to 8.2 (SD, 2.0) (P < .001), and among the patients with a pre-
PET AD diagnosis having a positive tau PET scan (supportive
of an AD diagnosis) where certainty increased from 8.0 (SD,
1.4) to 9.0 (SD, 0.9) (P < .001). In patients with a pre-PET AD
diagnosis, having a negative tau PET scan visual dropped read
certainty from 6.5 (SD, 2.0) to 5.7 (SD, 2.1) (P < .001). The in-
crease in certainty in patients with a pre-PET AD diagnosis was
present in participants irrespective of cognitive status: AD de-
mentia (n = 196; 8.2 [SD, 1.3] vs 8.7 [SD, 1.7]; P < .001 [Figure 3C
and 3D]), MCI (n = 179; 7.3 [SD, 1.8] vs 7.8 [SD, 2.1]; P < .001
[Figure 3E and 3F]), and SCD (n = 33; 6.2 [SD, 1.7] vs 6.8 [SD,
1.8]; P = .01 [Figure 3G and 3H]). Since pre-PET rating at the
highest and lowest certainty levels (ie, 0 and 10) can either not
increase or decrease in certainty, the study team presented
752 JAMA Neurology July 2023 Volume 80, Number 7 (Reprinted)
Downloaded From: https://jamanetwork.com/ on 08/29/2023
Figure 2. Changes in Diagnoses and Medication
A Change in diagnosis
a
AD Non-AD
Before 408 470
361 47 19 451
After 380 498
B Change in medication
Taking a Not taking
medication medication
Before 482 396
477 5 43 353
After 520 358
A, McNemar test, P < .001. B, McNemar, test P < .001. Participants with an
increase in medication (on top of preexisting cognitive medication) are
considered as being off of medication at baseline. AD indicates Alzheimer
disease.
the same data where these values have been removed
(Figure 3B-3H). Graphs showing paired data of the changes in
certainty in the AD diagnostic group are provided in eFigure 5
in Supplement 1.
The diagnostic certainty after the tau PET scan was lower
in participants with a pre-PET non-AD diagnosis compared with
a pre-PET AD diagnosis (6.7 [SD, 2.6] vs 8.2 [SD, 2.0]; P < .001),
but there was still a significant, yet subtle, increase in cer-
tainty in the non-AD group after receiving the tau PET result (6.3
[SD, 2.5] vs 6.7 [SD, 2.6]; P < .001; eFigure 6 in Supplement 1).
In the Aβ-positive group, the tau PET result was associ-
ated with increased overall certainty of diagnoses from 7.2 (SD,
2.0) to 7.8 (SD, 2.2) (P < .001). Within the Aβ-positive group,
significant increases in certainty were seen irrespective of cog-
nitive status (SCD, 6.0 [SD, 2.0] to 6.5 [SD, 2.3]; P = .01; MCI,
6.8 [SD, 2.1] to 7.4 [SD, 2.3]; P < .001; and dementia, 7.9 [SD,
1.6] to 8.5 [SD, 1.7]; P < .001). There was a significant change
of certainty in diagnosis also in the Ab-negative group, but the
change was smaller in magnitude (from 6.5 [SD, 2.5] to 6.7 [SD,
2.6]; P < .001).
Association of Tau PET With Treatment
At baseline, 506 of all 878 participants were receiving antide-
pressant medication or medication to enhance cognitive func-
tions (eg, acetylcholinesterase inhibitors or memantine), or had
tried medication but discontinued due to adverse effects or lack
of effect. A total of 308 of 408 patients with a pre-PET AD di-
agnosis (75%) were receiving medication to enhance cogni-
tive functions or had tried such a medication. This encom-
passed 185 of 196 patients with a pre-PET AD dementia
diagnosis (94%), 118 of 179 with pre-PET MCI due to AD diag-
nosis (66%), and 5 of 33 with SCD where AD was judged as the
underlying cause before PET (15%). In the total study popu-
lation, 17 participants started receiving medication to en-
hance cognitive function and 21 had an addition of a medica-
tion with a different mechanism (acetylcholinesterase
jamaneurology.com
 Clinical Utility of Tau Positron Emission Tomography in the Diagnostic Workup of Patients With Cognitive Symptoms Original Investigation Research

Figure 3. Change in Certainty in Participants With a Pre–Positron Emission Tomography (PET)

Alzheimer Disease (AD) Diagnosis

A All participants with AD B All participants with AD with extremes removed

125 120
Positive visual read
100 Negative visual read 90
Participants, No.

Participants, No.
Overlaid pos/neg
75
60
50

30
25

0 0
–10 –5 0 5 –10 –5 0 5
Change in certainty Change in certainty

C All participants with AD dementia D All participants with AD dementia with extremes removed

80 80

60 60
Participants, No.

Participants, No.

40 40

20 20

0 0
–10 –5 0 5 –10 –5 0 5
Change in certainty Change in certainty

E All participants with AD MCI F All participants with AD MCI with extremes removed
50 50

40 40
Participants, No.

Participants, No.

30 30

20 20

10 10

0 0
–8 –4 0 4 –8 –4 0 4
Change in certainty Change in certainty
Change was calculated as certainty
G All participants with AD SCD H All participants with AD SCD with extremes removed post-PET minus certainty pre-PET.
The results from subgroups with
20 20
different levels of cognitive
impairment are shown in
15 15 C, dementia; E, mild cognitive
Participants, No.

Participants, No.

impairment (MCI); and G, subjective

cognitive decline; (SCD). Since
10 10 participants beginning at certainty
level 10 cannot increase and 0 cannot
decrease, results where individuals
5 5
starting at 10 (for positive visual
reads) or 0 (for negative visual reads)
0 0 have been removed and resulting
–6 –3 0 3 6 –6 –3 0 3 6 graphs are shown for all participants
Change in certainty Change in certainty in B; and participants with D,
dementia; F, MCI; and H, SCD.

inhibitors added to memantine or the other way around) af- in 5 participants the medication to enhance cognitive func-
ter receiving the tau PET result. For 5 patients, the tau PET tions was discontinued (McNemar’s test, P < .001; Figure 2B).
result led to start or increase of antidepressant medication and In total, changes in cognitive or antidepressant medication

jamaneurology.com (Reprinted) JAMA Neurology July 2023 Volume 80, Number 7 753

Downloaded From: https://jamanetwork.com/ on 08/29/2023

 Research Original Investigation Clinical Utility of Tau Positron Emission Tomography in the Diagnostic Workup of Patients With Cognitive Symptoms
were made in 48 participants of the total population (5.5%).
In a sensitivity analysis, only including changes in medica-
tion to enhance cognitive functions, the study team found simi-
lar results (eFigure 7 in Supplement 1).
Discussion
The visual read of the tau PET led to a change in the diagnosis
of 7.5% of the total population with 11.5% of those with a pre-
PET AD diagnosis changing to a post-PET non-AD diagnosis and
4.0% of those with a pre-PET non-AD diagnosis changing to a
post-PET AD diagnosis. Similarly, the result led to a signifi-
cant change in medication with 48 participants receiving or
discontinuing medication after the tau PET examination (5.5%).
These numbers may seem small, but 75% of AD participants
(including 92% with AD dementia) were already receiving
medication at baseline.
We found that increased AD diagnosis certainty was asso-
ciated with tau PET in participants where AD was the primary
suspected pre-PET etiology, irrespective of cognitive status.
Certainty decreased in the event of an unexpected negative
scan in this patient population. The decrease was more marked
in participants with an objective cognitive impairment
(Figure 3C-F), whereas no change in certainty was seen with
a negative scan in the SCD population with a suspected pre-
PET AD etiology (Figure 3G and H). Ruling out underlying AD
based on a negative tau PET result in the absence of clear ob-
jective cognitive impairment is questionable.18,19 This is likely
the reason why no decrease in certainty was seen in the par-
ticipants with SCD with suspected underlying AD etiology and
a negative visual read. In line with this, a significant change
in diagnoses was not seen in the SCD group, but only seen in
the MCI and dementia groups.
The results of this study indicate an added value of in-
cluding tau PET on top of an already extensive clinical workup,
which included CSF AD biomarkers. CSF biomarkers are in-
creasingly used in many memory clinics because of the high
diagnostic accuracy and relatively low costs, and several CSF
AD biomarker assays have recently been approved for clinical
use by the US Food and Drug Administration.20,21 Still, lum-
bar puncture is invasive, but some novel blood-based biomark-
ers are reaching diagnostic accuracies that are similar to those
of CSF AD biomarkers,6,22 which is why such blood markers
will likely precede a decision to perform tau PET imaging in
most memory clinics in the coming years. Therefore, the cur-
rent prospective study will likely also be relevant to clinical set-
tings where high-performing blood AD biomarkers will be used
instead of CSF AD biomarkers.
The relatively large number of participants changing from
a pre–PET AD to a post-PET non-AD underlying diagnosis in
the present study suggests that relying solely on CSF biomark-
ers, such as Aβ42/Aβ40 or Aβ42/pTau, may result in an over-
estimation of the AD element in the etiology of the cognitive
decline. These fluid biomarkers change early in the disease de-
velopment and the biomarkers are pathological in a relatively
high proportion of cognitively unimpaired individuals.18,19
Therefore, using fluid Aβ and pTau biomarkers alone may lead
754 JAMA Neurology July 2023 Volume 80, Number 7 (Reprinted)
Downloaded From: https://jamanetwork.com/ on 08/29/2023
to difficulties to separate if subtle cognitive symptoms are de-
rived from an early-stage AD or whether they are due to other
causes, such as stress, depression, or another concomitant
non-AD neurodegenerative disease.19,23 On the other hand,
tau PET changes later in the AD disease course24 and is more
closely associated with cognitive deterioration than fluid AD
biomarkers.6,25-28
We found that including tau PET in the diagnostic workup
in a secondary memory clinic setting provided best informa-
tion when AD is the primary suspected pre-PET diagnosis.
Similarly, in a subanalysis of CSF Aβ-positive and CSF Aβ-
negative individuals, we found no clear added value of tau PET
in the CSF Aβ-negative subgroup, but a clear change in the CSF
Aβ-positive group. In the former group, there was a minor in-
crease in the diagnostic certainty, but no association with
change of diagnoses. We therefore suggest that tau PET should
mainly be used in individuals where fluid biomarkers indi-
cate presence of AD pathology.
It has been suggested that in case of cognitive impair-
ment at the MCI or dementia level, the tau PET results might
be used both to rule in and to rule out AD as the underlying
primary pathology,23 but in cognitively unimpaired patients
with SCD, the tau PET result might only be used to rule in AD
pathology as the underlying cause.16,25,29,30 We recommend
that tau PET is only used in patients with SCD when there is
an increased risk of an underlying AD pathology and where the
information may result in a beneficial change in patient man-
agement.
The association of Aβ PET with patient management and
diagnosis has been studied previously in the large US multi-
center IDEAS study.7 The authors found an association with
patient management with 60.2% to 63.5% of study partici-
pants having a change in management as a result of the Aβ PET
scan and diagnoses were changed in 25.1% and 10.5% with posi-
tive and negative Aβ PET scans, respectively. The largest ef-
fects on management were seen in patient medication where
43.6% to 44.9% of participants with a positive scan had a
change in treatment regimen.7 In our study, the changes in
treatment were more modest, probably reflecting that a large
proportion of the patients participating were receiving medi-
cation already at baseline after having received results from
fluid AD biomarkers in addition to the clinical assessments.
However, our results indicate that tau PET change manage-
ment, also when high-performing fluid AD biomarkers, has
been used during the early stages of the diagnostic workup.
A previous retrospective study8 evaluated the added value
of Aβ PET and tau PET for diagnosis and diagnostic certainty on
top of a clinical workup, including clinical and neuropsycho-
logical assessment and MRI, but no blood or CSF biomarkers.
The study included 136 memory clinic patients and found simi-
lar added values of Aβ and tau PET with an increase in diagnos-
tic certainty and a similar degree of changed diagnoses (28%
change) for the 2 PET methods.8 In our study, where fluid AD
biomarkers were already known before tau PET imaging was
performed, the change in diagnoses was as expected smaller
(7.5%). This was comparable with the added diagnostic value
of tau PET after having the Aβ PET result in the previous retro-
spective study (9%).8 The design of the previous study was very
jamaneurology.com
 Clinical Utility of Tau Positron Emission Tomography in the Diagnostic Workup of Patients With Cognitive Symptoms Original Investigation Research

different from the design of our current study. The diagnoses
and diagnostic certainty were determined based on summa-
rized clinical information, cognitive tests, biomarkers, and
imaging data in a retrospective fashion without seeing the
patients.8 The advantage of the study design applied in this
study is the prospective approach, a real-world memory clinic
setting, and the larger study size. That said, both studied pro-
vided rather similar results. Does an increased diagnostic cer-
tainty and changed diagnoses matter for the patients? With
emerging new treatments for AD being developed, an in-
creased diagnostic certainty is becoming more important, but
also in absence of a specific treatment a recent study has shown
delayed institutionalization, lower mortality, and reduced care
costs in participants receiving Aβ PET for diagnosis compared
with participants not undergoing the PET scan.31
Limitations
There are several limitations to the current study. First, all 3
secondary centers, including patients, are located in Sweden
and the study population is ethnically rather homogenous;
therefore generalizability to other ethnic groups and centers
may be lower. Second, despite the cohort being large, the num-
ber of participants in some subgroups, such as SCD with a pre-
PET suspected underlying AD, were relatively low and results
will need to be interpreted with this in mind. Third, the study
design did not include a control group. Fourth, even though
most participants with positive visual reads were Aβ positive
(98%), we cannot fully rule out that some early B reads might
represent primary age-related tauopathy. Additionally, when
the study started in 2017, high-performing blood-based bio-
markers were not available in clinical practice, which is why
CSF AD biomarkers were used instead before tau PET imaging.
Conclusions
In this cohort study, we found that tau PET added value to an
already extensive diagnostic workup, including high-
performing fluid AD biomarkers. The tau PET results were as-
sociated with an increased diagnostic certainty and change in
diagnoses and medication of the participants. The added value
of tau PET was most pronounced in CSF Aβ-positive partici-
pants and we suggest that the clinical use of tau PET be lim-
ited to a population where fluid AD biomarkers are abnormal.

ARTICLE INFORMATION
Accepted for Publication: March 24, 2023.
Published Online: May 22, 2023.
doi:10.1001/jamaneurol.2023.1323
Open Access: This is an open access article
distributed under the terms of the CC-BY License.
© 2023 Smith R et al. JAMA Neurology.
Author Contributions: Dr Smith had full access to
all of the data in the study and takes responsibility
for the integrity of the data and the accuracy of the
data analysis.
Concept and design: Smith, Klein, Jögi, Hansson.
Acquisition, analysis, or interpretation of data:
All authors.
Drafting of the manuscript: Smith.
Critical revision of the manuscript for important
intellectual content: Hägerström, Pawlik, Klein, Jögi,
Ohlsson, Stomrud, Hansson.
Statistical analysis: Smith, Klein.
Obtained funding: Smith, Jögi, Hansson.
Administrative, technical, or material support:
Smith, Hägerström, Jögi, Ohlsson, Hansson.
Supervision: Klein, Jögi, Hansson.
Conflict of Interest Disclosures: Dr Smith reported
grants from the Swedish Alzheimer Foundation
(AF-939981), Regionalt Forskningsstöd
(2021-1013), the Kockska Foundation, and the
Swedish federal government under the ALF
agreement (2020-YF0020), and non-financial
support from Roche during the conduct of the
study and speaker fees from Roche outside the
submitted work. Dr Pawlik reported grants from
Swedish government under the ALF agreement
during the conduct of the study. Dr Klein reported
being a full-time employee and stakeholder of
F. Hoffmann-La Roche during the conduct of the
study. Dr Hansson reported non-financial support
from Roche during the conduct of the study
(RO948) and consultant fees from Biogen, Eisai,
and BioArtic outside the submitted work. No other
disclosures were reported.
Funding/Support: The study was supported by the
Swedish Research Council (2016-00906), the Knut
and Alice Wallenberg foundation (2017-0383), the
Marianne and Marcus Wallenberg foundation
(2015.0125), the Strategic Research Area MultiPark
(Multidisciplinary Research in Parkinson’s disease)
at Lund University, the Swedish Alzheimer
Foundation (AF-939932; AF-939981), the Swedish
Brain Foundation (FO2021-0293), the Parkinson
Foundation of Sweden (1280/20), the Cure
Alzheimer’s fund, the Konung Gustaf V:s och
Drottning Victorias Frimurarestiftelse, the Skåne
University Hospital Foundation (2020-O000028),
Regionalt Forskningsstöd (2020-0314; 2021-1013),
the Kockska Foundation, and the Swedish federal
government under the ALF agreement
(2018-Projekt0279; 2020-YF0020).
Role of the Funder/Sponsor: The funding sources
had no role in design and conduct of the study;
collection, management, analysis, and
interpretation of the data; preparation, review, or
approval of the manuscript; and decision to submit
the manuscript for publication. The precursor of
18
F-RO948 was provided by Roche.
Data Sharing Statement: See Supplement 2.
REFERENCES
1. Buchhave P, Minthon L, Zetterberg H, Wallin AK,
Blennow K, Hansson O. Cerebrospinal fluid levels of
β-amyloid 1-42, but not of tau, are fully changed
already 5 to 10 years before the onset of Alzheimer
dementia. Arch Gen Psychiatry. 2012;69(1):98-106.
doi:10.1001/archgenpsychiatry.2011.155
2. Hardy J, Selkoe DJ. The amyloid hypothesis of
Alzheimer’s disease: progress and problems on the
road to therapeutics. Science. 2002;297(5580):
353-356. doi:10.1126/science.1072994
3. Villemagne VL, Burnham S, Bourgeat P, et al;
Australian Imaging Biomarkers and Lifestyle (AIBL)
Research Group. Amyloid β deposition,
neurodegeneration, and cognitive decline in
sporadic Alzheimer’s disease: a prospective cohort
study. Lancet Neurol. 2013;12(4):357-367.
doi:10.1016/S1474-4422(13)70044-9
4. Braak H, Braak E. Neuropathological stageing of
Alzheimer-related changes. Acta Neuropathol. 1991;
82(4):239-259. doi:10.1007/BF00308809
5. Tanner JA, Rabinovici GD. Relationship between
tau and cognition in the evolution of Alzheimer’s
disease: new insights from tau PET. J Nucl Med.
2021;62(5):612-613. doi:10.2967/jnumed.120.257824
6. Hansson O. Biomarkers for neurodegenerative
diseases. Nat Med. 2021;27(6):954-963.
doi:10.1038/s41591-021-01382-x
7. Rabinovici GD, Gatsonis C, Apgar C, et al.
Association of amyloid positron emission
tomography with subsequent change in clinical
management among Medicare beneficiaries with
mild cognitive impairment or dementia. JAMA.
2019;321(13):1286-1294. doi:10.1001/jama.2019.
2000
8. Altomare D, Caprioglio C, Assal F, et al.
Diagnostic value of amyloid-PET and tau-PET:
a head-to-head comparison. Eur J Nucl Med Mol
Imaging. 2021;48(7):2200-2211. doi:10.1007/s00259-
021-05246-x
9. Leuzy A, Chiotis K, Lemoine L, et al. Tau PET
imaging in neurodegenerative tauopathies-still a
challenge. Mol Psychiatry. 2019;24(8):1112-1134.
doi:10.1038/s41380-018-0342-8
10. Fleisher AS, Pontecorvo MJ, Devous MD Sr,
et al; A16 Study Investigators. Positron emission
tomography imaging with [18f]flortaucipir and
postmortem assessment of Alzheimer disease
neuropathologic changes. JAMA Neurol. 2020;77
(7):829-839. doi:10.1001/jamaneurol.2020.0528
11. Pontecorvo MJ, Keene CD, Beach TG, et al.
Comparison of regional flortaucipir PET with
quantitative tau immunohistochemistry in three
subjects with Alzheimer’s disease pathology:
a clinicopathological study. EJNMMI Res. 2020;10
(1):65. doi:10.1186/s13550-020-00653-x

jamaneurology.com (Reprinted) JAMA Neurology July 2023 Volume 80, Number 7 755

Downloaded From: https://jamanetwork.com/ on 08/29/2023

 Research Original Investigation Clinical Utility of Tau Positron Emission Tomography in the Diagnostic Workup of Patients With Cognitive Symptoms
12. Smith R, Puschmann A, Schöll M, et al.
18F-AV-1451 tau PET imaging correlates strongly
with tau neuropathology in MAPT mutation
carriers. Brain. 2016;139(9):2372-2379.
doi:10.1093/brain/aww163
13. Smith R, Wibom M, Pawlik D, Englund E,
Hansson O. correlation of in vivo [18f]flortaucipir
with postmortem Alzheimer disease tau pathology.
JAMA Neurol. 2019;76(3):310-317. doi:10.1001/
jamaneurol.2018.3692
14. Jie CVML, Treyer V, Schibli R, Mu L. Tauvi: the
first FDA-approved PET tracer for imaging tau
pathology in Alzheimer’s disease. Pharmaceuticals
(Basel). 2021;14(2):110. doi:10.3390/ph14020110
15. Smith R, Schöll M, Leuzy A, et al. Head-to-head
comparison of tau positron emission tomography
tracers [18F]flortaucipir and [18F]RO948. Eur J Nucl
Med Mol Imaging. 2020;47(2):342-354.
doi:10.1007/s00259-019-04496-0
16. Leuzy A, Smith R, Ossenkoppele R, et al.
Diagnostic performance of RO948 F 18 tau positron
emission tomography in the differentiation of
Alzheimer disease from other neurodegenerative
disorders. JAMA Neurol. 2020;77(8):955-965.
doi:10.1001/jamaneurol.2020.0989
17. Seibyl JP, DuBois JM, Racine A, et al. A Visual
Interpretation Algorithm for Assessing Brain
Tauopathy with 18-F MK-6240 Positron Emission
Tomography. J Nucl Med. 2022;64(3):444-451.
doi:10.2967/jnumed.122.264371
18. Mattsson-Carlgren N, Andersson E, Janelidze S,
et al. Aβ deposition is associated with increases in
soluble and phosphorylated tau that precede a
positive Tau PET in Alzheimer’s disease. Sci Adv.
2020;6(16):eaaz2387. doi:10.1126/sciadv.aaz2387
756 JAMA Neurology July 2023 Volume 80, Number 7 (Reprinted)
Downloaded From: https://jamanetwork.com/ on 08/29/2023
19. Ossenkoppele R, van der Kant R, Hansson O.
Tau biomarkers in Alzheimer’s disease: towards
implementation in clinical practice and trials. Lancet
Neurol. 2022;21(8):726-734. doi:10.1016/S1474-
4422(22)00168-5
20. Gobom J, Parnetti L, Rosa-Neto P, et al.
Validation of the LUMIPULSE automated
immunoassay for the measurement of core AD
biomarkers in cerebrospinal fluid. Clin Chem Lab Med.
2021;60(2):207-219. doi:10.1515/cclm-2021-0651
21. Hansson O, Seibyl J, Stomrud E, et al; Swedish
BioFINDER study group; Alzheimer’s Disease
Neuroimaging Initiative. CSF biomarkers of
Alzheimer’s disease concord with amyloid-β PET
and predict clinical progression: A study of fully
automated immunoassays in BioFINDER and ADNI
cohorts. Alzheimers Dement. 2018;14(11):1470-1481.
doi:10.1016/j.jalz.2018.01.010
22. Janelidze S, Bali D, Ashton NJ, et al.
Head-to-head comparison of 10 plasma
phospho-tau assays in prodromal Alzheimer’s
disease. Brain. 2022;awac333. doi:10.1093/brain/
awac333
23. Ossenkoppele R, Hansson O. Towards clinical
application of tau PET tracers for diagnosing
dementia due to Alzheimer’s disease. Alzheimers
Dement. 2021;17(12):1998-2008. doi:10.1002/alz.
12356
24. Janelidze S, Berron D, Smith R, et al.
Associations of plasma phospho-tau217 levels with
tau positron emission tomography in early
Alzheimer disease. JAMA Neurol. 2021;78(2):149-156.
doi:10.1001/jamaneurol.2020.4201
25. Jack CR, Wiste HJ, Botha H, et al. The bivariate
distribution of amyloid-β and tau: relationship with
established neurocognitive clinical syndromes. Brain.
2019;142(10):3230-3242. doi:10.1093/brain/awz268
26. Ossenkoppele R, Pichet Binette A, Groot C,
et al. Amyloid and tau PET-positive cognitively
unimpaired individuals are at high risk for future
cognitive decline. Nat Med. 2022;28(11):2381-2387.
doi:10.1038/s41591-022-02049-x
27. Ossenkoppele R, Smith R, Mattsson-Carlgren N,
et al. Accuracy of tau positron emission
tomography as a prognostic marker in preclinical
and prodromal Alzheimer disease: a head-to-head
comparison against amyloid positron emission
tomography and magnetic resonance imaging.
JAMA Neurol. 2021;78(8):961-971. doi:10.1001/
jamaneurol.2021.1858
28. Smith R, Cullen NC, Pichet Binette A, et al;
Alzheimer’s Disease Neuroimaging Initiative.
Tau-PET is superior to phospho-tau when
predicting cognitive decline in symptomatic AD
patients. Alzheimers Dement. Published online
December 14, 2022. doi:10.1002/alz.12875
29. Ossenkoppele R, Rabinovici GD, Smith R, et al.
Discriminative accuracy of [18f]flortaucipir positron
emission tomography for Alzheimer disease vs
other neurodegenerative disorders. JAMA. 2018;
320(11):1151-1162. doi:10.1001/jama.2018.12917
30. Pascoal TA, Therriault J, Benedet AL, et al.
18F-MK-6240 PET for early and late detection of
neurofibrillary tangles. Brain. 2020;143(9):2818-
2830. doi:10.1093/brain/awaa180
31. van Maurik IS, Broulikova HM, Mank A, et al.
A more precise diagnosis by means of amyloid PET
contributes to delayed institutionalization, lower
mortality, and reduced care costs in a tertiary
memory clinic setting. Alzheimers Dement. 2022.
doi:10.1002/alz.12846
jamaneurology.com