Third international workshop on interim-PET in lymphoma Menton (France), Palais de l’Europe, September 26-27th, 2011
Chauvie S
Imaging Exchange Diagnosis/report Diagnosis/report
exchange
Imaging
Review
Data
Thought
Imaging in clinical trial Reviewing STUDY START Qualify PET centers for the protocol (CTQ) Write PET protocol
Distribute images Gather images
Verify image quality Verify Protocol adherence
Gather review
Distribute review STUDY •3 STOP
Clinical Trial Qualification (CTQ) Qualify scanner varying make/ models of imaging equipment Review panel varying image acquisition protocols Promote imaging harmonization between centres
control / limit the multiple nuclear medicine / inter-variability image interpretation
varying image quality Qualify site imaging capabilities
Image Management ptspts pts pts
pts pts pts pts pts
Retrospective or Prospective
PET SITE PET SITE
t
pts pts pts pts pts
l a An
e z y
e ag an
PET SITE
lle c
M
Co
Sites
Reviewers
ew
Patients
vi Re
Store
MD
MD MD
WEB viewers: ◦ ◦ ◦ ◦ ◦ ◦
Learning curve Quality of monitors Limited access (e.g. ~ 5 reviewers/system) Speed of connection Need SW installation Cost
Dedicated HW: ◦ ◦ ◦ ◦ ◦ ◦
Do it yourself:
No learning curve Learning curve Unlimited access Unlimited access Optimal quality Optimal quality Fast connection Fast connection No need of SW&HW Need SW&HW installation installation Very costly ◦ Time consuming ◦ ◦ ◦ ◦ ◦
•6
Centres: which standardization? Patients: how to exchange images? Review: what viewer to use to ?
Clinical Study Retrospective, 3 centres 120 patients 2 reviewers Centres: which standardization? verify DICOM inter-operability Patients: how to exchange images? via ftp-http-CD-DVD-USB Review: what viewer to use to? do it yourself
Prospective Clinical Trial Prospective with time constraints, 100 centres, 700 patients, 7 reviewers Centres: which standardization? verify DICOM inter-operability assure standardization Patients: how to exchange images? via http Review: what viewer to use to? do it by yourself
CTQ HD0607
HD IVS Verify DICOM compatibility. Assure image quality for visual assessment
Verify DICOM compatibility. Soft CTQ: Send IQ phantom Send QA/QC program PET Protocol Assure image quality for visual assessment
2 examples
HD IVS
HD0607
HD0607
HD IVS
HD0607 PET Centre IVS PET Centre
baseline PET baseline PET interim PET interim PET
Reviewer 1
Reviewer 2
Reviewer 3
Reviewer 4
Reviewer 5
Reviewer 6
Reviewer 1
visualize PET
PET + PET -
Reviewer 3
Reviewer 4
Reviewer 5
Reviewer 6
download PET
PET + PET -
review
Reviewer 2
PET + PET -
PET+ PET-
PET + PET-
review within 3 days
PET+ PET-
HD0607 Clinical Centre
PET + PET -
HD0607 http s
http s
Q A
http s
HD IVS
VLAN
dedicated VPN
HD0607 Upload time to INFN server The average (median) time per scan upload and download were 14’33” (3’23”) and 6’56” (3’34”) respectively.
System load
HD IVS Upload time to Keosys server Bottleneck 5 PETCT/day Total time: 100 days 20 weeks
average (median) PET scan size were 133.3 MB (120.7 MB), ranging from 51.2 MB to 469.8 MB.
400 patients enrolled
336 patients with PET/CT scans
uploaded & quality controlled
261 patients with PET/CT scans approved & sent to review
Reason for PET scan exclusion •Absence of CT images •Absence of baseline PET •Absence of interim PET •CT slices missing •PET slices missing •Poor quality scans •Miscellaneous
22 25 1 3 10 6 8
•REVIEWERS •Sally Barrington – London – UK •Alberto Biggi- Cuneo – I •Michele Gregianin – Padova - I •Martin Hutchings- Copenhagen – DK •Lale Kostakoglu – New York – USA •Michel Meignan – Paris – F
Review results acquired and statistical analysed
360 patients enrolled
123 patients with PET/CT scans
uploaded & quality controlled
123 patients with PET/CT scans sent to review
Reason for PET scan exclusion •Absence of CT images •Absence of baseline PET •Absence of interim PET •CT slices missing •PET slices missing •Poor quality scans •Miscellaneous
0 0 0 0 0 0 0
•REVIEWERS • Alberto Biggi- Cuneo • Agostino Chiaravalloti - Roma • Federico Fallanca – Milano • Umberto Ficola - Palermo • Ugo Paolo Guerra – Bergamo • Michele Gregianin - Padova
Review results acquired and statistical analysed
®
Web-based Image and Diagnosis Exchange Network
STUDY START
Gather images
Verify Protocol adherence
Distribute images
Reviewing
Gather review STUDY STOP
Audit Statistical analysis •17
is a spin-off of the University of Torino and in progress with…
•To be user friendly: as an internet page •To permit fast image exchange: trough TCP/IP network •To assure security of patients’ data: anonymisation and internet banking security •To be cheap: freeware SW, no HW needed •To bypass “hospital network”: use internet •To be universal: as internet…
Login into site
https://magic5.to.inf.it/gitil
19
Upload in progress
20
Reviewer 1
Reviewer 2
Reviewer 3
Reviewer 4
Reviewer 5
Reviewer 6
Reviewer 7
reviewers receive e-mail and SMS with links
21
PET reviewers use their own SW to evaluate images
Within 3 working days 22
Reviewer inserts PET result +/- and their analysis
23
Haematologist receive e-mail and SMS with PET results
24
The average (median) time frame between the files upload by the submitting centre and the availability of the review results was 54h 24‘ (39h 42’). 7% of the scans were reviewed among in days 4 and 5; 7% of the cases were reviewed after the fifth day. In cases of a delayed scan review beyond three days the delay depended on the presence of a week-end of other holiday time.
IVS
Krippendorf’s Alpha = 0.758 binary score (4 threshold) = 0.352 1-5 scoring system
Cohen’s Kappa Binary 0.61–0.80 substantial >0.81 almost perfect
HD 0607 Krippendorf’s Alpha = 0.803 binary score (4 threshold) = 0.442 1-5 scoring system
0 is chance 1 is perfect agreement
Failure Free Survival 1,0 PET2 negative
0,8
PET2 positive
0,6 0,4 0,2 0,0 0
Time [months]
60
1,0
0,9 Time [months] 0,8 0,7 0,6 0,5 0,4
All
0,3
FFS GITIL (N=122)
Pet2 Pos
0,2
Pet2 Neg
0,1
Time [days]
0,0 0
90
180
270
360
450
540
Conclusion: ◦ Many approach exist ◦ Use the one tailored to your needs