Molecular Imaging Instrumentation Lab (Craig Levin)med.stanford.edu/.../ED/research/levinlab.pdf · 3 MIPS Stanford University Molecular Imaging Program at Stanford School of Medicine

11

MIPS Stanford UniversityMolecular ImagingProgram at Stanford

School of MedicineDepartment of Radiology

Angela Peter Craig Guillem Billie

Jin Frezghi Garry

Molecular Imaging Instrumentation Lab(Craig Levin)



I. High resolution radionuclide imaging:• positron emission tomography (PET)• single photon emission computed tomography (SPECT)

II. Optical tomography:• Bioluminescence tomography (BLT)• Fluorescence tomography (FLT)

Research interests (by imaging modality)

Research interests (by imaging technology)•New photon sensors•Low noise electronics•Data acquisition systems and software•Computer simulation and modeling•Image reconstruction algorithms•Image analysis algorithms•Integrating components into new imaging systems



Two new radionuclide camerasunder development

I. Miniature gamma ray camera forsurgical cancer staging

II. Dedicated breast PET system tohelp resolve difficulties withbreast cancer detection,diagnosis and staging

22



I. Cancer staging problem

Localization and biopsy of “sentinel” lymph node is important for cancer staging (e.g. melanoma and breast cancer) to avoid unnecessary dissection of the lymphatic system

Sentinel node negative Sentinel node negative Cancer has not spread Cancer has not spread

Identification of sentinel node requires:•Radiotracer that accumulates in node•Pre-operative Lymphoscintigraphy•Sensitive Radiation Detector Probe



Solid State Detector ProbeStandard Non-imaging Radiation DetectorProbe for Surgical Cancer Staging

Features:•Audio tone generator•LCD display•Digital count readout•Automated energy windowing

Neoprobe, Inc.



SUMMARY OF THE GAMMA PROBE METHODSummary of the Sentinel Node Method

• High accumulation in the sentinel node(s)• High detection sensitivity of the non imaging radiation detector• Location of sentinel node(s) can be identified prior to any incisions

• Poor labeling efficiency• Inappropriate size of colloid particle• Disruption of lymphatic flow• Inappropriate injection dose or volume

• Sentinel node is not identified during lymphoscintigraphy• Node is within or near the injection site or bolus activity• Node relatively deep within tissue• Node cannot be distinguished from other secondary nodes or

colloid migration path in lymphatic channels• Node is too small or its activity is too low

Limitations of the method

DESIRABLE CHARACTERISTICSDesirable characteristics

COULD BECOULD BE RESOLVED

WITH AWITH A

• Procedure is time consuming and at times quite invasive• Learning curve is significant

SMALLSMALLCAMERACAMERA

33



Borys R. Krynyckyi, MD, et al. “Clinical Breast Lymphoscintigraphy: Optimal Techniques for Performing Studies,Image Atlas, and Analysis of Image”, Radiographics. 2004 Jan-Feb;24(1):121-45; discussion 139-45.

Lymphoscintigraphy of the Sentinel Node

Challengingcase: node

could not bedetected by

non-imagingprobe since itwas too closeto injection

site



Gammarays

Small camera visualizesand measures gamma raycount rate variations frominjected radio-colloid (5x5 cm2 FOV)

Cam

era

Injection site(IS)

Sentinel node(SN)

2-D IMAGE

Miniature Gamma Ray Camera for ImprovedMiniature Gamma Ray Camera for ImprovedSurgical Staging of CancerSurgical Staging of Cancer

SNIS

Camera front-end assembly

Collimator NaI(Tl) PSPMT HV/readout

3 mm 4 mm

5 mm

6 mm7 mm

8 mm6 cm

Node phantom(spheres)

Node detection studiesNodes 3cm deep 4 cm deep

Acquis.Time5 sec

10 sec

Sentinel node imaging

5 cm

Can visualize 3 mm nodes in 5 seconds!



Detection: ~ 25% of screened cases are inconclusive (due to dense/distorted tissue)

Diagnosis: Mammographic breast cancer signatures are only 20-30% specific---> ~ 600,000 unnecessary biopsies are performedannually in this country ($1-3K)

Staging: Missed nodes and unnecessary axillary lymph node dissections---> non-invasive techniques are needed

Need to investigate sensitive, specific, non-invasive techniques

II. Difficulties with breast cancer detection,diagnosis and staging

44



Can radionuclide imaging help with breastcancer screening/diagnosis/staging?

•Can the lesion/node be easily detected (high sensitivity)?

•Does the radiotracer has a very high affinity/accumulation for

cancer/nodes (high specificity)?

•Is the cost of the imaging procedure reasonable?

•Can the study be performed reasonably quick?



Drawbacks of the Standard PET Camerafor Breast Imaging

80 cm



Drawbacks of the Standard PET Camerafor Breast Imaging

•Large and awkward for breast and axilla imaging

•Inability for close-proximity imaging ---> poor spatial resolution (6-12mm)

---> poor detection efficiency (<1%)

•Accepts activity from outside organs---> background haze

•Relative high cost per study

Detectorcrystals

Tumor~1 cm

55



PET camera dedicated to breast imaging

Potential Roles:

•Screening indeterminate cases?

•Reducing rate of false positives??

•Staging of axillary nodes?

•Detecting recurrence?

•Monitoring therapy?



PET camera dedicated to breast cancer imaging

1x1x10 mm3 LSOscintillation crystal

511 keVphotons

Standardchip andceramicpackage

Thin PSAPD chip on flexcircuit (~200µm thick)

Highly compact semiconductorlight sensors (PSAPD)

Heart (10 cm)

Breas

t

Torso

Detectorpanel

(10x15cm2)

30 cm20 cm

30 cm

Detector panel(10x15 cm2)

Tumors

10x15 cm2

Position along z-direction (mm) Position along z-direction (mm)

-50 0 50Position along z-direction (mm)-50 0 50-5

0

5

Coun

ts

-50

5

0 50-50

8 secacquisition

100 200 300

50100150200250300350400

0

1

2

3

4

5

1050 100 150 200 250 300

50

100

150

200

250

300

350

400

0

1

2

3

4

5

10

Coun

ts

12 secacquisition

19 secacquisition

100 200 300

50

100

150

200

250

300

350

400

0

1

2

3

4

5

0

10

Coun

ts

20

-100

1020

0 50-50Position along z-direction (mm)

Coun

ts

Hot tumors, warm breast, hot heart,warm torso

50 100 150 200 250 300

50

100

150

200

250

300

350

400

0

1

2

3

4

5

6

7

8

910

60 sec acquisitionHot tumors only

4 mm3.5 mm

3 mm 2.5 mm

Breast tumor detection simulations

Simulations predict: can see 3 mm tumors in 10 sec.

With background activity in breast, heart and torso No background Activity concentration ratio--Tumors:Breast:Heart:Torso=10:1:10:1

Documents

Molecular Imaging Instrumentation Lab (Craig Levin)med.stanford.edu/.../ED/research/levinlab.pdf · 3 MIPS Stanford University Molecular Imaging Program at Stanford School of Medicine