joint imaging-therapy symposium imaging, guidance, and new ... · • from dqe to tre - translation...

Joint Imaging-Therapy SymposiumImaging, Guidance, and New Technologies

in IG Interventions

Joint Imaging-Therapy SymposiumImaging, Guidance, and New Technologies

in IG Interventions

Jeff Siewerdsen, PhDDepartment of Biomedical Engineering

Johns Hopkins University

David Hawkes, PhDUniversity College

London UK

Paul Keall, PhDDepartment of Radiation Oncology

Stanford University

DQE �� FDK �� TREThese and Other TLAs in the Development

of New Technologies for Image-Guided Surgery

DQE �� FDK �� TREThese and Other TLAs in the Development

of New Technologies for Image-Guided Surgery

Jeff Siewerdsen, PhDDepartment of Biomedical Engineering

Johns Hopkins University

Johns Hopkins UniversitySchools of Medicine and Engineering

Depts. of Biomedical Engineering,Computer Science and Surgery

Overview• “Closing the loop” in image-guided surgery

- Operating rooms of the future

• Imaging modalities for IGS

- MR… US… CT… and cone-beam CT

• From DQE to TRE

- Translation of a new imaging technology from the laboratory to the OR

• Coming advances in IGI

- Evolution and revolution in image-guided procedures

- Meeting new standards in technology assessment

New Opportunities in IGIHigh-precision imaging / guidance / monitoring / adaptation:

- Maximize the performance of existing techniquesIncreased target ablationAvoidance of critical structuresMore efficient therapeutic delivery, workflow

- Expand the application of current techniquesManagement of otherwise “untreatable” diseaseIncreasingly patient-specific therapies

- Develop entirely new therapeutic approachesAdvanced delivery systems (e.g., robotics, PDT, …)Investigate the synergy of adaptive, multi-modality therapies (physical / molecular / cellular / pharmaceutical)

- Expose fundamental factors determining outcome

Current technologies for IGI do not provide a “closed loop.”

Planning and Navigation

The PatientThe Surgeon

Image-Guided Intervention (IGI)

Treatment Delivery

Intra-operative imaging helps to close the loop.

Image-Guided Intervention (IGI)

The Surgeon

Planning and NavigationIntra-Operative Imaging

Treatment Delivery

The Patient

Cycle of IGI

Planning andNavigationResponse

Imaging

Intervention

ORs of the FutureFor example: AMIGO: Advanced Multi-Modality Image-Guided OR

F Jolesz, C Timpani, et al., CIMIT, BWH

UltrasoundFluoroscopyEndoscopyMicroscopy

Patient transport

3T MRI PET-CT

• Emerging themes:- Computerized integration of systems / technologies

(e.g., planning and real-time navigation)- A wealth of data presented intra-operatively- Advanced therapy delivery systems (e.g., robotics)- Multi-modality imaging

• The immediate challenges:- Implementation, integration, and evaluation- Comparative effectiveness

• The longer-term challenges:- Cost- Entirely new technology advances, specific to the OR- Streamlined integration with therapy workflow

ORs of the Future

• Imaging technologies- Continuous advances in Dx imaging technology apply

to Tx imaging as well- However, challenges for Tx imaging go beyond direct

implementation in the arena of therapy

• Imaging for Tx guidance involves distinct:- Imaging tasks- Requirements (image quality, speed, accuracy, etc.)- Constraints (dose, patient access, etc.)

• Overall challenge:- Develop imaging strategies motivated directly by the

objectives and constraints of therapy guidance

Imaging for Therapy Guidance

X-ray Fluoro / CBCT

Imaging Modalities for IGICT Ultrasound

MR Nuclear MedicineOptical

Imaging Modalities for IGIKey Characteristics• Real-time

(or near-real-time)

• Radiation dose~1/10 – 1/2 of Dx CT

• Sub-mm resolution

• Soft-tissue visibility

X-ray Fluoro / CBCT

Cone-Beam CT

Projection dataMultiple projections

over ~180o

Volume reconstructionSub-mm spatial resolution

+ soft tissue visibility

Mobile C-Armfor Intraoperative Cone-Beam CT

Multiple projection images acquired over ~180o

2D Image acquisition- Nominal: 60 s- High-speed motor: 10 s

3D Image reconstruction- Nominal: 60 s- High-speed recon: 10 s

Radiation dose- ~1/10th that of Dx CT

MotorizedOrbit Replace XRII with

Flat-Panel Detector

GeometricCalibration

Tube + CollimatorModification (FOV)

Image Acquisition3D Reconstruction

Control System

Mobile Isocentric C-ArmSiemens PowerMobil

Mobile Isocentric C-ArmCone-Beam CT-Capable C-Arm

Pre-clinical platform for multi-mode Fluoro / CBCT guidance

Image Acquisition3D Reconstruction

Control System

Applications in IG Surgery

Platform for optimizing / integrating imaging and

navigation

• Orthopedic Surgery• Spine Surgery• Brachytherapy• Ear Surgery• Interventional Radiology• Urology• Lung Surgery• Breast Surgery• Head and Neck Surgery


In vivo studiesof image quality and geometric precision




Soft-tissue visualization and real-time planning

Resection ofsub-palpable lesions




Maximal target ablation and critical structure

avoidance


An IntegratedImage Guidance System

for Head and Neck Surgery

CBCT C-Arm Real-Time Tracking Endoscopy

Image-to-World Deformable Reg Endo-to-CBCT

Pre-Op CT, MR, etc. Surgical Planning

3D / 4D Multi-ModalityVisualization

Integration of Image Guidance Systems

Intr

a-O

pR

egP

re-O

p

?

Multi-Scale Demons Algorithm

1

0 5 10 15 20 25 300.93

0.94

0.95

0.96

0.97

0.98

0.99

Imag

e C

ross

-Cor

rela

tion

Iteration #

Bin x8

Bin x4Bin x2

Bin x1

11 sec

12 sec

18 sec52 sec

Def

orm

atio

n F

ield

Diff (I1 – I0Rigid)

Diff (I1 – I0Deformable)

RIG

IDD

EF

OR

MA

BLE

Intra-Op CBCT (POST-Resection)

Intra-Op CBCT (PRE-Incision)

3D Deformable Image Registration

Rig

id R

egD

emons

Reg

3D Deformable Image Registration:“Demons” Algorithm

0

1

2

3

4

5

6

7

8

Targ

et R

egis

trat

ion

Err

or

(mm

)

RigidDemons

Registration Performance (TRE)

Performance Evaluation10 PatientsRigid vs. Demons6 Anatomical Targets

TemporalBone

L RMandibleL R Spine Soft

Tissue

RigidDemons

Real-Time Tracking and Navigation

Stereoscopic IR Camera(NDI Polaris) Fiducials

TargetCentroid

Optimal fiducial arrangementsAutomatic registration of image and tracking systems

Tungsten BB(1.0 mm radius)

InfraredReflective Sphere(5.8 mm radius)

Multi-Modality Marker(IR-Xray)

Automatic Tracking RegistrationTarget Registration Error (TRE)Manual Registration

TRE = (1.29 ± 0.34) mmAutomatic Registration

TRE = (1.14 ± 0.20) mm� Comparable accuracy� Improved reproducibility

Markers fixed to cranium surfaceTRE = (0.83±0.20) mm

Markers off-surface(better arrangement for H&N)

TRE = (1.24±0.20) mm� NSS (p=0.79)

WorkflowManual Registration: minutesAutomatic Registration: seconds, automatic with each CBCT scan

Integrated CBCT + Endoscopy

PolarisVicra

C-Arm

3D Visualization + Video Endoscopy

CadaverHead

Tracked EndoscopePointer

Application in Sinus and Skull Base SurgeryPre-Clinical Studies

• Tomosynthesis- Image quality and dose- Localization accuracy

• Temporal Bone- Image quality (spatial resolution)- Visualization of cochlear implants

• Sinus ablation- Identification of CSF leak- Frontal recess, ethmoid, sphenoid

• Skull base- Completeness of target ablation- Avoidance of critical structures- Quantitative analysis of surgical

performance under CBCT guidance

CBCT-Guided Sinus Surgery

• Identification of CSF LeakCadaver specimenTotal sinus ablation

• Breach introducedDiameter: 1, 2, 4 mm1-2: Cribriform3-4: Lateral lamella5-6: Fovea ethmoidalis7-8: Planum

• Observer study5 surgeons, 1 radiologist5-point detectability scaleMin detectable breach

No CSF Breach

Coronal CBCT

Ethmoid Air Cell Ablationin proximity to

Fovea Ethmoidalis

CBCT-Guided Sinus Surgery1 mm Breach

2 mm Breach 4 mm Breach

5= Perfectly obvious4= Visible3= Visible, but challenging2= Could be overlooked1= Unable to identify

CBCT-Guided Sinus Surgery

0

1

2

3

4

5

0 1 2 3 4 5

Defect S ize (mm)

(d) Planum

Obs

erve

r Sc

ore

Defect S ize (mm)

Y = 0.68x + 2.02

0

1

2

3

4

5

0 1 2 3 4 5

Defect S ize (mm)

(d) Planum

Obs

erve

r Sc

ore

Defect S ize (mm)

Y = 0.68x + 2.02

Scor

e (D

etec

tabi

lity)

Detection of CSF Leak

Rating Scale:

Confident detection >~1.5 - 2 mm breach

Skull Base Surgery:Target Abation in the Clivus

Intra-Operative CBCT

TARGET volumeNORMAL volume

Critical

Skull Base Surgery:Target Abation in the Clivus

Intra-Operative CBCT


Post-Operative CBCT

TARGET RemainingNORMAL Remaining

Critical Critical

Hypothesis-testing framework for evaluation of surgical performance

Fraction of Tumor

Excised

Fraction of Normal

Remaining

= Sensitivity =

= Specificity =

FNTP

TP

+

TNFP

TN

+

Positive Predictive

Value= PPV =

FPTP

TP

+

Negative Predictive

Value= NPV =

FNTN

TN

+

Quantification of Surgical PerformanceQuantification of Surgical Performance

0.0

0.2

0.4

0.6

0.8

1.0

0.0 0.2 0.4 0.6 0.8 1.0

CBCT-Guided

Unguided(conventional)

1-Specificity(Fraction of Normal Excised)

Sen

sitivity

(Fra

ctio

n o

f Targ

et E

xcis

ed)

CBCT-Guided Skull Base Surgery

Normal TissueTarget Volume (excised)Target Volume (residual)

Translation to Clinical Trials

C-Arm Trials: Mandibulectomy

Sca

n 1

Sca

n 2

Sca

n 3

Sca

n 4

Sca

n 1

Target(Radionecrosis)

Sca

n 2

Sca

n 3

Pla

n C

utPlan

Cut

ResectionSca

n 4

FibulaReconstruction

Plates

C-Arm Trials: Invasive Tumor

Sca

n 1

Sca

n 2

Sca

n 3

Sca

n 4

Sca

n 1

ChondrosarcomaSca

n 2

Craniotomy

Sca

n 3

Tumorresection

Tumormargins

Sca

n 4

Closure

Packing

Looking Ahead:Opportunities in IGI

Multi-Modality Intra-Operative Imaging- Essential to ‘closing the loop’ in image-guided surgery

MR / US / CT / Cone-beam CT- More than simple translation from Dx to Tx

New development for the task of surgical guidance

Integration of Guidance Systems- Essential to preserving / improving surgical workflow

Intraoperative imagingImage registrationPlanning and real-time navigation3D visualizationTherapeutic delivery (e.g., robotics)

ORs of the FutureCurrently characterized by conglomeration of technology

- For example: MR/CT/US + Robot- Engineering challenge is beyond simple integration

Rather, a re-engineering effort (or new system altogether) is required according to the surgical task

ORs of the (more distant) future- Hopefully exceed conglomerate “IGORs” of the present

Technologies created specifically for the ORStreamlined integration with surgical workflow

An inter-disciplinary approach to new technologies in the operating theatre

- (Surgeons, Nurses, Anesthesia) + (Physicists, Engineers)- Radiation therapy offers a successful model example

Demonstrating True BenefitSuch IG systems and ORs will need to meet new standards of performance: comparative effectiveness

- Within technology (intra-metrics):Optimal implementation of ‘Technology X’

- Between technologies (inter-metrics):Performance of ‘Technology X’ vs ‘Technology Y’

- Clinical justification:Effectiveness of therapy under ‘Tech X’ vs standard of care

Quantification of surgical performance is key- Short of randomized clinical trials, approaches like

hypothesis-testing could be valuable

- Rigorous computerized planning

- Quantitative evaluation of planned (and delivered) Tx

AcknowledgementsUniversity Health Network

• JC Irish, MD – Surgical Oncology• G Bachar, MD• E Barker, MD• H Chan, PhD• MJ Daly, MSc• M Rafferty, MD• A Vescan, MD

University of Toronto• N Hamming, MHSc• S Nithiananthan

Funding Support• NIH R01-CA112163• NIH R01-CA127444• Siemens Healthcare (SP, Erlangen)• Princess Margaret Hospital Foundation

Thank you.Thank you.Thank you.Thank you.

Cone-Beam CT

Projection dataMultiple projections

over ~180o

Volume reconstructionSub-mm spatial resolution

+ soft tissue visibility


Platform for optimizing / integrating imaging and

navigation



In vivo studiesof image quality and geometric precision




Soft-tissue visualization and real-time planning

Resection ofsub-palpable lesions




Maximal target ablation and critical structure

avoidance


Skull Base Surgery:Target Abation in the ClivusIntra-Operative CBCT


Critical

Skull Base Surgery:Target Abation in the ClivusIntra-Operative CBCT


Post-Operative CBCT

TARGET RemainingNORMAL Remaining

Critical Critical

C-Arm Trials: Maxillectomy

Sca

n 1

Sca

n 2

Sca

n 3

Sca

n 4

Sca

n 1

Adenoid CysticCarcinoma

Sca

n 3

Resection

Sca

n 4

DentalProsthesis

joint imaging-therapy symposium imaging, guidance, and new ... · • from dqe to tre - translation...

Documents