State of the Standards 2018 Kevin ODonnell, MASc. Canon Medical - - PowerPoint PPT Presentation

State of the Standards 2018

Kevin O’Donnell, MASc. Canon Medical Research

“If you build it, they will come…”

Yeah, right.

Draft & Publish Standard Implement & Release in Product Install & Use in Practice SDOs Vendors Users

IHE

• Secretariat: RSNA & HIMSS

– www.ihe.org – secretary@ihe.net

• Scope:

– Profiling existing standards to address specific use cases in healthcare

• 13 Domains

– Radiology, Cardiology, IT Infrastructure, Lab/Pathology, Eyecare, …

IHE Process

• Add Profiles to IHE Technical Frameworks
• Steps:

–(Annual Cycle) –Profile Proposal –Public Comment –Trial Implementation –Final Text

IHE Radiology – What’s New

Recent TI Publications

• Management of Acquisition Protocols (MAP)

– Collect/Manage/Distribute CT Protocols

• Standardized Operational Log of Events (SOLE)

– Capture Operational Events related to Imaging

• Results Distribution (RD)

– Send Radiology Reports with key flags and metadata – (Foundation for FUNC – Follow-Up of Non-Critical actionable findings)

… … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … …

MAP Motivation

One Scanner can have 100 or more protocols… Quality of Care

– Good, consistent image quality depends on good protocols, used consistently

Efficient Workflow

– Hanging Protocols, Image Processing, Reporting Workflow depends on consistent series names, procedure codes, etc.

Dose Management

– Managing dose depends on managing protocols

Analytics

– Summarizing data depends on consistent tagging

… … … … … … … … … … … … … … … … … … … …

Management of Acquisition Protocols (MAP)

Centralized, cross-vendor review and approval

• Upload protocols
• Periodically review & approve
• Modify (within limits)
• Re-distribute protocols and

approvals

Protocol Archive Scanner A Scanner B Scanner C Protocol Manager (Workstation)

MAP Technology

DICOM CT Defined Procedure Protocol IOD

– Encodes a complete protocol (incl. private tags and constraints)

DICOM Protocol Approval IOD

– Encodes list of protocols and associated approvals

DICOMweb - NPI Service

– RESTful transactions to Store, Query, and Retrieve Non-patient Instances

SOLE Motivation

Help sites solve operational problems; improve performance

• “You can’t manage what you don’t measure”
• Scope initially limited:

– Imaging departments that use orders and schedules. – Activities between Order and Report.

• Build on SWIM, Syslog, DICOM Schema (like ATNA), REST

– Existing tooling, shared infrastructure

Use Case: Track Order to Ready-to-Read (Part I)

* Courtesy Rob Horn

Use Case: Track Ready-to-Read to Report-Available (Part II)

* Courtesy Rob Horn

Use Case: Bulk Analysis

* Courtesy Rob Horn

Results Distribution (RD)

Send reports to EMR (and external systems, e.g. for FUNC action)

• FOCUS ! Existing infrastructure -> HL7 v2.x ORU
• Enable SOME structured, coded data to begin to move industry
• Useful gap solution, not end game
• Important “baby steps” towards computer processable data

* Courtesy Teri Sippel, Vital Images

Profiled HL7 v2.5.1 ORU

* Courtesy Teri Sippel, Vital Images

Profiled OBX Segments

• OBX – DICOM Study Instance UID
• OBX – Finding
• OBX – Radiologist’s Recommendation
• OBX – Radiologist Requests Consultation
• OBX - Radiologist Requests Feedback
• OBX - Imaging Result Payload

Note: OBX “type” indicated by LOINC Code in OBX-3

* Courtesy Teri Sippel, Vital Images

ACR Actionable Findings Categories

Finding type ACR Actionable Finding Category Meaning Normal Normal As expected or unremarkable Non-actionable Non-Actionable Not quite normal, but no action to be taken (e.g., “the spleen is slightly enlarged”) Non-critical ACR Category 3 Medical attention required within days to months; “incidental” (e.g., lung nodule) Urgent ACR Category 2 Medical attention required within hours (e.g., diverticulitis) Emergent ACR Category 1 Medical attention required within minutes (e.g., large pneumothorax)

* Courtesy Teri Sippel, Vital Images

Standardized Mapping of Actionable Codes & Priority

* Courtesy Teri Sippel, Vital Images

IHE Radiology – What’s Next

Entering Trial Implementation

• Encounter-Based Imaging Workflow (EBIW)

– Part 1: Point-of-Care Ultrasound

• Web Image Access (WIA)

– Formerly MHD-I

Entering Public Comment later this year (Sneak Preview)

• Import and Display of External Priors (IDEP)

– Proactively find and import priors over the network

What is Encounter-Based Imaging?

EBI = imaging captured in the context of an encounter between a patient and a care-provider (typically without an imaging order)

Enterprise Imaging

Order-based Imaging

[CT, MR, PET, XA, MG, Echo]

(Radiology, Cardiology, etc.) Encounter-based Imaging

[PoCUS, Camera, iPhone,…] (Dermatology, Wound Care, Burn Unit, ER, etc., etc., etc., etc.)

So What’s the Key?

Key Metadata

• To manage the images, and readily find & use them again

– Patient Metadata

• Name, Date of Birth, Age, MRN, Other IDs, …

– Encounter Metadata

• Department, Specialty, Physician, Referring Physician,

Admission ID, Admission Date/Time, Reason for Admission, …

– Procedure Metadata

• Modality, Imaging Procedure, Body Part, Reason for Imaging, Accession #, Acquisition

Date/Time, Device, Operator, Acquisition Parameters (more for PoCUS than cameras), …

– Pixel Metadata

• Rows, Columns, Bit Depth, Pixel Size, Encoding, ...
• … turning a JPEG into a medical document

Encounter-Based Imaging Workflow (EBIW)

Use Case 1: Point-of-Care Ultrasound (ER)

• Scan wristband
• [Gets metadata]
• Scan Patient
• [Adds metadata]
• [Stores Images]
• [Notifies EMR]

Image Manager/ Archive PoC US Encounter Manager Result Aggregator Get Imaging Context Notify (Patient ID, Visit ID, Acq. Time, Procedure Code, Accession #, Body Part, etc.) EMR, MPI, PAM, PDQ, etc. EMR

IHE EBIW Actors & Transactions

DICOM

• Secretariat: MITA

– dicomstandard.org – dicom@dicomstandard.org

• Scope:

– data and workflow compatibility between imaging systems and HIT systems – diagnostic, therapeutic & associated data in medical disciplines that use digital images

• 31 Working Groups

– CT, MR, Dentistry, Physics, Web Technology, etc.

DICOM Process

• Add Supplements to DICOM Standard
• Steps:

–New Workitem Proposal –Public Comment –Letter Ballot –Final Text

DICOM – What’s New (2017)

• Sup 189 Advanced Blending Presentation State
• Sup 190 Volume Rendering Presentation State
• Sup 191 Patient Radiation Dose Report
• Sup 192 Protocol Approval
• Sup 194 RESTful Non-Patient Instances
• Sup 197 Ophthalmic OCT Angiography
• Sup 200 Transform NCI AIM & DICOM SR Measures
• Sup 198 Retire WADO-WS
• Sup 201 Retire Radiation Dose from MPPS

Patient Dose Surface

Advanced Blending Presentation State

(Example)

+ + +

Anatomical DTI fMRI (Relative Opacity 0.33) fMRI (Relative Opacity 0.33) fMRI (Relative Opacity 0.33) Relative Opacity 0.6 Relative Opacity 0.4

* Courtesy Wim Corbijn

(Relative Opacity 0.7) (Relative Opacity 0.3)

Volume Rendering - Example

* Courtesy WG-11

Graphic Annotation Volumetric Graphic Annotation Graphics Layer Projection

P-Values or PCS-Values P-Values or PCS-Values

VOI LUT Cropping Volume Rendering

RGBA

Classification Cropping Specification Index value, Volume Cropping Render Geometry Render Shading 3D Voxel Data 2D Pixel Data Render Display

Patient Dose SR

Patient Dose Map Patient Dose Surface Patient Model Registration IOD

Equipment Information

• Table dimension
• Attenuating material, …

Patient Model FOR

Radiation Dose SR

Radiation Exposure Information

• X-Ray exposure techniques
• Table, Gantry Angle, Beam Geometry, collimation
• Dose measure: CTDI, DAP, ...

Modality Image X-Ray Equipment

Patient Registration Information

• Patient position on table
• Fiducials
• Registration of Patient FOR and Equipment FOR

Patient Dose SR

Organ Dose Reporter System

* Courtesy WG-28

Modality FOR

Reactivated Working Groups

WG-14 Security

• Common members with MITA Cybersecurity
• Ready to apply & extend DICOM

WG-17 3D (Additive Manufacturing)

• Aligned with SME (“Society of Manufacturing Engineers”)

WG-23 Application Hosting (Data APIs)

• APIs for query & data access appropriate to training AIs, and

using AI to process a current study

• Machine Learning buzz continues at RSNA, HIMSS, SIIM

DICOM – What’s Next

• Sup 188 Multi-energy CT Image
• Sup 164 Contrast Administration SR
• Sup 183 Web Services Re-documentation
• Sup 193 REST Notifications
• Sup 203 DICOMweb Thumbnails

4 families many flavors

Virtual Mono- energetic Image (VMI) Material- Specific Image Color Overlay Image Color Blending Image Discrete Labeling Image Effective Atomic Number (Z) Image Electron Density Image Probability Map Image Color Map Image

Multi Energy Imaging Material Quantification Family Material Labeling Family Material Visualization Family

Standard CT Image

Objective Image Family

Iodine Map; Bone Density

Multi-Energy CT

Proportional Map Image Value based Map Image Material- Modified Image

Gout crystals Highlighted; Partially- Suppressed

Material- Removed Image

• Virt. Non-

Contrast;

• Virt. Non-Ca;

CT IOD Other IOD

* Courtesy WG-21

Multi-Energy CT

* Courtesy WG-11

Performed Administration SR

Contains

Programmed Plan Delivered Plan

Programmed at injector device Actually delivered by injector device

Steps

1 1

1-n

Phases

1-n

Steps

1-n

Phases

1-n

Pressure

• Vs. Time

Graphs Flow Rate

• Vs. Time

Graphs

Contrast Administration SR

DICOM – What’s Next

36

• Sup 204 TLS Security Profiles (FT 2018)
• Sup 206 CRYPTREC TLS Profile
• Sup 205 STL Encapsulation for 3D Manufacturing (FT 2018)
• Sup 208 X3D Encapsulation for 3D Manufacturing
• Sup 147 Second Generation Radiotherapy
• Sup 175-9: 2nd Generation Radiotherapy
• Sup 202 Real Time Video
• Sup 207 Conformity Assessment

DICOM Realtime Video

• Sample Use Case: Automated OR “Multi-display”
• Tile multiple videos on one monitor
• Layout (“hanging protocol”) depends on procedure, metadata

* Courtesy Working Group 13

QIBA

• Secretariat: RSNA

– www.rsna.org\qiba – qiba@rsna.org

• Scope:

– Improve the value and practicality of quantitative biomarkers by reducing variability across devices, clinical sites and time

• 4 Coordination Committees (CT, MR, PET, US)

– Collaboration with Europe (EIBALL) and Japan

QIBA – Industrializing Biomarkers

39 Treat W ait

Measure = 7±6

?

Problem Analysis Solution

Treat W ait

±2

!

Goal

Measure = 7 ±6

Sources of Variance Differences in:

• Patient Handling
• Acq. Protocols
• Reconstruction
• Segmentation

. . .

When all participating actors conform…

Protocol Reqs Recon Reqs Resolution Reqs Noise Reqs

• Segment. Reqs

Patient Prep & Operation Reqs

• Segment. Reqs

Calibration Reqs

QIBA Process

• Add Profiles to QIBA Library
• Steps:

– Biomarker Proposal – Groundwork – Public Comment – Consensus – Technical Confirmation – Claim Confirmation – Clinical Confirmation

QIBA – Recently Published

Technically Confirmed

• CT Tumor Volumetry

– Claim: A true change in tumor volume occurred with 95% confidence if measured change is > 24% / 29% / 39% respectively for tumors with initial diameters of 50-100 / 35-49 / 10-34 mm.

• FDG-PET/CT for Tumor Evaluation

– Claim: Tumor glycolytic activity (SUVmax) measurements have a within-subject coefficient of variation of 10-12%.

Consensus

• CT Lung Small Nodule Screening
• SPECT I-123 Ioflupane for Dopamine Transporters in Neurodegenerative Disease

QIBA – More Work

Public Comment

• fMRI BOLD for Pre-Op Mapping of Cortex
• PET Amyloid for Neurologic Disease
• MR DWI for Tumor Evaluation
• MR Elastography

Work in Progress

• UL Shearwave for Liver Elastography
• MR Proton Density Fat Fraction
• CT Lung Density for COPD