statistical process control for analysis of routine machine...
Post on 20-Sep-2020
3 Views
Preview:
TRANSCRIPT
Statistical Process Control Statistical Process Control for Analysis of Routine for Analysis of Routine
Machine QCMachine QC
Mac Clements, MSMac Clements, MS
Radiological Imaging Technology, Inc.Radiological Imaging Technology, Inc.
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 11
BackgroundBackground
�� Recent publications have stressed the Recent publications have stressed the need for managing QA systems.need for managing QA systems.�� Sentinel Event Alert #47 (2011)Sentinel Event Alert #47 (2011)
�� ““Implement a system for centralized quality and Implement a system for centralized quality and safety performance monitoring of inventoried safety performance monitoring of inventoried equipment under the supervision of a qualified equipment under the supervision of a qualified medical physicistmedical physicist””
�� AAPM Task Group 142 (2009)AAPM Task Group 142 (2009)�� ““A sudden and significant deviation from the A sudden and significant deviation from the
expected value should be called to the attention of expected value should be called to the attention of the MP, even if the measurement itself does not the MP, even if the measurement itself does not exceed the table tolerance value.exceed the table tolerance value.””
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 22
BackgroundBackground
�� ACR AccreditationACR Accreditation�� Therapy accreditation inspecting to Therapy accreditation inspecting to
TG142 standards (as of 2011)TG142 standards (as of 2011)
�� State legislation (TG142 State legislation (TG142 –– Ohio??)Ohio??)
�� RTOGRTOG�� November 2011 November 2011 –– updated its updated its
requirement and is testing to TG142 requirement and is testing to TG142 standards rather than TG40standards rather than TG40
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 33
Definition of Definition of ““ QualityQuality ””
Quality ControlQuality Control
�� Is the process Is the process ““in in controlcontrol”” or or ““out of out of controlcontrol””??
�� Consistency in Consistency in measurements.measurements.
Quality AssuranceQuality Assurance
�� Predictability in Predictability in measurements.measurements.
�� Opposite of the Opposite of the ““rearrear--view mirror approachview mirror approach””..
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 44
•• Minimizing process variationMinimizing process variation•• Understanding source of fluctuationsUnderstanding source of fluctuations
•• And how/when to fix itAnd how/when to fix it
Principles for Understanding DataPrinciples for Understanding Data 11
�� ““No data have meaning apart from their No data have meaning apart from their context.context.””
�� ““Every data set contains noise, and some data Every data set contains noise, and some data sets contain signals. Therefore, before you sets contain signals. Therefore, before you can detect a signal, you must first filter out the can detect a signal, you must first filter out the noise.noise.””
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 55
1. Understanding Variation; The Key to Managing Chaos. (2nd Edition) Donald Wheeler (2000)
Conventional QAConventional QA
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 66
Begin
End
Determine Specifications
Perform the Test
Result
Determine Reason
Document
Unacceptable resultAcceptable result
�� No understanding of variation No understanding of variation inside specificationsinside specifications
�� No understanding of trends in No understanding of trends in datadata
Pawlicki, T., Whitaker, M. “Variation and control of process behavior”, Int J Rad Onc Biol Phys (2008)
The Source of (Some) SpecificationsThe Source of (Some) Specifications
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 77
�� Without any context for the data, specification Without any context for the data, specification may give an incomplete picture.may give an incomplete picture.
AAPM
ACR
Manufacturer
1. Understanding Variation; The Key to Managing Chaos. (2nd Edition) Donald Wheeler (2000)
Specifications vs. ContextSpecifications vs. Context
�� Is the measurement on day 14 a signal?Is the measurement on day 14 a signal?
�� Is it something to be concerned about?Is it something to be concerned about?
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 88
Possible FailuresPossible Failures
�� Whenever a system attempts to Whenever a system attempts to detect a detect a ““signalsignal””, there are 2 possible , there are 2 possible failures:failures:�� 1: Failure to separate the signal from 1: Failure to separate the signal from
the background noisethe background noise
�� 2: Incorrect identification of noise as 2: Incorrect identification of noise as signalsignal
�� Both result in lost time (looking for Both result in lost time (looking for source of error, or making up for source of error, or making up for missed error)missed error)
�� How do we minimize both?How do we minimize both?©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 99
Process Behavior ChartsProcess Behavior Charts
�� Developed by Shewhart in 1930Developed by Shewhart in 1930’’ss
�� Used mostly in manufacturingUsed mostly in manufacturing
�� Recently been adopted into Recently been adopted into radiotherapy and diagnostic radiologyradiotherapy and diagnostic radiology
�� Designed to separate the two types of Designed to separate the two types of variation in any processvariation in any process�� Routine variationRoutine variation
�� Exceptional variationExceptional variation
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1010
Process Behavior ChartsProcess Behavior Charts
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1111
Process Behavior ChartsProcess Behavior Charts
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1212
Clinical ExampleClinical Example
�� Daily Flat/Sym OutputDaily Flat/Sym Output
�� Red dash = SPC limitsRed dash = SPC limits
�� Dotted line = 1Dotted line = 1σσ
�� Clinical specification limits Clinical specification limits 3% (Action 1) and 5% 3% (Action 1) and 5% (Action 2)(Action 2)
�� Limits calculated from first Limits calculated from first 10 measurements10 measurements
�� First First ““signalsignal”” at point 9 in at point 9 in process behavior chartprocess behavior chart
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1313
Pawlicki, Whitaker, and Boyer “SPC in Radiotherapy” Med Phys 32(9) (2005)
Clinical ExampleClinical Example
�� Introduction of Introduction of ““systematic errorssystematic errors””clearly indicated in process chart.clearly indicated in process chart.�� Change of output factor for ion Change of output factor for ion
chamberschambers
�� Change of SSD in setupChange of SSD in setup
�� Process Behavior Chart reProcess Behavior Chart re--evaluated evaluated after systematic error was corrected.after systematic error was corrected.
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1414
Clinical ExampleClinical Example
�� Daily Flat/Sym OutputDaily Flat/Sym Output
�� Limits calculated from Limits calculated from subsequent subsequent measurements (9measurements (9--19)19)
�� First First ““signalsignal”” at point at point 2121
�� Incorrect SSD (setup to Incorrect SSD (setup to buildup not surface of buildup not surface of ion chamber)ion chamber)
�� Corrected for next Corrected for next measurementmeasurement
�� Second Second ““signalsignal”” at at point 32point 32
�� Same mistake (on Same mistake (on therapist part)therapist part)
�� Immediately caught and Immediately caught and corrected (measurement corrected (measurement 33)33)
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1515
Clinical ExampleClinical Example
�� At no point were any measurements At no point were any measurements outside 1outside 1σσ
�� At no point were any measurements At no point were any measurements outside clinical specification action outside clinical specification action limits (3%, 5%)limits (3%, 5%)
�� Systematic changes quickly caughtSystematic changes quickly caught
�� Immediate indication of whether or not Immediate indication of whether or not process is process is ““in controlin control””
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1616
Use of Process Behavior ChartsUse of Process Behavior Charts
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1717
Assume process of routine QA displays a predictable behavior
Calculate process behavior limits (averages and ranges)
Compare observed average and range values with process
behavior limits
If routine QA readings are within process behavior limits, then process may be stable.
If routine QA readings are outside process behavior limits, then process is definitely unstable.
“Proof” of stable process is continued routine QA readings within process behavior limits.
Take action to identify and remove systematic errors in the process
Continue charting routine QA readings to confirm removal of systematic errors
Practical GuidelinesPractical Guidelines
�� PBC should never supersede clinical PBC should never supersede clinical requirements.requirements.�� i.e. If process is outside clinical i.e. If process is outside clinical
requirements, action should be taken requirements, action should be taken regardless of whether or not process is regardless of whether or not process is ““in in controlcontrol””
�� Limits can be calculated with little data, Limits can be calculated with little data, and revised as more data becomes and revised as more data becomes availableavailable�� Increases the sensitivity of the chartIncreases the sensitivity of the chart
�� If a deliberate change is made to the If a deliberate change is made to the process, limits should be recalculatedprocess, limits should be recalculated
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1818
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 1919
MultiMulti --Source Data ManagerSource Data Manager
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2020
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2121
Automated Routines For:Automated Routines For:
�� MLC Picket FenceMLC Picket Fence
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2222
Automated Routines For:Automated Routines For:
�� Beam Profile Measurements Beam Profile Measurements
�� Star / Spoke ShotStar / Spoke Shot
�� Stereotactic (WinstonStereotactic (Winston--Lutz)Lutz)
�� Asymmetric BeamAsymmetric Beam
�� RadiationRadiation--Light FieldLight Field
Use your EPID
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2323
Planar kV ImagingPlanar kV Imaging
PTW Normi4 FLU Leeds TOR18 FG
DISC Plus IBA PrimusL
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2424
Planar MV ImagingPlanar MV Imaging
PTW EPID QC QC3
Las Vegas RIT EPID
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2525
CBCTCBCT
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2626
ISOISO--CubeCube
ReferencesReferences
�� Pawlicki T, Whitaker M, Boyer A. Pawlicki T, Whitaker M, Boyer A. ““Statistical Process Statistical Process Control for radiotherapy quality assuranceControl for radiotherapy quality assurance”” Med Phys 32 Med Phys 32 (2005)(2005)
�� Pawlicki T, Whitaker M. Pawlicki T, Whitaker M. ““Variation and control of process Variation and control of process behaviorbehavior”” Int J Rad Onc Biol Phys 71 (2008)Int J Rad Onc Biol Phys 71 (2008)
�� Donald Wheeler Donald Wheeler ““Understanding variation: The key to Understanding variation: The key to managing chaosmanaging chaos”” 22ndnd Edition, SPC Press (2000) Edition, SPC Press (2000)
�� Able C, Bright M. Able C, Bright M. ““Quality control of external beam Quality control of external beam treatment delivery: mechanical parameterstreatment delivery: mechanical parameters”” Med Phys Med Phys 36 (2009)36 (2009)
�� Able C. Able C. ““Statistical Process Control: A new paradigm in Statistical Process Control: A new paradigm in radiotherapy QAradiotherapy QA”” www.ablemedicalphysics.comwww.ablemedicalphysics.com
�� W A Shewhart. W A Shewhart. ““Economic control of quality of Economic control of quality of manufactured product.manufactured product.”” Van Nostrand (1931)Van Nostrand (1931)
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2727
The New (And Improved!) RIT FamilyThe New (And Improved!) RIT Family
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2828
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 2929
AppendixAppendix
©© Radiological Imaging Technology Radiological Imaging Technology -- 20122012 3030
WeWe’’ve got you coveredve got you covered
�� Ongoing helpOngoing help
�� New recommendationsNew recommendations
�� This is a process / transitionThis is a process / transition
�� RegularRegular�� UpdatesUpdates
�� Online training sessionsOnline training sessions
�� ClassesClasses
�� Tech SupportTech Support
�� Come see us at booth Come see us at booth #10087#10087
top related