QIBA CT Volumetrics Group 1B: (Patient Image Datasets) Teleconference Nov 5, 2008.

Slides:



Advertisements
Similar presentations
Vision and Image Analysis Group (VIA) Anthony P. Reeves School of Electrical and Computer Engineering Cornell University © A. P. Reeves 2007.
Advertisements

Evaluation of segmentation. Example Reference standard & segmentation.
We processed six samples in triplicate using 11 different array platforms at one or two laboratories. we obtained measures of array signal variability.
Experimental Design, Response Surface Analysis, and Optimization
G. Alonso, D. Kossmann Systems Group
L.M. McMillin NOAA/NESDIS/ORA Regression Retrieval Overview Larry McMillin Climate Research and Applications Division National Environmental Satellite,
PET/CT Working Group Update Jayashree Kalpathy-Cramer Sandy Napel.
15 February Partial volume correction for liver metastases and lymph nodes1Institute for Medical Image Computing/16SPIE 2010 Partial volume correction.
Model-Based Organ Segmentation: Recent Methods Jiun-Hung Chen General Exam Paper
Statistics: Data Analysis and Presentation Fr Clinic II.
INFO 624 Week 3 Retrieval System Evaluation
Multi-Scale Analysis for Network Traffic Prediction and Anomaly Detection Ling Huang Joint work with Anthony Joseph and Nina Taft January, 2005.
MEASUREMENT (A Quantitative Observation) MEASUREMENTS always have 2 things: Number & Unit All measurements have error in them! A measurement consists of.
Ch. 3.1 – Measurements and Their Uncertainty
Basic principles Geometry and historical development
CT Quality Control for CT Scanners. Quality Control in CT A good idea? Yes Required for accreditation? Sometimes Improves image quality? Sometimes Depends.
QIBA CT Volumetrics - Cross-Platform Study (Group 1C) May 6, 2009 Interclinic Comparison of CT Volumetry Quantitative Imaging Biomarker Alliance.
Are the results valid? Was the validity of the included studies appraised?
Chapter 1: The Study of Life
Estimation of Demand Prof. Ravikesh Srivastava Lecture-8.
Chapter 14: Nonparametric Statistics
Four patients underwent a PET/CT scan with a Philips True Flight Gemini PET/CT scanner. We manually identified a total of 26 central-chest lesions on the.
Update on Lung Cancer Image Processing Rick Avila Karthik Krishnan Luis Ibanez Kitware, Inc. April 19, 2006.
QIBA CT Volumetrics - Cross-Platform Study (Group 1C) December 23, 2009 CT Cross-platform Sizing of Phantom Lesions Quantitative Imaging Biomarker Alliance.
Radiogenomics in glioblastoma multiforme
Slide - 1 Confidential. Slide - 2 Confidential Slide - 3 Confidential Interalgorithm Study using CT Images of synthetic nodules……….
How to Teach Statistics in EBM Rafael Perera. Basic teaching advice Know your audience Know your audience! Create a knowledge gap Give a map of the main.
Success depends upon the ability to measure performance. Rule #1:A process is only as good as the ability to reliably measure.
BUILDING EXTRACTION AND POPULATION MAPPING USING HIGH RESOLUTION IMAGES Serkan Ural, Ejaz Hussain, Jie Shan, Associate Professor Presented at the Indiana.
Designing Semantics-Preserving Cluster Representatives for Scientific Input Conditions Aparna Varde, Elke Rundensteiner, Carolina Ruiz, David Brown, Mohammed.
Computed Tomography Q & A
1 Inter-observer Agreement of The Response To Therapy AssessmentInter-observer Agreement of The Response To Therapy Assessment in Advanced Lung Cancer.
CCNA1 v3 Module 2 W04 – Sault College – Bazlurslide 1 Accuracy vs. Precision.
AUTOMATIZATION OF COMPUTED TOMOGRAPHY PATHOLOGY DETECTION Semyon Medvedik Elena Kozakevich.
S Demehri 1, M.K Kalra 2, M.L Steigner 1, F.J Rybicki 1, M.J. Lang, 3, S.G Silverman 1. 1.Department of Radiology, Brigham & Women's Hospital, Harvard.
QIBA CT Volumetrics - Cross-Platform Study (Group 1C) September 2, 2009 Interclinical Comparison of CT Volumetry Quantitative Imaging Biomarker Alliance.
QIBA CT Volumetrics Group 1B: (Patient Image Datasets) Update April 19, 2011.
QIBA CT Volumetrics Group 1B: (Patient Image Datasets)
Slide - 1 Confidential 3A Group Dr Maria Athelogou Grace Kim: Clarification in the study design: Since we have contours of lesions from many readers (for.
QIBA CT Volumetrics - Cross-Platform Study (Group 1C) March 18, 2009 Interclinic Comparison of CT Volumetry Quantitative Imaging Biomarker Alliance.
N Petrick RSNA QIBA Phantom Group November 13, QIBA Proposed Phase 1A Project V5.0 QIBA Phantom Subgroup.
1 METHODS FOR DETERMINING SIMILARITY OF EXPOSURE-RESPONSE BETWEEN PEDIATRIC AND ADULT POPULATIONS Stella G. Machado, Ph.D. Quantitative Methods and Research.
Simulation Study for Longitudinal Data with Nonignorable Missing Data Rong Liu, PhD Candidate Dr. Ramakrishnan, Advisor Department of Biostatistics Virginia.
Chapter 20 Classification and Estimation Classification – Feature selection Good feature have four characteristics: –Discrimination. Features.
R. Ty Jones Director of Institutional Research Columbia Basin College PNAIRP Annual Conference Portland, Oregon November 7, 2012 R. Ty Jones Director of.
Number (multiply and divide) perform mental calculations, including with mixed operations and large numbers multiply multi-digit numbers up to 4 digits.
Data Analysis, Presentation, and Statistics
The PET/CT Working Group: CT Segmentation Challenge Informatics Issues Multi-site algorithm comparison Task: CT-based lung nodule segmentation.
Brian Lukoff Stanford University October 13, 2006.
1 URBDP 591 A Analysis, Interpretation, and Synthesis -Assumptions of Progressive Synthesis -Principles of Progressive Synthesis -Components and Methods.
Chapter. 3: Retrieval Evaluation 1/2/2016Dr. Almetwally Mostafa 1.
Accuracy vs. Precision. 2 minutes: Write what you think accuracy and precision mean 1 minute: discuss at your table what you think.
MBF1413 | Quantitative Methods Prepared by Dr Khairul Anuar 8: Time Series Analysis & Forecasting – Part 1
Accuracy, Reliability, and Validity of Freesurfer Measurements David H. Salat
Statistical Concepts Basic Principles An Overview of Today’s Class What: Inductive inference on characterizing a population Why : How will doing this allow.
Uncertainty in Measurement How would you measure 9 ml most precisely? What is the volume being measured here? What is the uncertainty measurement? For.
WELCOME TO BIOSTATISTICS! WELCOME TO BIOSTATISTICS! Course content.
Automated GrowCut for Segmentation of Endoscopic Images Farah Deeba, Francis M. Bui, Khan A. Wahid Department Of Electrical And Computer Engineering University.
New Year 6 End of year expectations Number and Place Value Read, write, order and compare numbers up to 10,000,000 and determine the value of each digit.
Sampling procedures for assessing accuracy of record linkage Paul A. Smith, S3RI, University of Southampton Shelley Gammon, Sarah Cummins, Christos Chatzoglou,
Physica Medica 32 (2016) 1570–1574 報告人:王俊淵
Accuracy of RT Structure Set: An Inter-comparison of Four Treatment Planning Systems. Richa Sharma1, Kamlesh Passi2, PS Negi1, Sandhya Sood2, RK Grover1,
Chapter 3: Measurement: Accuracy, Precision, and Error
Model-Based Organ Segmentation: Recent Methods
Economics 20 - Prof. Anderson
Basic principles Geometry and historical development
2.3 Estimating PDFs and PDF Parameters
Learning Probabilistic Graphical Models Overview Learning Problems.
MS-SEG Challenge: Results and Analysis on Testing Data
Presentation transcript:

QIBA CT Volumetrics Group 1B: (Patient Image Datasets) Teleconference Nov 5, 2008

Charge 1. To agree on questions to be answered by these Reference Datasets 2. To identify requirements for those datasets, based on questions to be answered 3. To identify existing datasets that can be leveraged to provide desired datasets

Questions (overview) 1. What level of accuracy and precision can be achieved in measuring tumor volumes in patient datasets? 2. What level of reproducibility in estimating change can be achieved when measuring tumors in phantom datasets? 3. What is the minimum detectable level of change that can be achieved when measuring tumors in patient datasets under a “No Change” condition? 4. What level of reproducibility in estimating change can be achieved in measuring tumors in patient datasets with “Unknown Change” condition? 5. What is the effect of slice thickness on estimating change in tumors using patient datasets?

Questions 1. What level of accuracy and precision can be achieved in measuring tumor volumes in patient datasets? (a) Investigate both bias and variance? (b) Comparison between volume and RECIST here? (c) inter-observer variability (d) Intra-observer variability (NOTE: here observer should be interpreted broadly – could be algorithm, could be human, could be combination). For this question:  lesions with “known size” will be used, where datasets with estimates of boundaries would be used.  Only a single time point is needed  LIDC datasets (thick and thin slices, lots of lesions)  Annotated RIDER datasets? Does this count as truth (currently these have single dimension measurements; could be extended to volumes via additional readers)

Questions 2. What level of reproducibility in estimating change can be achieved in measuring tumors in phantom datasets? (a) RECIST change vs. volume change (b) Investigate just variance (not bias)? (c) inter-observer variability (d) Intra-observer variability (e) Change metric – absolute value? fractional change in volume/diameter? categorical variable? For this question:  Variety of lesions with known size  Compare “size” of different lesions somehow (TBD) Use different existing lesions and treat them as though they were the same lesion at different time points. Use different existing lesions and treat them as though they were the same lesion at different time points. Physically alter lesions over time and scan at both time points (Bob Ford’s water balloon experiment) Physically alter lesions over time and scan at both time points (Bob Ford’s water balloon experiment)

Questions 3. What is the minimum detectable level of change that can be obtained in measuring tumors in patient datasets under a “No Change” condition? (a) RECIST change vs. volume change (b) Investigate just variance? (c) inter-observer variability (d) Intra-observer variability (e) Change metric – absolute value? fractional change in volume/diameter? categorical variable? For this question:  Coffee break experiment – lesions with “No change” condition  Variety of lesions (true size may be unknown)  Patient datasets with same lesions at different time points.  MSK Coffee Break experiment data  Extend their experiment with additional readers (RECIST only? Volumes?)

Questions 4. What level of reproducibility in estimating change can be achieved in measuring tumors in patient datasets with “Unknown Change” condition? (a) RECIST change vs. volume change (b) Investigate just variance? (c) inter-observer variability (d) Intra-observer variability (would be good to have more than 2 time points; may not exist in RIDER). (e) Change metric – absolute value? fractional change in volume/diameter? categorical variable? (f) Look at effects of lesion size for a specific (thin) slice thickness For this question:  Variety of lesions (true size may be unknown)  Patient datasets with same lesions at different time points.  Lesions may or may not have changed size – change is unknown.  RIDER datasets – unannotated as of yet. We have identified about 20 lesions of various sizes. Possible cases from RadPharm.

Questions 5. What is the effect of slice thickness on estimating change in tumors using patient datasets? (a) RECIST change vs. volume change (b) Investigate just variance? (c) inter-observer variability (d) Intra-observer variability (e) Change metric – absolute value? fractional change in volume/diameter? categorical variable? (f) Look at interactions between slice thickness and lesion size (and other lesion characteristics such as shape, margin, etc.) For this question:  Also Variety of lesions (true size may be unknown)  Patient datasets with same lesions at different time points, all done originally with thin slices; create a thick slice series (average together adjacent images).  Lesions may or may not have changed size – change is unknown.  Estimate change on thin and thick series and compare  Thin Slice RIDER datasets that can be fused together. Other cases from RadPharm?

Questions  NOTE: In many of these cases, we anticipate that “ground truth” or actual change in volume/size/density is unknown; we also anticipate that clinical outcome (progression or regression) will also be unknown

Existing Resources  RIDER thick slice annotated Annotated by two readers with single dimension (e.g. RECIST) Annotated by two readers with single dimension (e.g. RECIST) Thick sections (5mm) Thick sections (5mm) 23 cases, 99 series 23 cases, 99 series Used a subset of these cases for NIST Biochange 2008 Used a subset of these cases for NIST Biochange 2008  RIDER thin slice Thin section (1.25 mm) with >= 2 visits Thin section (1.25 mm) with >= 2 visits Looking at mix of lesions now (each patient will have multiple lesions, but may not want to use each one) Looking at mix of lesions now (each patient will have multiple lesions, but may not want to use each one) 7 lesions > 10 mm (from 2 different patients)7 lesions > 10 mm (from 2 different patients) 11 lesions < 10 mm (from 2 different patients)11 lesions < 10 mm (from 2 different patients) No annotations No annotations

Existing Resources  RIDER – MSK Coffee Break Experiment (No Change Condition) 32 NSCLC patients 32 NSCLC patients Imaged twice on the same scanner w/in 15 minutes Imaged twice on the same scanner w/in 15 minutes Thin section (1.25 mm) images Thin section (1.25 mm) images Manual linear measurements performed by 3 readers; volume obtained from algorithm. Manual linear measurements performed by 3 readers; volume obtained from algorithm.

Existing Resources  RadPharm Will search their own database for cases having thin slice and multiple time points Will search their own database for cases having thin slice and multiple time points Then search for good candidate lesions Then search for good candidate lesions

Next Steps  Next Conf Call Continue to investigate Inventory of Patient Datasets (RIDER thin slice, RadPharm, etc.) Continue to investigate Inventory of Patient Datasets (RIDER thin slice, RadPharm, etc.) Continue to Refine Questions and Experimental Design (similar to what 1A has done) Continue to Refine Questions and Experimental Design (similar to what 1A has done)