Non-parametric and non-linear DSC-MRI p ost-processinng ... · Francisco, San Fr ical Surgery, UC...
1
No 1 Grad Fr Metho echo-p post-co native white m specim MRI fi used to Result predict linear: comple p<.01). presenc these b patient method Conclu gadolin CE, bu Abnorm study p vascula Refere Ackno Figure specim nCBV microv Figure vascul correct on-parametric Emma Esso duate Group in Bio rancisco, San Fran Francisco, S ods: 72 image-guid lanar imaging (flip ontrast T1-weighte resolution. An ave matter (NAWM) men was graded on indings. A random o determine if the b s: Predicting Vasc ted the tissue vasc nPH, and non-par ex vasculature. Ev . Identifying Abno ce of abnormal va blood volume mea with abnormal va d. usions: The blood nium significantly ut non-linear and n mal (simple or com provides histopath ature features with ences: [1] Paulson owledgements: NI e 2. Factor VIII I mens with delicate V, non-linear: nPH, vascular hyperplas e 1. Hemodynamic lature calculated u tion (left) and non and non-linea ck-Burns 1,2 , Joann oengineering, UC ncisco, CA, United San Francisco, CA ded tissue samples p angle=35°, TE/T ed SPGR, T2-weig erage hemodynam for normalization, n an ordinal scale m-effects ordinal re blood volume mea cular Morphology cular morphology rametric: nPH are ven adjusted for C ormal Vasculature asculature, while C sures was associat asculature in both d volume measure predicted the und non-parametric an mplex) vasculatur hological support hin the heterogeneo et al. (2008) Rad H Grant R01 CA1 IHC staining (top) e, simple, and co , and non-parametr sia. c curve from a bio sing non-linear ga n-parametric analys ar DSC-MRI p na J. Phillips 3,4 , An Berkeley/UC San d States, 3 Neurolog A, United States, 5 B s were collected fr TR=54–56/1250-1 ghted FLAIR). A 5 mic curve was calcu , using both the n (delicate, simple, egression model, ad asure(s) from each (delicate, simple, (non-linear: nCBV predictive of a gr CE, these same CB e (simple or comp CE was not (non- ted with approxim the CE and non-C es from both the erlying vascular m nalysis may assist re also existed bey of the non-param ous GBM, which m 249:2 [2] Weiskof 127612-01A1 and ) and correspondin omplex vasculatur ric: nPH are signif psy with complex mma-variate fit w sis (right). post-processin with trea nnette M. Molinaro Francisco, San Fr gical Surgery, UC Bioengineering an Intr vasc sam is t hype susc yet glom limi volu for a its u has strat vasc plann proc micr rom 35 patients ne 500ms, slice thick 5-mm diameter sp ulated from within non-linear and non complex) using F djusted for contras post-processing m , or complex): Adj V p<.05, nPH p<.0 reater degree of h BV estimates were lex): Non-linear: linear: nCBV p=. mately a 2.2-fold gr CE tissue, which is non-linear and n morphology determ in guiding sampl yond the CE-lesion metric and non-li may assist in guidi ff (1994) ISMRM NIH Grant P01 CA ng ΔR2* curve (b re (a-c). Increased ficant risk factors f ith leakage ng methods pr atment-naive G o 3 , Janine M. Lupo rancisco, CA, Uni San Francisco, Sa d Therapeutic Scie roduction: Tissue cularized human mples for tumor gra the presence of erplasia, tortuous ceptibility contrast there are known meruloid vasculat tation and have be ume (CBV) [1]. No addressing this lim use in the clinical p not been extensiv tegy-dependent C cular features for d ning. The goal o essing methods to rovasculature in pa ewly diagnosed w kness=3-4 mm, ma herical region at t n each specimen r n-parametric post- actor VIII immun st-enhancement (C method significantl justed for CE, non 03; non-parametri hyperplasia. Identif e marginally signi nCBV, non-linear 03, nPH p<.03; n reater likelihood o s correctly identifi on-parametric ana mined by Factor VI ling toward comp n and was accurat inear post-process ing biopsy location p 279, [3] Essock- A11816-01A2. bottom) of 3 d non-linear: for increased redict underly GBM o 2 , Soonmee Cha 2,3 ited States, 2 Depar an Francisco, CA, ences, UC San Fra e heterogeneity brain tumor, has ading that accurate complex glomer s lamina, and b t (DSC) MRI has limitations near r ture. Multiple een shown to grea on-linear gamma-v mitation in the rese practice. Non-para vely validated with CBV estimates w diagnosis and char of this study was o determine which atients with treatm with GBM. Preoper atrix=128x128, 0. the specimen targe region and an aver -processing metho nohistochemical (I CE) at the target lo ly predicted vascu n-linear: nCBV, n ic: nPH p<.05). Fi fying Complex V ificant predictors r nPH, and non-pa non-parametric: nP of presence of abn ied with the CBV alysis of DSC da III IHC analysis. C lex vasculature w tely identified by sing techniques a n and identifying p -Burns et al. (2011 ying vascular h 3 , Susan M. Chang rtment of Radiolog United States, 4 D ancisco, San Fran of glioblastoma s historically bee ely represent the tu ruloid vasculatur breakdown of th s been used to non regions of contras post-processing m atly influence the variate fit with lea earch setting, yet n ametric analysis is h histopathology. will help guide b racterize vascular to compare non-li h CBV estimate m ment naive GBM. rative 3T MR exa 1mmol/kg Gd-DT et location was ov rage curve was ca od [2-3] (see Figu HC) staining by a ocation and repeate ular morphology de non-linear: nPH, an igure 2 illustrates asculature: As exp of complex vascu arametric: nPH we PH p<.02; CE: p= normal vasculature measures from bo ata acquired with Complex vasculatu when it is present both the non-line s noninvasive me patients for targete 1) Neuro-oncol 13 histopatholog g 3 , and Sarah J. Ne gy and Biomedical Department of Path ncisco, CA, United (GBM), a high n a challenge fo umor biology. One re, characterized he blood brain ninvasively assess st agent extravasat methods exist fo resultant estimate akage correction is necessitates curve- s less computation Histopathological biopsy selection burden for anti-an inear and non-par ost accurately refl ams included T2* TPA) and anatomic erlaid on the align alculated within th ure 1). Vascular m an experienced pat ed specimen samp etermined by IHC nd non-parametric how greater non-l pected, CE was a ulature (CBV mea ere each significan =>.2). In general, a e. Figure 3 illustrat oth the non-linear a a 35° flip angle ure is far more like within the heterog ear and non-param ethods for charac ed anti-angiogenic 3:1 gy in patients elson 1,5 l Imaging, UC San hology, UC San d States hly malignant an or selecting biops e hallmark of GBM by microvascula barrier. Dynami s microvasculatur tion, common nea or addressing th es of cerebral bloo s a common metho -fitting which limi nally expensive, ye validation of thes toward aggressiv ngiogenic treatmen rametric DSC pos lects the underlyin DSC gradient-ech c imaging (pre- an ned DSC data at th e normal appearin morphology of eac thologist blinded t ples per patient, wa analysis. c: nPH significantl linear: nCBV, non strong predictor o asures: p<=.09, CE nt predictors of th a 1-unit increase i tes an example of and non-parametri and single-dose o ely to coincide wit geneous CE-lesion metric analysis. Th cterizing aggressiv c therapies. Figure 3. a. T1w-CE MRI of a patient with 2 biopsy targets (pink), from CE (left) and non-CE (right) tissue. b. Factor VIII IHC staining shows abnormal vasculature in both specimens, which is identified by elevated non- linear and non- parametric CBV estimates in c. nd sy M ar ic e, ar is od od ts et se ve nt t- ng ho nd he ng ch to as ly n- of E: he in a ic of th n. is ve - s 193 Proc. Intl. Soc. Mag. Reson. Med. 20 (2012)
Transcript of Non-parametric and non-linear DSC-MRI p ost-processinng ... · Francisco, San Fr ical Surgery, UC...
No
1GradFr
Methoecho-ppost-conative white mspecimMRI fiused to
Resultpredictlinear: complep<.01).presencthese bpatient method
ConclugadolinCE, buAbnormstudy pvascula
RefereAckno
FigurespecimnCBVmicrov
Figurevasculcorrect
on-parametric
Emma Essoduate Group in Biorancisco, San Fran
Francisco, S
ods: 72 image-guidlanar imaging (flip
ontrast T1-weighteresolution. An avematter (NAWM)
men was graded onindings. A randomo determine if the b
s: Predicting Vascted the tissue vascnPH, and non-par
ex vasculature. Ev. Identifying Abnoce of abnormal va
blood volume meat with abnormal vad. usions: The bloodnium significantly ut non-linear and nmal (simple or comprovides histopathature features withences: [1] Paulson owledgements: NI
e 2. Factor VIII Imens with delicateV, non-linear: nPH,vascular hyperplas
e 1. Hemodynamiclature calculated ution (left) and non
and non-linea
ck-Burns1,2, Joannoengineering, UC ncisco, CA, UnitedSan Francisco, CA
ded tissue samplesp angle=35°, TE/Ted SPGR, T2-weigerage hemodynamfor normalization,
n an ordinal scale m-effects ordinal reblood volume mea
cular Morphologycular morphology rametric: nPH areven adjusted for Cormal Vasculatureasculature, while Csures was associatasculature in both
d volume measurepredicted the undnon-parametric anmplex) vasculaturhological support
hin the heterogeneoet al. (2008) Rad H Grant R01 CA1
IHC staining (top)e, simple, and co, and non-parametrsia.
c curve from a biosing non-linear ga
n-parametric analys
ar DSC-MRI p
na J. Phillips3,4, AnBerkeley/UC San
d States, 3NeurologA, United States, 5B
s were collected frTR=54–56/1250-1ghted FLAIR). A 5
mic curve was calcu, using both the n(delicate, simple,
egression model, adasure(s) from each
(delicate, simple,(non-linear: nCBV predictive of a gr
CE, these same CBe (simple or compCE was not (non-ted with approximthe CE and non-C
es from both the erlying vascular m
nalysis may assist re also existed bey
of the non-paramous GBM, which m249:2 [2] Weiskof
127612-01A1 and
) and correspondinomplex vasculaturric: nPH are signif
psy with complex mma-variate fit wsis (right).
post-processinwith trea
nnette M. MolinaroFrancisco, San Fr
gical Surgery, UC Bioengineering an
Intrvascsamis thypesuscyet glomlimivolufor aits uhas stratvascplannprocmicr
rom 35 patients ne500ms, slice thick5-mm diameter spulated from within
non-linear and noncomplex) using Fdjusted for contraspost-processing m
, or complex): AdjV p<.05, nPH p<.0reater degree of hBV estimates werelex): Non-linear: linear: nCBV p=.
mately a 2.2-fold grCE tissue, which is
non-linear and nmorphology determ
in guiding samplyond the CE-lesionmetric and non-limay assist in guidiff (1994) ISMRM NIH Grant P01 CA
ng ΔR2* curve (bre (a-c). Increasedficant risk factors f
ith leakage
ng methods pratment-naive Go3, Janine M. Luporancisco, CA, UniSan Francisco, Sad Therapeutic Scie
roduction: Tissuecularized human
mples for tumor grathe presence of erplasia, tortuousceptibility contrastthere are known
meruloid vasculattation and have be
ume (CBV) [1]. Noaddressing this limuse in the clinical pnot been extensivtegy-dependent Ccular features for dning. The goal oessing methods to
rovasculature in paewly diagnosed wkness=3-4 mm, maherical region at tn each specimen rn-parametric post-actor VIII immunst-enhancement (C
method significantl
justed for CE, non03; non-parametri
hyperplasia. Identife marginally signinCBV, non-linear03, nPH p<.03; nreater likelihood os correctly identifi
on-parametric anamined by Factor VIling toward compn and was accuratinear post-processing biopsy locationp 279, [3] Essock-A11816-01A2.
bottom) of 3 d non-linear: for increased
redict underlyGBM o2, Soonmee Cha2,3
ited States, 2Deparan Francisco, CA,ences, UC San Fra
e heterogeneity brain tumor, has
ading that accuratecomplex glomer
s lamina, and bt (DSC) MRI haslimitations near rture. Multiple een shown to greaon-linear gamma-v
mitation in the resepractice. Non-para
vely validated withCBV estimates wdiagnosis and charof this study was o determine whichatients with treatm
with GBM. Preoperatrix=128x128, 0.the specimen targeregion and an aver-processing methonohistochemical (ICE) at the target loly predicted vascu
n-linear: nCBV, nic: nPH p<.05). Fifying Complex Vificant predictors r nPH, and non-pa
non-parametric: nPof presence of abnied with the CBV
alysis of DSC daIII IHC analysis. Clex vasculature wtely identified by sing techniques an and identifying p-Burns et al. (2011
ying vascular h
3, Susan M. Changrtment of Radiolog United States, 4Dancisco, San Fran
of glioblastoma
s historically beeely represent the turuloid vasculaturbreakdown of ths been used to nonregions of contraspost-processing matly influence the variate fit with leaearch setting, yet nametric analysis ish histopathology.
will help guide bracterize vascular to compare non-li
h CBV estimate mment naive GBM.
rative 3T MR exa1mmol/kg Gd-DTet location was ovrage curve was caod [2-3] (see FiguHC) staining by a
ocation and repeateular morphology de
non-linear: nPH, anigure 2 illustrates asculature: As expof complex vascuarametric: nPH wePH p<.02; CE: p=normal vasculature
measures from bo
ata acquired with Complex vasculatu
when it is present both the non-lines noninvasive mepatients for targete1) Neuro-oncol 13
histopatholog
g3, and Sarah J. Negy and Biomedical
Department of Pathncisco, CA, United
(GBM), a highn a challenge foumor biology. Onere, characterized he blood brain ninvasively assess
st agent extravasatmethods exist foresultant estimate
akage correction isnecessitates curve-s less computationHistopathological biopsy selection burden for anti-aninear and non-parost accurately refl
ams included T2* TPA) and anatomicerlaid on the align
alculated within thure 1). Vascular man experienced pated specimen sampetermined by IHC
nd non-parametrichow greater non-lpected, CE was a ulature (CBV meaere each significan
=>.2). In general, ae. Figure 3 illustratoth the non-linear a
a 35° flip angle ure is far more likewithin the heterog
ear and non-paramethods for characed anti-angiogenic3:1
gy in patients
elson1,5 l Imaging, UC Sanhology, UC San d States
hly malignant anor selecting biopse hallmark of GBM
by microvasculabarrier. Dynami
s microvasculaturtion, common neaor addressing thes of cerebral bloos a common metho-fitting which liminally expensive, ye
validation of thestoward aggressiv
ngiogenic treatmenrametric DSC poslects the underlyin
DSC gradient-echc imaging (pre- anned DSC data at the normal appearin
morphology of eacthologist blinded t
ples per patient, waanalysis.
c: nPH significantllinear: nCBV, nonstrong predictor o
asures: p<=.09, CEnt predictors of tha 1-unit increase ites an example of and non-parametri
and single-dose oely to coincide witgeneous CE-lesion
metric analysis. Thcterizing aggressivc therapies.
Figure 3. a. T1w-CE MRI of a
patient with 2 biopsy targets (pink), from CE (left) and
non-CE (right)tissue.
b. Factor VIIIIHC staining
shows abnormal
vasculature in both
specimens, which is
identified by elevated non-
linear and non-parametric
CBV estimatesin c.
nd sy M ar ic e, ar is
od od ts et se ve nt t-
ng
ho nd he ng ch to as
ly n-of E: he in a
ic
of th n. is
ve
-
s
193Proc. Intl. Soc. Mag. Reson. Med. 20 (2012)
![Semi-Flyweights - Linux Magazine · vious candidates are Open / LibreOffice Base [3][4], H2 [5], and DBDesigner 4 [6]. ... MAY 2012 ISSUE 138 LINUX-MAGAZINE.COM ... Kexi vs. Glom](https://static.fdocuments.us/doc/165x107/6040426ec171416b486f3e5e/semi-flyweights-linux-vious-candidates-are-open-libreoffice-base-34-h2.jpg)
