Abs ac
Re e ences
Acknowledgemen s
Accele a ed EPG-based myoca dial T1 mapping wi h PENGUIN
INTRODUCTION: Model-based Deep Lea ning (DL) allows accele a ing MRI econs uc ion and quan i a i e mapping.
He e, we p opose a DL a chi ec u e ha pe o ms myoca dial T1 mapping di ec ly om accele a ed k-space, which
models he signal wi h he Ex ended Phase G aph (EPG) o mula ion, o allow mo e accu a e quan i ica ion. The DL
a chi ec u e is a modi ica ion o PhasE G aph sigNal and G adien s QUan i a i e In e ence machiNe (PENGUIN)1, wi h
he ollowing no el ies: (a) quan i a i e mapping is pe o med di ec ly om unde sampled k-space; (b) signal in ensi y
cu es a e simula ed no only o a ange o T1 alues, bu also o a ange o egula hea a e (HR) alues.
METHODS: PENGUIN (Figu e 1a) combines a Recu en In e ence Machine2 wi h a dic iona y o EPG simula ed signal
e olu ion cu es ha p o ides he ne wo k wi h a p e-calcula ed signal model. PENGUIN pe o ms 𝐽=2 in e ence s eps
o ob ain 𝐽 es ima es o he T1 maps, conside ing he L1-no m loss unc ion e alua ed in he myoca dium. Ne wo ks
we e ained o accele a ion ac o s acc={4,8}, ADAM op imize , lea ning a e=1e-4, o 300 epochs, 𝐶=64 and 𝐶=256
channels o accele a ion ac o s 4 and 8, espec i ely. Da a we e ob ained om MICCAI’s 2023 CMR econs uc ion
challenge3. T1 mapping was conduc ed ollowing a 4-(1)-3-(1)-2 MOLLI sequence, sho axis (SA) iew only,
FOV=360×307 mm2, spa ial esolu ion=1.4×1.4 mm2, slice hickness=5.0 mm, TR=2.67 ms, TE=1.13 ms, pa ial
Fou ie =7/8, and GRAPPA ac o o 2, on a 3T MAGNETOM Vida Siemens scanne . K-space da a we e unde sampled
e ospec i ely, ollowing a adial k- sampling ajec o y wi h golden angle inc emen s. EPG-based simula ions we e
implemen ed o HR=[28,102] bpm, T1=0:1:2000 ms, T2=50 ms. G ound- u h T1 maps we e ob ained by pe o ming a
pa e n ecogni ion app oach o e he econs uc ed, ully sampled signal in ensi y images, h ough do p oduc
ma ching. Ze o- illed (ZF) and Comp essed Sensing (CS) econs uc ions we e pe o med o compa e wi h PENGUIN.
RESULTS & DISCUSSION: PENGUIN’s T1 maps (Figu e 1b-c) achie ed mean ela i e e o s o 10.1±11.0% and
15.9±15.4%, and mean MSSIM sco es o 0.999±0.001 and 0.998±0.001, o acc=4 and 8, espec i ely. PENGUIN is
less esou ce-consuming han CS econs uc ion, since he compu a ional bu den is mo ed o he p ep ocessing s age.
Ca a ina N Ca alho1,2 *, And eia S Gaspa 2, Ri a G Nunes2, Te esa M Co eia1,3
1Cen e o Ma ine Sciences (CCMAR), Fa o, Po ugal; 2Ins i u e o Sys ems and Robo ics - Lisboa and Depa men o
Bioenginee ing, Ins i u o Supe io Técnico, Uni e sidade de Lisboa, Lisbon, Po ugal; 3School o
Biomedical Enginee ing and Imaging Sciences, King's College London, London, Uni ed Kingdom
*[email p o ec ed]
Figu e 1 – (a) PENGUIN a chi ec u e o in e ence s ep 𝑗. The es ima e 𝑝𝑗 is gi en o he dic iona y o ob ain he signal-in ensi y and
co esponding de i a i e o each T1 alue in 𝑝𝑗. The g adien o he nega i e log-likelihood Δ𝐿𝑗 is calcula ed, conca ena ed o 𝑝𝑗 and
gi en as inpu o he ne wo k, which ou pu s he inc emen al upda e o he es ima ed maps. ℎ𝑗: memo y ec o s. (b) T1 maps o 5
dis inc es subjec s ( ows), acc=4; g ound u h (GT), Exponen ial Fi ing, ze o- illed (ZF), comp essed sensing (CS), and PENGUIN.
(c) T1 ela i e e o s and MSSIM sco es ob ained o all es ing images; * (p- alue<0.05) and *** (p- alue<0.001).
1. Ca alho C, Gaspa A, Nunes R, Co eia T. Di ing in o Ex ended Phase G aph-based Deep Lea ning o accu a e T2 mapping
wi h PENGUIN. In p oceedings o 2023 ISMRM & ISMRT Annual Mee ing & Exhibi ion; 2. Pu zky P, Welling M. Recu en In e ence
Machines o Sol ing In e se P oblems. June 2017. 3. Wang C, Lyu J, Wang S, e al. CMRxRecon: A publicly a ailable k-space
da ase and benchma k o ad ance deep lea ning o ca diac MRI. Sci Da a. 2024;11(1):687.
a)
b)
c)
This wo k was suppo ed by: NVIDIA GPU ha dwa e g an ; “la Caixa” Founda ion and FCT, I.P [LCF/PR/HR22/00533]; FCT
(SFRH/BD/120006/2016, PTDC/EMD/EMD/29686/2017; UIDP/50009/2020), P og ama Ope acional Regional de Lisboa 2020 (LISBOA-
01-0145-FEDER-029686), LARSyS unding (DOI: 10.54499/LA/P/0083/2020,10.54499/UIDP/50009/2020,10.54499/UIDB/50009/2020.
This esea ch was suppo ed by FCT h ough p ojec s UIDB/04326/2020 (DOI:10.54499/UIDB/04326/2020), UIDP/04326/2020
(DOI:10.54499/UIDP/04326/2020) and LA/P/0101/2020 (DOI:10.54499/LA/P/0101/2020).
