Accuracy of subject-specific prediction of end-systolic time in MRI across a range of RR intervals

Background Prediction of End-Systole time is of utmost importance for cardiac MRI to correctly associate acquired k-space lines during reconstruction of cine acquisitions. This prediction is usually based on the patient’s heart rate using Weissler’s formula, which was calibrated by linear regression within a population and cannot account for individual variability. Objective We propose an automatic method to build a personalized model that better predicts end-systole. Methods A phase contrast sequence was modified to acquire only central k-space line with 6.6ms temporal resolution, in a slice passing through the aorta during 128 heartbeats in 35 subjects. Segmentation of aorta and detection of end of systolic ejection was automatic. Duration of electromechanical systole duration as function of heart rate was determined for each subject separately. Results In comparison with the global models, the adapted cardiac model predicted significantly better both echocardiographic end-systolic reference (bias = 0ms vs 17ms, p<0.001) and MRI measurements (bias = 6.8ms vs 17ms). Favorable impact was shown on the cine reconstruction of the 5 subjects with the higher cardiac variability (p = 0.02). Conclusions Personalization of cardiac model to the subject is feasible in MRI and reduces the error of prediction of systole.


Context:
The STANISLAS cohort is a monocentric familial longitudinal cohort composed of 1006 famillies (4295 subjects) living in/near Nancy. They were included in 1993-1995 by the Centre de Médecine Préventive (CMP) of Vandoeuvre lès Nancy and underwent periodical (every 5 years) systematic health exam under the direction of the Caisse Nationale d'Assurance Maladie (CNAM). The fourth visit of this cohort will begin in 2011 in order to collect data 17 years after the inclusion and to allow all the potential users of the cohort to continue their work.
The ESCIF ancillary study, proposed by Dr Laurent Bonnemains (CIC-IT, Nancy), concerns 76 volunteers of the 4th visit of the Stanislas Cohort. Its aim is to compare cardiac functional indices acquired with MRI and with trans-thoracic echocardiography.
The IADI laboratory has developed several technics aiming at acquiring MRI images of moving organs either at high temporal resolution or with motion correction such as with the post processing technic named GRICS Generalized Reconstruction by Inversion of Coupled System. The use of GRICS or other technics allow us to compute from MRI acquisitions various cardiac functional indices which were up to now only used in echocardiography. At this moment, the feasibility of these technics has been shown on 5 healthy volunteers. Secondary Objective of the ancillary study  To assess the interobserver reproducibility of these indices.

Methodology, Organization:
For the ancillary study ESCIF, a specific MRI cardiac examination will be proposed to 76 participants of the Stanislas cohort.

RESEARCH
Magnetic Resonnance Imaging (MRI) is a non-invasive technic using no radiations. It is based on a high static magnetic field coupled with radio-frecquency waves. It can produce anatomical and functionnal images. It can be considered nowadays as a reference technic for these purposes. Since the beginning of MRI around 1985, it has evolved a lot. A lot of parameters can be tuned to optimize the diagnostic capacity of the exam (type and parameters of the sequence, antenna, strength of the field, reconstruction method, …) One of these major improvement in the field of post-treatment was brought by the IADI lab which proposed recently the GRICS technic (Generalized Reconstruction by Inversion of Coupled System).This technic allows the reconstruction of moving objects with motion correction based on external data generally coming from external captors such as belts or accelerometers or based on internal data coming from the MRI acquisition itself (such as with navigators).
The motion detection and correction is therefore retrospective and works on data acquired freely without any constraint on the type of motion.
High temporal resolution phase contrast acquisitions could allow the measurement of indices yet only accessible with tissue-doppler imaging in echocardiographysuch as the Myocardial Performance Index (MPI). These functional indices (MPI and others) are at the moment used in clinical practice in echography but not in MRI. Their use in MRI could constitute an important improvement of the diagnostic quality of our exams.

PRINCIPAL OBJECTIVE
To compare cardiac functional indices measured with MRI using technics derived from GRICS with equivalent indices measured with echocardiography.

SECONDARY OBJECTIVE
To assess the interobserver reproducibility of these indices.

PRINCIPAL EVALUATION CRITERION
Paired comparison (MRI vs echocardiography) of cardiac functional indices for each subject: for example, velocity of tricuspid or mitral annulus, MPI and myocardial strains for both ventricles. The list of MRI and echographic parameters to be compared is indicative and will be completed when knowledge will evolved and when the exams will be processed.

SECONDARY EVALUATION CRITERION
Bland and Altman comparisons of two assessments of each parameter.

INCLUSIONS
The volunteers accepting to participate to the fourth visit (at 17 years) of the STANISLAS cohort will be proposed to participate into the ESCIF ancillary study, whenever the MRI scanners will be available the day of their visit. Subjects will be informed during a previous phone call, planned by the main Stanislas study.

The day of the 4 th STANISLAS visit (if the MRI scanners are available)
-The investigator will inform the volunteer of the objectives and methods of the ancillary study.
-The information letter and consent form will be presented and given to the STANISLAS volunteer, with the MRI non contra-indication questionnary.

At the end of the echocardiography (comprised into the main STANISLAS study)
-The volunteer group will be chosen (group 1 if the echography is normal, group 2 if at least one index is not normal). The volunteer will give his consent to participate into the ancillary study (if the inclusions for this group are not closed). The consent form will be signed by the volunteer and by the investigator. -MRI non contra-indication questionary will be validated after discussion with the investigator. -Concerning women able to become pregnant, in case there is a possibility of incoming pregnancy, a urinary pregnancy test will be performed. The investigator will check this test and the woman will be excluded of the study if the result is positive.

THE MRI CARDIAC EXAM
The exam will be performed using one of the General Electric MRI scanners of the CHU Nancy Brabois. These scanners are used every day for clinical practice and/or for research protocols. They have been homologated and more than 20000 of them are used worldwide.
As for a conventional cardiac MRI exam, the subject will be prepared and installed by the MRI radiographer. The subject will then be put inside the bore, once the different captors will have proven that they work (two pneumatic belts and one ECG captor).
The MRI protocol will be composed of differents sequences, based either on phase contrast cine MRI, or SSFP cine MRI. The duration of the examination will be around 45 minutes.
However, the volunteer will be allowed to leave the bore whenever he will request. Indeed, a communication system will allow a periodic conversation between the volunteer and the radiographer. The functional cardiac indices measured in MRI with technics based on the GRICS method will be compared to their equivalent assessed during the echocardiographic exam a few minutes before.
A second assessment of these indices will be performed later to allow the assessment of inter-observer reproducibility of both methods.

CALENDAR OF THE MAIN STUDY
Beginning of the main study: 05/07/2011 Theoretical delay necessary to include everybody : 48 month

Participation of each volunteer : 1 day
Theoretical end of the study : 05/07/2015

DATA COLLECTED
The main data collected will be: -Sex, weight, height -Inclusion criteria and non inclusion criteria -MRI and echocardiographic data The echocardiographic and MRI raw data an images as well as the captors data will be stored in a ARCHIMED database declared to the french governement agency (CNIL).

POPULATION DE L'ETUDE ANCILLAIRE
This study will comprise 76 subjects (cf population size justification) consecutively selected from the main STANISLAS cohort. This subjects will be included into two groups of 38 subjects. Group one subjects will have a complete echocardiographic exam considered normal (ejection fraction, maximal S wave at the tricuspid and mitral annulus, ratio E/Ea, MPI, longitudinal strain for both ventricles), whereas group 2 subjects will have at least one abnormally in this exam. This design was chose to obtain widely spread values for the tested indices to maximize the sensitivity of the concordance study.
Subjects will be enrolled in the study consecutively according to the availability of the MRI scanners, until the size of each group will be reached.
The thresholds chosen to discriminate group 1 from group 2 will be chosen by consensus at the beginning of the study (the state of the art may change rapidly in this field of knowledge).

INCLUSION CRITERIA
-Contentment for the main study: Stanislas 4th visit.
-Contentment for the ancillary study.

EXCLUSION CRITERIA
-when volonteers have withdrawn their consent -when the MRI exam was not complete, -if a non-inclusion criterion appears

CONSERVATION ACCESS TO CLINICAL AND PARACLINICAL DATA
Every data concerning volunteers of the study will be anonymized by the CIC-P.
It will be possible after anonymization to share these data in accordance with existing regulations concerning national or international research teams.

SECURITY EVALUATION
MRI is an imaging technic presenting no risk for human being when security rules (especially those concerning metallic items) are respected. The ESCIF study will be performed on two MRI scanners (1.5 and 3T) within the CHU of Nancy-Brabois. These MRI scanners are homologated and used world widely for clinical or research purpose (more than 20000 scanners in the world). These machines are regularly updated with a support maintenance contract. All the limits imposed by norms will be respected and automatically controlled during the acquisitions. There will be no invasive act (no gadolinium injection). In case of unexpected discovery of any anatomical anomaly, a medical consultation will be proposed to the volunteer with a specialist physician of the CHU or with the usual GP.
The expected effects of the MRI can comprise: headache, heat sensation… The adverse events will be noted by the investigators during the duration of the examination. They will be noted on the CRF. Documents written by national and international societies of Radiology concerning MRI will be used to define the expected or unexpected character of an adverse event.

STATISTIC METHODS
Data will be described using mean ± standard-deviation, median [25% percentile -75% percentile], maximum and minimum. The main objective analysis will be performed with a Bland-Altman comparison of each parameter assessed both in MRI and in echocardiography. The reproducibility of each method will also be assessed with Bland-Altman method and will be completed by the computation of ICC (inter technic and inter observer) with a mixed ANOVA using the volunteer as random factor.

SIZE OF THE POPULATION
With 60 volunteers, the mean of the differences could be assessed with a precision (half 95% interval) < 0.26 SD, and the limits of its 95% confidence interval with a precision < 0.10 SD.
The total number of subject to include was augmented to 76 to take into account the possible missing data due to unavoidable events during MRI acquisition or discovered during echocardiographic post treatment.
Two groups of 38 subjects with/without echocardiographic anomaly will be recruited to maximize the opportunity to obtain widely spread parameters.