Data on prognostication models comparison for neurological recovery after cardiac arrest using proton chemical shift imaging (1H-CSI)

We report in this data article the statistical comparison of three models for neurological prognostication 6 months after cardiac arrest: M1 associated SAPS II and coma Glasgow score at MRI, M2 associated SAPS II, coma Glasgow score, and FLAIR-DWI “deep gray nuclei”score, M3 associated SAPS II, coma Glasgow score, FLAIR-DWI “deep gray nuclei”score, and Lenticular cores NAA/Cr ratio. These data are related to “Value of assessment of multivoxel proton chemical shift imaging to predict long term outcome in patients after out-of-hospital cardiac arrest: A preliminary prospective observational study” (Quintard et al., 2018) [1].


Specifications
The significance of the data is potentially helpful for predicting recovery after cardiac arrest.

Data
Twenty-nine patients were included in the analysis [1]. The neurologic follow-up, evaluated as Cerebral Performance category (CPC, Additional File 1) was specified in the protocol to be at 180 7 14 days, but the time to follow-up was in some cases several weeks longer for logistic reasons. Twenty-one patients had a bad outcome (CPC 3-5) at 6 months and 9 a favorable one (CPC 1-2) ( Table 1). Data report multivariate analysis (Table 2) of 3 prognosis models: M1 associated SAPS II and Glasgow coma score at MRI, M2 associated SAPS II, coma Glasgow score, and FLAIR-DWI "deep gray nuclei" score, M3 associated SAPS II, coma Glasgow score, FLAIR-DWI "deep gray nuclei" score, and Lenticular cores NAA/Cr ratio. Receiver-Operating-Characteristic curves for the three multivariate logistic regression models were realized (Fig. 1). Structural MRI acquisition took about 20 min, depending on the number of slices required to cover the brain. Multivoxel 1H-CSI was acquired using the point-resolved proton spectroscopy sequence (PRESS) with the following parameters: slice thickness, 20 mm; TR/TE, 1000/144 ms; matrix, 16 Â 16; FOV, 240 mm; number of excitations, 1. The axial 1H-MRS covering a region from the head of striatum to the thalamus. This acquisition took 11 min. The MR images were retrieved in DICOM (Digital Imaging and Communications in Medicine) format from picture archiving and communication system servers at the hospitals (Table 3).

Image processing
Brain MR images were reviewed by a board-certified blinded radiologist (ME.A, with 10 years of postfellowship experience) to evaluate for structural abnormalities. Participants with any evidence of territorial stroke of any age, intracranial hemorrhage, or intracranial mass lesions were excluded; in addition, any subject with head motion greater than 2 mm across the imaging session was excluded. MR imaging morphologic analysis was accomplished using a validated score on the basis of a visual rating. In the FLAIR-DWI scoring (Additional file 2) proposed by Hirsch et al. [2] developed in a similar population with CA, the anatomic location and degree of signal abnormality on FLAIR images and DWI sequences is rated from 0 (normal) to 4 (severe), and composite tissue signal change is then calculated in cortical, basal ganglia and thalamus, brainstem, and cerebellar structures. The sum of scores in cortical, basal ganglia, and all structures is referred to as the "cortex", "deep gray nuclei", and Table 1 Outcomes in the cohort.

All patients
Favorable outcome at follow-up* (CPC 1-2) Unfavorable outcome at follow-up* (CPC 3-5) The neurologic follow-up was specified in the protocol to be at 180 7 14 days, but the time to follow-up was in some cases several weeks longer for logistic reasons. † Cause of death missing in four cases in the derivation cohort and in one case in the validation cohort. ICU denotes intensive care unit.  "overall score", respectively. The sum or all scores recorded by using FLAIR and DWI is known as the DWI-FLAIR score, respectively [2]. Multivoxel 1H-CSI data processing was performed by an expert radiologist allowing to measure of the area under the curve by standard manufacturer software dedicated to MR spectroscopy postprocessing (Advantage Windows for General Electric). The volume of interest was placed carefully to avoid contact with cerebral spinal fluid on non-angled FLAIR images after coregistration of spectroscopic data and FLAIR volume. The spectra were analyzed for determining the concentration of metabolites in the lenticular cores (right and left side) and thalami (right and left side): NAA (at 2.0 ppm), Cho (at 3.2 ppm) and Cr (at 3.0 ppm). The quality of the selected spectra was visually inspected and was considered acceptable only if Cho and Cr signals were clearly separated. Spectra was also rejected if MR spectroscopists detected artefacts such as large baseline distortions, exceptionally broadened metabolite peaks, insufficient removal of the water line, large phase errors, and signals originating from outside the voxel as conventionally used in clinical practice. In all cases, peak surface area was used to calculate metabolite ratios for metabolite quantification. The worst value of bilateral lenticular cores and thalami were analyzed except in cases of some non-interpretable voxels. In this case, the sole side with spectra of good quality was considered.

Calculation of odds ratios
Because there were categorical and continuous variables with different units, odds ratios (with 95%CI) were computed by taking the exponent of the absolute value of the estimated parameters (and 95% CI), the latter being multiplied by a factor that accounts for the unit used (i.e., 0.1 in the case of 1H-MRS parameters). We tested the null hypothesis of an estimated parameter being equal to zero with the use of the likelihood ratio test (LR test) with one degree of freedom. This corresponds to the null hypothesis of an odds ratio being equal to 1, i.e., no predictive value. Hence, a variable was considered to be predictive if P o 0.05. DWI, diffusion-weighted imaging; FLAIR, fluid-attenuated inversion recovery; NAA/Cr, N-acetyl aspartate over creatine ratios and SAPS II, Simplified Acute Physiology Score II. † Sensitivity significantly different than the one of the NAA/Cr Lenticular cores (P o 0.05).

Transparency document. Supporting information
Transparency data associated with this article can be found in the online version at https://doi.org/ 10.1016/j.dib.2018.10.066.