Cerebrospinal fluid T1 value phantom reproduction at scan room temperature

Abstract The T1 value of pure water, which is often used as a phantom to simulate cerebrospinal fluid, is significantly different from that of in‐vivo cerebrospinal fluid. The purpose of this study was to develop a phantom with a T1 value equivalent to that of in‐vivo cerebrospinal fluid under examination room temperature (23°C–25°C). In this study, 1.5 and 3.0 T magnetic resonance imaging scanners were used. We examined the signal intensity change in relation to pure water temperature, the T1 values of acetone‐diluted solutions (0–100 v/v%, in 10 steps), and the correlation coefficients obtained from volunteers and the prepared phantoms. The T1 value was close to the value reported in the literature for cerebrospinal fluid when the acetone‐diluted solution was 70 v/v% or higher at scan room temperature. The value at that time was 3532.81–4704.57 ms at 1.5 T and it ranged from 4052.41 to 5701.61 ms at 3.0 T. The highest correlation with the values obtained from the volunteers was r = 0.993 with pure acetone at 1.5 T and r = 0.991 with acetone 90 v/v% at 3.0 T. The relative error of the best phantom‐volunteer match was 32.61 (%) ± 6.71 at 1.5 T and 46.67 (%) ± 4.31 at 3.0 T. The T1 value measured by the null point method did not detect a significant difference between in vivo CSF and acetone 100 v/v% at 1.5 T and acetone 90 v/v% at 3.0 T. The T1 value of cerebrospinal fluid in the living body at scan room temperature was reproduced with acetone. The optimum concentration of acetone for cerebrospinal‐fluid reproduction was pure acetone at 1.5 T and 90 v/v% at 3.0 T.

report has reproduced very long T1 and T2 values, such as those found in cerebrospinal fluid (CSF).
Images in which the CSF is suppressed with normal MRI examination are clinically commonly used. T2-fluid-attenuated inversion recovery (FLAIR), T1-FLAIR, and white matter attenuated inversion recovery (WAIR) using double inversion recovery (DIR) are clinically significant representative sequences. [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22] The most important factor affecting contrast in these inversion recovery sequences is the T1 value. The substance most commonly used as a phantom to reproduce CSF is pure water (PW). PW is treated water with a reverse osmosis membrane to remove most of the organic matter and ion components. Moreover, it is known that suppression of CSF is insufficient in the FLAIR image in postmortem MRI examinations. [23][24][25][26][27] This is probably because the T1 value changes due to the fact that the postmortem temperature is lower than the biological temperature. Tsukiashi et al. measured the T1 value against the water temperature and reported the change in the temperature and in the T1 value. 28 From this, it is certain that the T1 value largely fluctuates due to changes in temperature. In addition, there is a possibility that the T1 value of PW under examination room temperature (RT) deviates markedly from that of in-vivo CSF. Therefore, it is highly likely that CSF simulated with PW at RT cannot approximate the T1 value of the CSF in the living body. The difference in the T1 value between in-vivo CSF and the PW phantom causes a large error when the imaging condition is changed, and therefore PW is unsuitable as a phantom to simulate CSF. Therefore, we attempted to develop a phantom able to reproduce the T1 value of CSF at RT.

2.A | Materials
All images were acquired using a 3.0 T MRI scanner (Ingenia; Philips Healthcare, Amsterdam, Netherlands) and 1.5 T MRI scanner (Achieva d-stream; Philips Healthcare). The solutions used were acetone (Matsuba, Ltd., Osaka, Japan) and PW (KENEI Pharmaceutical Co. Ltd., Osaka, Japan). A 10 mL (1.0 mL scale) graduated cylinder (AS ONE Co., Osaka, Japan) was used as the dispensing tool and an alcohol thermometer (Shinwa Rules Co., Niigata, Japan) was used for temperature measurement. The signal intensity was acquired using the ImageJ software (National Institutes of Health, Bethesda, MD). 29

2.B | Dilution of acetone and the T1 value
Acetone was diluted with PW in 10 v/v% increments, and the T1 values were measured. The calculated values were obtained from the average of six measurements. The coefficient of variation (CV) of the six measurements was also calculated at the same time.

| DISCUSSION
Here, we developed a phantom that reproduced the T1 value of CSF at RT (23°C-25°C). PW at RT has a much lower T1value than that of in-vivo CSF and thus is not suitable as a phantom for CSF reproduction. We solved this problem using acetone.
It is well known that the T1 value varies with temperature. 28 Therefore, by constructing a system that can be adjusted so that the phantom temperature can be maintained at a level a researcher desires, it is possible to develop a phantom that has a similar T1 value as that of in-vivo CSF. However, the thermoregulatory equipment cannot be used in many facilities. Acetone-water mixtures can match in-vivo the T1 of CSF without additional thermoregulatory equipment.
The T1 value measurement results revealed that acetone had a T1 value much higher than that of PW at RT. Therefore, we believed that the same T1 value as that of in-vivo CSF could be reproduced at RT by diluting acetone with PW. To date, the relaxation of wateracetone series has not been well studied. Therefore, it remains unclear why the T1 value of acetone is longer than that of PW.
However, we assume that the cause may be the difference in chemical shift with PW. In the scarce available literature, there is a statement on chemical shifts related to water and acetone series. 37 According to that study, it was inferred that the difference in the peak of the water-acetone system is approximately 2.4 ppm. 37 That is, the chemical shift of acetone is larger than that of water. It is known that the rate of relaxation based on the mechanism of the magnetic resonance phenomenon slows as the proportion of spins belonging to the resonance frequency band decreases in the Larmor frequency distribution, resulting in a longer T1 value. 38 . Therefore, it was considered that the resonance frequency of acetone became dominant as the concentration of acetone increased in the wateracetone system, and the T1 value increased. Acetone is a type of it possesses is positively charged. As a result, the π electron of the C = O bond moves to the oxygen atom, and the oxygen atom acquires a formal negative charge. If these reactions are carried out in a solvent with a hydroxy group (−OH), such as H2O, a proton H+ will usually be added to this negative charge. This reaction is termed "nucleophilic addition" and is considered to occur in a mixture of acetone and water. There is reported oxygen molecules in acetone form hydrogen bonds with surrounding water molecules and interact strongly with each other. 39 However, the linewidth of acetone and water is narrow based on the spectral peak reported in the literature 36 ; it is expected to have a relatively slow chemical exchange rate, as it is observed as an independent peak. If the exchange between those two pools is slow, the acetone protons and water protons may retain their own relaxation time to certain degree.
Therefore, it should be noted that there is a possibility that T1 relaxation cannot be accurately characterized with a single exponential function.
A T1 value close to those reported in the literature 30  The T1 value of the acetone-water system observed for longterm fluctuation did not vary greatly. This suggesting that stable use is possible in the long-term.
It is hence possible to reproduce the T1 value of in-vivo CSF using acetone, which cannot be performed using PW. We believe that our findings may contribute to MRI research focusing on the CSF.
The purpose of this study was to reproduce the T1 value of CSF at RT. As a limiting factor, the proton density and T2 value may have been imperfect.

| CONCLUSION
The CSF T1 value in the living body at scanner RT was reproduced with acetone. The optimum acetone concentration that reproduced the CSF T1 value was pure acetone at 1.5 T and acetone 90 v/v% at 3.0 T.
The findings of this study may be useful for MRI studies that focus on the brain.