Efficacy of fat quantification methods used in MRI to distinguish between normal,benign,and malignant bone marrow pathologies in children

Abstract

Background

Fat quantification methods in magnetic resonance imaging (MRI) have been studied to differentiate bone marrow pathologies in adult patients; however, scarce literature is available in pediatric patients.

Purpose

To evaluate the efficacy of the T1 signal intensity value (T1-SIV), out-of-phase/in-phase signal ratio (OP/IP SR), and fat fraction (FF) to differentiate between normal, benign, and malignant pathological processes.

Material and Methods

A total of 48 pediatric patients with lumbar and pelvic MRI were classified into three groups according to bone marrow pathology (group 1, normal; group 2, benign pathology/reconversion; group 3, malignant). The efficacy of T1-SIV, OP/IP SR, and FF values in differentiating these pathologies was evaluated using Kruskal–Wallis or analysis of variance and followed by Bonferroni or Dunn–Bonferroni tests. Cutoff values for malignant infiltration were defined using ROC analysis.

Results

Although these values were significantly different in all three groups (P = 0.001–0.008), this difference was not sufficient to discriminate between all groups. Subgroup analyses showed significant differences in T1-SIV between groups 1–3, in OP/IP SR between groups 1–3, 2–3, and 1–2, in FF between groups 1–2 and 1–3 in various regions (P = 0.001–0.049). Cutoff values had a sensitivity and specificity of 90%–100% for OP/IP SR and FF.

Conclusion

T1-SIV, OP/IP SR, and FF may potentially distinguish normal from pathological bone marrow. OP/IP SR and FF values detected malignant infiltration with high sensitivity and specificity in this study. However, only OP/IP SR may significantly differentiate benign and malignant bone marrow pathologies which needs to be confirmed in the future study with a larger patient population.

Keywords

Fat quantification pediatric bone marrow fat fraction Dixon

Introduction

Bone marrow mainly consists of myeloid tissue, fat cells, and trabecular matrix. It can be classified into yellow and red marrow depending on its cellular distribution. While yellow marrow is composed mostly of fat, red marrow contains similar proportions of water and fat and is the active component of bone marrow. In newborns, the bone marrow consists mainly of red marrow. The conversion from red to yellow marrow starts soon after birth and continues until the bone marrow forms an adult pattern (1 –5). However, when red marrow cannot provide enough hematopoiesis, the yellow marrow may convert back to red marrow, referred to as reconversion (4). The reconversion proceeds in the reverse order of conversion. Further, in contrast to the predictable pattern of conversion, reconversion involves an irregular and asymmetrical process. Malignant processes infiltrating the bone marrow proliferate by destroying the yellow and red marrow, which results in the loss of fat component in both yellow and red marrow (6,7).

Magnetic resonance imaging (MRI) is accepted as an ideal method for the radiological assessment of the bone marrow. This assessment routinely includes T1-weighted (T1W), short-tau inversion recovery (STIR), and T2-weighted (T2W) sequences (8). However, normal bone marrow, reconversion-like benign bone marrow pathologies, and malignant bone marrow infiltration may not be properly characterized using the above sequences. It can be challenging to distinguish these processes, especially in pediatric cases, since the red marrow has a low fat and high water content.

In the literature, there are several studies investigating fat quantification methods in MRI to quantify the fat content in bone marrow and differentiate normal, benign, and malignant pathologic bone marrow (3,4,8,9). The MRI methods and measurements used for this purpose include T1 signal intensity value (T1-SIV) measurement, out-of-phase (OP)/in-phase (IP) signal ratio in images obtained using the 2-point or 3-point Dixon method, and fat fraction (FF) obtained from multi-echo Dixon images (10 –14). In malignant bone marrow infiltration, the signal from fat is reduced relative to normal bone marrow, because fat cells are usually diminished due to tumoral infiltration. Most studies on these techniques have been performed in adult patients, and the literature on pediatric patients is limited (3,12).

The aim of the present study was to evaluate the efficacy of T1-SIV, OP/IP SR, and FF used for fat quantification in MRI of the bone marrow in pediatric patients to differentiate normal, benign, and malignant pathological bone marrow.

Material and Methods

This study was approved by Bursa Uludag University Faculty of Medicine Clinical Research Ethics Committee (decision number: 2022-2/1).

Patients

Our study was a single-center retrospective study. Between March 2020 and January 2022, 48 patients aged 0–18 years who underwent lumbar and pelvic MRI performed with a standardized protocol were included in the study. All patients and their parents were informed about the main technical features of the hip and lumbar MRI examinations to be performed in the study. Subsequently, written and verbal informed consent was obtained from every responsible parent. Patients without a reliable clinical and/or pathologic diagnosis, patients without suitable images for evaluation, patients with focal bone marrow lesions, and patients treated with radiotherapy and/or chemotherapy were not included in the study. Consecutive sampling method was used for patient selection. All patients included in the study had pelvic and lumbar MRI of good diagnostic quality. The evaluated patients were divided into three groups based on bone marrow findings. Patients in group 1 (control group) were selected from a cohort that hematopathologic, neoplastic, inflammatory and infectious processes were excluded by clinical evaluations and laboratory examinations, and who was evaluated hip and lumbar MRI for mechanical low back pain. All patients in this group presented to the hospital with lumbar and hip pain and none of them had any known chronic disease. No pathologic findings were found in the laboratory values of the patients. MRI examinations revealed either no pathologic findings or minimal lumbar disc herniation. None of the patients in this group had MRI findings suggestive of focal or diffuse bone marrow pathology. Group 2 (benign pathology/reconversion group) included six patients with benign bone marrow pathologies. The diagnoses of patients in the benign pathology/reconversion group were based on conventional sequence images and clinical–laboratory findings (7,10). All patients in the benign group had low hemoglobin levels. Ten patients with malignant bone marrow infiltration were included in group 3 (malignant group). These patients were diagnosed by pediatric hematologists through bone marrow aspiration from the iliac crest and histopathological examination of biopsy materials. Only patients with diffuse bone marrow involvement were included in the pathologic groups.

MRI technique

All MRI scans were performed with a 1.5-T MRI system (Aera; Siemens, Erlangen, Germany), using an 11-channel spine coil for the lumbar region and a 16-channel body coil for the pelvic region. The imaging area included all lumbar vertebrae on lumbar MRI. On pelvic MRI, the images extended from the lower lumbar vertebrae to the proximal femur. Imaging was performed in the coronal plane of the pelvic and sagittal planes of the lumbar region. No contrast material was used in any examination. The sequences and parameters are listed in Table 1.

Table 1.

Magnetic resonance imaging sequences and their parameters.

	T1W TSE lumbar spine	T1W VIBE Dixon lumbar spine	qDixon lumbar spine	T1W TSE pelvis	T1W VIBE Dixon pelvis	qDixon pelvis
TR (ms)	572	7.1	15.6	457	7.35	7.35
TE (ms)	11	2.39–4.77	2.38–14.28	10	2.39–4.77	2.38–14.28
ΔTE (ms)			2.38			2.38
FA (°)	150	10	4	160	10	4
FOV (mm)	280 × 280	280 × 280	280 × 280	300 × 300	300 × 300	383 × 272
Matrix	384/257	384/392	160/126	320/240	448/287	352/190
Slice thickness (mm)	4	2	2	4	4	2
NEX	1	2	1	2	2	1
No. of echoes	1	2	6	1	2	6

FA, flip angle; FOV, field of view; NEX, number of excitations; qDixon, quantitative Dixon; T1W, T1-weighted; TE, echo time; TR, repetition time; VIBE Dixon, volumetric interpolated breath-hold examination Dixon.

Imaging interpretation

Reviewer information

The images were evaluated by three radiologists, including two observers, one with 25 years of experience in pediatric radiology and the other with five years of experience in neuroradiology, and one operator with five years of experience as a radiology trainee. The observers agreed on the measurement locations and the operator performed the measurements at the workstation under the supervision of the observers. All measurements were evaluated by the operator and observers together. The observers and operator were blinded to the study data.

Imaging review preparation

T1W turbo spin-echo images, OP and IP images obtained from the T1 volumetric interpolated breath-hold examination (VIBE) Dixon sequence, and proton density fat fraction (PDFF) map images generated automatically by the device using the quantitative Dixon (qDixon) sequence for each case in the study were examined. Images were obtained using a dedicated program for the MR scanner (Syngo VH22B; Siemens, Erlangen, Germany).

Imaging review process

In total, six circular regions of interest (ROIs) were placed by the operator on the L4 and L5 vertebral bodies, bilateral iliac wings, and bilateral femur proximal metaphyses, where the observers agreed with a consensus. The reason for choosing these localizations was to be able to measure from areas with red bone marrow. Circular ROIs have a mean size of 0.7–1 cm² and the size varies according to the age and localizations. SIV was measured from the T1W, OP, and IP images, and FF was determined from the PDFF map. OP/IP SR (OP SIV/IP SIV) was calculated using measurements obtained from the out-of-phase and in-phase images (3,4,15 –19).

Statistical analysis

The data were examined for conformity to a normal distribution using the Shapiro–Wilk test. In intergroup comparisons, an analysis of variance (ANOVA) was performed if the distribution conformed to a normal distribution and met the assumptions of the ANOVA; otherwise, the Kruskal–Wallis test was preferred. In the presence of statistical differences between the groups, subgroup comparisons were made using the Bonferroni test after ANOVA and Dunn–Bonferroni test after the Kruskal–Wallis test. Descriptive values were measured as mean ± standard deviation in the case of parametric tests and median (range) in the case of non-parametric tests. Receiver operating characteristic (ROC) curve analysis was applied to determine whether the measured variables had a significant cutoff value in the identification of malignancy. The statistical analyses were conducted using SPSS version 22 software (IBM Corp., Armonk, NY, USA) and MedCalc 19.5.6 software (MedCalc Software Ltd., Ostend Belgium). The level of statistical significance was set at P = 0.05.

Results

Study population

The female/male ratio and mean age were 17/15 and 158 months in group 1, 4/2 and 192 months in group 2, and 6/4 and 138.5 months in group 3. In the control group, 10 patients had lumbar disc herniation without stenosis of the spinal canal, and MRI of the remaining 22 patients showed no pathology. Among the patients in the benign pathologic group, five were diagnosed with chronic anemia-bone marrow reconversion and one with chronic malnutrition/nutritional rickets. In the malignant pathologic group, three patients had lymphoma (two Hodgkin's lymphoma, one peripheral T-cell lymphoma), three had neuroblastoma, and four had leukemia (two T-ALL, two AML).

Imaging finding analyses

Three measurements (T1-SIV, OP/IP SR, and FF) from six regions (L4-L5 lumbar vertebrae, bilateral iliac wings, and bilateral femur proximal metaphyses) of the patients were statistically compared. The Kruskal–Wallis test revealed a statistically significant difference in all the values obtained for each localization (P <0.05). Since significant results were found in the Kruskal–Wallis test, subgroup comparisons were performed using the Dunn–Bonferroni test. When the T1-SIV obtained from six regions was examined between groups, statistically significant differences were found between group 1 and group 3 in four localizations excluding the iliac wings, and between group 1 and group 2 in four localizations excluding the vertebrae (P = 0.002–0.024). OP/IP SR showed significant differences between groups 1 and 3 in all regions, between groups 2 and 3 in the proximal parts of the femur, and between groups 1 and 2 in all locations except the femur (P = 0.000–0.047). The FF showed significant differences in all regions between groups 1 and 2 and between groups 1 and 3 (P = 0.000–0.049). No significant differences were found between groups 2 and 3 in T1-SIV and FF for any localization (P >0.05). For these localizations, the median T1-SIV was in the range of 116.00–202.00, 49.50–91.50, and 77.50–98.50; the median OP/IP SR was in the range of 0.20–0.42, 0.43–0.74, and 1.03–1.08; and the median FF was in the range of 35.00–56.50, 17.50–37.00, and 4.00–6.00 in groups 1 (Fig. 1), 2 (Fig. 2), and 3 (Fig. 3), respectively (Tables 2 and 3).

Fig. 1.

A 14-year-old female patient with hip pain. The patient has no known chronic disease and no pathologic MRI findings. Measurements on pelvic (a) T1W, (b) PDFF, (c) IP, and (d) OP images showed right femur T1-SIV: 127, left femur T1-SIV: 130; right femur FF: 47, left femur FF: 46; right femur OP/IP SR: 0.8, left femur OP/IP SR: 0.84. IP, in-phase; OP, out-of-phase; PDFF, proton density fat fraction; SR, signal ratio; T1-SIV, T1 signal intensity value; T1W, T1-weighted.

Fig. 2.

A 15-year-old female patient. She was diagnosed with chronic malnutrition/nutritional rickets. Measurements on lumbar (a) T1W, (b) PDFF, (c) IP, and (d) OP images revealed L4 T1-SIV: 123, L5 T1-SIV: 139; L4 FF: 13, L5 FF: 12; L4 OP/IP SR: 0.67, L5 OP/IP SR: 0.75. IP, in-phase; OP, out-of-phase; PDFF, proton density fat fraction; SR, signal ratio; T1-SIV, T1 signal intensity value; T1W, T1-weighted.

Fig. 3.

Pelvic (a) T1W, (b) PDFF, (c) IP, and (d) OP images of a 10-year-old male patient diagnosed with leukemia showed right femur T1-SIV: 55, left femur T1-SIV: 56; right femur FF: 1, left femur FF: 3; right femur OP/IP SR: 1.07, left femur OP/IP SR: 1.09. IP, in-phase; OP, out-of-phase; PDFF, proton density fat fraction; SR, signal ratio; T1-SIV, T1 signal intensity value; T1W, T1-weighted.

Table 2.

Comparison of quantitative magnetic resonance imaging findings in all three groups.

	Group 1	Group 2	Group 3	P value
T1-SIV
L4 vertebra	133 (30–254)	80.5 (35–123)	77.5 (24–100)	0.005
L5 vertebra	132.5 (32–264)	87 (35–139)	80 (26–102)	<0.001
Right iliac wing	116 (31–248)	49.5 (16–84)	77.5 (16–143)	0.001
Left iliac wing	126 (36–254)	53.5 (10–84)	80 (26–102)	0.008
Right femur	200 (76–385)	91 (47–168)	98.5 (16–159)	<0.001
Left femur	202 (74–386)	91.5 (41–136)	98.5 (15–147)	0.003
OP/IP SR
L4 vertebra	0.28 (0.12–0.62)	0.72 (0.32–1.04)	1.03 (1–1.2)	<0.001
L5 vertebra	0.30 (0.13–0.64)	0.74 (0.30–1.00)	1.08 (1–1.19)	<0.001
Right iliac wing	0.20 (0.10–0.48)	0.46 (0.32–0.76)	1.05 (1–1.33)	<0.001
Left iliac wing	0.21 (0.12–0.46)	0.49 (0.26–0.67)	1.03 (0.9–1.11)	<0.001
Right femur	0.38 (0.16–0.80)	0.44 (0.22–0.54)	1.05 (0.92–1.12)	<0.001
Left femur	0.42 (0.15–0.84)	0.43 (0.28–0.61)	1.03 (0.96–1.13)	<0.001
FF
L4 vertebra	35 (20–53)	18 (8–33)	4.5 (1–12)	<0.001
L5 vertebra	35 (20–53)	17.5 (10–34)	4.5 (1–14)	<0.001
Right iliac wing	38.5 (20–55)	24 (7–31)	6 (3–10)	<0.001
Left iliac wing	40 (22–57)	23 (6–33)	5 (2–12)	<0.001
Right femur	56.5 (42–77)	32.5 (20–45)	6 (1–11)	<0.001
Left femur	51 (44–78)	37 (22–44)	4 (2–8)	<0.001

Values are given as median (range). Statistically significant p values are indicated in bold. Group 1, control group; group 2, benign pathology/reconversion group; group 3, malignant group.

FF, fat fraction; IP, in-phase; OP, out-of-phase; SR, signal ratio; T1-SIV, T1 signal intensity value.

Table 3.

Comparison of groups in a subgroup analysis.

	Group 1 – group 2 (P value)	Group 1 – group 3 (P value)	Group 2 – group 3 (P value)
T1-SIV
L4 vertebra	0.170	0.009	1.000
L5 vertebra	0.245	0.007	1.000
Right iliac wing	0.024	0.147	1.000
Left iliac wing	0.017	0.149	0.944
Right femur	0.023	0.002	1.000
Left femur	0.009	0.003	1.000
OP/IP SR
L4 vertebra	0.008	<0.001	0.875
L5 vertebra	0.032	<0.001	0.434
Right iliac wing	0.031	<0.001	0.428
Left iliac wing	0.047	<0.001	0.350
Right femur	1.000	<0.001	0.003
Left femur	1.000	<0.001	0.004
FF
L4 vertebra	0.049	<0.001	0.352
L5 vertebra	0.043	<0.001	0.406
Right iliac wing	0.034	<0.001	0.448
Left iliac wing	0.021	<0.001	0.609
Right femur	0.009	<0.001	0.731
Left femur	0.007	<0.001	0.790

Group 1, control group; group 2, benign pathology/reconversion group; group 3, malignant group.

FF, fat fraction; IP, in-phase; OP, out-of-phase; SR, signal ratio; T1-SIV, T1 signal intensity value. Statistically significant p values are indicated in bold.

ROC analysis

A ROC analysis was performed to estimate the T1-SIV, OP/IP SR, and FF cutoff values with the highest sensitivity and specificity for the diagnosis of bone marrow malignant infiltration. Values below the cutoff values for T1-SIV and FF and values above the cutoff values for OP/IP SR were defined for malignant infiltration. The cutoff values for each region were 8.00–14.00 (area under the ROC curve [AUC] = 0.986–1.000, sensitivity = 100%, specificity = 92.11%–100%, P <0.05) for FF and 0.76–0.87 for OP/IP SR (AUC = 0.986–1.000, sensitivity = 100%, specificity = 97.37%–100%, P <0.05). For T1-SIV, no statistically significant cutoff values were determined for measurements made from both iliac wings (P >0.05). The cutoff values for the L4–L5 vertebrae and both femurs were in the range of 88.00–159.00 (Table 4 and Fig. 4).

Fig. 4.

Receiver operating characteristic curves for fat fraction, in-phase/out-of-phase signal ratio, and T1 signal intensity value.

Table 4.

ROC analysis to determine the cut-off values to be used for the diagnosis of malignant infiltration.

	Cutoff value	AUC (95% CI)	P value	Sensitivity (%)	Specificity (%)
T1-SIV
L4 vertebra	88.00	0.774 (0.630–0.882)	0.001	90 (55.5–99.7)	71.05 (54.1–84.6)
L5 vertebra	90.00	0.788 (0.646–0.893)	0.001	90 (55.5–99.7)	70.3 (51.3–81.5)
Right iliac wing	143.00	0.653 (0.500–0.786)	0.1195	100 (69.2–100.00)	35.14 (20.2–52.5)
Left iliac wing	152.00	0.650 (0.497–0.783)	0.1247	100 (69.2–100.00)	32.43 (18.00–49.8)
Right femur	159.00	0.800 (0.659–0.901)	0.001	100 (69.2–100.00)	55.26 (38.3–71.4)
Left femur	157.00	0.780 (0.643–0.925)	0.001	100 (69.2–100.00)	55.26 (38.3–71.4)
OP/IP SR
L4 vertebra	0.87	0.986 (0.900–1.000)	0.001	100 (69.2–100.00)	97.37 (86.2–99.9)
L5 vertebra	0.86	0.999 (0.923–1.000)	0.001	100 (69.2–100.00)	97.37 (86.2–99.9)
Right iliac wing	0.76	1.000 (0.926–1.000)	0.001	100 (69.2–100.00)	100 (90.7–100)
Left iliac wing	0.77	1.000 (0.926–1.000)	0.001	100 (69.2–100.00)	100 (90.7–100)
Right femur	0.80	1.000 (0.926–1.000)	0.001	100 (69.2–100.00)	100 (90.7–100)
Left femur	0.84	1.000 (0.926–1.000)	0.001	100 (69.2–100.00)	100 (90.7–100)
FF
L4 vertebra	12.00	0.997 (0.921–1.000)	0.001	100 (69.2–100.00)	97.37 (86.2–99.9)
L5 vertebra	14.00	0.992 (0.911–1.000)	0.001	100 (69.2–100.00)	92.11 (78.6–98.3)
Right iliac wing	10.00	0.993 (0.914–1.000)	0.001	100 (69.2–100.00)	97.37 (86.2–99.9)
Left iliac wing	12.00	0.986 (0.900–1.000)	0.001	100 (69.2–100.00)	94.74 (82.3–99.4)
Right femur	11.00	1.000 (0.926–1.000)	0.001	100 (69.2–100.00)	100 (90.7–100)
Left femur	8.00	1.000 (0.926–1.000)	0.001	100 (69.2–100.00)	100 (90.7–100)

AUC, area under the ROC curve; CI, confidence interval; FF, fat fraction; IP, in-phase; OP, out-of-phase; ROC, receiver operating characteristic; SR, signal ratio; T1-SIV, T1 signal intensity value. Statistically significant p values and cuttoff values are indicated in bold.

Discussion

In our study, we found that T1-SIV, OP/IP SR, and FF are valuable for differentiating normal bone marrow from benign and malignant pathological bone marrow. For OP/IP SR and FF, the cutoff values determined in various regions had a high sensitivity and specificity. Lower rates of specificity were observed at the T1-SIV cutoff values. However, the only method that showed significant efficacy in the discrimination of malignant and benign pathologic bone marrow was OP/IP SR measurement from the femora.

The main MRI method used for the quantitative evaluation of bone marrow is the Dixon techniques (19). The standard Dixon technique uses a single in-phase and out-of-phase echo. In this method, the fat content may not be measured accurately because of the T2* relaxation effects and the complex pattern of the fat spectrum. The multi-echo qDixon (PDFF) technique was developed to avoid these disadvantages. This technique is preferred over older methods in FF measurement in current practice because it avoids T1 heterogeneity, T2* relaxation effects, fat spectral complexity, and aliasing problems (19,20). Moreover, in our study, we applied the multi-echo qDixon technique to evaluate FF.

Regarding the measurements from all regions in the PDFF map, significant differences were observed between the malignant and control group FFs, similar to other studies (12,20,21 –23). However, FFs were not useful in differentiating malignant from benign pathology/reconversion groups in any region. Most studies using PDFF have evaluated focal lesions in the adult population (21 –23). Some studies have analyzed the effectiveness of PDFF in separating benign and malignant vertebral lesions (24 –26). In these studies, FFs were found from the lowest to the highest in malignant collapsing lesions, benign collapsing lesions, and normal bone marrow. However, none of these studies were performed in pediatric patients. In one study that included children, the FFs of children with diffuse malignant bone marrow infiltration were lower than those of children in the control group (12). In this study, unlike our study, only a comparison was made between the malignant and normal groups because there was no benign group. Although our results overlap with these findings, this study differs from ours because it did not compare malignant and benign pathologies in children. In addition, no other study in the literature has compared the efficacy of FF with OP/IP SR and T1-SIV. We believe that FF is not sufficient to adequately distinguish benign from malignant vertebral lesions in pediatric patients.

In our study, the only method that showed significant efficacy in the differentiation of benign and malignant bone marrow infiltration was OP/IP SR measurement from the femora. This may be caused by bone marrow conversion from peripheral to central in children or earlier development of conversion in the proximal femur than in the axial skeleton. The superiority of OP/IP SR over other methods may be due to the signal decrease in OP images compared to IP which reflects fat at the microscopic level, whereas FF and T1-SIV reflect fat at the macroscopic level (27). For OP and IP imaging, we chose the VIBE Dixon sequence because of its shorter imaging time. In our study, as in the study by Donners et al. (28), the OP/IP SR ratios from T1W VIBE Dixon sequences overlap with other studies using Dixon sequences for bone marrow in the literature (3,4,15 –18). In other studies, a statistically significant difference was found between the malignant and benign groups in the OP/IP SR (3,4,15 –18). In most of these studies, unlike in our study, the sites of measurement were not specified separately, and focal lesions were evaluated (3,15 –18). Only Akman et al. (4) included bone marrow reconversion as a benign lesion. The benign pathologies examined in other studies differed from those in our study. In the study by Akman et al. in 2019 (4), OP/IP SR was calculated only from the vertebrae, and the values showed a statistically significant difference in the malignant-control and malignant-reconversion groups. However, this study was performed in adult patients, and no similar study has been conducted in pediatric patients. In the literature, there is only one study in pediatric patients that evaluated bone marrow infiltration using the Dixon method. In this study (3), as mentioned before, the diagnoses of patients in the benign and malignant groups were different from those in our study.

In the ROC analysis performed to set the cutoff values to be used in malignant infiltration diagnosis, the highest AUC and most significant P values were observed in the OP/IP SR. In addition, FF showed a high AUC and significant P values. Both methods showed a sensitivity of 100% and specificity of 90%–100%. Both methods showed a higher diagnostic performance in the appendicular skeleton than in the axial skeleton. This result may be explained by the fact that conversion is completed earlier in the appendicular skeleton than in the axial skeleton. Our cutoff values are in line with the results of other studies on malignant lesions in the literature (3,4,15 –17,21 –26). The T1-SIV did not show significant efficacy in these two regions in the ROC analysis. It had a high sensitivity and low specificity in the remaining four regions. This is because T1-SIV is a relative value that is affected by factors such as the magnetic field strength and sequence parameters (29). The results of the present study on T1W overlap with the accepted general information about T1W in the radiologic examination of the bone marrow (1,5 –8).

The present study has some limitations. The most important limitations are the limited number of patients and the fact that our study was a single-center study. Future studies with more extensive populations are recommended to confirm these results. The absence of histological examination of bone marrow abnormalities in the benign group may be another limitation. Similar studies in patients with diffuse bone marrow abnormalities, including different benign pathologies, are necessary to confirm our results. Another limitation is that the number of patients in the control group was higher than the number of patients in the other two groups. Although MRI is frequently preferred in lumbar and hip pain, it is not always used for the diagnosis of bone marrow infiltration. Therefore, our patients in the control group, although similar in age to the other two groups, are more in number. One more limitation is the age of our patients. We know that bone marrow distribution and imaging findings are different in early childhood compared to adolescence and adulthood. In our study, the mean age of the patient groups was in the range of 11.5–16 years and the youngest patient was 5 years old. We think that similar studies on bone marrow involvement in children with a younger age group will contribute to the literature.

The present study showed that T1-SIV, OP/IP SR, and FF were successful in differentiating normal bone marrow from pathologic bone marrow in at least one but not all regions. The OP/IP SR and FF values can diagnose malignant infiltration with high sensitivity and specificity. However, only OP/IP SR showed significant efficacy in diagnosing malignant and benign bone marrow pathologies.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Duygu Erkal Tonkaz

Mehmet Tonkaz

References

Foster

Chapman

Johnson

. MRI of the marrow in the paediatric skeleton. Clin Radiol 2019;59:651–673.

Major

Anderson

. Muskuloskeletal MRI. 3rd edn. Philedelphia, PA: Elsevier, 2020.

Winfeld

Ahlawat

Safdar

. Utilization of chemical shift MRI in the diagnosis of disorders affecting pediatric bone marrow. Skeletal Radiol 2016;45:1205–1212.

Akman

Ata Korkmaz

Sarı

. Efficacy of chemical shift MRI for differentiating diffuse red bone marrow reconversion and hematological malignancies. Turk J Med Sci 2019;49:644–652.

Patel

Brian

Chauvin

. Pearls and pitfalls in imaging bone marrow in pediatric patients. Semin Ultrasound CT MR 2020;41:472–487.

Nouh

Eid

. Magnetic resonance imaging of the spinal marrow: basic understanding of the normal marrow pattern and its variant. World J Radiol 2015;7:448–458.

Małkiewicz

Dziedzic

. Bone marrow reconversion - imaging of physiological changes in bone marrow. Pol J Radiol 2012;77:45–50.

Shah

Hanrahan

. MRI of spinal bone marrow: part I, techniques and normal age-related appearances. AJR Am J Roentgenol 2011;197:1298–1308.

Del Grande

Farahani

Carrino

, et al. Bone marrow lesions: a systematic diagnostic approach. Indian J Radiol Imaging 2014;24:279–287.

10.

Wismer

Rosen

Buxton

, et al. Chemical shift imaging of bone marrow: preliminary experience. AJR Am J Roentgenol 1985;145:1031–1037.

11.

Van Vucht

Santiago

Lottmann

, et al. The Dixon technique for MRI of the bone marrow. Skeletal Radiol 2019;48:1861–1874.

12.

Samet

Deng

Schafernak

, et al. Quantitative magnetic resonance imaging for determining bone marrow fat fraction at 1.5 T and 3.0 T: a technique to noninvasively assess cellularity and potential malignancy of the bone marrow. Pediatr Radiol 2021;51:94–102.

13.

Samet

Fayad

. Advanced problem solving MRI techniques to aid in the diagnosis of bone and soft tissue musculoskeletal lesions. Adv Clin Radiol 2021;3:217–226.

14.

Sieron

Drakopoulos

Loebelenz

, et al. Correlation between fat signal ratio on T1-weighted MRI in the lower vertebral bodies and age, comparing 1.5-T and 3-T scanners. Acta Radiol Open 2020;9:2058460120901517.

15.

Erly

Outwater

. The utility of in-phase/opposed-phase imaging in differentiating malignancy from acute benign compression fractures of the spine. AJNR Am J Neuroradiol 2006;27:1183–1188.

16.

Zampa

Cosottini

Michelassi

, et al. Value of opposed-phase gradient-echo technique in distinguishing between benign and malignant vertebral lesions. Eur Radiol 2002;12:1811–1818.

17.

Disler

McCauley

Ratner

, et al. In-phase and out-of-phase MR imaging of bone marrow: prediction of neoplasia based on the detection of coexistent fat and water. AJR Am J Roentgenol 1997;169:1439–1447.

18.

Eito

Waka

Naoko

, et al. Vertebral neoplastic compression fractures: assessment by dual-phase chemical shift imaging. J Magn Reson Imaging 2004;20:1020–1024.

19.

Zhan

Olsen

Zhang

, et al. Detection of hepatic steatosis and iron content at 3 tesla: comparison of two-point Dixon, quantitative multi-echo Dixon, and MR spectroscopy. Abdom Radiol (NY) 2019;44:3040–3048.

20.

Idilman

Aniktar

Idilman

, et al. Hepatic steatosis: quantification by proton density fat fraction with MR imaging versus liver biopsy. Radiology 2013;267:767–775.

21.

Kim

Kannengiesser

Paek

, et al. Differentiation between focal malignant marrow-replacing lesions and benign red marrow deposition of the spine with T2*-corrected fat-signal fraction map using a three-echo volume interpolated breath-hold gradient echo Dixon sequence. Korean J Radiol 2014;15:781–791.

22.

Lee

Hahn

, et al. Fat fraction estimation of morphologically normal lumbar vertebrae using the two-point mDixon turbo spin-echo MRI with flexible echo times and multipeak spectral model of fat: comparison between cancer and non-cancer patients. Magn Reson Imaging 2016;34:1114–1120.

23.

Yoo

Hong

Kim

, et al. Measurement of fat content in vertebral marrow using a modified Dixon sequence to differentiate benign from malignant processes. J Magn Reson Imaging 2017;45:1534–1544.

24.

Jung

Jeon

Kwack

, et al. Differentiation of vertebral metastases from focal hematopoietic marrow depositions on MRI: added value of proton density fat fraction. AJR Am J Roentgenol 2021;216:734–741.

25.

Kwack

Lee

Jeon

, et al. Comparison of proton density fat fraction, simultaneous R2*, and apparent diffusion coefficient for assessment of focal vertebral bone marrow lesions. Clin Radiol 2020;75:123–130.

26.

Schmeel

Enkirch

Luetkens

, et al. Diagnostic accuracy of quantitative imaging biomarkers in the differentiation of benign and malignant vertebral lesions: combination of diffusion-weighted and proton density fat fraction spine MRI. Clin Neuroradiol 2021;31:1059–1070.

27.

van Vucht

Santiago

Pressney

, et al. Role of in-phase and out-of-phase chemical shift MRI in differentiation of non-neoplastic versus neoplastic benign and malignant marrow lesions. Br J Radiol 2021;94:20200710.

28.

Donners

Hirschmann

Gutzeit

, et al. T2-weighted Dixon MRI of the spine: a feasibility study of quantitative vertebral bone marrow analysis. Diagn Interv Imaging 2021;102:431–438.

29.

Dale

Brown

Semelka

. MRI: Basic principles and applications. 5th edn. Chichester: Wiley Blackwell, 2015.