Skip to main content
NCBI home page
North American Spine Society Journal logoLink to North American Spine Society Journal
. 2023 Mar 1;14:100204. doi: 10.1016/j.xnsj.2023.100204

Correlation between bone density measurements on CT or MRI versus DEXA scan: A systematic review

Amer Ahmad a, Charles H Crawford III a,b, Steven D Glassman a,b, John R Dimar II a,b, Jeffrey L Gum a,b, Leah Y Carreon b,⁎
PMCID: PMC10119682  PMID: 37090222

Abstract

Background

Novel methods of bone density assessment using computed tomography (CT) and magnetic resonance imaging (MRI) have been increasingly reported in the spine surgery literature. Correlations between these newer measurements and traditional Dual-Energy X-ray Absorptiometry (DEXA) is not well known. The purpose of this study is to perform an updated systematic review of correlations between bone mineral density (BMD) from CT or MRI and DEXA.

Methods

Articles published between 2011 and 2021 that reported correlations between the CT-HU or MRI measurements to DEXA t-scores or BMD of lumbar spine or hip were included in this systematic review.

Results

A total of 25 studies (15 CT, 10 MRI) met the inclusion criteria with a total number of 2,745 patients. The pooled correlation coefficient of spine CT-HU versus spine DEXA, spine CT-HU versus hip DEXA and spine CT-HU versus lowest t-score were 0.60, 0.50 and 0.60 respectively. Regarding spine DEXA parameters, the pooled r2 for spine CT-HU versus spine t-score was 0.684 and spine CT-HU versus spine BMD was 0.598. Furthermore, in patients undergoing spine surgery in four studies, the pooled correlation between spine CT and spine DEXA was (r2: 0.64). In MRI studies, the pooled r2 of spine MRI versus spine DEXA and spine MRI versus hip DEXA were -0.41 and -0.44 respectively.

Conclusions

CT-HU has stronger correlations with DEXA than MRI measurements. Lumbar CT-HU has the highest pooled correlation (r2 = 0.6) with both spine DEXA and lowest skeletal t-score followed by lumbar CT-HU with hip DEXA (r2 = 0.5) and lumbar MRI with hip (r2 = 0.44) and spine (r2 = 0.41) DEXA. Both imaging modalities achieved only a moderate correlation with DEXA. Few studies in both modalities have investigated the correlation in spine surgery populations and the available data shows that the correlations are worse in the degenerative spine population. A careful interruption of CT HU and MRI measurement when evaluation of BMD as they only moderately correlated with DEXA scores. At this time, it is unclear which modality is a better predictor of mechanical complications and clinical outcomes in spine surgery patients.

Keywords: Hounsfield unit, Computed tomography scan, Quantitative CT scan, Magnetic resonance imaging, Dual-energy xray absorptiometry, Correlation, Osteoporosis, Bone mineral density, T-score, Lumbar spine

Introduction

With the increase of the aging population, osteoporosis has become a common health problem with low detection and treatment rates [1], [2], [3], [4]. Evaluating bone strength is important in patients undergoing instrumented lumbar spine surgeries as it may be associated with mechanical failure and other complications [5]. Currently, bone mineral density (BMD) is considered the best measure for bone quality. Thus, having an accurate method to measure BMD in spine surgery is important for preoperative planning and optimization [6], [7], [8].

Dual-Energy X-ray Absorptiometry (DEXA) scans are considered the gold standard for BMD assessment [9,10], yet it has some disadvantages as it tends to overestimate the BMD in patients with degenerative spines, aortic calcifications or with high Bone Mass index (BMI) [3,[11], [12], [13], [14], [15], characteristics which are commonly seen among patients seeking spine surgery treatment.

Computed tomography (CT) and magnetic resonance imaging (MRI) are frequently used in the preoperative assessment of spine surgery patients and recently they are increasingly used as alternatives to estimate BMD [16,17]. The purpose of this study is to perform an updated systematic review to compare between BMD estimates from lumbar CT and MRI in term of correlation with the more traditional DEXA scans.

Material and methods

A systematic search was conducted on October 2021 for articles published from 2011 to 2021 in PubMed and Google scholar data bases using the following terms: “Hounsfield units”, “computed tomography”, “Quantitative CT scan”, “MRI”, “magnetic resonance imaging”, “bone mineral density”, “osteoporosis”, “lumbar spine”, “DEXA”, “DXA” and “correlation”. A total of 1,131 full text articles were identified. Cohort studies written in English that reported the correlation between either the HU/MRI measurements of lumbar spine or specific level and DEXA t-score or BMD in patients older than 18-year-old regardless of CT/MRI protocol used were included. Duplicate studies, Biomechanical and cadaver studies or studies that predict the lumbar BMD using the CT or MRI without reporting the correlation coefficient with DEXA scan were excluded (Fig. 1).

Fig. 1.

Fig

Open in a new tab

Literature review workflow

The data from each included CT scan and MRI studies were collected in Excel spread sheet by the Author and included: study design, principal author, year of publication, total number of patients (N), patient's demographics, inclusion and exclusion criteria, CT and MRI protocols and regions, Measurement of Hounsfield unit and MRI methods, DEXA scores, the mean duration between the CT/MRI and DEXA and the correlation coefficient between the CT/MRI and DEXA. Data was analyzed by two independent reviewers.

Correlation studies included in this review were categorized into 5 groups: spine CT with spine DEXA, spine CT with hip DEXA, spine CT with lowest t-score, spine MRI with spine DEXA and spine MRI with hip DEXA. The pooled correlation coefficient weighted by the sample size was calculated for each group. In addition, a separate pooled correlation coefficient was calculated for CT HU in patients undergoing spine surgery.

Results

A total of 26 studies (16 CT scan, 10 MRI) met inclusion criteria for the review with a total number of 2,745 patients. Among the CT scan correlation studies, additional one study was excluded after further review due to the inconsistency of the spine level used for measuring the HU in breast cancer patients; when L1 HU from chest CT was not available for the measurements due to compression fracture in some patients, either T12 or L2 were used as alternative level without being specified [18].

CT scan studies

All the 15 CT scan studies were retrospective with total number of 2,027 patients. The correlation of HU with spine DEXA was reported in thirteen studies (N = 1,979), HU with hip DEXA in 3 studies (N:456) and HU with lowest skeletal t-score in 3 studies (N: 455). Some studies correlated the HU for each lumbar vertebra and others correlated the HU mean value for the lumbar spine (L1–L4) (Table 1) as it has been shown no significant difference between lumbar vertebrae HU values [19]. The pooled correlation coefficient of spine CT vs spine DEXA, spine CT versus hip DEXA and spine CT versus lowest t-score were 0.60, 0.50 and 0.60, respectively. Regarding spine DEXA parameters, HU was correlated with BMD only in 3 studies, with t-score only in 3 studies and with both measurements in 9 studies. The pooled r2 for spine CT vs spine t-score was 0.684, spine CT versus spine BMD was 0.598. Furthermore, four CT studies correlated the spine CT with spine DEXA in patients undergoing spine surgery with pooled correlation (r2: 0.64).

Table 1.

CT scan correlation studies

Study CT scan CT HU Region of interest (ROI) DEXA Max. Duration between CT & DEXA patients’ population Mean age (years) Total number of patients (N) Year of publication Study design
Kim et al. [24] Lumbar CT* Largest trabecular ROI at mid axial of vertebral body spine DEXA BMD, hip DEXA BMD 3 mo Patients undergoing lumbar spine surgery in single center 68.1 180 2019 retrospective
Cohen et al. [25] Abdominal & Lumbar CT* Trabecular ROI on mid-axial and mid-sagittal of vertebral body lowest skeletal T score 6 mo Arab, Ashkenazi and Sephardic jew in single center 64 246 2021 retrospective
Da Zou et al. [5] Lumbar CT* Trabecular ROI on mid axial of vertebral body spine DEXA T score & BMD 1 mo Patients undergoing lumbar degenerative spine surgery in single center Undefined 334 2018 retrospective
Chia et al. [27] Contrast enhanced CT scan* Mean of trabecular ROI measured at 3 different locations on axial image spine DEXA T score, lowest skeletal T score 3 wks Patients with age 50 and above who underwent CECT for any medical condition in single center Undefined 50 2021 retrospective
Islamian et al. [21] Abdominal & Lumbar CT* Trabecular ROI on mid axial of vertebral body spine DEXA BMD 3 mo Patients with spine fracture from minor trauma who underwent both CT and DEXA within 3 mo in single center 60.2 61 2016 retrospective
Alawi et al. [28] Abdominopelvic & Lumbar CT* Mean of trabecular ROI measured at 3 different locations on axial image spine DEXA T score & BMD 2 y Pre or postmenopausal women who underwent DEXA and CT within 2 years in single center 61.1 78 2021 retrospective
Choi et al. [26] Lumbar CT* Trabecular ROI on mid axial of vertebral body spine DEXA T score & BMD 3 mo Patients undergoing spine surgery in single center 67.5 110 2016 retrospective
Schcreiber et al. [20] Abdominopelvic & Lumbar CT* Mean of trabecular ROI measured at 3 different locations on axial image spine DEXA T score & BMD 12 mo Spinal trauma or compression fracture in single center 71.3 25 2011 retrospective
Lee et al. [1] Lumbar CT Mean of trabecular ROI measured at 3 different locations on axial image spine DEXA T score 12 mo Female patients above age 40 with low back pain, single center Undefined 128 2013 retrospective
Elarjani et al. [33] Lumbar CT Trabecular ROI on mid axial vertebral body and mean of 5 trabecular ROI measured at different locations on sagittal image spine DEXA T score & BMD 1 y Undefined 60.2 100 2021 retrospective
Kohan et al. [30] Lumbar CT Mean of trabecular ROI measured at 3 different locations on axial image spine DEXA BMD, hip DEXA BMD Undefined White female patients undergoing ASD surgery in single center Undefined 48 2017 retrospective
Kim et al. [23] Chest LDCT* Volumetric reconstruction analysis of multiple ROIs on axial image spine DEXA BMD, hip DEXA BMD 30 d patients above age 50 who underwent LDCT in single center 65.9 224 2017 retrospective
Amin et al. [10] Abdominopelvic & Lumbar CT Mean of trabecular ROI measured at 3 different locations on axial image lowest skeletal T score 12 mo Predominantly Asians from different ancestries, single center Undefined 159 2021 retrospective
Burke et al. [34] Abdominal CT* Mean of 3 trabecular ROI on mid axial vertebral body by 3 separate readers spine DEXA T score & BMD 6 mo Patients over age 50, had MDCT for other clinical indications 71 171 2016 retrospective
Li et al. [19] Abdominal CT* Trabecular ROI on mid sagittal of vertebral body spine DEXA T score & BMD 6 mo Chinese patients who underwent CT and DEXA within 6 mo in single center 67 109 2018 Retrospective
Open in a new tab
⁎

CT tubal voltage: 120 kvp.

Lumbar CT without contrast was the most used for HU measurements followed by abdominal CT without contrast. There was a variation among the scanning parameters; tube current (range: 30-330mA) and slice thickness (range: 1–5 mm) which were specified in nine studies only [20], [21], [22], [23], [24], [25], [26], [27], [28]. Axial CT was the most common plane used as ROI for HU measurements. The duration between CT scan and DEXA used as a part of inclusion criteria for patients was defined in all except for one study and it varies from 3 weeks to 2 years [Table 2].

Table 2.

Correlation coefficients between Spine CT and DEXA (T-score, BMD)

CT scan
Study DEXA score L1 vertebra L2 vertebra L3 vertebra L4 vertebra Lumbar spine (L1–4)
Kim et al. [24] Spine BMD 0.552 0.535 0.542 - 0.489‡
Femur neck BMD 0.349 0.469 0.374 - 0.393
Da Zou et al. [5] Spine T-score 0.667* 0.767† 0.64* 0.767† 0.658* 0.717† 0.667* 0.764† -
Spine BMD 0.665* 0.771† 0.647* 0.764† 0.662* 0.732† 0.627* 0.77† -
Chia et al. [27] Spine T-score 0.683 - - - -
Islamian et al. [21] Spine BMD - - - - 0.766
Alawi at al. [28] Spine T-score 0.544 0.6 0.611 0.6 -
Spine BMD 0.581 0.623 0.653 0.612 -
Choi et al. [26] Spine T-score 0.3* 0.701† 0.457* 0.709† 0.433* 0.709† 0.447* 0.649† 0.398* 0.734†
Spine BMD 0.313* 0.684† 0.499* 0.693† 0.454* 0.709† 0.455* 0.639† 0.426* 0.721†
Schreiber et al. [20] Spine T-score - - - - 0.48
Spine BMD - - - - 0.44
Lee et al. [1] Spine T-score 0.673 0.794 0.766 0.713 -
Spine BMD 0.657 0.774 0.737 0.673 -
Elarjani et al. [33] Spine T-score 0.592§ 0.504ǁ 0.482§ 0.519++ 0.460§ 0.458ǁ 0.471§ 0.369ǁ -
Spine BMD 0.559§ 0.468ǁ 0.482§ 0.504++ 0.453§ 0.450ǁ 0.456§ 0.353ǁ -
Kohan et al. [30] Spine BMD - - - - 0.463
Femur neck BMD - - - - 0.303
Kim et al. [23] Spine BMD 0.726 - - - -
Femur neck BMD 0.503 - - - -
Total hip BMD 0.665 - - - -
Burke et al. [34] Spine T-score 0.392 - - - -
Spine BMD 0.437 - - - -
Li et al. [19] Spine T-score - - - - 0.62
Spine BMD - - - - 0.61
Open in a new tab
⁎

Correlations in degenerative spine group.

†

Correlations in nondegenerative spine group.

‡

L1-3 mean value.

§

Correlation with Axial CT HU

ǁ

Correlation with Sagittal CT HU.

The patients among the studies varied in ethnicity, number, inclusion, and exclusion criteria. Most of the cohorts were female (1,193 female, 398 male). Four studies only evaluated the correlation in patients undergoing spine surgery [24,26,29,30]. Patients with lumbar fractures, infections, tumors, previous spine instruments, vertebroplasty or severe spinal degeneration were excluded in most studies.

MRI studies

Seven studies were prospective and three were retrospective with total number of 1,024 patients. Eight studies reported correlations between spine MRI with spine DEXA (N = 812) and two studies with hip DEXA (N = 212). The pooled r2 of spine MRI vs spine DEXA and spine MRI vs hip DEXA were -0.41 and -0.44 respectively (Table 3).

Table 3.

MRI correlation studies

Study MRI measurement technique for ROI MRI sequence (s) for ROI DEXA Max. duration between MRI and DEXA Patients’ population Control group Mean age (years) Total number of patients Year of publication MRI measurements Level Study design
Ergen et al. [38] BMFF (using T2*-IDEAL technique) T1W spine echo sequence (TR:660 ms, TE: 8.5ms) and STIR sequence (TR: 3500 ms, TE: 42 ms) Spine DEXA BMD 3 wk Female patients with low back pain from single center NA 49.3 45 2014 L1–4 prospective
Agrawal et al. [37] BMFF and ADC (using DWI and MR Spectroscopy sequences) T1W, T2W spine echo sequence Spine DEXA T score & BMD 18 mo Indian postmenopausal women who underwent DEXA in recruited randomly from single center NA 52.4 50 2015 L3 prospective
Shen et al. [41] BMAT T1W whole body MRI Spine DEXA BMD, hip DEXA BMD Non specified African American, Caucasian recruited from CARDIA study NA Undefined 76 2012 L1–5 prospective
Shih et al. [35] LWR, lipid LW, water LW (using proton MR spectroscopy sequence) T1W, T2W spine echo sequence Spine DEXA BMD 2 wk Female patients who referred to orthopedic or osteoporosis clinic NA 58 52 2004 L3 prospective
Saad et al. [42] M score (calculated from SNR) T1W spine echo sequence (TR: 400-600 ms, TE:7 ms) Spine DEXA T score & BMD 6 mo Postmenopausal women with low back pain in single center Healthy female of matched age with normal BMI 59.4 50 2019 L1–4 retrospective
Shayganfar et al. [40] M score (calculated from SNR) T1W spine echo sequence (TR:400 ms, TE: 16 ms) Spine DEXA T score 6 mo Iranian postmenopausal women who underwent DEXA in single center Healthy female aged between 20 and29 y 59.1 82 2019 L1–4 prospective
Shih et al. [36] Peaked enhanced ratio (BMP) derived from time-Signal intensity curve T1W spine echo sequence (TR:600 ms, TE: 12 ms) Spine DEXA BMD 2 wk Female patients who referred to orthopedic or osteoporosis clinic NA 57 62 2004 L1–5 prospective
Bandirali et al. [39] M score T1W spine echo sequence (TR: 600 ms, TE: 11 ms) Spine DEXA T score 6 mo Caucasian female patients with low back pain in single center Healthy Caucasian female aged between 20 and 29 years with normal BMI 65 226 2015 L1–4 retrospective
Ehresman et al. [31] VBQ score T1W spine echo sequence Hip DEXA T score, lowest skeletal T score 2 y Patients undergoing degenerative spine surgery in single center NA Undefined 68 2019 L1-4 retrospective
Chang et al. [32] PD (using synthetic MRI sequences; T1 map, T2 map, PD map) and VBQ score, T1 intensity T1W spine echo sequence Spine DEXA T score 3 mo Patients undergoing degenerative spine surgery in single center NA 61.9 62 2021 L1–4 prospective
Open in a new tab

In most studies, 1.5 Tesla Lumbar MRI without contrast was used. One study used IV contrast to measure the peak enhancement ratio as a parameter for bone marrow perfusion in the vertebral body to correlate with BMD. Another study used three Tesla machine for measuring the synthetic MRI quantitative parameters of bone physical properties. Different MRI sequences with different measurements used for the correlation: Signal-to-noise ratio (SNR) and M-score (3 studies), vertebral bone marrow fat content (4 studies), Vertebral Bone Quality (VBQ) scores, which is calculated from dividing the average signal intensities (SIs) of lumbar spine by cerebrospinal fluid (CSF) signal intensity (2 studies) and peak vertebral enhancement ratio (1 study). The duration between MRI and DEXA varied from 2 weeks to 2 years among the studies (Table 4).

Table 4.

Correlation coefficients between spine MRI and DEXA (T-score, BMD)

study MRI measurement* DEXA L1 vertebra L2 vertebra L3 vertebra L4 vertebra Lumbar spine
Ergen et al. [38] BMFF Spine BMD - - -0.420 - -
Agrawal et al. [37] BMFF Spine T-score - - -0.450 - -
Spine BMD - - -0.345 -
Shen et al. [41] BMAT Spine BMD - - - - -0.45
Hip BMD - - - - -0.399
Shih et al. [35] Lipid LW Spine BMD - - -0.67 - -
Saad et al. [42] M score Spine T-score - - - - -0.48
Spine BMD - - - - -0.37
Shayganfar et al. [40] M score Spine T-score - - - - -0.551
Shih et al. [36] Peaked enhanced ration Spine BMD - - - - 0.63
Bandirali et al. [39] M score Spine T-score - - - - -0.682
Ehresman et al. [31] VBQ Femur neck T-score - - - - -0.51
Total hip T -score - - - - -0.41
Chang et al. [32] Proton Density Spine T-score - - - - -0.565
VBQ score - - - - -0.651
Open in a new tab
⁎

All MRI measurements have negative correlation with DEXA except for “Peaked enhanced ratio” which has a positive correlation.

Most MRI correlation studies were on female patients with different ethnicity and inclusion criteria. Mean age among the cohorts ranged from 49.3 to 65 years. Two studies only evaluated the correlation in patients undergoing degenerative spine surgeries [31,32].

Discussion

We included in our review the studies that correlated CT scan or MRI to DEXA measurements in both spine and nonspine cohorts and measured the pooled correlation weighted by the sample size for each study. Our systematic review showed that CT Hounsfield unit has stronger correlations with DEXA than MRI measurements. Lumbar CT has the highest pooled correlation (r2 = 0.6) with both spine DEXA and lowest skeletal t-score followed by lumbar CT with hip DEXA (r2 = 0.5) and lumbar MRI with hip (r2 = 0.44) and spine (r2 = 0.41) DEXA. Both imaging modalities achieved only a moderate correlation with DEXA BMD and t-scores.

The correlation studies so far either investigated the ability of CT scan or MRI as opportunistic tools for osteoporosis screening in patients with different morbidities [10,[19], [20], [21], [22], [23],25,27,28,[33], [34], [35], [36], [37], [38], [39], [40], [41], [42] or as alternatives for DEXA in predicting bone quality in spine surgery population [24,26,[29], [30], [31], [32]. Few studies in both modalities (4 CT, 2 MRI) have investigated the correlation in spine surgery patients (Table 5). Among the four CT studies, Spine CT-HU with spine DEXA showed the same moderate pooled correlated (r2: 0.64) [24,26,29,30]. The pooled correlation could not be calculated for spine patients in MRI studies as there are only two studies, each one of them correlated spine MRI with different DEXA region [31,32].

Table 5.

Correlations between CT scan/ MRI and DEXA of lumbar spine in patients undergoing lumbar spine surgery

Study Modality Spine DEXA BMD Spine DEXA T score Hip DEXA BMD Hip DEXA T score
Kim et al. [24] Axial CT HU 0.489 - 0.393 -
Da Zou et al. [5]* Axial CT HU 0.650† 0.658† - -
0.760‡ 0.754‡
Choi et al. [26] Axial CT HU 0.426† 0.398† - -
0.721‡ 0.734‡
Kohan et al. [30] Axial CT HU 0.463 - 0.303 -
Ehresman et al. [31] MRI VBQ - - - -0.510 (Femur neck)
-0.410 (Total hip)
Chang et al. [32] MRI VBQ - -0.651 - -
MRI PD -0.565
Open in a new tab
⁎

The mean values of this study are calculated.

†

correlations in degenerative spine group.

‡

correlations in non-degenerative spine group.

CT scans and/or MRIs are routinely done as a part of preoperative evaluation in patients undergoing spine surgery. BMD assessment is important for surgical planning in such patients especially when using instrumentation as it can be proxy for bone strength, healing, and fusion rates. DEXA scan is still considered the gold slandered for BMD assessment and bone quality evaluation [9,10]. The inherent inaccuracy of DEXA measurements in patients with degenerative spine and the routine use of CT scan and/or MRI before spine surgery paved the way to study the potential of using these modalities as alternatives for BMD assessment in such patients.

In 2011, Schreiber et al introduced the Hounsfield unit for the first time as a measuring tool for BMD using Region of interest (ROI) on conventional CT scan without exposing patients to higher radiation doses compared with Quantitative CT scan [13,20]. More studies have used different CT protocols for BMD measurements in different populations to validate this method further in terms of reliability and applicability. According to the pooled correlation analysis, Spine CT showed moderate correlation with both spine and hip DEXA. Further correlation with the two spine DEXA measurements (t-score and BMD) were calculated. t-score showed a better correlation (r2: 0.684) with HU comparing with BMD (r2: 0.598). In addition, as the lowest t-score from spine and hip DEXA is now recommended by WHO for osteoporosis screening and treatment [43], we calculated from the available studies the pooled correlation for HU and the lowest skeletal t-score which showed the same moderate result as with hip or spine DEXA alone (r2: 0.60).

Among the CT studies, there was a variation in the correlations between lumbar spine HU mean values and DEXA measurements. The strongest correlation was 0.766 [21] while the lowest was 0.303 [30]. This variation could be a result of the inconsistency between the studies in terms of cohort's spine degenerative status, the durations between the images or the variations of CT calibrations (slice thickness and tuba currency) used. These variations can affect in a way or another HU measurements and DEXA differently, hence the variation in the correlation between these modalities among studies. On the other hand Using different HU ROI methods can probably not result in such variation, as the literature showed no significant difference between these different methods [13].

Despite the moderate correlation, CT scan has advantages over DEXA in spine surgery patients. It provides a three-dimensional (3D) estimate for trabecular BMD without being affected by cortical degenerative changes (sclerosis and osteophytes) or aortic calcifications which are common findings among these patients. This may explain the better correlation between CT and DEXA in nondegenerative spine populations (Table 5). In addition, the trabecular bone is affected the most by osteoporosis and correlated better with bone mechanical strength [44] thus can predict the fracture risk and surgical outcomes more accurately in such patients.

MRI has also been investigated as a possible surrogate for bone quality evaluation. Multiple quantitative methods have investigated measuring the trabecular bone microstructure or bone marrow fat content based on differences in signal intensities within bone tissues [31,32,[35], [36], [37], [38], [39], [40], [41], [42]. Changes in these parameters has a relevant negative correlation with osteoporosis and bone quality.

M-score, a novel MRI score simulating DEXA t-score calculation, has been introduced by Bandirali et al. [39] for the first time in 2015. It has been evaluated further by other studies [40,42] which showed a better correlation with BMD (pooled r2: -0.58) comparing with other MRI measurements. Another promising measurement is the peaked enhanced ratio (r2: 063). It measures the IV contrast uptake within the vertebral body as a reflection of bone marrow perfusion which in turn is affected by aging and osteoporosis [36]. The disadvantage of this method is that it requires contrast, which cannot be used routinely for BMD assessment.

The pooled correlation was calculated from different MRI measurements which could not be representing the actual pooled correlation for each of them. The paucity of studies of each certain method can justify calculating pooled correlation from these different measurements.

As with CT studies, MRI studies showed that same inconsistency regarding the cohorts and duration between the images which may again add to the variation in the correlations between MRI and DEXA among the studies. In addition, most patients in MRI studies are female and the mean age is younger when compared to CT studies (49.3–65 years vs. 60.1–71 years) which may make the correlations of MRI not representative for spine population as alternative imaging for bone density evaluation.

As with CT scans, MRI measurements are also not be affected by degenerative cortical changes. Moreover, MRI lacks the radiation risk which make it even more desirable. On the other hand, claustrophobia and metallic implants are unique limitations for this modality. The existing literature has several limitations: Most MRI and CT cohorts were females, which means the results could not be necessarily applied to the general populations. Focusing on such population can be justified since the guidelines for DEXA screeningw and osteoporosis treatment are designed for pre- and post-menopausal women only and no consistent ones for male patients yet [6].

Both MRI and CT studies lack the consistency in cohort's populations, imaging protocols and durations between the imaging. In addition, Studies targeting spine surgical patients are still few and more investigation is needed not only to understand how effective these modalities are in predicting bone strength, but also to acknowledge the reliability in predicting surgical outcomes and complications in such patients. Finally, both CT scans and MRIs have limitations, despite showing superiority over DEXA in BMD measurement in degenerative spine, they still have limited application in pathologies that affect the cancellous bone (eg,: tumors, infections, or fractures). Vertebroplasty and previous spinal instrumentation also can affect the measurements in both modalities.

In conclusion, CT-HU has stronger moderate correlation with DEXA than MRI. Both modalities are superior to DEXA in degenerative spine which gives them a great potential in evaluating bone quality in spine surgery populations. There are inconsistencies among correlation studies regarding cohorts, imaging timing and protocols which can be responsible for the heterogeneity of the results. Studies targeting spine surgical patients are still few and more investigation is needed to understand the correlation better between these modalities and clinical outcomes.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

No funding was received for the design, in the collection, analysis, and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication.

Footnotes

FDA device/drug status: Not applicable.

Author disclosures: AA: Research Support (Investigator Salary): Pfizer (D, Paid directly to institution/employer); Research Support (Investigator Salary): TSRH (B, Paid directly to institution/employer); Research Support (Investigator Salary): Alan L. & Jacqueline B. Stuart Spine Research (C, Paid directly to institution/employer); Research Support (Investigator Salary): Cerapedics (D, Paid directly to institution/employer); Research Support (Investigator Salary): Scoliosis Research Society (E, Paid directly to institution/employer); Research Support (Investigator Salary): Medtronic (E, Paid directly to institution/employer). CHC: Royalties: Alphatec (C–D); Consulting: Medtronic (D); Consulting: Nuvasive (D); Research Support (Investigator Salary): Pfizer (D, Paid directly to institution/employer); Research Support (Investigator Salary): TSRH (B, Paid directly to institution/employer); Research Support (Investigator Salary): Alan L. & Jacqueline B. Stuart Spine Research (C, Paid directly to institution/employer); Research Support (Investigator Salary): Cerapedics (D, Paid directly to institution/employer); Research Support (Investigator Salary): Scoliosis Research Society (E, Paid directly to institution/employer); Research Support (Investigator Salary): Medtronic (D, Paid directly to institution/employer). SDG: Royalties: Medtronic (F); Consulting: Medtronic (F); Consulting: K2M/Stryker: (D); Scientific Advisory Board/Other Office: American Spine Registry; Research Support (Investigator Salary): Pfizer (D, Paid directly to institution/employer); Research Support (Investigator Salary): TSRH (B, Paid directly to institution/employer); Research Support (Investigator Salary): Alan L. & Jacqueline B. Stuart Spine Research (C, Paid directly to institution/employer); Research Support (Investigator Salary): Cerapedics (D, Paid directly to institution/employer); Research Support (Investigator Salary): Scoliosis Research Society (E, Paid directly to institution/employer); Research Support (Investigator Salary): Medtronic (E, Paid directly to institution/employer) JRD: Consulting: Stryker, Depuy, Medtronic (E); Speaking and/or Teaching Arrangements : Stryker, Depuy, Medtronic (E); Research Support (Investigator Salary): Pfizer (D, Paid directly to institution/employer); cont'd on next page…Research Support (Investigator Salary): TSRH (B, Paid directly to institution/employer); Research Support (Investigator Salary): Alan L. & Jacqueline B. Stuart Spine Research (C, Paid directly to institution/employer); Research Support (Investigator Salary): Cerapedics (D, Paid directly to institution/employer); Research Support (Investigator Salary): Scoliosis Research Society (E, Paid directly to institution/employer); Research Support (Investigator Salary): Medtronic (E, Paid directly to institution/employer). JLG: Royalties: Acuity (F); Royalties: Nuvasive (D); Stock Ownership: Intrinsic Spine: Cingulate therapeutics (<1% ownership); Consulting: Medtronic (F); Consulting: Acuity (F); Consulting: Stryker (C); Consulting: Nuvasive (D); Consulting: Mazor (B); Consulting: DePuy (B); Speaking and/or Teaching Arrangements: Baxter (A); Speaking and/or Teaching Arrangements: Broadwater (B); Speaking and/or Teaching Arrangements: Pacira (A); Board of Directors: National Spine Health Foundation; Scientific Advisory Board/Other Office: Stryker (C); Scientific Advisory Board/Other Office: Medtronic (F); Research Support (Investigator Salary): Pfizer (D, Paid directly to institution/employer); Research Support (Investigator Salary): TSRH (B, Paid directly to institution/employer); Research Support (Investigator Salary): Alan L. & Jacqueline B. Stuart Spine Research (C, Paid directly to institution/employer); Research Support (Investigator Salary): Cerapedics (D, Paid directly to institution/employer); Research Support (Investigator Salary): Scoliosis Research Society (E, Paid directly to institution/employer); Research Support (Investigator Salary): Medtronic (E, Paid directly to institution/employer). LYC: Consulting: National Spine Health Foundation (C); Consulting: Orthopedic Research Foundation (B); Scientific Advisory Board/Other Office: University of Louisville Institutional Review Board (Nonfinancial); Scientific Advisory Board/Other Office: The Spine Journal (Nonfinancial); Scientific Advisory Board/Other Office: Spine (Nonfinancial); Scientific Advisory Board/Other Office: Spine Deformity (Nonfinancial); Scientific Advisory Board/Other Office: American Spine Registry (Nonfinancial); Research Support (Investigator Salary): Pfizer (D, Paid directly to institution/employer); Research Support (Investigator Salary): TSRH (B, Paid directly to institution/employer); Research Support (Investigator Salary): Alan L. & Jacqueline B. Stuart Spine Research (C, Paid directly to institution/employer); Research Support (Investigator Salary): Cerapedics (D, Paid directly to institution/employer); Research Support (Investigator Salary): Scoliosis Research Society (E, Paid directly to institution/employer); Research Support (Investigator Salary): Medtronic (E, Paid directly to institution/employer); Research Support (Investigator Salary): SDU Faculty Scholarship (E, Paid directly to institution/employer); Research Support (Investigator Salary): Johnson & Johnson (E, Paid directly to institution/employer); Research Support (Investigator Salary): Cerapedics (F, Paid directly to institution/employer); Research Support (Investigator Salary): IRSs Kursus - og rejsepulje (B, Paid directly to institution/employer); Research Support (Investigator Salary): TrygFonden (F, Paid directly to institution/employer); Region Syddanmark PhD Puljen (E, Paid directly to institution/employer); Research Support (Investigator Salary): SLB Forskningsrad (E, Paid directly to institution/employer); Research Support (Investigator Salary): Sygeforsikring Donation (F, Paid directly to institution/employer); Research Support (Investigator Salary): Sundhedsstyrelsen (F, Paid directly to institution/employer); Research Support (Investigator Salary): SLB Forskningsrad Projektstotte (C, Paid directly to institution/employer).

References

  • 1.Lee B.J., Koo H.W., Yoon S.W., et al. Usefulness of trabecular ct attenuation measurement of lumbar spine in predicting osteoporotic compression fracture: is the L4 trabecular region of interest most relevant? Spine (Phila Pa 1976) 2021;46(3):175–183. doi: 10.1097/BRS.0000000000003756. [DOI] [PubMed] [Google Scholar]
  • 2.Lee S.J., Binkley N, Lubner MG, et al. Opportunistic screening for osteoporosis using the sagittal reconstruction from routine abdominal CT for combined assessment of vertebral fractures and density. Osteoporos Int. 2016;27(3):1131–1136. doi: 10.1007/s00198-015-3318-4. [DOI] [PubMed] [Google Scholar]
  • 3.Gausden E.B., Nwachukwu B.U., Schreiber J.J., et al. Opportunistic use of CT imaging for osteoporosis screening and bone density assessment: a qualitative systematic review. J Bone Joint Surg Am. 2017;99(18):1580–1590. doi: 10.2106/JBJS.16.00749. [DOI] [PubMed] [Google Scholar]
  • 4.Pickhardt P.J., Pooler B.D., Lauder T., et al. Opportunistic screening for osteoporosis using abdominal computed tomography scans obtained for other indications. Ann Intern Med. 2013;158(8):588–595. doi: 10.7326/0003-4819-158-8-201304160-00003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Zou D., Jiang S, Zhou S., et al. Prevalence of osteoporosis in patients undergoing lumbar fusion for lumbar degenerative diseases: a combination of DXA and Hounsfield units. Spine (Phila Pa 1976) 2020;45(7):E406–E410. doi: 10.1097/BRS.0000000000003284. [DOI] [PubMed] [Google Scholar]
  • 6.Pennington Z., Ehresman J., Lubelski D., et al. Assessing underlying bone quality in spine surgery patients: a narrative review of dual-energy X-ray absorptiometry (DXA) and alternatives. Spine J. 2021;21(2):321–331. doi: 10.1016/j.spinee.2020.08.020. [DOI] [PubMed] [Google Scholar]
  • 7.Ahern D.P., McDonnell J.M., Riffault M., et al. A meta-analysis of the diagnostic accuracy of Hounsfield units on computed topography relative to dual-energy X-ray absorptiometry for the diagnosis of osteoporosis in the spine surgery population. Spine J. 2021;21(10):1738–1749. doi: 10.1016/j.spinee.2021.03.008. [DOI] [PubMed] [Google Scholar]
  • 8.Silva B.C., Leslie W.D., Resch H., et al. Trabecular bone score: a noninvasive analytical method based upon the DXA image. J Bone Miner Res. 2014;29(3):518–530. doi: 10.1002/jbmr.2176. [DOI] [PubMed] [Google Scholar]
  • 9.Hocaoglu E., Inci E., Vural M. Could computed tomography Hounsfield unit values of lumbar vertebrae detect osteoporosis? Curr Med Imaging. 2021;17(8):988–995. doi: 10.2174/1573405617999210112193545. [DOI] [PubMed] [Google Scholar]
  • 10.Amin M.F.M., Zakaria W.M.W., Yahya N. Correlation between Hounsfield unit derived from head, thorax, abdomen, spine and pelvis CT and t-scores from DXA. Skeletal Radiol. 2021;50(12):2525–2535. doi: 10.1007/s00256-021-03801-z. [DOI] [PubMed] [Google Scholar]
  • 11.Yu E.W., Thomas B.J., Brown J.K., et al. Simulated increases in body fat and errors in bone mineral density measurements by DXA and QCT. J Bone Miner Res. 2012;27(1):119–124. doi: 10.1002/jbmr.506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Salzmann S.N., Shirahata T., Yang J., et al. Regional bone mineral density differences measured by quantitative computed tomography: does the standard clinically used L1-L2 average correlate with the entire lumbosacral spine? Spine J. 2019;19(4):695–702. doi: 10.1016/j.spinee.2018.10.007. [DOI] [PubMed] [Google Scholar]
  • 13.Zaidi Q., Danisa O.A., Cheng W. Measurement techniques and utility of Hounsfield unit values for assessment of bone quality prior to spinal instrumentation: a review of current literature. Spine (Phila Pa 1976) 2019;44(4):E239–E244. doi: 10.1097/BRS.0000000000002813. [DOI] [PubMed] [Google Scholar]
  • 14.Wichmann J.L., Booz C., Wesarg S., et al. Dual-energy cT-based phantomless in vivo threedimensional bone mineral density assessment of the lumbar spine1. Radiology. 2014;271(3):778–784. doi: 10.1148/radiol.13131952. [DOI] [PubMed] [Google Scholar]
  • 15.Cherif R., Vico L., Laroche N., et al. Dual-energy X-ray absorptiometry underestimates in vivo lumbar spine bone mineral density in overweight rats. J Bone Miner Metab. 2018;36(1):31–39. doi: 10.1007/s00774-017-0813-z. [DOI] [PubMed] [Google Scholar]
  • 16.Flanigan P.M., Mikula A.L., Peters P.A., et al. Regional improvements in lumbosacropelvic Hounsfield units following teriparatide treatment. Neurosurg Focus. 2020;49(2):E11. doi: 10.3171/2020.5.FOCUS20273. [DOI] [PubMed] [Google Scholar]
  • 17.Sollmann N., Löffler M.T., Kronthaler S., et al. MRI-based quantitative osteoporosis imaging at the spine and femur. J Magn Reson Imaging. 2021;54(1):12–35. doi: 10.1002/jmri.27260. [DOI] [PubMed] [Google Scholar]
  • 18.Park S.H., Jeong Y.M., Lee H.Y., et al. Opportunistic use of chest CT for screening osteoporosis and predicting the risk of incidental fracture in breast cancer patients: a retrospective longitudinal study. PLoS One. 2020;15(10) doi: 10.1371/journal.pone.0240084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Li Y.L., Wong K.H., Law M.W-M., et al. Opportunistic screening for osteoporosis in abdominal computed tomography for Chinese population. Arch Osteoporos. 2018;13(1):76. doi: 10.1007/s11657-018-0492-y. [DOI] [PubMed] [Google Scholar]
  • 20.Schreiber J.J., Anderson P.A., Rosas H.G., et al. Hounsfield units for assessing bone mineral density and strength: a tool for osteoporosis management. J Bone Joint Surg Am. 2011;93(11):1057–1063. doi: 10.2106/JBJS.J.00160. [DOI] [PubMed] [Google Scholar]
  • 21.Islamian P.J., Garoosi I, Fard K.A., et al. Comparison between the MDCT and the DXA scanners in the evaluation of BMD in the lumbar spine densitometry. Egypt J Radiol Nucl Med. 2016;47(3):961–967. [Google Scholar]
  • 22.Lee S., Chung C.K., Oh S.H., et al. Correlation between bone mineral density measured by dual-energy x-ray absorptiometry and hounsfield units measured by diagnostic CT in lumbar spine. J Korean Neurosurg Soc. 2013;54(5):384–389. doi: 10.3340/jkns.2013.54.5.384. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kim Y.W., Kim J.H., Yoon S.H., et al. Vertebral bone attenuation on low-dose chest CT: quantitative volumetric analysis for bone fragility assessment. Osteoporos Int. 2017;28(1):329–338. doi: 10.1007/s00198-016-3724-2. [DOI] [PubMed] [Google Scholar]
  • 24.Kim K.J., Kim D.H., Lee J.I., et al. Hounsfield units on lumbar computed tomography for predicting regional bone mineral density. Open Med (Wars) 2019;14:545–551. doi: 10.1515/med-2019-0061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Cohen A., et al. Opportunistic screening for osteoporosis and osteopenia by routine computed tomography scan: a heterogeneous, multiethnic, middle-eastern population validation study. Eur J Radiol. 2021;136 doi: 10.1016/j.ejrad.2021.109568. [DOI] [PubMed] [Google Scholar]
  • 26.Choi M.K., Kim S.M., Lim J.K. Diagnostic efficacy of Hounsfield units in spine CT for the assessment of real bone mineral density of degenerative spine: correlation study between T-scores determined by DEXA scan and Hounsfield units from CT. Acta Neurochir (Wien) 2016;158(7):1421–1427. doi: 10.1007/s00701-016-2821-5. [DOI] [PubMed] [Google Scholar]
  • 27.Chia K.K., Haron J., Nik Malek N.F.S. Accuracy of computed tomography attenuation value of lumbar vertebra to assess bone mineral density. Malays J Med Sci. 2021;28(1):41–50. doi: 10.21315/mjms2021.28.1.6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Alawi M., Begum A., Harraz M., et al. Dual-energy x-ray absorptiometry (DEXA) scan versus computed tomography for bone density assessment. Cureus. 2021;13(2):e13261. doi: 10.7759/cureus.13261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Zou D., Li W., Deng C., et al. The use of CT Hounsfield unit values to identify the undiagnosed spinal osteoporosis in patients with lumbar degenerative diseases. Eur Spine J. 2019;28(8):1758–1766. doi: 10.1007/s00586-018-5776-9. [DOI] [PubMed] [Google Scholar]
  • 30.Kohan E.M., Nemani V.M., Hershman S., et al. Lumbar computed tomography scans are not appropriate surrogates for bone mineral density scans in primary adult spinal deformity. Neurosurg Focus. 2017;43(6):E4. doi: 10.3171/2017.9.FOCUS17476. [DOI] [PubMed] [Google Scholar]
  • 31.Ehresman J., Pennington Z., Schilling A., et al. Novel MRI-based score for assessment of bone density in operative spine patients. Spine J. 2020;20(4):556–562. doi: 10.1016/j.spinee.2019.10.018. [DOI] [PubMed] [Google Scholar]
  • 32.Chang H.K., Hsu T.W., Ku J., et al. Simple parameters of synthetic MRI for assessment of bone density in patients with spinal degenerative disease. J Neurosurg Spine. 2021:1–8. doi: 10.3171/2021.6.SPINE21666. [DOI] [PubMed] [Google Scholar]
  • 33.Elarjani T., Warner T., Nguyen K., et al. Quantifying bone quality using computed tomography Hounsfield units in the mid-sagittal view of the lumbar spine. World Neurosurg. 2021;151:e418–e425. doi: 10.1016/j.wneu.2021.04.051. [DOI] [PubMed] [Google Scholar]
  • 34.Burke C.J., Didolkar M.M., Barnhart H.X., et al. The use of routine non density calibrated clinical computed tomography data as a potentially useful screening tool for identifying patients with osteoporosis. Clin Cases Miner Bone Metab. 2016;13(2):135–140. doi: 10.11138/ccmbm/2016.13.2.135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Shih T.T., Chang C.J., Hsu C.Y., et al. Correlation of bone marrow lipid water content with bone mineral density on the lumbar spine. Spine (Phila Pa 1976) 2004;29(24):2844–2850. doi: 10.1097/01.brs.0000147803.01224.5b. [DOI] [PubMed] [Google Scholar]
  • 36.Shih T.T.-F., Liu H.C., Chang C.J., et al. Correlation of MR lumbar spine bone marrow perfusion with bone mineral density in female subjects. Radiology. 2004;233(1):121–128. doi: 10.1148/radiol.2331031509. [DOI] [PubMed] [Google Scholar]
  • 37.Agrawal K., Agarwal Y., Chopra R.K., et al. Evaluation of MR Spectroscopy and diffusion-weighted MRI in postmenopausal bone strength. Cureus. 2015;7(9):e327. doi: 10.7759/cureus.327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Ergen F.B., Gulal G., Yildiz A.E., et al. Fat fraction estimation of the vertebrae in females using the T2*-IDEAL technique in detection of reduced bone mineralization level: comparison with bone mineral densitometry. J Compu Assist Tomogr. 2014;38(2):320–324. doi: 10.1097/RCT.0b013e3182aa4d9d. [DOI] [PubMed] [Google Scholar]
  • 39.Bandirali M., Leo G.D., Papini G.D.E., et al. A new diagnostic score to detect osteoporosis in patients undergoing lumbar spine MRI. Eur Radiol. 2015;25(10):2951–2959. doi: 10.1007/s00330-015-3699-y. [DOI] [PubMed] [Google Scholar]
  • 40.Shayganfar A., Khodayi M., Ebrahimian S., et al. Quantitative diagnosis of osteoporosis using lumbar spine signal intensity in magnetic resonance imaging. Br J Radiol. 2019;92(1097) doi: 10.1259/bjr.20180774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Shen W., Scherzer R., Gantz M., et al. Relationship between MRI-measured bone marrow adipose tissue and hip and spine bone mineral density in African-American and Caucasian participants: the CARDIA study. J Clin Endocrinol Metab. 2012;97(4):1337–1346. doi: 10.1210/jc.2011-2605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Saad M.M., Ahmed A.T., Mohamed K.E., et al. Role of lumbar spine signal intensity measurement by MRI in the diagnosis of osteoporosis in post-menopausal women. Egypt J Radiol Nucl Med. 2019;50(1):1–7. [Google Scholar]
  • 43.Sebro R., Ashok S.S. A statistical approach regarding the diagnosis of osteoporosis and osteopenia from DXA: are we underdiagnosing osteoporosis? JBMR Plus. 2021;5(2):e10444. doi: 10.1002/jbm4.10444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Kulkarni A.G., Thonangi Y., Pathan S., et al. Should Q-CT be the gold standard for detecting spinal osteoporosis? Spine (Phila Pa 1976) 2022;47(6):E258–E264. doi: 10.1097/BRS.0000000000004224. [DOI] [PubMed] [Google Scholar]

Articles from North American Spine Society Journal are provided here courtesy of Elsevier

RESOURCES