Original Research
FOCUS ON: Musculoskeletal Imaging
August 22, 2014

JOURNAL CLUB: Quantitative Evaluation of Benign and Malignant Vertebral Fractures With Diffusion-Weighted MRI: What Is the Optimum Combination of b Values for ADC-Based Lesion Differentiation With the Single-Shot Turbo Spin-Echo Sequence?

Abstract

OBJECTIVE. The purpose of our study was to determine the optimum combination of b values for calculating the apparent diffusion coefficient (ADC) using a diffusion-weighted (DW) single-shot turbo spin-echo (TSE) sequence in the differentiation between acute benign and malignant vertebral body fractures.
SUBJECTS AND METHODS. Twenty-six patients with osteoporotic (mean age, 69 years; range, 31.5–86.2 years) and 20 patients with malignant vertebral fractures (mean age, 63.4 years; range, 24.7–86.4 years) were studied. T1-weighted, STIR, and T2-weighted sequences were acquired at 1.5 T. A DW single-shot TSE sequence at different b values (100, 250, 400, and 600 s/mm2) was applied. On the DW images for each evaluated fracture, an ROI was manually adapted to the area of hyperintense signal intensity on STIR-hypointense signal on T1-weighted images. For each ROI, nine different combinations of two, three, and four b values were used to calculate the ADC using a least-squares algorithm. The Student t test and Mann-Whitney U test were used to determine significant differences between benign and malignant fractures. An ROC analysis and the Youden index were used to determine cutoff values for assessment of the highest sensitivity and specificity for the different ADC values. The positive (PPV) and negative predictive values (NPV) were also determined.
RESULTS. All calculated ADCs (except the combination of b = 400 s/mm2 and b = 600 s/mm2) showed statistically significant differences between benign and malignant vertebral body fractures, with benign fractures having higher ADCs than malignant ones. The use of higher b values resulted in lower ADCs than those calculated with low b values. The highest AUC (0.85) showed the ADCs calculated with b = 100 and 400 s/mm2, and the second highest AUC (0.829) showed the ADCs calculated with b = 100, 250, and 400 s/mm2. The Youden index with equal weight given to sensitivity and specificity suggests use of an ADC calculated with b = 100, 250, and 400 s/mm2 (cutoff ADC, < 1.7 × 10−3 mm2/s) to best diagnose malignancy (sensitivity, 85%; specificity, 84.6%; PPV, 81.0%; NPV, 88.0%).
CONCLUSION. ADCs calculated with a combination of low to intermediate b values (b = 100, 250, and 400 s/mm2) provide the best diagnostic performance of a DW single-shot TSE sequence to differentiate acute benign and malignant vertebral body fractures.
In previous studies, quantitative diffusion-weighted (DW) MRI has shown its potential to differentiate between benign and malignant vertebral body fractures. The diagnostic potential and optimum sequence parameters of diffusion-weighted echo-planar imaging (DW-EPI) sequences, in particular, have been researched [14] because of the capability of DW-EPI to shorten the acquisition time. However, because of some problems associated with this technique, such as limited spatial resolution, sensitivity to eddy currents and local susceptibility gradients, and chemical-shift artifacts [5], previous studies revealed contradictory results regarding the diagnostic capability of quantitatively evaluated DW-EPI sequences to differentiate between benign and malignant vertebral body fractures [1, 4, 69]. Quantitatively evaluated DW fast spin-echo and DW single-shot turbo spin-echo (TSE) sequences [6, 7, 10, 11] have shown promising results but also a notable overlap between benign and malignant fractures.
The purpose of this study was to determine the optimum combination of b values for calculating the apparent diffusion coefficient (ADC) using a DW single-shot TSE sequence to differentiate acute benign and malignant vertebral body fractures.

Subjects and Methods

Patients

After we obtained approval by the local ethics committee and informed consent, 46 consecutive patients (27 women and 19 men; mean age, 66.6 years; range, 24.7–86.4 years) who were admitted from the emergency department or the orthopedic department with acute benign vertebral body fractures (group 1: 18 women and eight men; mean age, 69 years; range, 31.5–86.2 years) and acute malignant vertebral body fractures (group 2: nine women and 11 men; mean age, 63.4 years; range, 24.7–86.4 years) were included in the study.
We enrolled patients who were over 18 years old with a clinical presentation of back pain for less than 3 months at the level of the vertebral fracture and bone marrow edema on STIR at the fracture site. Pregnancy, contraindications to MRI, diffuse hematologic disorders, contraindications to gadolinium-containing contrast agents, and mental incapacity to voluntarily take part in the study were the exclusion criteria. The patients were assigned to group 1 or group 2 by histology (n = 24), follow-up MRI (n = 11), PET/CT (n = 3), or clinical follow-up including CT after more than 6 months (n = 8).
In group 1, five patients suffered from at least one newly or formerly diagnosed malignant disease (breast cancer, uterine cancer, renal cancer, non-Hodgkin lymphoma); in these and two other patients without known malignant disease, the presence of a malignant fracture was excluded by histology (n = 7). The other patients of this group underwent follow-up MRI (disappearance of edema, no morphologic signs of malignancy in the fractured vertebral body combined with clinical followup) (n = 11) and CT after more than 6 months without morphologic signs of malignancy combined with clinical follow-up (n = 8) to exclude a malignant fracture. In the benign group, one fracture affected the upper thoracic spine (T1–T6), 12 fractures affected the lower thoracic spine (T7–T12), and 14 fractures affected the lumbar spine. In group 2, the primary neoplasms were renal cell carcinoma (n = 1), thyroid carcinoma (n = 1), multiple myeloma (n = 6), breast cancer (n = 5), hypopharyngeal cancer (n = 1), nonseminoma (n = 1), bladder cancer (n = 1), adenocarcinoma (n = 3), and lung cancer (n = 1).
The diagnoses of the malignant origin of the fractures were confirmed by histopathologic examination of specimens obtained during surgery (n = 4), CT-guided biopsy (n = 13), or PET/CT showing a definite pathologic SUV after a time of 7–20 months (n = 3). In the malignant group, two fractures affected the upper thoracic spine (T1–T6), seven fractures affected the lower thoracic spine (T7–T12), and 11 fractures affected the lumbar spine.

MRI

The measurements were performed with a 32-channel 1.5-T MRI scanner (Magnetom Avanto, Siemens Healthcare) using a quadrature spine surface coil. For lesion localization and proper slice positioning, we used a morphologic MRI protocol, which consisted of a T1-weighted (TR/TE, 531/12), STIR (TR/TE, 3790/61; inversion time, 180 ms) and T2-weighted (TR/TE, 4420/118) TSE sequence. Twenty-one sagittal slices with a slice thickness of 3 mm were acquired using a 44 × 44 cm2 FOV and a matrix size of 384 × 384. The total acquisition time for these morphologic images was 6 minutes 30 seconds.
The patients were examined with a fat-saturated DW single-shot TSE sequence (TR/TE, 3000/72; 128 × 92 matrix; flip angle, 180°; bandwidth, 735 Hz/pixel) using b values of 100, 250, 400, and 600 s/mm2. The diffusion weightings were applied in a diagonal direction (gradients in all three dimensions were applied simultaneously). Ten averages were taken to improve the signal-to-noise ratio (SNR) because of the generally low signal intensity of bone marrow. The acquisition time for this DW sequence was 2 minutes 13 seconds.

Postprocessing

Two radiologists with more than 12 and 8 years of experience in musculoskeletal imaging interpreted the images using software developed in-house (PMI 0.4, Platform for Research in Medical Imaging, written in IDL 6.4, ITT Visual Information Solutions). This software was used for image viewing, selection of the ROIs, calculating the different ADC values, and generating the exemplary ADC map (Fig. 1F). For image evaluation, the readers were blinded for age, sex, and possibly underlying malignant diseases. The fracture with the highest signal intensity on the STIR sequence at the level of back pain was selected, assuming that recent benign fractures would show a stronger edema pattern than older ones [12] and that recent malignant fractures also show high signal intensity. In patients with multiple fractures, we chose only one acute fracture for statistical reasons, and both readers evaluated the same vertebral body. In group 1, 12 of 26 patients had other fractures (four had one other fracture and eight had two other fractures) that showed lower signal intensity than the fracture chosen. In group 2, six of 20 patients had other fractures (three had one other fracture, one had two other fractures, and two had three other fractures) that were of smaller extent and at a different spinal level of maximum pain than the fracture chosen. In group 2, 17 of 20 patients had other focal bone marrow lesions in nonfractured vertebral bodies.
Fig. 1A —84-year-old woman with benign fracture of L1 vertebra.
A, STIR image shows hyperintense edema in fractured vertebral body.
Fig. 1B —84-year-old woman with benign fracture of L1 vertebra.
B, Single-shot turbo spin-echo image shows diffusion-weighting of b = 100 s/mm2 and exemplarily highlighted ROI (red).
Fig. 1C —84-year-old woman with benign fracture of L1 vertebra.
C, Corresponding images show diffusion-weightings of 250 (C), 400 (D), and 600 (E) s/mm2.
Fig. 1D —84-year-old woman with benign fracture of L1 vertebra.
D, Corresponding images show diffusion-weightings of 250 (C), 400 (D), and 600 (E) s/mm2.
Fig. 1E —84-year-old woman with benign fracture of L1 vertebra.
E, Corresponding images show diffusion-weightings of 250 (C), 400 (D), and 600 (E) s/mm2.
Fig. 1F —84-year-old woman with benign fracture of L1 vertebra.
F, Example of apparent diffusion coefficient (ADC) map calculated with all b values together shows ADC in ROI is 1.61 × 10−3 mm2/s).
Each ROI was manually adapted to the area of hyperintense signal intensity on STIR-hypointense signal intensity on T1-weighted images. Each ROI was exactly copied to the DW images and corrected for distortions if necessary (Fig. 1). For each ROI, all possible combinations of two, three, and four b values were used to calculate the ADC using a least-squares algorithm [13].

Statistical Evaluation

All analyses were performed using SPSS, version 12, software, and a p value of < 0.05 was defined to indicate statistical significance. Statistical differences were calculated using the Student t test for normally distributed data and the Mann Whitney U test for nonnormally distributed data. Additionally, an ROC analysis was performed and the AUC and the Youden index with the cutoff values for the best sensitivity and specificity were given. The positive (PPV) and negative predictive values (NPV) also were determined.

Results

The ADCs calculated with the combination of b = 100, 250, 400, and 600 s/mm2 in the same cohort were evaluated as part of a previous study by the same authors [6]. All other combinations of b values have been exclusively evaluated for this study. The results are summarized in Tables 1 and 2 and Figures 2 and 3. All calculated ADCs, except for the combination of b = 400 and 600 s/mm2, showed statistically significant differences between benign and malignant vertebral body fractures, with benign fractures having higher ADCs than malignant ones. The use of higher b values resulted in lower ADCs than calculated with low b values (e.g., mean ADC100/250 malignant = 1.67 × 10−3 mm2/s and mean ADC100/250 benign = 2.00 × 10−3 mm2/s vs mean ADC400/600 malignant = 1.23 × 10−3 mm2/s and mean ADC400/600 benign = 1.16 × 10−3 mm2/s). The calculation of the ADCs with three or four higher b values instead of two also decreased the mean ADC (e.g., mean ADC100/250 malignant = 1.67 × 10−3 mm2/s and mean ADC100/250 benign = 2.00 × 10−3 mm2/s vs mean ADC100/250/400 malignant = 1.36 × 10−3 mm2/s and mean ADC100/250/400 benign = 1.8 × 10−3 mm2/s vs mean ADC100/250/400/600 malignant 1.31 = × 10−3 mm2/s and mean ADC100/250/400/600 benign = 1.64 × 10−3 mm2/s).
TABLE 1: Summary of Apparent Diffusion Coefficients (ADCs) Calculated With Different Combinations of b Values in Patients With Osteoporotic and Malignant Vertebral Fractures
Fracture Typeb Value Combinations (s/mm2)
100/250100/400100/600100/250/400100/250/400/600250/400
MeanMedianSDMeanMedianSDMeanMedianSDMeanMedianSDMeanMedianSDMeanMedianSD
Osteoporotic (n = 26)2.002.050.41 (0.38/2.43)1.761.840.32 (0.69/2.21)1.561.610.27 (0.77/1.93)1.801.870.33 (0.67/2.24)1.641.680.31 (0.76/2.13)1.511.510.32 (0.91/2.07)
Malignant (n = 20)1.671.610.64 (0.72,3.23)1.311.420.41 (0.59/2.23)1.261.270.30 (0.71/2.03)1.361.470.43 (0.60/2.29)1.311.320.36 (0.64/2.14)1.061.080.45 (0.15/1.96)
pa0.0032, no normal distribution using Mann-Whitney U test0.0001, normal distribution using unpaired Student t test0.001, normal distribution using unpaired Student t test0.0003, normal distribution using unpaired Student t test0.002, normal distribution using unpaired Student t test0.0002, normal distribution using unpaired Student t test
ROC analysis                  
 AUC0.7560.8500.8150.8290.7810.786
 Youden index0.59620.64620.59620.69620.55770.5577
 Cutoff ADC (× 10−3 mm2/s)≤ 1.81≤ 1.58≤ 1.4≤ 1.7≤ 1.48≤ 1.3
 Sensitivity (%)758075857575
 Specificity (%)84.684.684.684.680.880.8
 PPV (%)78.980.078.981.075.075.0
 NPV (%)81.484.681.588.080.880.8

Note—Data in parentheses are minimum/maximum. PPV = positive predictive value, NPV = negative predictive value.

a
All p values statistically significant.
TABLE 2: Summary of Apparent Diffusion Coefficients (ADCs) Calculated With Different Combinations of b Values in Patients With Osteoporotic and Malignant Vertebral Fractures
Fracture type250/600250/400/600400/600
MeanMedianSDMeanMedianSDMeanMedianSD
Osteoporotic (n = 26)1.411.440.28 (0.85/2.12)1.391.410.26 (0.85/1.78)1.231.220.35 (0.07/1.71)
Malignant (n = 20)1.051.040.40 (0.21/1.84)1.121.130.38 (0.23/1.86)1.161.060.43 (0.73/2.71)
p0.0008a, normal distribution using unpaired Student t test0.0062a, normal distribution using unpaired Student t test0.57, normal distribution using unpaired Student t test
 ROC analysis         
  AUC0.7760.7220.659
  Youden index0.50770.39620.3462
  Cutoff ADC (× 10−3 mm2/s)  ≤ 1.15  ≤ 1.13  ≤ 1.03
  Sensitivity (%)705550
  Specificity (%)80.884.684.6
  PPV(%)73.773.371.4
  NPV (%)77.871.068.7

Note—Data in parentheses are minimum/maximum. PPV = positive predictive value, NPV = negative predictive value.

a
Statistically significant.
Fig. 2 —Graph shows comparison of benign (blue) and malignant (red) vertebral fractures. Box plots summarize apparent diffusion coefficients (ADCs) determined with different combinations of b values. Arrowheads and circles indicate outliers.
Fig. 3 —Graph shows ROC curves for differentiation of benign and malignant vertebral body fractures for apparent diffusion coefficients (ADCs) determined with b = 100 and 400; b = 100, 250, and 400; and b = 100, 250, 400, and 600 s/mm2. Cutoff values for best differentiation (circles) are shown.
The ROC analysis revealed the highest AUC for the ADCs calculated with b = 100 and 400 s/mm2 (AUC = 0.85) (Tables 1 and 2 and Figs. 2 and 3B) and the second highest AUC for the ADCs calculated with b = 100, 250, and 400 s/mm2 (AUC = 0.829) (Tables 1 and 2 and Figures 2 and 3). The analysis of the Youden index with equal weight given to sensitivity and specificity suggests the use of an ADC calculated with the b values of 100, 250, and 400 s/mm2 lower than 1.7 × 10−3 mm2/s to best diagnose malignancy (sensitivity, 85%; specificity, 84.6%; PPV, 81%; NPV, 88%) (Tables 1 and 2).

Discussion

DW imaging is based on the Brownian motion of water molecules in the interstitial tissue. The b value determines the duration and strength of diffusion gradients combining several physical factors, specifically the gyromagnetic ratio (γ), the amplitude of the diffusion gradient pulses (G), the duration of the pulses (δ), and the time between the two diffusion pulses (Δ): b = γ2 × G2 × δ2 (Δ − δ / 3) [14]. The ADC is a quantitative measure of diffusion with regard to several modulating and hindering mechanisms, such as blood flow, restriction in closed spaces, and tortuosity. It can be calculated by collecting images with at least two or more different b values according to the formula: ADC = ln [S2 / S1] / (b1b2) where S1 and S2 are the signal intensities after application of b1 and b2. The ADC can be influenced by other effects on the signal intensity depending on the strength of the applied b values, such as perfusion effects or low SNR [1517].
Although malignant vertebral body fractures are expected to show restricted diffusion (low ADC) due to dense tumor cell packing and restricted extracellular space [18], acute osteoporotic fractures show an increased diffusion (high ADC) due to increased proton mobility in the bone marrow edema [3, 4, 9].
Although the underlying water diffusion properties are independent from sequence parameters and scanners, various studies have shown some variability of the measured ADC values of benign osteoporotic or traumatic fractures, which vary from 0.32 to 2.23 × 10−3 mm2/s, and of malignant fractures or metastases, which vary from 0.19 to 1.04 × 10−3 mm2/s [17]. These variations may partly be explained by the choice of inappropriate measurement parameters, such as b values that are too low or spatial resolutions that are too high, resulting in low SNRs.
The DW pulse sequence used in this study was a single-shot TSE sequence, which is also known as single-shot fast spin-echo or rapid acquisition with relaxation enhancement sequence; a closely related technique involves the use of DW half-Fourier acquisition single-shot TSE sequences. These spin-echo sequences avoid the frequently gross geometric distortions of single-shot echo-planar imaging (EPI) techniques. The cervical and thoracic spine is often particularly affected by geometric distortions in EPI due to magnetic field inhomogeneity in the area of soft tissue to bone or soft tissue to air interfaces [19]. However, the spin-echo signal intensity of the single-shot TSE sequence is insensitive to field inhomogeneity at the cost of lower SNR and slightly increased image blurring in structures with short T2. Increasing the number of averages and the receiver bandwidth can mitigate these effects to a certain degree.
There are different physiologic and physical conditions that influence the determination of the ADC. Whereas low b values include more perfusion effects [15, 16], the use of higher b values decreases the contribution of perfusion and sets the weighting toward diffusion. Relatively high b values greater than approximately 600 s/mm2, on the other hand, may underestimate diffusion due to signal intensities in the same range as the noise level [17]. DW images acquired at low b values (< 150 s/mm2) show good image contrast enhancement that originates from the T2 shine-through effect [20], whereas the use of high b values is limited by the associated decrease in signal intensity and contrast enhancement [1].
Malignant fractures are expected to have restricted diffusion capability because of dense cell packing, whereas in benign fractures normal bone marrow and edema are expected to result in an increase in diffusion. The edematous changes in acute benign vertebral body fractures are probably caused by disrupted microvessels and exudation of fluid into the interstitium caused by increased perfusion, whereas in malignant vertebral fractures the size of the interstitial space might be limited because of the densely packed cells presumably associated with a minor degree of perfusion compared with acute benign fractures [18].
Our results may corroborate these hypotheses. The ADCs in benign fractures were significantly higher than in malignant fractures at most of the possible combinations of b values, which is compatible with the assumption that dense cell packing lowers the ADC. With the use of higher b values, the ADCs in our study also decreased; this can be explained by the lower contribution of perfusion effects and the greater contribution of diffusion at higher b values. The combined use of b = 400 and 600 s/mm2 to calculate the ADC did not show significant differences between benign and malignant fractures and also showed a low AUC (0.659). This may be due to the low SNR of the acquired images because the combination of b = 250, 400, and 600 s/mm2 provides ADCs showing highly significant (p = 0.0062) differences between benign and malignant fractures but also low AUC (0.722). ADCs calculated with low b values (b = 100 and 250 s/mm2) have a lower relative contribution of diffusion (in contrast to perfusion effects) to the signal attenuation, limiting the diagnostic capability in differentiating benign and malignant vertebral fractures, expressed by a low AUC (0.756).
The best diagnostic performance was detected for the ADCs calculated with b = 100 and 400 s/mm2 and also with b = 100, 250, and 400 s/mm2. These low-to-intermediate b values may provide a favorable compromise between adequate diffusion weighting and signal intensity. Although two b values (b = 100 and 400 s/mm2) are sufficient to calculate an ADC, the addition of a third b value (b = 250 s/mm2) may partially reduce the higher SD introduced by the higher b value (b = 400 s/mm2). Omitting much higher b values eliminates low-signal-intensity effects. It is also possible that minor perfusion effects, which might occur at low b values, such as b = 100 and 250 s/mm2, contribute to the ADC at these combinations. Perfusion is known to be different in benign and malignant vertebral body fractures; therefore, it is likely that these perfusion effects at lower b values add to the specificity to discriminate benign from malignant vertebral body fractures [2124]. The highest NPV (88%) with a reasonable balance of sensitivity (85%) and specificity (84.6%) was detected for the combination of b = 100, 250, and 400 s/mm2, with a cutoff ADC value of ≤ 1.7 ×10−3 mm2/s, indicating malignancy. Thus, a fracture detected as benign, is truly benign with a certainty of 88% in that case. Therefore, choosing an adequate combination of b values for calculation of ADCs can improve the diagnostic capability of DW MRI in differentiating acute benign and malignant vertebral fractures.
A possible limitation of our study is the lack of histologic proof of the benign or malignant cause of the vertebral body fracture. For ethical reasons, biopsy or surgical sampling was not possible in all patients. However, clinical and imaging follow-up was performed in all patients to exclude malignancy. Another possible limitation is that investigating different malignant entities might cause a mixed behavior at different diffusion-weightings because of the different tissue composition, perfusion, and size of the extracellular and intracellular spaces, which might influence the signal characteristics of the fractures.

Conclusion

ADCs calculated with the combination of low-to-intermediate b values (b = 100, 250, and 400 s/mm2) provide the best diagnostic performance of a DW single-shot TSE sequence to differentiate acute benign and malignant vertebral body fractures, providing a well-balanced ratio of diffusion-weighting, perfusion effects, and SNR. The choice of an adequate combination of b values can reduce the overlap of ADCs of benign and malignant vertebral body fractures and can improve the diagnostic significance of quantitative DW MRI.

Acknowledgment

We thank Melvin D'Anastasi for linguistic revision of the manuscript.

References

1.
Tang G, Liu Y, Li W, Yao J, Li B, Li P. Optimization of b value in diffusion-weighted MRI for the differential diagnosis of benign and malignant vertebral fractures. Skeletal Radiol 2007; 36:1035–1041
2.
Oztekin O, Ozan E, Hilal Adibelli Z, Unal G. Abul Y. SSH-EPI diffusion-weighted MR imaging of the spine with low b values: is it useful in differentiating malignant metastatic tumor infiltration from benign fracture edema? Skeletal Radiol 2009; 38:651–658
3.
Balliu E, Vilanova JC, Pelaez I, et al. Diagnostic value of apparent diffusion coefficients to differentiate benign from malignant vertebral bone marrow lesions. Eur J Radiol 2009; 69:560–566
4.
Chan JH, Peh WC, Tsui EY, et al. Acute vertebral body compression fractures: discrimination between benign and malignant causes using apparent diffusion coefficients. Br J Radiol 2002; 75:207–214
5.
Le Bihan DJ. Differentiation of benign versus pathologic compression fractures with diffusion-weighted MR imaging: a closer step toward the “holy grail” of tissue characterization? Radiology 1998; 207:305–307
6.
Geith T, Schmidt G, Biffar A, et al. Comparison of qualitative and quantitative evaluation of diffusion-weighted MRI and chemical-shift imaging in the differentiation of benign and malignant vertebral body fractures. AJR 2012; 199:1083–1092
7.
Oner AY, Aggunlu L, Akpek S, Tali T, Celik A. Diffusion-weighted imaging of the appendicular skeleton with a non-Carr-Purcell-Meiboom-Gill single-shot fast spin-echo sequence. AJR 2007; 189:1494–1501
8.
Herneth AM, Philipp MD, Naude J, et al. Vertebral metastases: assessment with apparent diffusion coefficient. Radiology 2002; 225:889–894
9.
Herneth AM, Naude J, Philipp M, Beichel R, Trattnig S, Imhof H. The value of diffusion-weighted MRT in assessing the bone marrow changes in vertebral metastases. [in German] Radiologe 2000; 40:731–736
10.
Zhou XJ, Leeds NE, McKinnon GC, Kumar AJ. Characterization of benign and metastatic vertebral compression fractures with quantitative diffusion MR imaging. AJNR 2002; 23:165–170
11.
Raya JG, Dietrich O, Birkenmaier C, Sommer J, Reiser MF, Baur-Melnyk A. Feasibility of a RARE-based sequence for quantitative diffusion-weighted MRI of the spine. Eur Radiol 2007; 17:2872–2879
12.
Yamato M, Nishimura G, Kuramochi E, Saiki N, Fujioka M. MR appearance at different ages of osteoporotic compression fractures of the vertebrae. Radiat Med 1998; 16:329–334
13.
Stejskal EO, Tanner JE. Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. J Chem Phys 1965; 42:288
14.
Le Bihan D. Intravoxel incoherent motion imaging using steady-state free precession. Magn Reson Med 1988; 7:346–351
15.
Le Bihan D, Breton E, Lallemand D, Aubin ML, Vignaud J, Laval-Jeantet M. Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging. Radiology 1988; 168:497–505
16.
Yeung DK, Wong SY, Griffith JF, Lau EM. Bone marrow diffusion in osteoporosis: evaluation with quantitative MR diffusion imaging. J Magn Reson Imaging 2004; 19:222–228
17.
Dietrich O, Biffar A, Reiser MF, Baur-Melnyk A. Diffusion-weighted imaging of bone marrow. Semin Musculoskelet Radiol 2009; 13:134–144
18.
Baur A, Dietrich O, Reiser M. Diffusion-weighted imaging of bone marrow: current status. Eur Radiol 2003; 13:1699–1708
19.
Dietrich O, Biffar A, Baur-Melnyk A, Reiser M. Technical aspects of MR diffusion imaging of the body. Eur J Radiol 2010; 76:314–322
20.
Herneth AM, Friedrich K, Weidekamm C, et al. Diffusion weighted imaging of bone marrow pathologies. Eur J Radiol 2005; 55:74–83
21.
van der Woude HJ, Verstraete KL, Hogendoorn PC, Taminiau AH, Hermans J, Bloem JL. Musculoskeletal tumors: does fast dynamic contrast-enhanced subtraction MR imaging contribute to the characterization? Radiology 1998; 208:821–828
22.
Moulopoulos LA, Maris TG, Papanikolaou N, Panagi G, Vlahos L, Dimopoulos MF. Detection of malignant bone marrow involvement with dynamic contrast-enhanced magnetic resonance imaging. Ann Oncol 2003; 14:152–158
23.
Hawighorst H, Libicher M, Knapp M, Moehler T, Kauffmann GW, van Kaick G. Evaluation of angiogenesis and perfusion of bone marrow lesions: role of semiquantitative and quantitative dynamic MRI. J Magn Reson Imaging 1999; 10:286–294
24.
Stäbler A, Baur A, Bartl R, Munker R, Lamerz R, Reiser MF. Contrast enhancement and quantitative signal analysis in MR imaging of multiple myeloma: assessment of focal and diffuse growth patterns in marrow correlated with biopsies and survival rates. AJR 1996; 167:1029–1036

APPENDIX 1: AJR JOURNAL CLUB

Study Guide

Quantitative Evaluation of Benign and Malignant Vertebral Fractures With Diffusion-Weighted MRI: What Is the Optimum Combination of b Values for ADC-Based Lesion Differentiation With the Single-Shot Turbo Spin-Echo Sequence?

Alan Mautz1, Joseph J. Budovec2, Margaret Mulligan2
1The Aroostook Medical Center, Presque Isle, ME
2Medical College of Wisconsin, Milwaukee, WI
Introduction
1. What was the purpose of this study? Do the authors provide a suitable rationale for conducting the study? How would you state the authors' hypothesis?
2. What are the clinical ramifications of the study, assuming that the authors' hypothesis is not refuted? What is this study's null hypothesis?
Methods
3. What were the inclusion criteria for the study? What were the exclusion criteria? If you were designing a similar study, would you include additional exclusion criteria?
4. Would another investigator be able to reproduce the study based on the authors' description?
5. What statistical analyses were used to evaluate the study's findings?
6. What are the limitations of this study? Are these limitations adequately discussed?
Results
7. Was the research question answered? Were the hypotheses resolved?
8. Are the b values cited by the study as the most optimal for apparent diffusion coefficient (ADC) calculation likely to be well-suited for clinical practice?
9. Do the results of the study corroborate the study's hypothesis?
Physics
10. Briefly review how diffusion-weighted imaging is performed, how ADC maps are generated, and how ADCs obtained at differing b values allow the generation of diffusion-weighted imaging?
Discussion
11. The study reports both positive predictive values and negative predictive values for ADCs, allowing differentiation of benign and malignant vertebral fractures. Are the values sufficient to allow definitive diagnosis?
12. The study acknowledges that some of the patients with benign vertebral fractures had additional nonacute vertebral fractures and that some patients with malignant vertebral fractures had additional marrow lesions in nonfractured vertebrae. In studies of this nature, should readers be blinded to additional but related findings?
13. When designing a study, what is the best way to blind the image interpreters?
14. What criteria does your institution or practice use to differentiate benign from malignant vertebral fractures? Are the results of this study readily transferable to your practice? Should diffusion-weighted imaging be used routinely in spine imaging?
*
Please note that the authors of the Study Guide are distinct from those of the companion article.

Background Reading

1.
Oztekin O, Ozan E, Hilal Adibelli Z, Unal G, Abal Y. SSH-EPI diffusion-weighted MRI of the spine with low b values: is it useful in differentiating malignant meta-static tumor infiltration from benign fracture edema? Skeletal Radiol 2009; 38:651–658
2.
Herneth AM, Phillip MO, Naude J, et al. Vertebral metastases: assessment with apparent diffusion coefficient. Radiology 2002; 225: 889–894

FOR YOUR INFORMATION

This article has been selected for AJR Journal Club activity. The accompanying Journal Club study guide can be found on the following page.

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: 582 - 588
PubMed: 25148160

History

Submitted: July 30, 2013
Accepted: October 18, 2013

Keywords

  1. diffusion-weighted MRI
  2. malignant vertebral fracture
  3. osteoporotic vertebral fracture

Authors

Affiliations

Tobias Geith
Institute of Clinical Radiology, LMU University of Munich-Campus Grosshadern, Marchioninistrasse 15, 81377 Munich, Germany.
Gerwin Schmidt
Institute of Clinical Radiology, LMU University of Munich-Campus Grosshadern, Marchioninistrasse 15, 81377 Munich, Germany.
Andreas Biffar
Institute of Clinical Radiology, LMU University of Munich-Campus Grosshadern, Marchioninistrasse 15, 81377 Munich, Germany.
Olaf Dietrich
Institute of Clinical Radiology, LMU University of Munich-Campus Grosshadern, Marchioninistrasse 15, 81377 Munich, Germany.
Hans Roland Duerr
Department of Orthopedic Surgery, LMU University of Munich-Campus Grosshadern, Munich, Germany.
Maximilian Reiser
Institute of Clinical Radiology, LMU University of Munich-Campus Grosshadern, Marchioninistrasse 15, 81377 Munich, Germany.
Andrea Baur-Melnyk
Institute of Clinical Radiology, LMU University of Munich-Campus Grosshadern, Marchioninistrasse 15, 81377 Munich, Germany.

Notes

Address correspondence to T. Geith ([email protected]).

Funding Information

This work was supported by Deutsche Forschungsgemeinschaft (DFG) grant no. DI 1413/1-1.

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media