Skip to main content
NCBI home page
The British Journal of Radiology logoLink to The British Journal of Radiology
. 2019 Jul 24;92(1102):20181051. doi: 10.1259/bjr.20181051

Comparison of 3T diffusion-weighted MRI and 18F-FDG PET/CT in musculoskeletal tumours: quantitative analysis of apparent diffusion coefficients and standardized uptake values

So-Yeon Lee 1, Won-Hee Jee 1,2,1,2,✉, Ie Ryung Yoo 3, Joon-Yong Jung 1, Soo-A Im 1, Yang-Guk Chung 4, Jin Hyoung Kang 5
PMCID: PMC6774587  PMID: 31322913

Abstract

Objective:

To determine whether the apparent diffusion coefficient (ADC) on 3T MR imaging including diffusion-weighted MR imaging (DWI) correlate with the standardized uptake value (SUV) on 18F-FDG PET/CT in musculoskeletal tumours.

Methods:

This retrospective cohort study included 57 patients (36 males, 21 females, mean age 54 years, range 12–90 years) with pathologically confirmed soft tissue (n = 32) and bone (n = 25) tumours who underwent 3T MR imaging including DWI and whole-body 18F-FDG PET/CT before treatment. 14 patients had follow-up MR imaging and 18F-FDG PET/CT after treatment. The minimum (ADCmin) and mean (ADCmean) ADCs of musculoskeletal tumour, ADC of normal skeletal muscle (ADCmus), SUVmax and SUVmean of musculoskeletal tumour were obtained. Correlation between ADCs and SUVs was assessed using Pearson correlation coefficients (r). ADCmin and SUVmax were compared between pretreatment and posttreatment by t-test.

Results:

There was inverse correlation between SUVmax and the ratio ADCmin/ADCmus (r = - 0.505 to - 0.495, p ≤ 0.001) and between SUVmean and the ratio ADCmean/ADCmus (r = - 0.501 to - 0.493, p = 0.001). After treatment ADC was significantly increased whereas SUV was significantly decreased (p = 0.001). There was significant correlation in percent change between the initial and follow-up values of ADCmin and SUVmax (r = 0.750 to 0.773, p ≤ 0.005). The ADCmin was increased by 163% and SUVmax was decreased by 61% in 11 patients with treatment response.

Conclusion:

ADC at 3T MR DWI and SUV at 18F-FDG PET/CT have an inverse correlation in musculoskeletal tumours.

Advances in knowledge:

Our study showed that ADC at 3T DWI and SUV at 18F-FDG PET/CT had an inverse correlation in musculoskeletal tumours.

Introduction

Musculoskeletal sarcoma accounts for 0.9% of newly diagnosed cancers in adults, with soft tissue sarcoma approximately four times more common than bone.1 The age-adjusted 5-year survival rate is 65.3% for soft tissue sarcoma and 66.6% for bone sarcoma.2 It is important for radiologists and physicians to survey systemic and local recurrence as well as diagnosis to obtain optimal outcome.

Positron emission tomography (PET)/computerized tomography (CT) imaging is a nuclear medicine technique which combines functional information from PET and anatomic information from CT scan. The radio-labeled tracer 2-deoxy-2-[18F]fluoro-D-glucose (18F-FDG) is commonly used for oncology imaging, since the distribution of 18F-FDG is a good reflection of the distribution of glucose uptake. 18F-FDG is a glucose analog, and then actively transported into the cell by a group of structurally related glucose transport proteins (GLUT). Tumour cells display increased number of glucose transporters, particularly GLUT-1 and GLUT-3.3 Tumour cells are metabolically active with high mitotic rate. These mechanisms allow for tumour cells to retain higher levels of 18F-FDG when compared to normal tissues. The standardized uptake value (SUV) is commonly used as a semi-quantitative measure of 18F-FDG accumulation relative to normal tissue. The SUV on 18F-FDG PET/CT correlate with glucose metabolism of tumour and increase in malignancy. 18F-FDG PET/CT helps differentiate malignant musculoskeletal tumours from non-malignant ones.4–6 In addition, 18F-FDG PET/CT has been used for monitoring treatment response in musculoskeletal tumours, because these functional imaging techniques can detect changes of tumour microenvironment which are not detectable by tumour size reduction on conventional imaging.7–9 In general, decreasing SUV indicates treatment response in the musculoskeletal tumour.7

MRI including diffusion-weighted imaging (DWI) is increasingly used in oncology for the detection, differentiation between benignity and malignancy, and treatment-response evaluation of various cancers.10 Apparent diffusion coefficient (ADC) is used for quantitative analysis. ADC on DWI depends on cellularity and extracellular space with cell membrane integrity.10 ADC is lower in malignant tumours than non-malignant lesions.11–13 Prior studies have found that DWI is useful for differentiation of malignant musculoskeletal tumours from non-malignant ones using qualitative and quantitative analyses.11–14 In addition, DWI may be an effective tool in assessing the response to treatment in musculoskeletal tumours.14–18

For the evaluation of treatment response in the musculoskeletal tumours, 18F-FDG PET/CT has been widely used.7–9 However, there is an accumulation of radiation dose due to repeated 18F-FDG PET/CT follow-up after treatment. The advantage of DWI on MR imaging is lack of radiation hazard as well as no requirement of contrast-enhancement, as compared with 18F-FDG PET/CT and contrast-enhanced MR imaging. Both ADC and SUV correlate with cellularity.10,15–19 Thus, we hypothesized that ADC on DWI might inversely correlate with SUV on 18F-FDG PET/CT in musculoskeletal tumours and could help assess the response of treatment. In that case, MR imaging may be particularly useful for repeated short-term follow-up for early assessment of treatment response. However, there have been a few controversial reports regarding correlation between ADC and SUV in musculoskeletal tumours.20–23

The purpose of our study was to retrospectively determine whether ADC on 3T DWI correlates with SUV on 18F-FDG PET/CT in musculoskeletal tumours.

Methods and materials

This study was approved by our institutional review board and the requirement for informed consents was waived for this retrospective study.

Patient population

A total of 531 consecutive patients underwent 3T MR imaging including DWI for musculoskeletal mass in our institution from April 2009 to March 2013. Among them 474 patients were excluded according to exclusion criteria (Figure 1). Thus, 57 patients (36 males, 21 females, mean age 54 years, range 12–90 years) with pathologically confirmed soft tissue (n = 32) and bone (n = 25) tumours were included in this study (Table 1). Mean interval between MR imaging and 18F-FDG PET/CT was 6 days (range, 0–28 days). 14 patients had undergone both 3T MR imaging including DWI and 18F-FDG PET/CT after treatment. Mean interval between follow-up MR imaging and follow-up 18F-FDG PET/CT was 34 days (range, 0–95 days).

Figure 1. .

Figure 1. 

Open in a new tab

Flow diagram of the study. Note- MRI; DWI = diffusion-weighted imaging; 18F-FDG PET/CT = fluorine-18-fluorodeoxyglucose positron emission tomography/CT; FU = follow-up

Table 1. .

Histological Types of Included Musculoskeletal Tumours

Classification Histological type Number of cases
Bone tumour
 Non-malignant Benign histiocytoma 1
  Enchondroma 1
 Malignant Metastasis 13
  Osteosarcoma 4
  Lymphoma 3
  Solitary plasmacytoma of bone 1
  Acute lymphocytic leukemia 1
  Chondrosarcoma 1
 Total 25
Soft tissue tumour
 Non-malignant Neurilemmoma 2
  Myoepithelioma 1
 Malignant Metastasis 10
  Leiomyosarcoma 3
  Liposarcoma 3
  Lymphoma 3
  Extraskeletal osteosacrcoma 2
  Malignant melanoma 2
  Angiosarcoma 1
  Extraskeletal Ewing sarcoma 1
  Fibrosarcoma 1
  Malignant solitary fibrous tumour 1
  Malignant peripheral nerve sheath tumour 1
  Poorly differentiated sarcoma 1
 Total 32
Total 57
Open in a new tab

MR imaging protocols

MR imaging was performed at 3T (Verio; Siemens Medical Systems, Erlangen, Germany) with various coils depending on the anatomic regions. The standard MR imaging protocols included longitudinal fat-suppressed T2-weighted turbo spin-echo (TSE) sequence, axial T1 (repetition time msec/echo time msec, 680–870/11–21)- and T2 (4000–5600/63–83)- weighted TSE sequences, and longitudinal and axial fat-suppressed contrast-enhanced T1-weighted TSE sequences. MR parameters were as follows; FOV 80–220 mm; section thickness 2–5 mm; no intersection gap; matrix size 512 × 256; turbo factors, three for T1-weighted TSE, 13 for T2-weighted TSE; number of excitation, 1. Before contrast-enhancement, a single-shot spin-echo echo-planar DWI sequence with an EPI factor of 56 was obtained in the axial plane. A parallel imaging technique using GRAPPA (GeneRalized Autocalibrating Partially Parallel Acquisitions) was combined with an acceleration factor of 2. MR parameters for DWI are as follows; matrix size, 64 × 45–120 × 128; 500–8700/71–85; NEX 3–5. Various combination of b values were used; b values of 0, 400, 800 sec/mm2 (n = 11) and b values of 0, 300, 800, 1400 sec/mm2 (n = 46). ADC map was obtained common b values of 0 and 800 sec/mm2. Pixel-based ADC maps were created based on mono-exponential calculation from DWI using a commercial software and workstation (Leonardo MR Workplace; Siemens Medical Solution, Erlangen, Germany).

18F-FDG PET/CT Protocols

All patients fasted for at least six hours prior to 18F-FDG PET/CT study. 18F-FDG was injected intravenously (370–555 MBq) after confirming a blood glucose level less than 130 mg dl–1. At 30 min after the 18F-FDG injection, PET/CT studies were acquired on combined PET/CT in-line systems, either Biograph Duo or Biograph Truepoint (Siemens Medical Systems, Knoxville, TN). All patients were in a supine position with their arms raised. CT was scanned from the top of the head to the upper thigh, then PET followed immediately over the same body region. CT parameters were as follows: 130 kVp, 80 mA and 5 mm slice thickness; 120 kVp, 50 mA and 5 mm slice thickness.

MR imaging analysis

Two musculoskeletal radiologists (W.H.J,, S.Y.L with 17 and 6 years of experience in musculoskeletal radiology, respectively) retrospectively interpreted MR images and DWI independently. The readers were blinded to the imaging reports, clinical history and results of pathologic examination. MR images were randomly ordered. The solid portion of the tumour was selected after correlation with standard MR imaging. Sites of hemorrhage, necrosis, or calcification were carefully avoided. For quantitative analyses, ADCmin and ADCmean were independently measured by two readers.12,24 ADCmin was measured by manually drawn regions of interest (ROI) on the ADC map within solid portion that presented hyperintense signal on DWI with high b value on a picture archiving and communication system (PACS). For selecting the lowest value of ADC, ROI was drawn three to five times and the minimum of them were recorded as ADCmin. ADCmean was defined as mean ADC value obtained from ROI drawing entire mass on one axial plane except for peripheral most portions in order to avoid partial-volume effects. In addition mean ADC for normal appearing skeletal muscle (ADCmus) was obtained. ADCmean/mus was calculated from ADCmean of the tumour divided by ADCmus, and ADCmin/mus was defined as ADCmin of the tumour divided by ADCmus. Both ADCmean/mus and ADCmin/mus were referred to as normalized ADC.

18F-FDG PET/CT Analysis

One nuclear physician (I.R.Y., 11 years of experience in nuclear medicine) retrospectively interpreted PET/CT after correlation with MR imaging. All 18F-FDG PET/CT images were reviewed at a workstation with fusion software (Syngo, Siemens) that provided multiplanar reformatted images and displayed PET images after attenuation correction, CT images and 18F-FDG PET/CT fusion images. Using manually drawing ROI mean (SUVmean) and maximum SUV value (SUVmax) of the tumour were obtained. For selecting the highest value of SUV, ROI were drawn three to five times and the maximum of them were recorded as SUVmax. SUVmean was defined as mean SUV value obtained from ROI drawing entire mass on one axial plane except for peripheral most portions in order to avoid partial-volume effects.

Statistical analysis

Interobserver measurement reliability for the ADC measurement was assessed by using intraclass correlation coefficient (ICC) as follows25: An ICC value of less than 0.40 was indicative of poor, 0.40–0.75 fair to good, more than 0.75 excellent agreement. Correlation between ADCs and SUVs was assessed using Pearson correlation coefficients (r). An r values from 0 to 0.25 or from 0 to −0.25 were indicative of the absence of correlation, from 0.25 to 0.50 or from −0.25 to −0.50, poor, from 0.50 to 0.75 or −0.50 to −0.75, moderate to good, from 0.75 to 1 or from −0.75 to −1 very good to excellent correlation.26 ADCmin and SUVmax were compared between pretreatment and posttreatment by t-test. As for evaluation of quantitative analysis of follow-up imaging, we calculated the difference and percent change.27 Differences for ADCs were calculated by subtraction of initial values from follow-up values, whereas differences for SUVs were calculated by subtraction of follow-up values from initial values, therefore positive value means treatment response for both ADC and SUV.

Percent change was defined as the ratio between difference and initial value. For all tests, p values of less than 0.05 were considered indicative of statistically significant difference. All statistical analyses were performed using the commercial software (SPSS, version 19, SPSS, Chicago, III and MedCalc Software, version 11.3.0.0, Mariakerke, Belgium).

Results

MR imaging analysis

ADCmin and ADCmean of non-malignant musculoskeletal tumours were 1846 ± 319 and 1881 ± 320㎛2/sec for reader 1 and 1861 ± 364 and 1945 ± 375㎛2/sec for reader 2, respectively. ADCmin and ADCmean of malignant musculoskeletal tumours were 953 ± 365 and 1033 ± 392 ㎛2/sec for reader 1 and 939 ± 370㎛2/sec and 1042 ± 393 for reader 2, respectively. Table 2 shows detailed data including ADCmean/mus and ADCmin/mus. Interobserver agreement for ADC measurement was excellent (ICC = 0.985–0.988).

Table 2. .

ADCs in Musculoskeletal Tumours

Classification ADCmin (㎛2/sec) ADCmean (㎛2/sec) ADCmin/mus ADCmean/mus
Non-malignant bone tumour
 Reader 1 1779.0 ± 128.7 1822.0 ± 151.3 1.37 ± 0.02 1.41 ± 0.01
 Reader 2 1743.0 ± 107.5 1900.0 ± 309.7 1.31 ± 0.09 1.41 ± 0.04
Non-malignant soft tissue tumour
 Reader 1 1891.0 ± 433.6 1920.3 ± 433.6 1.34 ± 0.47 1.36 ± 0.47
 Reader 2 1940.0 ± 486.1 1974.3 ± 480.2 1.46 ± 0.49 1.48 ± 0.50
Malignant bone tumour
 Reader 1 955.9 ± 301.3 1109.6 ± 295.9 0.69 ± 0.21 0.73 ± 0.19
 Reader 2 953.1 ± 334.0 1181.5 ± 310.8 0.70 ± 0.21 0.77 ± 0.19
Malignant soft tissue tumour
 Reader 1 951.5 ± 414.5 1004.4 ± 424.4 0.67 ± 0.31 0.70 ± 0.31
 Reader 2 927.8 ± 402.1 989.0 ± 412.4 0.63 ± 0.28 0.67 ± 0.29
Open in a new tab

ADC, apparent diffusion coefficient; ADCmean, mean ADC; ADCmean/mus, ADCmean divided by ADC of normal skeletal muscle; ADCmin, minimum ADC; ADCmin/mus, ADCmin divided by ADC of normal skeletal muscle.

Data are mean and standard deviation.

18F-FDG PET/CT Analysis

SUVmax and SUVmean of non-malignant bone tumour were 3.45 ± 1.91 and 3.20 ± 2.26, respectively. SUVmax and SUVmean of non-malignant soft tissue tumour were 3.51 ± 2.38 and 2.08 ± 1.64, respectively. SUVmax and SUVmean of malignant bone tumour were 8.46 ± 5.18 and 4.89 ± 2.83, respectively. SUVmax and SUVmean of malignant soft tissue tumour were 10.77 ± 7.22 and 6.26 ± 3.95, respectively.

Correlation of ADC and SUV before treatment

The location of the most cellular area on MR imaging with DWI matched the most metabolic area on 18F-FDG PET/CT within a lot of tumors (Figure 2). ADCs and SUVs were inversely correlated (Table 3). There were poor inverse correlations between the ADCmin and the SUVmax (r = −0.431, p = 0.001 for reader 1, r = −0.501, p = 0.001 for reader 2) and between the ADCmean and the SUVmean (r = −0.430, p = 0.003 for reader 1, r = −0.458, p = 0.002 for reader 2) (Figure 3). The poor inverse correlations were observed between the ADCmin/mus and the SUVmax (r = −0.505, p < 0.001 for reader 1, r = −0.495, p = 0.001 for reader 2) and between the ADCmean/mus and the SUVmean (r = −0.493, p = 0.001 for reader 1, r = - 0.501, p = 0.001 for reader 2).

Figure 2. .

Figure 2. 

Open in a new tab

A 76-year-old female with fibrosarcoma in the thigh. a An axial fat-suppressed contrast-enhanced T1-weighted magnetic resonance (MR) image shows intense enhancement of the tumour in the vastus lateralis muscle (arrow). b On the corresponding axial apparent diffusion coefficient (ADC) map (b = 0 and 800 sec/mm2), minimum ADC (ADCmin, blue ROI) and mean ADC (ADCmean, yellow ROI,) were measured as 792.5 ㎛2/sec and 902.5㎛2/sec, respectively (arrow), suggesting malignancy. c An axial fusion image of T2-weightedimage and diffusion-weighted imaging (DWI) (b = 800 sec/mm2) at same level shows hyperintense signal intensity of the tumour (arrow). d An axial fusion image of 18F-FDG PET/CT shows high 18F-FDG uptake of the tumour, where maximum standardized uptake values (SUV) (SUVmax) and mean SUV (SUVmean) of the tumour (arrow) are 8.1 and 5.9, respectively. Note lateral portion of the tumour show more increased SUV than that of medial portion of the tumour, which is corresponding to that of DWI.

Table 3. .

Correlation Between ADCs and SUVs in Musculoskeletal Tumours Before Treatment

MRI with DWI Parameters SUVmax SUVmean
Correlation Coefficienta p Correlation Coefficienta p
ADCmin
 Reader 1 −0.431 (-0.622 to -0.192) 0.001 −0.435 (-0.646 to -0.163) 0.003
 Reader 2 −0.501 (-0.692 to -0.243) 0.001 −0.475 (-0.674 to -0.210) 0.001
ADCmean
 Reader 1 −0.421 (-0.639 to -0.145) 0.004 −0.430 (-0.643 to -0.157) 0.003
 Reader 2 −0.450 (-0.657 to -0.180) 0.002 −0.458 (-0.662 to -0.190) 0.002
ADCmin/mus
 Reader 1 −0.505 (-0.695 to -0.248) <0.001 −0.503 (-0.694 to -0.246) <0.001
 Reader 2 −0.495 (-0.688 to -0.236) 0.001 −0.503 (-0.694 to -0.245) <0.001
ADCmean/mus
 Reader 1 −0.492 (-0.687 to -0.232) 0.001 −0.493 (-0.687 to -0.234) 0.001
 Reader 2 −0.489 (-0.684 to -0.228) 0.001 −0.501 (-0.692 to -0.243) 0.001
Open in a new tab

ADC, apparent diffusion coefficient; ADCmean, mean ADC; ADCmin, minimum ADC; DWI, diffusion-weighted imaging; SUV, standardized uptake values; SUVmax, maximum SUV; SUVmean, mean SUV.

a

95% confidence interval within parenthesis.

Figure 3.

Figure 3.

Open in a new tab

a Correlation of minimum apparent diffusion coefficient (ADC) (ADCmin) and maximum standardized uptake values (SUV) (SUVmax) of musculoskeletal tumours. b Correlation of mean ADC (ADCmean) and mean SUV (SUVmean) of musculoskeletal tumours.

ADC and SUV after treatment

There were significant differences of both SUVs and ADCs between pre-treatment and post-treatment images (Figure 4). ADCmin was significantly increased after treatment: from 886 ± 254 ㎛2/sec to 1703 ± 703㎛2/sec for reader 1 (p = 0.001); and from 843 ± 278 ㎛2/sec to 1756 ± 777㎛2/sec for reader 2 (p = 0.001). SUVmax was significantly decreased after treatment: from 3.72 ± 2.63 to 3.48 ± 1.93 (p = 0.001). ADCmin and SUVmax of the treated tumours were not significantly correlated with each other; r = −0.360, p = 0.206 by reader 1, r = −0.385, p = 0.174, respectively.

Figure 4. .

Figure 4. 

Open in a new tab

A 19-year-old male with acute lymphocytic leukemia. Pre-treatment (a, b) and post-treatment (c, d) a An axial fusion image of T2-weighted image and diffusion-weighted imaging (DWI) (b = 800 sec/mm2) at same level shows hyperintense signal intensity of the tumour (arrow and arrowhead). b On axial fusion image of 18F-FDG PET/CT, maximum standardized uptake values (SUV) (SUVmax) of the tumour (arrow and arrowhead) is 12.2. c An axial fusion image of T2-weighted image and DWI (b = 800 sec/mm2) at same level shows hyperintense signal intensity of the tumour (arrow). However, previously noted extraosseous high signal is not present. d On post-treatment axial fusion image of 18F-FDG PET/CT, SUVmax decreased to 0.9, suggesting treatment-response.

Both MR imaging and 18F-FDG PET/CT showed treatment responses in 11 lesions with increase in ADCmin (percent change, 147±102% for reader 1, 178±135% for reader 2) and decrease in SUVmax (percent change, 61±23%). The ADC of one lesion returned to the ADC range of normal bone marrow, decreasing ADC min (−31% by reader 1,–33% for reader 2), whereas SUVmax of the lesion decreased (87%). In another tumour, ADCmin of the lesion was decreased (−32% by reader 1,–26% by reader 2), although SUVmax was decreased (13%), due to reparative sclerosis after treatment. One lesion showed no response on both MR imaging and 18F-FDG PET/CT with decrease in ADCmin (−12% by reader 1,–8% by reader 2) and increase in SUVmax (−61%).

The difference of ADCmin and SUVmax showed moderate to good correlation: r = 0.613 for reader 1 (p = 0.020) and r = 0.594 for reader 2 (p = 0.025). Percent change of ADCmin and SUVmax showed moderate to good correlation: r = 0.623 for reader 1 (p = 0.017) and r = 0.617 for reader 2 (p = 0.019) (Figure 5). In addition, we re-calculated correlation coefficients in 12 lesions except for two bone tumours with decreased ADC due to treatment response (Table 4). The difference of ADCmin and SUVmax showed moderate to good correlation: r = 0.650 for reader 1 (p = 0.022) and r = 0.622 for reader 2 (p = 0.031). Percent change of ADCmin and SUVmax showed very good to excellent correlation: r = 0.773 for reader 1 (p = 0.003) and r = 0.750 for reader 2 (p = 0.005).

Figure 5. .

Figure 5. 

Open in a new tab

a Correlation of the differences of minimum apparent diffusion coefficient (ADC) (ADCmin) and maximum standardized uptake values (SUV) (SUVmax) of musculoskeletal tumours after treatment. b Correlation of the percent changes of ADCmin and SUVmax of musculoskeletal tumours after treatment.

Table 4. .

Correlation Between ADCs and SUVs in Musculoskeletal Tumours After Treatment

SUVmaxa ADCmina Correlation Coefficienta p
Difference (㎛2/sec)b 6.74 ± 6.00 (2.93 to 10.56)
 Reader 1 1002.2 ± 649.5 (589.5 to 1414.9) 0.650
(0.121 to 0.891)		 0.022
 Reader 2 1106.8 ± 719.8 (649.4 to 1564.1) 0.622
(0.075 to 0.088)		 0.031
Percent change (%)c 51.00 ± 41.66 (24.51 to 77.45)
 Reader 1 134.1 ± 107.9 (65.5 to 202.6) 0.773
(0.358 to 0.933)		 0.003
 Reader 2 162.6 ± 139.6 (73.9 to 251.2) 0.750
(0.309 to 0.926)		 0.005
Open in a new tab

ADC, apparent diffusion coefficient; ADCmin, minimum ADC; DWI, diffusion-weighted imaging; SUV, standardized uptake values; SUVmax, maximum SUV.

a

Data are mean, standard deviation and 95% confidence interval.

b

Differences for ADCs were calculated by subtraction of initial values from follow-up values, whereas differences for SUVs were calculated by subtraction of follow-up values from initial values, therefore positive value means treatment response for both ADC and SUV.

c

Percent change was defined as the ratio between difference and initial value.

Discussion

Our study showed that ADC at 3T DWI and SUV at 18F-FDG PET/CT had an inverse correlation in musculoskeletal tumours. All ADC parameters including ADCmin, ADCmean and normalized ADC were inversely correlated with both SUVmax and SUVmean in musculoskeletal tumours before treatment. The high SUV and low ADC were observed in the malignant musculoskeletal tumours. In terms of correlation between ADC and SUV in musculoskeletal tumours there have been a few reports.20–23

Rakheja et al20 assessed the correlation of SUV and ADC for 52 nonosseous lesions and 17 bone metastatic lesions in 24 patients with known primary malignancies using 18F-FDG PET/MRI system with 3.0T magnet and b values of 0, 350, 750 sec/mm2. They found a weakly inverse correlation between 18F-FDG PET/CT SUVmax and ADCmin (r = - 0.29, p = 0.008), between 18F-FDG PET/MRI SUVmax and ADCmin (r = −0.21, p = 0.040), and between 18F-FDG PET/MRI SUVmean and ADCmean (r = −0.18, p = 0.070) in all osseous and nonosseous lesions. As for subgroup analysis, there was no correlation between 18F-FDG PET/MRI SUVmax and ADCmin (r = −0.12, p = 0.024), or between 18F-FDG PET/MRI SUVmean and ADCmean (r = −0.05, p = 0.014) in bone metastatic lesions. We assume that this difference may be due to different patient status. Namely, most of the included patients of Rakheja’s study20 had undergone treatment and variably responded to treatment in that study. In our study, ADCmin and SUVmax in the posttreatment state were not correlated. Moreover, they did not use normalized ADC. In our study, normalized ADC showed higher correlation with SUV than ADC value without normalization. Included malignant tumours were histologically variable and small in number, which is one of the other causes for the different results. Sagiyama et al22 assessed the correlation of SUV and ADC for 10 low/intermediate-grade tumours or 25 high-grade tumours using 18F-FDG PET/MRI system and b values of 0, 800 sec/mm2. This study found similar mean correlation coefficients for SUV in high grade sarcoma (−0.41  ±  0.25) by voxel-based analysis in comparison with our study. Park et al23 assessed the correlation of SUV on 18F-FDG PET/CT and intravoxel incoherent motion (IVIM) parameters with 11 b values (0, 10, 15, 20, 25, 50, 80, 120, 200, 300, and 800 sec/mm2) in 19 patients with vertebral bone metastases. They found positive correlation between perfusion fraction (f) and SUVs, and between perfusion-related D (D*) and SUVmean (ρ=0.321, p = 0.041).

Malignant bone tumours may show reparative sclerosis or increased fat within the tumour due to the post-treatment response.28 In our study there was one lesion that showed increased fat component on MR imaging in the treated tumour. This lesion showed decreased ADC due to increased fat component.28,29 This could be easily assessed as treatment response based on standard MR imaging, in spite of decreased ADC on DWI. Another lesion showed sclerotic change and decreased ADC in follow-up images. We could suspect the decrease of ADC based on standard MR imaging. Thus, review of standard MR imaging is essential to assess DWI.

The differences and percent changes of ADC and SUV were significantly correlated in our study, whereas ADC and SUV on posttreatment state were not correlated. This might be due to various follow-up intervals of imaging, small case numbers, and heterogeneous tumour types. Byun et al21 evaluated correlation between ADCmean from 3T MR imaging with b values of 0, 800 sec/mm2 and SUVmax for 27 patients with osteosarcoma. Percent change of SUV and ADC negatively correlated with each other (r = −0.593, p = 0.001). We found similar correlation in musculoskeletal tumours. We found higher correlation between ADC and SUV in the percent change rather than the difference. The absolute ADC values are affected by MR parameters such as b values and magnetic strength, then percent change may be recommended for follow-up.30 MR imaging may be particularly useful for repeated short-term follow-up for early assessment of treatment response instead of 18F-FDG PET/CT. Because repeated short-term follow-up with 18F-FDG PET/CT increases the radiation exposure.

The ADC is a quantitative measure of brownian motion of water. Diffusion is limited by an abundance of cell membranes and reduced extracellular space. Therefore, low ADC values are observed in tumours with high cellularity. Both SUV and ADC reflect characteristics of tumours on a cellular level.31–34 SUV relates mitotic and metabolic activity, while ADC relates on integrity of cell membrane and volume of extracellular space.31–34

Chemotherapy and radiation therapy induces cell necrosis and destruction of cell membrane integrity, which increases diffusion of the tumours. 18F-FDG uptake in the tumour after a treatment correlates with the number of viable cancer cells. Because cellular changes precede morphologic changes, treatment response can be assessed earlier on DWI and PET/CT than conventional imaging. There are some confounders for analysis of ADC and SUV. Tumours with prominent angiogenesis showed increased ADC because vascular structures enhance diffusion of water molecule. After a treatment response by antiangiogenic agent, ADC is decreased.35,36 Inflammatory tissues show hypermetabolic activity, SUV can increase after radiation therapy.37

There are several limitations in our study. We retrospectively included patients who had undergone both 18F-FDG PET/CT and MR imaging. Therefore, selection bias existed and the number of non-malignant lesions was small. Included tumours were histologically variable and each musculoskeletal tumour has variable tissue components. We used a non-blinded comparison of the lesions, however, the non-blinded side-by-side-comparison permitted optimal methods for correlation for the heterogeneous tissue components. There was a verification bias because only pathologically confirmed tumours were included. In this study, only two b values were used, because these were the common set of values among the variable DWI sequences in our institution. Finally, MR imaging and 18F-FDG PET/CT were not simultaneously obtained. However the interval between MR imaging and 18F-FDG PET/CT was not significantly long in both pre-treatment and post-treatment image sets.

In conclusion, ADC on DWI inversely correlates with SUV on 18F-FDG PET/CT and would help assess the treatment response in musculoskeletal tumours. This study supports a basis DWI may be used for repeated short-term follow-up to assess the early response of treatment instead of 18F-FDG PET/CT.

Contributor Information

So-Yeon Lee, Email: capella27@gmail.com.

Won-Hee Jee, Email: whjee12@gmail.com.

Ie Ryung Yoo, Email: iryoo@catholic.ac.kr.

Joon-Yong Jung, Email: JJdragon112@gmail.com.

Soo-A Im, Email: sooahim@hanmail.net.

Yang-Guk Chung, Email: ygchung@catholic.ac.kr.

Jin Hyoung Kang, Email: jinkang@catholic.ac.kr.

REFERENCES

  • 1. American Cancer Society Cancer Facts & Figures 2014. Atlanta, GA: The British Institute of Radiology.; 2014. [Google Scholar]
  • 2.Howlader N, Noone AM, Krapcho M.et al. Cancer Statistics Review, 1975e2011. Bethesda: The British Institute of Radiology.; 2013. [Google Scholar]
  • 3.Marom EM, Aloia TA, Moore MB, Hara M, Herndon JE, Harpole DH, et al. Correlation of FDG-PET imaging with GLUT-1 and GLUT-3 expression in early-stage non-small cell lung cancer. Lung Cancer 2001; 33(2-3): 99–107. doi: 10.1016/S0169-5002(00)00250-6 [DOI] [PubMed] [Google Scholar]
  • 4.Nose H, Otsuka H, Otomi Y, Terazawa K, Takao S, Iwamoto S, et al. Correlations between F-18 FDG PET/CT and pathological findings in soft tissue lesions. J Med Invest 2013; 60(3-4): 184–90. doi: 10.2152/jmi.60.184 [DOI] [PubMed] [Google Scholar]
  • 5.Shin D-S, Shon O-J, Han D-S, Choi J-H, Chun K-A, Cho I-H. The clinical efficacy of (18)F-FDG-PET/CT in benign and malignant musculoskeletal tumors. Ann Nucl Med 2008; 22: 603–9. doi: 10.1007/s12149-008-0151-2 [DOI] [PubMed] [Google Scholar]
  • 6.Charest M, Hickeson M, Lisbona R, Novales-Diaz J-A, Derbekyan V, Turcotte RE. Fdg PET/CT imaging in primary osseous and soft tissue sarcomas: a retrospective review of 212 cases. Eur J Nucl Med Mol Imaging 2009; 36: 1944–51. doi: 10.1007/s00259-009-1203-0 [DOI] [PubMed] [Google Scholar]
  • 7.Benz MR, Evilevitch V, Allen-Auerbach MS, Eilber FC, Phelps ME, Czernin J, et al. Treatment monitoring by 18F-FDG PET/CT in patients with sarcomas: interobserver variability of quantitative parameters in treatment-induced changes in histopathologically responding and nonresponding tumors. J Nucl Med 2008; 49: 1038–46. doi: 10.2967/jnumed.107.050187 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Iagaru A, Masamed R, Chawla SP, Menendez LR, Fedenko A, Conti PS. F-18 FDG PET and PET/CT evaluation of response to chemotherapy in bone and soft tissue sarcomas. Clin Nucl Med 2008; 33: 8–13. doi: 10.1097/RLU.0b013e31815c4fd4 [DOI] [PubMed] [Google Scholar]
  • 9.Evilevitch V, Weber WA, Tap WD, Allen-Auerbach M, Chow K, Nelson SD, et al. Reduction of glucose metabolic activity is more accurate than change in size at predicting histopathologic response to neoadjuvant therapy in high-grade soft-tissue sarcomas. Clin Cancer Res 2008; 14: 715–20. doi: 10.1158/1078-0432.CCR-07-1762 [DOI] [PubMed] [Google Scholar]
  • 10.Koh D-M, Collins DJ. Diffusion-Weighted MRI in the body: applications and challenges in oncology. AJR Am J Roentgenol 2007; 188: 1622–35. doi: 10.2214/AJR.06.1403 [DOI] [PubMed] [Google Scholar]
  • 11.Lee S-Y, Jee W-H, Jung J-Y, Park MY, Kim S-K, Jung C-K, et al. Differentiation of malignant from benign soft tissue tumours: use of additive qualitative and quantitative diffusion-weighted MR imaging to standard MR imaging at 3.0 T. Eur Radiol 2016; 26: 743–54. doi: 10.1007/s00330-015-3878-x [DOI] [PubMed] [Google Scholar]
  • 12.Sung JK, Jee W-H, Jung J-Y, Choi M, Lee S-Y, Kim Y-H, et al. Differentiation of acute osteoporotic and malignant compression fractures of the spine: use of additive qualitative and quantitative axial diffusion-weighted MR imaging to conventional MR imaging at 3.0 T. Radiology 2014; 271: 488–98. doi: 10.1148/radiol.13130399 [DOI] [PubMed] [Google Scholar]
  • 13.Ahlawat S, Khandheria P, Subhawong TK, Fayad LM. Differentiation of benign and malignant skeletal lesions with quantitative diffusion weighted MRI at 3T. Eur J Radiol 2015; 84: 1091–7. doi: 10.1016/j.ejrad.2015.02.019 [DOI] [PubMed] [Google Scholar]
  • 14.Dudeck O, Zeile M, Pink D, Pech M, Tunn P-U, Reichardt P, et al. Diffusion-Weighted magnetic resonance imaging allows monitoring of anticancer treatment effects in patients with soft-tissue sarcomas. J Magn Reson Imaging 2008; 27: 1109–13. doi: 10.1002/jmri.21358 [DOI] [PubMed] [Google Scholar]
  • 15.Hayashida Y, Yakushiji T, Awai K, Katahira K, Nakayama Y, Shimomura O, et al. Monitoring therapeutic responses of primary bone tumors by diffusion-weighted image: initial results. Eur Radiol 2006; 16: 2637–43. doi: 10.1007/s00330-006-0342-y [DOI] [PubMed] [Google Scholar]
  • 16.Oka K, Yakushiji T, Sato H, Hirai T, Yamashita Y, Mizuta H. The value of diffusion-weighted imaging for monitoring the chemotherapeutic response of osteosarcoma: a comparison between average apparent diffusion coefficient and minimum apparent diffusion coefficient. Skeletal Radiol 2010; 39: 141–6. doi: 10.1007/s00256-009-0830-7 [DOI] [PubMed] [Google Scholar]
  • 17.Gaeta M, Benedetto C, Minutoli F, D'Angelo T, Amato E, Mazziotti S, et al. Use of diffusion-weighted, intravoxel incoherent motion, and dynamic contrast-enhanced MR imaging in the assessment of response to radiotherapy of lytic bone metastases from breast cancer. Acad Radiol 2014; 21: 1286–93. doi: 10.1016/j.acra.2014.05.021 [DOI] [PubMed] [Google Scholar]
  • 18.Woolf DK, Padhani AR, Makris A. Assessing response to treatment of bone metastases from breast cancer: what should be the standard of care? Ann Oncol 2015; 26: 1048–57. doi: 10.1093/annonc/mdu558 [DOI] [PubMed] [Google Scholar]
  • 19.Rakheja R, Makis W, Skamene S, Nahal A, Brimo F, Azoulay L, et al. Correlating metabolic activity on 18F-FDG PET/CT with histopathologic characteristics of osseous and soft-tissue sarcomas: a retrospective review of 136 patients. AJR Am J Roentgenol 2012; 198: 1409–16. doi: 10.2214/AJR.11.7560 [DOI] [PubMed] [Google Scholar]
  • 20.Rakheja R, Chandarana H, DeMello L, Jackson K, Geppert C, Faul D, et al. Correlation between standardized uptake value and apparent diffusion coefficient of neoplastic lesions evaluated with whole-body simultaneous hybrid PET/MRI. AJR Am J Roentgenol 2013; 201: 1115–9. doi: 10.2214/AJR.13.11304 [DOI] [PubMed] [Google Scholar]
  • 21.Byun BH, Kong C-B, Lim I, Choi CW, Song WS, Cho WH, et al. Combination of 18F-FDG PET/CT and diffusion-weighted MR imaging as a predictor of histologic response to neoadjuvant chemotherapy: preliminary results in osteosarcoma. J Nucl Med 2013; 54: 1053–9. doi: 10.2967/jnumed.112.115964 [DOI] [PubMed] [Google Scholar]
  • 22.Sagiyama K, Watanabe Y, Kamei R, Hong S, Kawanami S, Matsumoto Y, et al. Multiparametric voxel-based analyses of standardized uptake values and apparent diffusion coefficients of soft-tissue tumours with a positron emission tomography/magnetic resonance system: preliminary results. Eur Radiol 2017; 27: 5024–33. doi: 10.1007/s00330-017-4912-y [DOI] [PubMed] [Google Scholar]
  • 23.Park S, Yoon J-K, Chung N-S, Kim SH, Hwang J, Lee HY, et al. Correlations between intravoxel incoherent motion diffusion-weighted MR imaging parameters and 18F-FDG PET/CT metabolic parameters in patients with vertebral bone metastases: initial experience. Br J Radiol 2018; 91: 20170889. doi: 10.1259/bjr.20170889 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Bonarelli C, Teixeira PAG, Hossu G, Meyer J-B, Chen B, Gay F, et al. Impact of Roi positioning and lesion morphology on apparent diffusion coefficient analysis for the differentiation between benign and malignant nonfatty soft-tissue lesions. AJR Am J Roentgenol 2015; 205: W106–W113. doi: 10.2214/AJR.14.13865 [DOI] [PubMed] [Google Scholar]
  • 25.Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979; 86: 420–8. doi: 10.1037/0033-2909.86.2.420 [DOI] [PubMed] [Google Scholar]
  • 26.Dawson B, Trapp RG. Basic & clinical biostatistics. 4th edition New York: The British Institute of Radiology.; 2004. [Google Scholar]
  • 27.Ciray I, Lindman H, Aström KG, Bergh J, Ahlström KH. Early response of breast cancer bone metastases to chemotherapy evaluated with MR imaging. Acta Radiol 2001; 42: 198–206. doi: 10.1034/j.1600-0455.2001.042002198.x [DOI] [PubMed] [Google Scholar]
  • 28.Reischauer C, Froehlich JM, Koh D-M, Graf N, Padevit C, John H, et al. Bone metastases from prostate cancer: assessing treatment response by using diffusion-weighted imaging and functional diffusion maps--initial observations. Radiology 2010; 257: 523–31. doi: 10.1148/radiol.10092469 [DOI] [PubMed] [Google Scholar]
  • 29.Messiou C, Collins DJ, Morgan VA, Bianchini D, de Bono JS, de Souza NM. Use of apparent diffusion coefficient as a response biomarker in bone: effect of developing sclerosis on quantified values. Skeletal Radiol 2014; 43: 205–8. doi: 10.1007/s00256-013-1768-3 [DOI] [PubMed] [Google Scholar]
  • 30.Sasaki M, Yamada K, Watanabe Y, Matsui M, Ida M, Fujiwara S, et al. Variability in absolute apparent diffusion coefficient values across different platforms may be substantial: a multivendor, multi-institutional comparison study. Radiology 2008; 249: 624–30. doi: 10.1148/radiol.2492071681 [DOI] [PubMed] [Google Scholar]
  • 31.Eary JF, O'Sullivan F, Powitan Y, Chandhury KR, Vernon C, Bruckner JD, et al. Sarcoma tumor FDG uptake measured by PET and patient outcome: a retrospective analysis. Eur J Nucl Med Mol Imaging 2002; 29: 1149–54. doi: 10.1007/s00259-002-0859-5 [DOI] [PubMed] [Google Scholar]
  • 32.Eary JF, Conrad EU, Bruckner JD, Folpe A, Hunt KJ, Mankoff DA, et al. Quantitative [F-18]fluorodeoxyglucose positron emission tomography in pretreatment and grading of sarcoma. Clin Cancer Res 1998; 4: 1215–20. [PubMed] [Google Scholar]
  • 33.Sugahara T, Korogi Y, Kochi M, Ikushima I, Shigematu Y, Hirai T, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999; 9: 53–60. doi: [DOI] [PubMed] [Google Scholar]
  • 34.Lang P, Wendland MF, Saeed M, Gindele A, Rosenau W, Mathur A, et al. Osteogenic sarcoma: noninvasive in vivo assessment of tumor necrosis with diffusion-weighted MR imaging. Radiology 1998; 206: 227–35. doi: 10.1148/radiology.206.1.9423677 [DOI] [PubMed] [Google Scholar]
  • 35.Lewin M, Fartoux L, Vignaud A, Arrivé L, Menu Y, Rosmorduc O. The diffusion-weighted imaging perfusion fraction F is a potential marker of sorafenib treatment in advanced hepatocellular carcinoma: a pilot study. Eur Radiol 2011; 21: 281–90. doi: 10.1007/s00330-010-1914-4 [DOI] [PubMed] [Google Scholar]
  • 36.Schraml C, Schwenzer NF, Martirosian P, Bitzer M, Lauer U, Claussen CD, et al. Diffusion-Weighted MRI of advanced hepatocellular carcinoma during sorafenib treatment: initial results. AJR Am J Roentgenol 2009; 193: W301–W307. doi: 10.2214/AJR.08.2289 [DOI] [PubMed] [Google Scholar]
  • 37.Hicks RJ, Mac Manus MP, Matthews JP, Hogg A, Binns D, Rischin D, et al. Early FDG-PET imaging after radical radiotherapy for non-small-cell lung cancer: inflammatory changes in normal tissues correlate with tumor response and do not confound therapeutic response evaluation. Int J Radiat Oncol Biol Phys 2004; 60: 412–8. doi: 10.1016/j.ijrobp.2004.03.036 [DOI] [PubMed] [Google Scholar]

Articles from The British Journal of Radiology are provided here courtesy of Oxford University Press

RESOURCES