Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Correlations between Functional Imaging Markers Derived from PET/CT and Diffusion-Weighted MRI in Diffuse Large B-Cell Lymphoma and Follicular Lymphoma

  • Xingchen Wu ,

    Xingchen.Wu@uta.fi

    Affiliations Department of Oncology, Tampere University Hospital, Tampere, Finland, Medical Imaging Centre, Department of Radiology, Tampere University Hospital, Tampere, Finland, Medical School, University of Tampere, Tampere, Finland

  • Hannu Pertovaara,

    Affiliation Department of Oncology, Tampere University Hospital, Tampere, Finland

  • Pasi Korkola,

    Affiliation Medical Imaging Centre, Department of Nuclear Medicine, Tampere University Hospital, Tampere, Finland

  • Prasun Dastidar,

    Affiliations Medical Imaging Centre, Department of Radiology, Tampere University Hospital, Tampere, Finland, Medical School, University of Tampere, Tampere, Finland

  • Ritva Järvenpää,

    Affiliation Medical Imaging Centre, Department of Radiology, Tampere University Hospital, Tampere, Finland

  • Hannu Eskola,

    Affiliations Medical Imaging Centre, Department of Radiology, Tampere University Hospital, Tampere, Finland, Department of Electronics and Telecommunications, Tampere University of Technology, Tampere, Finland

  • Pirkko-Liisa Kellokumpu-Lehtinen

    Affiliations Department of Oncology, Tampere University Hospital, Tampere, Finland, Medical School, University of Tampere, Tampere, Finland

Correlations between Functional Imaging Markers Derived from PET/CT and Diffusion-Weighted MRI in Diffuse Large B-Cell Lymphoma and Follicular Lymphoma

  • Xingchen Wu, 
  • Hannu Pertovaara, 
  • Pasi Korkola, 
  • Prasun Dastidar, 
  • Ritva Järvenpää, 
  • Hannu Eskola, 
  • Pirkko-Liisa Kellokumpu-Lehtinen
PLOS
x

Abstract

Objectives

To investigate the correlations between functional imaging markers derived from positron emission tomography/computed tomography (PET/CT) and diffusion-weighted magnetic resonance imaging (DWI) in diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL). Further to compare the usefulness of these tumor markers in differentiating diagnosis of the two common types of Non-Hodgkin's lymphoma (NHL).

Materials and Methods

Thirty-four consecutive pre-therapy adult patients with proven NHL (23 DLBCL and 11 FL) underwent PET/CT and MRI examinations and laboratory tests. The maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), and metabolic tumor burden (MTB) were determined from the PET/CT images. DWI was performed in addition to conventional MRI sequences using two b values (0 and 800 s/mm2). The minimum and mean apparent diffusion coefficient (ADCmin and ADCmean) were measured on the parametric ADC maps.

Results

The SUVmax correlated inversely with the ADCmin (r = −0.35, p<0.05). The ADCmin, ADCmean, serum thymidine kinase (TK), Beta 2-microglobulin (B2m), lactate dehydrogenase (LD), and C-reactive protein (CRP) correlated with both whole-body MTV and whole-body MTB (p<0.05 or 0.01). The SUVmax, TK, LD, and CRP were significantly higher in the DLBCL group than in the FL group. Receiver operating characteristic curve analysis showed that they were reasonable predictors in differentiating DLBCL from FL.

Conclusions

The functional imaging markers determined from PET/CT and DWI are associated, and the SUVmax is superior to the ADCmin in differentiating DLBCL from FL. All the measured serum markers are associated with functional imaging markers. Serum LD, TK, and CRP are useful in differentiating DLBCL from FL.

Introduction

Non-Hodgkin's lymphoma (NHL) represents a heterogeneous group of lymphoid malignancies that display varying patterns of biological behavior and response to treatment [1]. Prognosis of patients with NHL is affected by the stage, grade, and histological subtype. The most common subtypes of NHL affecting adults are diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL), which together account for more than 50% of the incidences of the disease [2]. Prognostic tumor markers may help to identify high-risk patients who might benefit from more aggressive therapy. Positron emission tomography/computed tomography (PET/CT) with the use of 2-deoxy-2-[18F]fluoro-D-glucose (18F-FDG) is an established imaging modality that has been proven to be of benefit in the management of malignant lymphomas [3]. Among the PET parameters, standardized uptake value (SUV) is currently the most commonly used semi-quantitative index of 18F-FDG metabolic rate. SUV reflects tumor glucose metabolism, and is commonly represented by the mean (SUVmean) or maximum (SUVmax) value. In PET image analysis, SUVmax has the advantage of being relatively operator independent. However, the measurement of SUVmax has been confined to detection of the most obvious metabolic activities of the tumor at a single site, but not the overall tumor activity. SUVmean is the average value generated from the entire tumor, but differences in operator contouring of tumor will yield varying values. In addition, both SUVmax and SUVmean represent only the metabolic activity per gram of tissue, but they are not able to reflect tumor dimensions and volume. In contrast, metabolic tumor burden (MTB) (i.e., total lesion glycolysis) is a newly proposed tumor marker in which both tumor activity and volume are integrated. MTB is the product of SUVmean and metabolic tumor volume (MTV). Lymphoma patients can have single or multiple lesions depending on the stage of the disease. In order to take the number of lesion into consideration, SUVsum (summation of SUVmax for all tumors), whole-body MTV (MTVwb; summation of MTV for all tumors) and whole-body MTB (MTBwb; summation of MTB for all tumors) were calculated in the present study and used as indexes that could potentially reflect overall tumor activity or malignant process of the entire body [4].

Cancer is not only characterized by pathological metabolism, but also by the higher cellularity and morphological changes of tumor cell and tissue [5]. Therefore, restricted water diffusion has been found to be a common feature of tumors. Recent studies have shown that diffusion-weighted magnetic resonance imaging (DWI) is a valuable imaging modality for detecting metastasis and cancer relapse [6]. DWI with apparent diffusion coefficient (ADC) mapping provides information on tumor tissue aggressiveness. ADC value has been applied to distinguish benign from malignant lymph nodes [7], [8], and it has also been used to assess treatment response in various malignancies including lymphoma [6], [9]. Tumors in NHL are typically heterogeneous and can have different histological grades or subtypes in different tumors of the same patient or even in a single tumor [8], [10]. Various components may influence the mean ADC (ADCmean) of the tumor/tumors. In contrast, the minimum ADC (ADCmin) is the most malignant site within the heterogeneous tumor/tumors [11]. In order to compare the diagnostic significance of DWI, both ADCmin and ADCmean were measured in the present study.

Both PET/CT and DWI are established imaging modalities in tumor assessment including tumor aggressiveness, treatment response, and prognosis, but they measure different aspects of tumor pathophysiology. A few studies have recently compared FDG-PET/CT and DWI in patients with different cancers, and an inverse correlation between SUVmax and ADCmin has been revealed [12][16]. There are also several well-known prognostic clinical and laboratory predictors for newly diagnosed malignant lymphoma, such as International Prognostic Index (IPI), Follicular Lymphoma IPI (FLIPI), and elevated levels of serum thymidine kinase (TK), Beta 2-microglobulin (B2m), lactate dehydrogenase (LD) and C-reactive protein (CRP). The aim of this study was to investigate the relationships between functional imaging markers derived from PET/CT and DWI, as well as serum tumor markers in DLBCL and FL. In addition, the study was desired to compare the usefulness of these tumor markers in differentiating diagnosis of the two common types of aggressive and indolent NHL.

Materials and Methods

Patients

Patients were enrolled from our prospective study investigating the potential of PET/CT and MRI for early chemotherapy response evaluation in patients with NHL. The inclusion criteria were: at least 18 years old, histologically proven DLBCL or FL, WHO performance scale (Zubrod score) better than 4. The exclusion criteria were: concomitant previous malignant disease, primary central nervous system lymphoma, pregnancy or lactation, psychosis, diabetes, human immunodeficiency virus infection or acquired immunodeficiency syndrome, or other serious medical conditions that would prevent imaging examination. The study was approved by the Ethics Committee of Tampere University Hospital, and all patients gave written informed consent prior to study entry.

All patients underwent anamnestic and physical examination, standard laboratory tests including the measurement of serum tumor markers such as TK, B2m, LD, and CRP, as well as CT scans of the chest, abdomen, and pelvis. In addition, unilateral bone marrow aspiration and trephine biopsy were performed on each patient. Pathological samples were reviewed by our expert hematopathologists and classified according to the WHO/Revised European-American Lymphoma classification of lymphoid neoplasm. An experienced physician selected the target tumor mass of interest (the region containing the largest tumor or the greatest number of >1 cm lymph nodes) for DWI analysis based on clinical presentation and CT examination. Clinical prognostic indexes, such as Ann Arbor stage and IPI/FLIPI were also evaluated.

FDG-PET/CT image acquisition

All patients underwent an integrated PET/CT (Discovery STE 16, GE Healthcare, Milwaukee, WI, USA) examination. The PET/CT imaging covered a volume from the skull base to the upper thigh, and was acquired 72±16 (Mean ± SD) minutes after intravenous injection of the 18F-FDG tracer (369±22 MBq) under fasting conditions (Patients were informed to fast at least 6 hours, which was confirmed by an interview). The acquisition was in the 3-dimensional (3D) mode with a 128×128 matrix and 70 cm field of view (FOV), 3 minutes per bed position. The PET images were reconstructed using the 3D VUE Point reconstruction algorithm (GE Healthcare) with 2 iterations and 28 subsets. The postfilter used was 6.0 mm FWHM. The acquisition parameters of the CT scanner were: tube voltage, 120 kV; tube current automatic exposure control range, 100–440 mA; noise index, 18.5 HU; rotation speed, 35 mm/rot; pitch, 1.75∶1. The CT images were reconstructed to slice thicknesses of 1.25 mm and 5.0 mm. The total examination time for PET/CT was approximately 30 minutes.

PET/CT image analysis

The PET/CT images were evaluated visually and quantitatively. The SUVmax, SUVmean, and MTV were measured from each site (tumor or group of tumors). For each PET/CT dataset, the tumor with the most intense 18F-FDG uptake among all foci was carefully identified, and the SUVmax was measured on the fused PET/CT images using the AW Volume Share™ workstation (GE Healthcare) [17].

For each tumor or group of tumors, the MTV was estimated in a 3D manner by selecting volume of interest (VOI) on the axial image, and the size of VOI was manually regulated on the corresponding coronal and sagittal images to include the entire active tumor in the VOI, and an isocontour threshold of 42% of the SUVmax was determined between the background and the maximal pixel value. The SUVmax, SUVmean, and MTV in the VOI were computed automatically by the program [17].

MRI Protocol

MR imaging was acquired using a 3 Tesla MR System (Siemens Trio-Tim, Erlangen, Germany) with the manufacturer's body and spine array coils. Additionally, a neck coil was used for the cervical region examination. The MR imaging consisted of a whole-body examination from the level of the skull base to the floor of the pelvis in the coronal plane using a parallel acquisition technique. High resolution axial images and DW images were acquired from the target tumor/tumors. The MRI protocol included a coronal T1-weighted turbo spin echo (TSE) imaging, a coronal T2-weighted inversion-recovery imaging, an axial T1-weighted 3D volumetric interpolated breath-hold examination (VIBE) with fat suppression once before and once after gadolinium (Gd)-DOTA (0.2 ml/kg Dotarem®) injection, and an axial T2-weighted TSE imaging once with and once without fat suppression. Before contrast administration, DWI was acquired using a single-shot echo-planar sequence with fat suppression in the axial plane with two b values (0 and 800 s/mm2). The diffusion-weighting gradients were applied in all three orthogonal directions. The DWI was performed during normal respiration and the MR parameters were different depending on the location of the target tumor/tumors [18]. After acquisition of the DWI data sets, pixel-by-pixel ADC maps were reconstructed automatically for each patient.

ADC value measurement

The ADC value of the target tumor/tumors was measured directly on the parametric ADC maps. A region of interest (ROI) was manually placed on every slice of the entire tumor/tumors that appeared as areas of low signal intensity. In order to ensure proper positioning of the ROI on the ADC maps, the corresponding T2-weighted images and contrast-enhanced T1-weighted images were reviewed side-by-side. Any necrotic areas were excluded from the analysis. For ADC measurement the open-source software ImageJ (created by Wayne Rasband, freely downloadable at the NIH website: http://rsb.info.nih.gov/ij/) was used, which lists the intensity of each pixel within every ADC slice in a single ROI output file. In this manner, the minimum and mean ADC values were calculated. The ADCmin was defined as the lowest ADC value within all of the slices of the target tumor/tumors. Accordingly, the ADCmean was defined as the mean value of all of the target tumor/tumors pixels in all of the slices.

Statistical analysis

The statistical analyses were performed using SPSS software. Mann-Whitney U test was used to compare functional imaging and serum markers between the DLBCL and FL groups. The Spearman's correlation coefficient was used to evaluate the correlations between the ADCmin or ADCmean and SUVmax, SUVsum, MTVwb, or MTBwb, as well as correlations between imaging markers derived from PET/CT or DWI and stage, IPI categories, or serum tumor markers. P values less than 0.05 were considered significant. Receiver operating characteristic (ROC) curve was used to determine the cut-off values of SUVmax, TK, LD, and CRP with the use of the best combination of sensitivity and specificity to differentiate between DLBCL and FL.

Results

Patient characteristics

Thirty-four pretherapy patients with DLBCL or FL (17 male and 17 female; mean age, 63 years; range, 32 to 86 years) underwent PET/CT and MRI examinations within two days. Twenty-three patients had DLBCL including 3 patients with concomitant DLBCL and FL, and 11 patients had FL: 10 with FL grade II and one with FL grade III. The clinical characteristics of the 34 study participants and their tumors are shown in Table 1.

thumbnail
Table 1. Demographic characteristics, tumor pathology, and clinical staging of 34 patients with DLBCL or FL.

http://dx.doi.org/10.1371/journal.pone.0084999.t001

The SUV and ADC value between DLBCL and FL groups

The SUVmax was significantly higher in the DLBCL group than that of the FL group (p<0.01) (Table 2 and an example image Figure 1). The only patient with FL grade III had a SUVmax of 30.2, which is much higher than the SUVmax of those with FL grade II. Serum levels of TK, LD, and CRP were significantly higher in the DLBCL group than in the FL group. However, there was no significant difference in the ADCmin, ADCmean, or serum B2m value between the two groups (Table 2).

thumbnail
Figure 1. Diffusion-weighted MRI and PET/CT images showing the abdominal region tumor in a 76- year old male patient with diffuse large B-cell lymphoma.

(a) B0 image. (b) Diffusion-weighted image with b value 800 s/mm2 showed the hyperintensity tumor, but it was not able to depict diffuse spleen involvement. (c) The corresponding ADC map showed the hypointensity tumor with ADCmin 0.34×10−3 mm2/s and ADCmean 0.68×10−3 mm2/s. (d) Axial CT image. (e) FDG-PET image. (f) The fused PET/CT image showed the active tumor and spleen involvement with SUVmax 23.9.

http://dx.doi.org/10.1371/journal.pone.0084999.g001

thumbnail
Table 2. Comparison of PET/CT and DWI indexes and serum biomarkers in the DLBCL and FL groups.

http://dx.doi.org/10.1371/journal.pone.0084999.t002

To further investigate whether the SUVmax, serum LD, TK, or CRP can differentiate DLBCL from FL, we used ROC curve analysis (Figure 2). The area under the ROC curve (AUC) was 0.796, 0.763, 0.787, or 0.787; respectively, which suggested that these markers were reasonable predictors for differentiating diagnosis. The SUVmax cut-off value of ≥10.5 (normal value <2.5) provided a fair balance, with sensitivity 87% and specificity 55% to detect a DLBCL. When a TK cut-off value of ≥10.5 (U/l) (normal range 0–0.8 U/l) or a LD cut-off value of ≥200.5 (U/l) (normal range 105–205 U/l) was used, yielded sensitivity 83% and specificity 64% to detect a DLBCL. When a CRP cut-off value of ≥1.85 (mg/l) (normal range 0–10 mg/l) was used, yielded sensitivity 87% and specificity 64% to detect a DLBCL. A higher cut-off value would capture more FLs. Conversely, a lower cut-off value would capture a higher percentage of DLBCLs.

thumbnail
Figure 2. Receiver operating characteristic curve analysis of SUVmax, serum TK, LD, and CRP in 34 patients with DLBCL or FL.

The area under the ROC curve (AUC) was 0.796, 0.763, 0.787, or 0.787 for the SUVmax, serum TK, LD, or CRP; respectively. When the SUVmax≥10.5 was used as a cut-off value to differentiate DLBCL from FL, yielded sensitivity 87% and specificity 55%. When a TK cut-off value of ≥10.5 (U/l) or a LD cut-off value of ≥200.5 (U/l) was used to differentiate DLBCL from FL, yielded sensitivity 83% and specificity 64% to detect a DLBCL. When a CRP cut-off value of ≥1.85 (mg/l) was used to differentiate DLBCL from FL, yielded sensitivity 87% and specificity 64% to detect a DLBCL.

http://dx.doi.org/10.1371/journal.pone.0084999.g002

Correlations between SUV and ADC

The SUV evaluated from PET/CT and the ADC determined from DWI were compared. The SUVmax correlated inversely with the ADCmin (r = −0.35, p<0.05) in all cases (Figure 3a). No correlation was detected in the DLBCL or FL group when subgroup analysis was performed. The SUVsum correlated inversely with the ADCmean (r = −0.42, p<0.05) in all cases (Figure 3b). The inverse correlation was limited in the FL group (r = −0.73, p<0.05) when subgroup analysis was performed. No correlation was found between the SUVmax and ADCmean or SUVsum and ADCmin.

thumbnail
Figure 3. Scatter plots showing the correlations between the SUVmax and ADCmin, (a), and between the SUVsum and ADCmean (b) in 34 patients with DLBCL or FL.

The SUVmax correlated inversely with the ADCmin (r = −0.35, p<0.05) (Figure 3a), and the SUVsum correlated inversely with the ADCmean (r = −0.42, p<0.05) (Figure 3b) in all cases.

http://dx.doi.org/10.1371/journal.pone.0084999.g003

Correlations between ADCmin or ADCmean and MTVwb or MTBwb

The ADCmin correlated inversely with MTVwb in all cases (r = −0.54, p<0.01) (Figure 4a). These values also correlated inversely in the DLBCL (r = −0.45, p<0.05) and FL group (r = −0.74, p<0.05); respectively. The ADCmin correlated inversely with MTBwb (r = −0.47, p<0.01) in all cases (Figure 4b). The inverse correlation was limited in the FL group (r = −0.69, p<0.05) when the analysis was performed in the subgroups.

thumbnail
Figure 4. Scatter plots showing the correlations between the ADCmin and the MTVwb or MTBwb in 34 patients with DLBCL or FL.

The ADCmin correlated inversely with the MTVwb (r = −0.54, p<0.01) (Figure 4a), and it also correlated inversely with the MTBwb (r = −0.47, p<0.01) (Figure 4b) in all cases.

http://dx.doi.org/10.1371/journal.pone.0084999.g004

The ADCmean correlated inversely with MTVwb in all cases (r = −0.43, p<0.05) (Figure 5a). The inverse correlation was limited in the FL group (r = −0.75, p<0.01) when the analysis was performed in the subgroups. The ADCmean correlated inversely with MTBwb in all cases (r = −0.35, p<0.05) (Figure 5b). The inverse correlation was limited in the FL group (r = −0.65, p<0.05) when subgroup analysis was performed.

thumbnail
Figure 5. Scatter plots showing the correlations between the ADCmean and the MTVwb or MTBwb in 34 patients with DLBCL or FL.

The ADCmean correlated inversely the MTVwb (r = −0.43, p<0.05) (Figure 5a), and it also correlated inversely with the MTBwb (r = −0.35, p<0.05) (Figure 5b) in all cases.

http://dx.doi.org/10.1371/journal.pone.0084999.g005

Correlations between imaging markers and serum markers

The data revealed good correlations between all the measured serum tumor markers (i.e., TK, B2m, LD, and CRP) and MTVwb or MTBwb regardless of disease types (p<0.01; respectively) (Table 3). Subgroup analyses showed that TK, LD, or CRP correlated with both MTVwb and MTBwb (p<0.01; respectively) and B2m correlated with both MTVwb and MTBwb (p<0.05; respectively) in the DLBCL group. TK or B2m correlated with both MTVwb and MTBwb (p<0.05; respectively) in the FL group, but no correlation was detected between LD or CRP and MTVwb or MTBwb.

thumbnail
Table 3. Correlations (Spearman's rho) between clinical stage, IPI, or functional imaging markers and serum markers in 34 patients with DLBCL or FL.

http://dx.doi.org/10.1371/journal.pone.0084999.t003

In addition, TK correlated inversely with ADCmin (r = −0.35, p<0.05), LD correlated with both ADCmin (r = −0.48, p<0.01) and SUVmax (r = 0.41, p<0.05), and CRP correlated with SUVmax (r = 0.41, p<0.05) in all cases (Table 3). Subgroup analysis showed that LD correlated inversely with ADCmin in the DLBCL cases (r = −0.48, p<0.05), and LD correlated with SUVmax in the FL cases (r = 0.63, p<0.05).

Correlations between functional imaging markers, serum markers, and IPI categories

There were moderate correlations between the IPI and MTBwb (r = 0.41, p<0.05) or SUVsum (r = 0.42, p<0.05) in the DLBCL group. No correlation was detected between the IPI and ADC value. No correlation was detected between the FLIPI and imaging biomarkers in the FL group. Serum B2m correlated with IPI in the DLBCL group (p<0.05). Serum B2m, LD, or TK correlated with Ann Arbor stage in all cases (p<0.05, respectively) (Table 3).

Discussion

The IPI, a strong predictor of survival in aggressive NHL, was determined from five factors: age, tumor stage, serum LD concentration, performance status, and number of sites of extranodal involvement. It has been used as the standard clinical tool in the selection of appropriate treatment strategies for individual patients. However, there are still significant differences in outcomes within the same prognostic categories. Therefore, more efficient prognostic markers or models to stratify patients with different survival outcomes are needed. Imaging biomarkers are important for detection and characterization of cancers as well as for monitoring the response to therapy. Integrated PET/CT, with the advantage of combining functional and anatomical information and better attenuation correction, is regarded as current standard reference for the management of lymphomas [3]. Several studies have demonstrated a relation between higher FDG uptake and more aggressive course of malignancy in NHL [19], [20]. High SUV is correlated with rapid cellular proliferation in different subtypes of NHL [19], [21][24].

Diffusion-weighted MRI is based on a different approach. In DWI, the random thermal motion of free water molecules known as the Brownian motion can be visualized in vivo, and the ADC value varies according to the microstructure and pathophysiological state of the tissues. Within biological tissue, unrestricted free diffusion of molecules does not exist due to cell membranes, organelles and molecular boundaries. In malignant tissue, the microstructural environment that has increased cellularity and dense tumor cell membranes, larger cell nuclei and more abundant macromolecular proteins as well as reduced extracellular space is known to act as a diffusion barrier leading to a decrease of water mobility. On the other hand, necrosis and apoptotic processes may lead to a decrease of cellularity and loss of cell membrane integrity. This, in turn, increases the amount of free water diffusion across the cell membrane and in the extracellular space [6], [8], [15].

In this study, we evaluated the SUV and ADC in patients with DLBCL and FL, and detected an inverse correlation between the SUVmax and ADCmin. This is in agreement with a recent study in Hodgkin's lymphoma (HL) [25]. In addition, several investigators have reported a significant inverse correlation between SUVmax and ADCmin in various tumors including rectal cancer [13], lung cancer [12], gastrointestinal stromal tumor [14], cervical cancer [15], and endometrial cancer [16]. However, there were also contradictory results. For example, an inverse correlation between the ratio of ADCmin/ADCmean and SUVmax/SUVmean was reported in 33 patients with uterine cervical cancer, but neither ADCmin nor ADCmean correlated with SUVmax or SUVmean [26]. The possible explanation for this finding is that different pathological types of cervical cancer patients including squamous cell carcinoma, adenocarcinoma, and adenosquamous carcinoma were present in the study cohort. In our study, there was a better correlation between the PET/CT and DWI derived markers in the FL group compared with the DLBCL group. This could be explained by the fact that the FL group was relatively homogenous in pathology; all 11 patients had FL grade II, except for one with FL grade III. In contrast, the DLBCL group had obvious pathological variability, including varying differentiating levels of DLBCL and concomitant DLBCL and FL cases. Cancer grows not only with a high proliferation rate, resulting in abnormally high number of cells, but also with architectural alterations in tumor cells and tissue. Changes in nuclear structure are among the most universal of these, a larger cell nuclear size usually means a more aggressive tumor [5]. The decreased ADC value in malignant tumors may be a result of their increased cellularity, larger nuclei with more abundant macromolecular proteins (aggressiveness), and decreased extracellular space. Thus, the ADC value correlates with tumor aggressiveness in specific tumor histological subtype [27]. In contrast, the SUVmax reflects the highest tumor metabolic rate, regardless of the underling microstructure changes. Thus, the SUVmax and ADCmin are independent functional imaging markers that may complement each other in the management of lymphomas, and to use both types of data may improve the accuracy of diagnoses [15].

The SUVsum reflects the total tumor metabolic activity of the whole body and the ADCmean indicates the overall tumor cellularity and aggressiveness. It is not surprising that an inverse correlation between the SUVsum and ADCmean was also detected in this study. To our knowledge, this is the first report of the relation between SUVsum and ADCmean, and this finding needs to be verified in future studies. In addition, both the ADCmin and the ADCmean correlated inversely with MTVwb and MTBwb, which reflect the total metabolic tumor volume or total amount of tumor glycolysis in the patient's body. These correlations could be explained by the fact that a lower ADC value indicates increased aggressiveness. In general, more aggressive tumors proliferate more rapidly and with a higher risk of metastasis, and accordingly the MTVwb and MTBwb are also greater. No correlation was detected between SUVmax and ADCmean. This is in agreement with our previous lesion-wise comparison of SUVmax and ADCmean in DLBCL [10], since the SUVmax represents the most malignant site of the heterogeneous tumors, whereas ADCmean indicates the overall tumor aggressiveness.

Both DLBCL and FL had a wide range of FDG avidity. Our result showed that SUVmax was a useful marker in differentiating between DLBCL and FL, although overlap between the two types of disease existed. A SUVmax cut-off value of ≥10.5 was indicative of aggressive disease. The only patient with FL grade III had much higher SUVmax compared with those with FL grade II. This is in agreement with previous studies showing that indolent FL is associated with low-grade FDG uptake, and more intense FDG accumulation has been observed in more aggressive B-cell lymphomas [19], [20]. Although typically biopsies are obtained from areas that are easily accessible, an unexpectedly high SUVmax may suggest a transformation of an indolent NHL to a more aggressive disease and a new biopsy of the site with the highest SUVmax should be considered. Therefore, our results suggest that PET scanning may also be helpful in directing biopsies and changing clinical management. In contrast, DWI with ADC value measurement is inferior to PET/CT derived SUVmax in differentiating DLBCL from FL and is therefore not able to guide biopsy. This finding is concordant with a previous study in a group of 16 indolent and 16 aggressive NHL patients [28]. The possible explanation is that ADC value reflects both tissue cellularity and tumor aggressiveness, but aggressive tumors do not always have higher cellularities. e.g., we have demonstrated that FL had a higher cellularity than DLBCL [17]. Thus far, the most promising oncologic application of DWI seems to be in tumor detection and treatment response evaluation [6].

In our study, serum TK, LD, and CRP were significantly higher in the DLBCL group than in the FL group. ROC analysis showed that they are useful markers in differentiating the two common types of indolent and aggressive lymphoma. Serum LD represents a surrogate quantitative measure of tumor burden and aggressiveness in NHL [29], and it is a one of the components of the IPI. A high serum B2m level is an independent adverse prognostic factor in malignant lymphoma, which is also related to the tumor burden [30], [31]. CRP is an acute-phase reactant, and elevated baseline serum CRP levels have also been found to be a poor prognostic factor in cancers of many types [32], [33]. High serum CRP might reflect a high metastatic potential, as it is known to promote metastatic spread by stimulating angiogenesis, increasing vascular permeability, and acting as an endothelial cell mitogen [34]. Elevated serum TK predicts high proliferation of tumor cells in lymphoma [35]. Our study showed that serum LD, TK, CRP or B2m correlated with both MTVwb and MTBwb. Therefore, these serum tumor markers could serve as clinically useful markers in malignant NHL, since they are widely available and relatively inexpensive. In addition, a recent study showed that they are independent prognostic markers [36]. The IPI correlated with quantitative PET/CT functional markers (MTBwb and SUVsum) in the DLBCL group. This indicates that the IPI remains a useful and simple clinical tool in evaluating the risk and prognosis of aggressive NHL.

There was a limitation in our study, the PET/CT was a whole body examination, but the DWI was performed only in the target tumor/tumors. As such, ADCmin might not be the minimum ADC of the whole body, since the largest tumor was not always the one with the highest malignancy. Additionally, the study included only a small cohort, and future studies with larger patient cohorts are needed to confirm our findings.

Conclusions

Glucose metabolism with PET/CT and ADC value with DW-MRI are different indexes for the characterization of lymphomas. The functional imaging markers derived from DWI and FDG-PET/CT are associated, and FDG-PET/CT is superior to DWI in differentiating DLBCL from FL. The measured serum tumor markers such as LD, TK, CRP, and B2m are associated with functional imaging markers, and LD, TK, and CRP are useful in differentiating DLBCL from FL.

Author Contributions

Conceived and designed the experiments: XW HP PK PD RJ HE PLKL. Performed the experiments: XW HP PK PD RJ. Analyzed the data: XW HP PK PD RJ. Contributed reagents/materials/analysis tools: XW PK PD RJ. Wrote the paper: XW. Critically read the manuscript: HP PK PD RJ HE PLKL.

References

  1. 1. Flowers CR, Armitage JO (2010) A decade of progress in lymphoma: advances and continuing challenges. Clin Lymphoma Myeloma Leuk 10: 414–423.
  2. 2. Jaffe ES (2009) The 2008 WHO classification of lymphomas: implications for clinical practice and translational research. Hematology Am Soc Hematol Educ Program 2009: 523–531.
  3. 3. Okada M, Sato N, Ishii K, Matsumura K, Hosono M, et al. (2010) FDG PET/CT versus CT, MR imaging, and 67Ga scintigraphy in the posttherapy evaluation of malignant lymphoma. Radiographics 30: 939–957.
  4. 4. Park S, Moon SH, Park LC, Hwang DW, Ji JH, et al. (2012) The impact of baseline and interim PET/CT parameters on clinical outcome in patients with diffuse large B-cell lymphoma. Am J Hematol 87: 937–940.
  5. 5. Zink D, Fischer AH, Nickerson JA (2004) Nuclear structure in cancer cells. Nat Rev Cancer 4: 677–687.
  6. 6. Patterson DM, Padhani AR, Collins DJ (2008) Technology insight: water diffusion MRI–a potential new biomarker of response to cancer therapy. Nat Clin Pract Oncol 5: 220–233.
  7. 7. Holzapfel K, Duetsch S, Fauser C, Eiber M, Rummeny EJ, et al. (2009) Value of diffusion-weighted MR imaging in the differentiation between benign and malignant cervical lymph nodes. Eur J Radiol 72: 381–387.
  8. 8. Wang J, Takashima S, Takayama F, Kawakami S, Saito A, et al. (2001) Head and neck lesions: characterization with diffusion-weighted echo-planar MR imaging. Radiology 220: 621–630.
  9. 9. Wu X, Kellokumpu-Lehtinen PL, Pertovaara H, Korkola P, Soimakallio S, et al. (2011) Diffusion-weighted MRI in early chemotherapy response evaluation of patients with diffuse large B-cell lymphoma–a pilot study: comparison with 2-deoxy-2-fluoro- D-glucose-positron emission tomography/computed tomography. NMR Biomed 24: 1181–1190.
  10. 10. Wu X, Korkola P, Pertovaara H, Eskola H, Jarvenpaa R, et al. (2011) No correlation between glucose metabolism and apparent diffusion coefficient in diffuse large B-cell lymphoma: a PET/CT and DW-MRI study. Eur J Radiol 79: e117–121.
  11. 11. Hirano M, Satake H, Ishigaki S, Ikeda M, Kawai H, et al. (2012) Diffusion-weighted imaging of breast masses: comparison of diagnostic performance using various apparent diffusion coefficient parameters. AJR Am J Roentgenol 198: 717–722.
  12. 12. Regier M, Derlin T, Schwarz D, Laqmani A, Henes FO, et al. (2012) Diffusion weighted MRI and 18F-FDG PET/CT in non-small cell lung cancer (NSCLC): does the apparent diffusion coefficient (ADC) correlate with tracer uptake (SUV)? Eur J Radiol 81: 2913–2918.
  13. 13. Gu J, Khong PL, Wang S, Chan Q, Law W, et al. (2011) Quantitative assessment of diffusion-weighted MR imaging in patients with primary rectal cancer: correlation with FDG-PET/CT. Mol Imaging Biol 13: 1020–1028.
  14. 14. Wong CS, Gong N, Chu YC, Anthony MP, Chan Q, et al. (2012) Correlation of measurements from diffusion weighted MR imaging and FDG PET/CT in GIST patients: ADC versus SUV. Eur J Radiol 81: 2122–2126.
  15. 15. Nakamura K, Joja I, Kodama J, Hongo A, Hiramatsu Y (2012) Measurement of SUVmax plus ADCmin of the primary tumour is a predictor of prognosis in patients with cervical cancer. Eur J Nucl Med Mol Imaging 39: 283–290.
  16. 16. Nakamura K, Joja I, Fukushima C, Haruma T, Hayashi C, et al. (2013) The preoperative SUVmax is superior to ADCmin of the primary tumour as a predictor of disease recurrence and survival in patients with endometrial cancer. Eur J Nucl Med Mol Imaging 40: 52–60.
  17. 17. Wu X, Dastidar P, Pertovaara H, Korkola P, Järvenpää R, et al. (2011) Early Treatment Response Evaluation in Patients with Diffuse Large B-Cell Lymphoma-A Pilot Study Comparing Volumetric MRI and PET/CT. Mol Imaging Biol 13: 785–92.
  18. 18. Wu X, Pertovaara H, Dastidar P, Vornanen M, Paavolainen L, et al. (2013) ADC measurements in diffuse large B-cell lymphoma and follicular lymphoma: a DWI and cellularity study. Eur J Radiol 82: e158–164.
  19. 19. Lapela M, Leskinen S, Minn HR, Lindholm P, Klemi PJ, et al. (1995) Increased glucose metabolism in untreated non-Hodgkin's lymphoma: a study with positron emission tomography and fluorine-18-fluorodeoxyglucose. Blood 86: 3522–3527.
  20. 20. Schöder H, Noy A, Gönen M, Weng L, Green D, et al. (2005) Intensity of 18fluorodeoxyglucose uptake in positron emission tomography distinguishes between indolent and aggressive non-Hodgkin's lymphoma. J Clin Oncol 23: 4643–4651.
  21. 21. Wu X, Pertovaara H, Korkola P, Vornanen M, Eskola H, et al. (2012) Glucose metabolism correlated with cellular proliferation in diffuse large B-cell lymphoma. Leuk Lymphoma 53: 400–405.
  22. 22. Shou Y, Lu J, Chen T, Ma D, Tong L (2012) Correlation of fluorodeoxyglucose uptake and tumor-proliferating antigen Ki-67 in lymphomas. J Cancer Res Ther 8: 96–102.
  23. 23. Papajík T, Mysliveček M, Sedová Z, Buriánková E, Procházka V, et al. (2011) Standardised uptake value of 18F-FDG on staging PET/CT in newly diagnosed patients with different subtypes of non-Hodgkin's lymphoma. Eur J Haematol 86: 32–37.
  24. 24. Tang B, Malysz J, Douglas-Nikitin V, Zekman R, Wong RH, et al. (2009) Correlating metabolic activity with cellular proliferation in follicular lymphomas. Mol Imaging Biol 11: 296–302.
  25. 25. Punwani S, Prakash V, Bainbridge A, Taylor SA, Bandula S, et al. (2010) Quantitative diffusion weighted MRI: a functional biomarker of nodal disease in Hodgkin lymphoma? Cancer Biomark 7: 249–259.
  26. 26. Ho KC, Lin G, Wang JJ, Lai CH, Chang CJ, et al. (2009) Correlation of apparent diffusion coefficients measured by 3T diffusion-weighted MRI and SUV from FDG PET/CT in primary cervical cancer. Eur J Nucl Med Mol Imaging 36: 200–208.
  27. 27. Matsubayashi RN, Fujii T, Yasumori K, Muranaka T, Momosaki S (2010) Apparent Diffusion Coefficient in Invasive Ductal Breast Carcinoma: Correlation with Detailed Histologic Features and the Enhancement Ratio on Dynamic Contrast-Enhanced MR Images. J Oncol DOI: 10.1155/2010/821048.
  28. 28. Kwee TC, Ludwig I, Uiterwaal CS, van Ufford HM, Vermoolen MA, et al. (2011) ADC measurements in the evaluation of lymph nodes in patients with non-Hodgkin lymphoma: feasibility study. Magma 24: 1–8.
  29. 29. Jagannath S, Velasquez WS, Tucker SL, Fuller LM, McLaughlin PW, et al. (1986) Tumor burden assessment and its implication for a prognostic model in advanced diffuse large-cell lymphoma. J Clin Oncol 4: 859–865.
  30. 30. Johnson PW, Whelan J, Longhurst S, Stepniewska K, Matthews J, et al. (1993) Beta-2 microglobulin: a prognostic factor in diffuse aggressive non-Hodgkin's lymphomas. Br J Cancer 67: 792–797.
  31. 31. Hagberg H, Killander A, Simonsson B (1983) Serum beta 2-microglobulin in malignant lymphoma. Cancer 51: 2220–2225.
  32. 32. Hashimoto K, Ikeda Y, Korenaga D, Tanoue K, Hamatake M, et al. (2005) The impact of preoperative serum C-reactive protein on the prognosis of patients with hepatocellular carcinoma. Cancer 103: 1856–1864.
  33. 33. Crumley AB, McMillan DC, McKernan M, Going JJ, Shearer CJ, et al. (2006) An elevated C-reactive protein concentration, prior to surgery, predicts poor cancer-specific survival in patients undergoing resection for gastro-oesophageal cancer. Br J Cancer 94: 1568–1571.
  34. 34. Grivennikov SI, Greten FR, Karin M (2010) Immunity, inflammation, and cancer. Cell 140: 883–899.
  35. 35. Topolcan O, Holubec L Jr (2008) The role of thymidine kinase in cancer diseases. Expert Opin Med Diagn 2: 129–141.
  36. 36. Cao Y, Shi YX, Chen JO, Tan YT, Cai YC, et al. (2012) Serum C-reactive protein as an important prognostic variable in patients with diffuse large B cell lymphoma. Tumour Biol 33: 1039–1044.