Whole-body DWIBS MR sequences provided high diagnostic quality images comparable to PET images (figure 1). There has been image distortion in abdomen and peritoneal reflections due to high susceptibility difference in these regions. In one patient DWIBS revealed possible para-aortic and mesenteric adenopathies that showed no activity on PET scans consistent with absence of involvement of the disease process (figure 2). This resulted in over-staging of tumor load by DWIBS where STIR was inconclusive. In agreement with previous authors (21), we observed that at b=1000, metastasis had higher measurements that normal tissues with the exception of spleen. Consequently, lymphomatous involvement of spleen could not be assessed by diffusion studies. ADC mapping were useful in delineating lesions of less serious nature as inflammation. Figure (1) reveals right lung parenchymatous infection of intermediate value b=400, that was not apparent on PET images. Yet, ADC mapping were variable and sometimes were not conclusive regarding differentiating between benign and malignant tissues. DWIBS showed multiple nodal involvement that were not confirmed by molecular uptake of FDG on PET studies. These nodes were likely small <1 cm3 to morphologically categorize them as metastasis. Also, the reverse may be true, as we had a patient with multiple nodal involvement with disclosed vertebral metastasis only on PET/CT study not revealed by either DWIBS or STIR MR sequences (figure 2).
Regarding STIR sequence, it revealed more skeletal deposits and metastatic sites that did DWIBS achieving a higher concordance with tracer uptake on PET scans regarding the detected number of lesions and thus lymphoma staging.
The median (interquartile range) number of lesions detected by DWIBS, STIR and PET/CT was 4 (2-6), 4.5 (3-7), and 5 (3.0-7.5) lesions, respectively. On the other hand, the median (interquartile range) stage of lymphoma as determined by DWIBS, STIR and PET/CT was 2 (2.0-3.25), 2.5 (2.0-5.0), and 3.5 (2.0-5.25), respectively (Table 1).
DWIBS had 81.2% (22.5%) concordance with PET/CT as regards the number of lesions detected, while the concordance of STIR with PET/CT was 95.0% (8.8%) (P < 0.05). The concordance between DWIBS and PET/CT as regards the stage of lymphoma was 45.5%, whereas the concordance between STIR and PET/CT was 81.8% (P < 0.05) (Table 2).
There was a strong correlation between DWIBS and PET/CT (correlation coefficient = 0.935, P < 0.001), and between STIR and PET/CT (correlation coefficient = 0.984, P < 0.001) as regards the number of lesions detected (table 3). Likewise, there was a strong correlation between DWIBS and PET/CT (correlation coefficient = 0.804, P < 0.001), and between STIR and PET/CT (correlation coefficient = 0.871, P < 0.001) as regards the staging of lymphoma (Tables 4).
Table (1) Number of lesions and lymphoma stage as determined by the three radiologic techniques.
| | number of lesions by DWIBS | number of lesions by STIR | number of lesions by PET/CT | stage by DWIBS | stage by STIR | stage by PET/CT |
Median | 4.00 | 4.50 | 5.00 | 2.00 | 2.50 | 3.50 |
Percentiles | 25th | 2.00 | 3.00 | 3.00 | 2.00 | 2.00 | 2.00 |
50th | 4.00 | 4.50 | 5.00 | 2.00 | 2.50 | 3.50 |
75th | 6.00 | 7.00 | 7.25 | 3.25 | 5.00 | 5.25 |
Data are median (interquartile range).
Table (2) Concordance between DWIBS or STIR and PET-CT as regards number of lesions and stage of lymphoma
Variable | DWIBS | STIR | P value |
Concordance with PET-CT as regards number of lesions detected (%) | 81.15 (22.5) | 94.97 (8.81) | < 0.05 |
Concordance with PET-CT as regards stage of lymphoma | | | < 0.05 |
Concordant | 10 [45.5%] | 18 [81.8%] | |
Non-concordant | 12 [54.5%] | 4 [18.2%] | |
Data are mean (SD) or number [%].
Table (3) Correlation between number of lesions detected by DWIBS or STIR and number of lesions detected by PET-CT
| | number of lesions by DWIBS | number of lesions by STIR |
number of lesions by PET/CT | Correlation Coefficient | 0.935 | 0.984 |
P value | < 0.001 | < 0.001 |
Table (4) Correlation between stage of lymphoma as determined by DWIBS or STIR and stage of lymphoma as determined by PET-CT
| | stage by DWIBS | Stage by STIR |
stage by PET/CT | Correlation Coefficient | 0.804 | 0.871 |
P value | < 0.001 | < 0.001 |