To the Editor:
The recent review about magnetic resonance imaging (MRI) of the bone marrow in hematologic malignancies has suggested that MRI had little or no value in the evaluation of myelofibrosis (MF).1 Although MR findings did not also appear beneficial in the differential diagnosis of myelofibrosis, our data showed that MRI may have some importance in predicting the prognosis of this disorder.
Thirteen patients with MF were evaluated (11 patients with primary MF and 2 patients with secondary MF) and diagnoses of MF were made according to previously defined criteria.2 Before the MR examination, bone marrow aspirates and biopsies from posterior iliac crest were performed. Clinical prognostic staging of the patients were determined on the basis of their hemoglobin value and percentage of white blood cell precursors in peripheral circulation, as it has been proposed by Visani et al.2 Myelofibrosis in bone marrow biopsies were evaluated in three different stages according to the criteria defined by Ward and Block.3
The MR investigations were performed with a 0.5-Tesla MR scanner (Gyroscan-T5-Holland, Philips). Coronal T1-weighted images from the lower lumbar vertebrae, sacrum, iliac wings, and proximal femur were obtained in contiguous 5-mm slices in a 256 × 256 matrix with TR, 400 milliseconds (ms); TE, 20 ms; and a number of signal acquisitions, 2. Short TI inversion recovery (STIR) coronal images of the same regions were obtained in 6-mm slices in a 256 × 256 matrix with TR, 1,800 to 2,200 ms; TE, 30 to 40 ms; and T1, 140 to 150 ms. In T1-weighted images, fatty tissue with high signal intensity appeared bright, whereas cellular bone marrow had a low signal intensity with dark appearance. In STIR sequences, the signal from fatty bone marrow was suppressed, so that the cellular marrow was brighter. Other MRI sequences, such as water-saturated T2-weighted images (TR/TE: 3,200 to 5,000/102 ms; 5-mm slice thickness and 1-mm gap; NEX: 3; and 256 × 160 matrix size), fat-saturated T2-weighted images (TR/TE: 5,000 to 6,500/102 ms; 5-mm slice thickness and 1-mm gap; NEX: 2; 256 × 160 matrix size), and fat-saturated T1-weighted images (TR/TE: 360/9; 5-mm slice thickness and 1-mm gap; NEX: 3; 256 × 160 matrix size) were also used for detecting the extension of the cellularity and the fibrosis. However, these sequences did not provide any additional information with respect to the bone marrow cellularity in comparison with the STIR techniques. MR sequences were classified according to the bone marrow cellularity, resting fatty marrow, and marrow fibrosis. The patterns were categorized as follows.
Pattern Ia.
Low signal intensity in iliac wings and lower lumbar vertebrae and high signal intensity in femoral epiphysis and/or thoracanteric region in T1-weighted images and high signal intensity in all STIR sequences were accepted as pattern Ia. That pattern was compatible with hypercellular bone marrow and early phase MF.
Pattern Ib.
Low signal intensity in all regions with T1-weighted images and high signal intensity in STIR sequences were also defined to be compatible with hypercellular bone marrow and early phase MF. Although pattern Ia and Ib were comparable in predicting the phase of MF, pattern Ia images also advocated still remaining fatty bone marrow, especially in the femoral epiphyses.
Pattern II.
Low signal intensity was apparent in all regions both in T1-weighted images and in STIR sequences. That pattern was accepted to predict late phase MF with extended fibrosis and low hematopoietic activity.
Correlation between the MRI patterns and prognostic staging, histopathologic staging, and serum LDH levels were determined by Spearman test.
The results are summarized in Table 1. Histopathologic staging in MF according to bone marrow biopsies was not found to be correlated with clinical prognostic staging and MRI findings (P > .05), as it has been reported by Varki et al.4 However, defined MRI patterns were well correlated with the prognostic staging and serum LDH levels (P = .026 and P = .007, respectively). Kaplan et al5suggested that MRI findings in MF could characterize the phase, severity, and progression of the disease, but that controversial data assessed correlation between the MRI findings, serum LDH levels, and spleen size. However, in our study, we evaluated this correlation with a confirmed clinical prognostic staging.2
Clinical Findings, Prognostic Staging, and MRI Pattern of the Patients With Myelofibrosis
| Patient No. . | Age/ Sex . | Diagnosis . | Hb (g/dL) . | Platelet Count (×109/μL) . | Myeloid Precursors (%) . | LDH (IU/L) . | Histopathologic Staging* . | Prognostic Staging-151 . | MRI Pattern . |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 66/F | Secondary | 5.3 | 27 | 3 | 1,209 | 2 | 2 | 1A |
| 2 | 64/F | Secondary | 7.8 | 56 | 15 | 2,123 | 3 | 3 | 1A |
| 3 | 30/F | Primary | 10.9 | 488 | 6 | 519 | 1 | 1 | 1A |
| 4 | 37/M | Primary | 8.7 | 490 | 23 | 1,617 | 3 | 3 | 2 |
| 5 | 66/M | Primary | 7.9 | 242 | 14 | 1,366 | 1 | 3 | 2 |
| 6 | 59/M | Primary | 11.6 | 320 | 7 | 920 | 1 | 1 | 1A |
| 7 | 58/M | Primary | 9.3 | 74 | 29 | 980 | 1 | 3 | 1B |
| 8 | 61/M | Primary | 12.8 | 310 | 4 | 853 | 1 | 1 | 1B |
| 9 | 34/F | Primary | 8.4 | 230 | 17 | 1,450 | 2 | 3 | 2 |
| 10 | 57/M | Primary | 5.3 | 5 | 34 | 2,300 | 3 | 3 | 2 |
| 11 | 52/M | Primary | 12.6 | 240 | 8 | 736 | 1 | 1 | 1A |
| 12 | 66/M | Primary | 7.5 | 698 | 10 | 528 | 2 | 2 | 1B |
| 13 | 48/F | Primary | 10.3 | 204 | 24 | 1,785 | 2 | 3 | 1B |
| Patient No. . | Age/ Sex . | Diagnosis . | Hb (g/dL) . | Platelet Count (×109/μL) . | Myeloid Precursors (%) . | LDH (IU/L) . | Histopathologic Staging* . | Prognostic Staging-151 . | MRI Pattern . |
|---|---|---|---|---|---|---|---|---|---|
| 1 | 66/F | Secondary | 5.3 | 27 | 3 | 1,209 | 2 | 2 | 1A |
| 2 | 64/F | Secondary | 7.8 | 56 | 15 | 2,123 | 3 | 3 | 1A |
| 3 | 30/F | Primary | 10.9 | 488 | 6 | 519 | 1 | 1 | 1A |
| 4 | 37/M | Primary | 8.7 | 490 | 23 | 1,617 | 3 | 3 | 2 |
| 5 | 66/M | Primary | 7.9 | 242 | 14 | 1,366 | 1 | 3 | 2 |
| 6 | 59/M | Primary | 11.6 | 320 | 7 | 920 | 1 | 1 | 1A |
| 7 | 58/M | Primary | 9.3 | 74 | 29 | 980 | 1 | 3 | 1B |
| 8 | 61/M | Primary | 12.8 | 310 | 4 | 853 | 1 | 1 | 1B |
| 9 | 34/F | Primary | 8.4 | 230 | 17 | 1,450 | 2 | 3 | 2 |
| 10 | 57/M | Primary | 5.3 | 5 | 34 | 2,300 | 3 | 3 | 2 |
| 11 | 52/M | Primary | 12.6 | 240 | 8 | 736 | 1 | 1 | 1A |
| 12 | 66/M | Primary | 7.5 | 698 | 10 | 528 | 2 | 2 | 1B |
| 13 | 48/F | Primary | 10.3 | 204 | 24 | 1,785 | 2 | 3 | 1B |
*Pathologic staging of myelofibrosis: 1, a cellular phase with panmyelosis, megakaryocytic and erytroid proliferation, and minimal increase in reticulin fibers; 2, myelofibrosis, with only 30% of the marrow space showing hematopoietic tissue; 3, myelofibrosis with osteosclerosis, wide-spread fibrosis and a minimal amount of hematopoietic tissue.
Clinical prognostic staging2: 1, good-risk, with Hb >10 g/dL and percentage of myeloid precursors <10%; 2, intermediate-risk, with Hb <10 g/dL and percentage of myeloid precursors <10%; 3, high-risk, with >10% of myeloid precursors, independent of Hb level.
In conclusion, MRI appeared as a useful and noninvasive method for the determination of phase and severity of MF, and the defined MRI patterns were comparable with the clinical prognostic staging.
Dr Alpdogan et al studied 13 patients with myelofibrosis with magnetic resonance (MR) images of the lower lumbar spine, pelvis, and proximal femurs. They concluded that MR patterns derived from changes in the normal appearance of the bone marrow in the above areas correlated with serum LDH levels and with the clinical prognostic staging of the disease but not with the histopathologic staging. Kaplan et al1-1 in 1992 reported similar findings in 14 patients with chronic myeloproliferative disease; they found that absence of fatty marrow in both the greater trochanter and femoral capital epiphysis was associated with significantly higher serum LDH values and lower serum cholesterol values. Alpdogan et al studied a larger number of patients with myelofibrosis and included findings on STIR images that may help differentiate hypercellular from fibrotic marrow.
In our review on hematologic malignancies of the bone marrow, the data of Kaplan et al1-1 are discussed in detail in the chapter on “Miscellaneous malignant bone marrow disorders.”2 I believe that there is no doubt that myelofibrosis cannot be distinguished from other hematologic malignancies based on morphological changes on MR images of the bone marrow. It is also well established by now that MR images, by providing the opportunity to examine a large volume of the bone marrow, may complement the bone marrow biopsy in assessing the extent, severity, and prognosis of most of the hematologic malignancies. However, before bone marrow MR imaging is used to replace some of the repeated bone marrow biopsies that patients with myeloproliferative diseases require, it is my opinion that larger numbers of patients need to be studied and that research should be guided towards obtaining quantitative measurements that can serve as more accurate indices of tumor burden.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal