Magnetic resonance imaging of bone marrow in oncology, Part 1
- 3.3k Downloads
Magnetic resonance imaging plays an integral role in the detection and characterization of marrow lesions, planning for biopsy or surgery, and post-treatment follow-up. To evaluate findings in bone marrow on MR imaging, it is essential to understand the normal composition and distribution of bone marrow and the changes in marrow that occur with age, as well as the basis for the MR signals from marrow and the factors that affect those signals. The normal distribution of red and yellow marrow in the skeleton changes with age in a predictable sequence. Important factors that affect MR signals and allow detection of marrow lesions include alterations in fat–water distribution, destruction of bony trabeculae, and contrast enhancement. This two-part article reviews and illustrates these issues, with an emphasis on the practical application of MR imaging to facilitate differentiation of normal marrow, tumor, and treatment-related marrow changes in oncology patients.
KeywordsMagnetic resonance imaging Marrow Metastases Tumor Sarcoma Treatment effects
Bone marrow is one of the largest organs in the body after the osseous skeleton, skin, and body fat [1, 2], and is mainly composed of fat and water [1, 3]. Marrow can weigh up to 3 kg in an adult . As such, at least some marrow is visible on nearly every CT and MR image, and can pose diagnostic challenges at each stage of a cancer patient’s clinical course—from the initial detection of the tumor through staging and subsequent post-treatment evaluation, to surveillance for tumor recurrence. To a large extent, MRI is superior to other imaging modalities in evaluating bone marrow because it is remarkably sensitive in detecting lipid; therefore, MRI can be used to detect processes that alter the relative amounts of fat and water in bone marrow. In order to interpret the appearances of marrow seen on MRI and to distinguish normal from abnormal, it is important to understand the normal components and composition of bone marrow, which vary greatly with age and anatomic location within the skeleton [4, 5, 6]
Components and composition of normal bone marrow
On gross examination, bone marrow appears red (hematopoietic marrow) or yellow (fatty marrow) depending on its predominant components. Despite its unclear physiological role, fat is the major component of both yellow marrow and (to a lesser extent) red marrow. In the iliac crests of adults, approximately 50–60% of marrow is hematopoietic . The cellular composition of red marrow consists of 60% hematopoietic cells and 40% adipocytes; its chemical composition is 40–60% lipid, 30–40% water, and 10–20% proteins [1, 2]. In contrast, yellow marrow is nearly entirely composed of adipocytes (95%), with its chemical composition being 80% lipid, 15% water, and 5% protein .
Conversion of bone marrow
In normal adult marrow, yellow marrow may reconvert to red marrow in the event of a functional demand for increased hematopoiesis. Benign conditions triggering reconversion include heavy smoking, long distance running, obesity, and middle age in women [8, 9]. Chronic disorders resulting in anemia (such as hemoglobinopathies and chronic infection) are also associated with diffuse red marrow hyperplasia . Diffuse hematopoietic marrow hyperplasia may be difficult to distinguish from diffuse marrow disease on MRI .
The reconversion process proceeds in the exact reverse sequence from that of the initial conversion (Fig. 2): namely, from the central to the peripheral skeleton; and in the long bones, from the proximal metaphyses and then the distal metaphyses to the diaphyses . Reconversion occurs more quickly in flat bones (such as the sternum, spine, and scapula) because they retain cellular marrow throughout life .
Factors affecting MR signal intensity of marrow
Vanel et al. [12, 13] have summarized three main factors that affect the signal intensity of bone marrow on MRI: the fat and water content of the marrow, the presence of bony trabeculae, and the use of contrast material. The MRI appearance of marrow is also dependent on the particular MR technique employed .
Fat and water content
Bone marrow is rich in fat and water, which together contribute the bulk of the marrow signal seen on MRI. Alteration in the balance of lipid versus water in marrow is a principal factor affecting the signal intensity of marrow on MRI.
Studies have suggested the potential of the chemical-shift MR technique as a complementary tool in predicting the likelihood that a marrow lesion is neoplastic. Disler et al.  studied the relative signal intensity ratios (defined as the ratio of signal intensity of a lesion versus control tissue on opposed-phase imaging divided by the corresponding ratio on in-phase imaging) of lesions in appendicular and pelvic bones. They found that the average relative signal intensity ratio for neoplastic lesions was 1.03, versus 0.61 for non-neoplastic lesions; using a relative signal intensity ratio higher than 0.81 resulted in a sensitivity and specificity of 95% for detection of a tumor. Zajick et al.  studied proportional change in signal intensity between in-phase and opposed-phase imaging of neoplastic and non-neoplastic lesions of the spine. They suggested that differences in signal intensity loss among normal, benign and malignant lesions were significant, with a decrease in signal intensity greater than 20% on opposed-phase imaging compared with in-phase imaging potentially being useful in distinguishing benign and malignant lesions. Erly et al.  similarly reported significant differences in signal intensity between benign and malignant spinal compression fractures on in-phase and opposed-phase imaging.
Another tool, the fat-suppression technique, facilitates detection of marrow lesions by suppressing signals from surrounding fat-containing marrow [18, 19]. A radiofrequency pulse of the same frequency as that of lipid is applied before the excitation pulse in order to selectively suppress the signals from lipid without affecting signals from other tissues—including those with a T1 relaxation time similar to that of lipid (such as some hematomas and contrast-enhanced tissues). Use of fat suppression is essential when using a fast (or turbo) SE T2-weighted sequence, as fatty marrow has moderately high signal that otherwise can mask the presence of concomitant tumor deposits on that sequence. However, this selective fat suppression technique is sensitive to susceptibility artifacts and requires excellent magnetic field homogeneity and a high magnetic field strength. In the presence of metal or air or when using a large field of view, the resultant fat suppression frequently is inhomogeneous; in fact, suppression of the water signal can inadvertently occur, thus obscuring pathologic conditions.
Alternatively, the short-tau inversion recovery (STIR) pulse sequence is less affected by inhomogeneous magnetic fields. In STIR, a 180° inversion pulse is delivered at the beginning of the sequence to invert the longitudinal magnetization; the differences in longitudinal magnetization of various tissues thus can be increased, improving T1-weighted contrast between tissues with relatively similar T1 relaxation times . Selecting the inversion time (TI) such that the longitudinal magnetization of lipid has recovered to zero at the time of echo formation eliminates the signal from fatty tissue, thereby increasing the conspicuity of marrow abnormalities; at 1.5 T, use of a TI in the range of 150–170 ms will achieve this result. STIR is a highly sensitive technique in marrow lesion detection and is considered superior to T1-weighted SE in some studies ; however, the findings on STIR images often lack specificity because the STIR sequence suppresses signals from any other tissue that has a T1 relaxation time similar to that of fat. Additionally, in oncologic imaging, STIR may lead to overestimation of the extent of a marrow lesion due to its demonstration of surrounding marrow edema pattern as high signal that may be similar to that of the lesion itself.
Conflicting results have been reported for DWI in marrow imaging. In a study of benign and pathologic spinal compression fractures using SSFP DWI, Baur et al.  reported that all malignant compression fractures were hyperintense to normal marrow, whereas most benign compression fractures were hypointense to normal marrow. However, a subsequent study by Castillo et al.  showed no advantage of DWI in the detection and characterization of vertebral metastases compared with T1-weighted imaging; metastases were hypointense to normal marrow in 3 out of 5 patients with focal disease and 5 out of 10 patients with multiple lesions. In a study of lymphoma in the iliac bones by Yasumoto et al. , normal red marrow appeared hyperintense relative to lymphoma on DWI, with a sensitivity of 77% and specificity of 92.5%. Takahara et al.  suggested a feasible multiple thin-slice whole-body DWI technique and showed that MRI, like PET/CT, might potentially play a role in tumor survey and follow-up. However, more studies are needed to evaluate the clinical utility of DWI in marrow imaging.
Distinguishing normal and abnormal marrow on MRI: normal variations in marrow, including marrow heterogeneity
Focal or diffuse red marrow generally has relatively low signal intensity on fat-suppressed T2-weighted SE images (Fig. 9b,d). In adults, red marrow deposits generally would not be present in epiphyses or apophyses (Fig. 9c) unless severe reconversion was already present throughout the diaphyses and metaphyses of the affected bone. In the setting of advanced marrow reconversion, the borders of regions of red marrow become more sharply defined and their signals further decrease on T1-weighted SE imaging; limited contrast enhancement in red marrow regions on T1-weighted SE imaging is another finding that suggests a benign nature [11, 35].
Magnetic resonance imaging is an excellent noninvasive modality for evaluating bone marrow and detecting marrow lesions, as it provides information at the level of cellular and chemical composition, in addition to gross morphologic data. Knowledge of normal marrow components and composition and their variation, as well as of factors that alter MR signal intensity, is important for optimal interpretation of MR images. The signal intensity, morphology, and location of marrow findings on MRI can be used to provide more accurate diagnoses, to guide treatment, and to follow therapy-related changes. Various MR imaging techniques are available to accentuate the different chemical and cellular compositions of normal marrow and marrow diseases. Although MRI is more sensitive than specific in detecting marrow changes, integrating all the clinical and radiologic data can result in more useful interpretations. The MRI evaluation of tumor and post-treatment changes in marrow, as well as some promising complementary MR techniques under evaluation, will be discussed in part 2 of this article.
- 2.Snyder WS, Cook MJ, Nasset ES, Karhausen LR, Howells CP, Tipton IH. International Commission on Radiological Protection. Report of the task group on reference man. Oxford: Pergamon; 1975. pp. 79–98.Google Scholar
- 4.Piney A. The anatomy of the bone marrow: with special reference to the distribution of the red marrow. BMJ 1922; 28: 792–795.Google Scholar
- 8.Shellock FG, Morris E, Deutsch AL, Mink JH, Kerr R, Boden SD. Hematopoietic bone marrow hyperplasia: high prevalence on MR images of the knee in asymptomatic marathon runners. Am J Roentgenol 1992; 158: 335–338.Google Scholar
- 9.Poulton TB, Murphy WD, Duerk JL, Chapek CC, Feiglin DH. Bone marrow reconversion in adults who are smokers: MR imaging findings. Am J Roentgenol 1993; 161: 1217–1221.Google Scholar
- 15.Disler DG, McCauley TR, Ratner LM, Kesack CD, Cooper JA. In-phase and out-of-phase MR imaging of bone marrow: prediction of neoplasia based on the detection of coexistent fat and water. Am J Roentgenol 1997; 169: 1439–1447.Google Scholar
- 19.Mitchell DG, Vinitski S, Rifkin MD, Burk DL Jr. Sampling bandwidth and fat suppression: effects on long TR/TE MR imaging of the abdomen and pelvis at 1.5 T. Am J Roentgenol 1989; 153: 419–425.Google Scholar
- 20.Bushong S. Magnetic resonance imaging: physical and biological principles. 3rd edn. St. Louis: Mosby; 2003.Google Scholar
- 21.Mirowitz SA, Apicella P, Reinus WR, Hammerman AM. MR imaging of bone marrow lesions: relative conspicuousness on T1- weighted, fat-suppressed T2-weighted, and STIR images. Am J Roentgenol 1994; 162: 215–221.Google Scholar