| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Departments of Diagnostic Radiology (L.J.H., S.M., K.M.J.) and Internal Medicine (B.A.B.), Yale University School of Medicine, 333 Cedar St, New Haven, CT 06510. From the 1996 RSNA scientific assembly. Received March 16, 1998; revision requested May 12; revision received August 4; accepted October 13. Address reprint requests to L.J.H.
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Nineteen patients with newly diagnosed breast cancer were imaged by using a 1.5-T echo-planar MR system. By using a table sweep method, 180 contiguous axial images were obtained from the cranial vertex through the feet with T2-weighted spin-echo and inversion-recovery sequences. Results were compared with those of conventional imaging. Therapeutic decisions based on echo-planar MR imaging and conventional imaging results were compared. Diagnostic truth was determined by means of tissue diagnosis, further imaging findings, and follow-up findings (median, 18 months).
RESULTS: Staging with total-body echo-planar MR imaging was correct in 18 patients (95%)—eight with metastases and 10 without—while staging with conventional imaging was correct in 15 patients (79%). In one patient, both echo-planar MR imaging and conventional imaging findings incorrectly indicated probable metastases. In one patient thought to have bone metastases at conventional imaging, echo-planar MR imaging findings were normal, which was correct. Two patients with stage IV disease were not suspected to have disease at conventional imaging: One had liver involvement and the other had skeletal metastases. The therapeutic decisions in these two patients were altered by the echo-planar MR imaging results.
CONCLUSION: Total-body echo-planar MR imaging was at least as accurate as conventional imaging for staging newly diagnosed breast cancer and was faster, simpler, and completely noninvasive.
Index terms: Bone neoplasms, radionuclide studies, 30.12172, 30.33, 40.12172, 40.33 • Breast neoplasms, 00.32, 00.33 • Neoplasms, CT, **.1211, .302 • Neoplasms, MR, **.121411, **.121413, **.30 • Neoplasms, radionuclide studies, **.12172, **.30 • Neoplasms, staging, **.30
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Common staging methods have known limitations. False-negative bone scintigraphic findings have been reported with radiographically visible lesions (3,4), and common benign processes may have radionuclide accumulation indistinguishable from that in metastases. Some staging guidelines for breast cancer recommend liver function tests and chest radiography (5,6), despite limited sensitivity and specificity. More costly computed tomography (CT) of the abdomen and chest exposes patients to intravenous contrast agents and ionizing radiation.
Magnetic resonance (MR) imaging is not commonly used in routine breast cancer staging in spite of evidence that it is sensitive and specific for the detection of metastases in general. This is in part because of concerns regarding cost, long imaging time, and limited availability. For the detection of breast cancer skeletal metastases, conventional MR imaging is reported to be more sensitive than bone scintigraphy (7–11) and more accurate than iliac crest biopsy (12). MR imaging is comparable with CT for the detection of liver involvement (13,14) and thoracic adenopathy (15,16). Its ability to depict lung metastases may be limited by susceptibility and motion artifact, although there is a report that 0.5-T MR imaging is at least as sensitive as CT (17).
Total-body MR imaging with an echo-planar technique has been described recently (18) and can be accomplished in as little as 18 seconds. If this method provides accurate essential staging data while reducing time, cost, ionizing radiation exposure, and exposure to intravenous contrast agents, a different approach to the detection of metastatic disease might be realized. The purpose of this study of total-body echo-planar MR imaging was to assess the ability to use it to correctly diagnose metastases in a group of patients with newly diagnosed breast cancer, to compare its performance with that of conventional imaging, and to evaluate its effect on therapeutic decisions.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Patient Population
The requirements for patient inclusion were newly diagnosed biopsy-proved breast cancer, referral for bone scintigraphy, no prior history of malignancy, no contraindications to MR imaging, and a willingness to enroll. Potential patients were identified by an attending oncologist or by reviewing nuclear medicine scheduling records. In all cases, the clinical decision to perform staging work-up was made prior to consideration for inclusion in this study. Our institution's human investigation committee approved the protocol.
Informed consent was obtained from 25 women undergoing evaluation for treatment planning; 19 underwent imaging. One patient withdrew because of feeling overwhelmed by her diagnosis, one patient withdrew after being notified of positive bone scintigraphic results, one patient could not undergo imaging because of metal within a chest wall tissue expander, and three patients could not undergo imaging because of scheduling conflicts.
Patients were aged 31–70 years (Table 1). Primary tumor sizes ranged from 1.4 cm to the entire breast, and the most common histologic diagnosis was invasive ductal carcinoma. The primary tumor was not definitely identified in one patient with invasive ductal carcinoma involving two intramammary lymph nodes, the largest of which was 1.2 cm in diameter.
|
Total-Body Echo-planar MR Imaging
Studies were performed by using a 1.5-T magnet (Signa version 4.8; GE Medical Systems, Milwaukee, Wis) outfitted with resonant gradients for echo-planar imaging (Instascan; Advanced NMR, Wilmington, Mass). By using the previously reported technique for total-body echo-planar MR imaging (18), two total-body sweeps were performed: T2-weighted spin-echo MR imaging (2,000/60 [repetition time msec/echo time msec] in 18 patients and 2,000/80 in one patient) and inversion-recovery MR imaging (2,000/60/140 [repetition time msec/echo time msec/inversion time msec] in 18 patients and 1,500/60/140 in one patient). Successive single-shot images were obtained at the isocenter as the patient table exited the magnet at 5 mm/sec. Each sweep consisted of acquisition of 180 contiguous 10-mm-thick axial images from the cranial vertex through the feet. The field of view was 40 x 20 cm. The acquisition matrix was 128 x 128 voxels, with the read direction interpolated to 256 pixels in postprocessing. No intravenous contrast agents were administered. Each imaging sweep took approximately 6 minutes.
The echo-planar MR imaging studies were interpreted independently by two radiologists (S.M.C., K.M.J.) blinded to the extent and results of the conventional radiologic evaluation. Image review was performed by scrolling through the stacks of 180 images displayed on a computer monitor. The likelihood of metastasis was graded on a five-point scale: "1" indicated metastasis definitely absent, "2" indicated metastasis probably absent, "3" indicated indeterminate findings, "4" indicated metastasis probably present, and "5" indicated metastasis definitely present. Likelihood was recorded for the skeletal system (skull, spine, pelvis, ribs, upper extremities, lower extremities), lung, pleura, liver, lymph nodes, spleen, omentum, adnexa, and brain. An overall impression was also recorded. Discordant results were resolved through consensus. A third radiologist helped to obtain consensus for two specific findings (one in the brain, one in a long bone).
Conventional Imaging
The conventional staging work-up in each patient was ordered by the attending surgeon and/or oncologist without knowledge of the echo-planar MR imaging results. Bone scintigraphy was performed in all patients (Table 2). Fourteen patients underwent chest radiography. Chest and/or abdominal CT was performed in 13 patients. Bone radiography was performed in seven patients, six with skeletal symptoms. Abdominal ultrasonography (US) was performed in only two patients.
|
Studies were double read; both radiologists were blinded to the echo-planar MR imaging results. The second reader (L.J.H.) assigned findings to the same five-point scale used for the echo-planar MR imaging interpretations. Likelihood of metastasis at anatomic sites and an overall impression were recorded. Discordant observations were resolved through consensus.
Effect on Therapeutic Decision
A panel of two oncologists, neither of whom was the patient's primary oncologist, rendered a pair of therapeutic decisions for each patient on the basis of findings from each of the two imaging arms as follows. First, the relevant clinical history, physical examination, and histologic results were presented. Then, the numeric scores for the likelihood of metastases in each anatomic site and the overall likelihood were presented from the imaging arm with the lowest overall likelihood score. In cases of equal overall scores, the presentation order of the imaging arms was random. The oncologists were always blinded as to which imaging arm was being described. Treatment decisions were selected from at least one of the following categories: clinical follow-up, surgery, radiation therapy, endocrine therapy, adjuvant chemotherapy, or chemotherapy. Consensus of the two oncologists was required.
Next, the results of the other arm were presented, and the oncologists were asked to select a treatment decision again by considering the findings from that arm only.
Images were not provided. To disguise the conventional imaging results, organ systems not evaluated by means of conventional staging were reported as having a score of 1. If a biopsy had been performed as a consequence of an arm's imaging findings, the histologic results were provided with that arm's results.
Determination of the True Diagnosis
The true presence of metastatic disease was established by using biopsy results in eight patients (iliac crest biopsy in four patients, lung biopsy in one, biopsy of the pleura in one, liver biopsy in one, skull biopsy in one), findings of immediate additional imaging in three patients (brain MR imaging findings in one patient, humeral and thoracolumbar spine MR imaging findings in one patient, and skull CT findings in one patient), clinical follow-up findings in all 19 patients (median, 21 months), and follow-up imaging in 16 patients.
Statistical Methods
For the null hypothesis that a given imaging arm performed better than chance, the Fisher exact test was used and the two-tailed P value was estimated. A P value less than .05 indicated a statistically significant difference.
For the null hypothesis that the echo-planar MR imaging arm performed as well as or better than the conventional arm, the two arms were compared by using a one-tailed sign test based on the binomial distribution. A P value less than .05 indicated a statistically significant difference.
The software used was JMP-IN version 3.1.5 (SAS Institute, New York, NY).
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
|
|
Discordance between Findings of the Two Imaging Arms
For three patients (patients 6, 12, and 13), the two imaging arms had discordant results concerning the likelihood of metastatic disease.
Total-body echo-planar MR imaging studies in patient 6 (Fig 3) were interpreted as showing probable skeletal metastases (likelihood score of 4), with a focal lesion in the left humerus and multiple vertebral body lesions. The findings of conventional imaging (bone scintigraphy, chest radiography, abdominal and pelvic CT, cervical spine and shoulder radiography, and brain MR imaging) were interpreted as metastasis probably absent (likelihood score of 2). At bone scintigraphy, uptake limited to the cervical and thoracic spine was attributed to degenerative changes. Widespread skeletal metastases were confirmed at dedicated MR imaging of the humerus and thoracic spine. The patient declined biopsy and was treated with tamoxifen citrate. Follow-up MR imaging findings at 10 months remained consistent with multiple osseous metastases.
|
|
|
Did One Imaging Arm Perform Better Than the Other?
With total-body echo-planar MR imaging, disease was correctly staged in 18 patients (95%)—in 10 (91%) of the 11 patients without metastases and in eight (100%) of the eight with metastases. In the one patient with disease incorrectly staged with echo-planar MR imaging, conventional gadolinium-enhanced brain MR imaging was performed. With conventional imaging, disease was correctly staged in 15 patients (79%)—in nine (82%) of the 11 without metastases and in six (75%) of the eight with metastases. In the four patients in whom disease was incorrectly staged by using the conventional arm, there were two false-negative findings in the skeleton and liver, which affected the therapeutic decision. The two false-positive studies led to an unnecessary lung biopsy in one patient and a skull biopsy in the other.
Was Treatment Affected By the Imaging Arm?
Total-body echo-planar MR imaging and conventional imaging resulted in the same therapeutic decision in all patients except the patient with unexpected liver metastasis (patient 12) (Fig 4) and the patient with the unsuspected bone disease (patient 6) (Fig 3); both the liver metastasis and the bone disease were detected with echo-planar MR imaging alone. On the basis of the conventional imaging results, the therapeutic decision in patient 12 was surgery and adjuvant chemotherapy. Total-body echo-planar MR imaging detection of multiple liver metastases resulted in a decision to perform chemotherapy instead. For patient 6, adjuvant chemotherapy followed by hormonal therapy was the therapeutic decision based on the conventional imaging arm findings. In the presence of bone metastases, hormonal therapy alone was selected instead. Although detection of liver metastases failed in patient 9 with conventional imaging, skeletal metastases were detected, and the therapeutic decision for induction chemotherapy was not affected.
Are the Results Statistically Significant?
For the null hypothesis that the performance of a given imaging arm was different from chance, the P value for the echo-planar MR imaging arm was less than .001; the P value for the conventional arm was .027.
To test the null hypothesis that the echo-planar MR imaging arm was as good as or better than the conventional arm in the sense of more correct staging results, a two-by-two contingency table was constructed, as in Table 4. A one-tailed sign test based on the binomial distribution was used, with a P value less than .05 considered to indicate significance; the P value was .126, meaning that the probability that the echo-planar MR imaging arm was actually worse than the conventional arm was 12.6%. The statistical power was low because of the small number of patients.
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The cornerstone of treatment planning for breast cancer is accurate staging, although the extent and timing of the metastatic evaluation can vary considerably depending on tumor size, axillary node status, patient symptoms, protocol requirements, and physician discretion, as in our study. Bone is the most common site of breast cancer metastases, but extraosseous metastases involving the liver, lung, and pleura have been reported as the first sites of distant disease in 25% of cases (19). Glynne-Jones and colleagues (20) reported that when bone scintigraphy was used as the sole staging study, only 78% of patients with stage IV disease were identified. In their series, bone scintigraphy combined with chest radiography increased the sensitivity of staging to 86%; however, bone scintigraphy combined with liver US had a sensitivity of 95%. In our study, the stage IV disease in patient 12 (Fig 4) was not detected with the conventional imaging combination of bone scintigraphy and chest radiography. Total-body echo-planar MR imaging enabled comprehensive multiorgan assessment, depicted the clinically unsuspected liver metastases, and substantially affected therapy.
Bone scintigraphy, the standard method for assessing skeletal involvement, has limitations. It is well known that radionuclide accumulation at the site of benign skeletal processes can be indistinguishable from metastases, as occurred in one patient with a false-positive bone scan in our study. Recent MR imaging literature suggests that skeletal metastases may be understaged with bone scintigraphy (7,10,21). Avrahami and colleagues (11) reported the MR imaging demonstration of spine metastases in 50% of patients with symptomatic tumors and equivocal bone scintigraphic findings, normal radiographs, and normal spine CT scans. In one patient (Fig 3) with a false-negative bone scan in our study, absence of radionuclide accumulation in the humerus was presumably due to lack of appreciable osteoblastic response, while spine uptake secondary to metastases was mistaken for benign degenerative disease. Had this patient undergone preoperative total-body echo-planar MR imaging, an axillary node dissection with its associated morbidity would have been avoided.
Detection of breast cancer tumor cells within bone marrow provides important and possibly unique prognostic information for the oncologist by enabling the prediction of substantially higher rates of distant relapse (22,23). Micrometastatic deposits in the marrow are scintigraphically occult. If these cells proliferate, lesions may become scintigraphically evident once bone changes occur. Conventional MR imaging has shown promise in depicting occult marrow involvement in breast cancer. Fourteen of 23 patients with findings reported by Sanal and colleagues (12) had MR imaging abnormalities of the spine, pelvis, or proximal part of the femur that were suggestive of metastases, seven of the 14 with normal bone scans and six without corresponding abnormalities. Six of the 14 patients had marrow involvement confirmed with histologic analysis and monoclonal antibody immunostaining, while four of the remaining eight developed metastatic disease demonstrable by using conventional methods within 18 months. None of their patients with operable disease and negative MR imaging findings developed metastases. Total-body echo-planar MR imaging may provide a noninvasive survey for occult marrow disease and therefore affect prognosis and possibly therapy.
The confirmation of metastatic disease in patients with breast cancer almost always influences therapeutic decisions, as was demonstrated in two patients in whom total-body echo-planar MR imaging depicted otherwise undiagnosed metastases. Although the failure to correctly diagnose liver metastases by using conventional imaging in another patient did not affect therapy (because bone metastases were correctly diagnosed), hepatic involvement does portend a much poorer prognosis (24). Also, judging therapeutic response by following "measurable" hepatic rather than "evaluatable, but nonmeasurable" skeletal metastases is preferred (25).
The only false-positive echo-planar MR imaging study (in patient 2) was interpreted as showing probable brain metastases. Suboptimal positioning of the head resulted in an oblique projection, and large sulci were mistaken for posterior parietal lesions. The conventional images in the same patient were suggestive of lung metastasis, also a false-positive finding.
A theoretic advantage of total-body echo-planar MR imaging staging is that one radiologist interprets the entire staging study. This averts multiple and sometimes discordant readings and allows the radiologist to interpret each finding in light of the others. Because the echo-planar MR imaging sequences used for this study each lasted only 6 minutes, cost comparisons of echo-planar MR staging with conventional staging should be favorable.
Limitations
The principal limitation of the present study is the small number of subjects. The study included only patients with newly diagnosed breast cancer, and the findings cannot be generalized to currently or previously treated patients.
The timing and selection of conventional imaging studies were not standardized, but instead varied with the judgment of the clinicians. The intention was to compare our technique with conventional imaging as the latter is actually practiced in standard care.
Because a perfect magnetic field shim cannot be obtained for a structure as varied in shape as the human body, the shim at each anatomic level was a compromise. This caused loss of signal in the soft tissue of the extremities but did not substantially compromise the views of the humeri and femurs. The entire upper extremities were not in the field of view, although in all patients at least the proximal third of the humerus was included. Aliasing in the phase-encoding direction degraded images in some patients.
Metastatic lung involvement that manifests as pulmonary nodules alone, although not occurring in our study, may be missed with the present echo-planar MR imaging technique unless the nodules are large enough to overcome the magnetic susceptibility effects of surrounding air.
Future Directions
Echo-planar MR imaging is increasingly available to clinicians (26). In addition, it is likely that the table-sweep technique could be adapted to use very fast gradient-echo conventional imaging (27). Technical developments in total-body echo-planar MR imaging that promise to improve image quality are under way (28).
The potential cost-benefit gain from this method of total-body echo-planar MR imaging may be considerable. A single rapid examination should cost less than multiple conventional examinations in both money and time. More difficult to quantify, but important to our patients, are the costs of understaging disease, which include unnecessary surgery and inappropriate therapy that potentially lead to inaccurate analysis of response to therapy. Costs of overstaging include unnecessary biopsy and considerable anxiety for the patient.
Total-body echo-planar MR imaging shows promise in the staging of breast cancer. We were able to correctly stage disease in nearly all of our patients with newly diagnosed breast cancer, which allowed for more accurate therapeutic planning. The ability to image multiple organ systems with a highly sensitive modality rapidly, noninvasively, and at a single sitting, with images interpreted by a single radiologist, offers considerable advantages. Because it may prove to be more sensitive than scintigraphy for the diagnosis of bone marrow involvement, it may also give prognostic information in early-stage disease. This technique has the potential to substitute for conventional radiologic staging and to improve diagnostic accuracy.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantor of integrity of entire study, L.J.H.; study concepts and design, L.J.H., K.M.J.; definition of intellectual content, L.J.H.; literature research, L.J.H.; clinical studies, L.J.H., B.A.B., S.M., K.M.J.; data acquisition and analysis, L.J.H., K.M.J.; statistical analysis, K.M.J.; manuscript preparation, L.J.H.; manuscript editing, L.J.H., K.M.J.; manuscript review, L.J.H., B.A.B., S.M., K.M.J.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. Brauck, M. O. Zenge, F. M. Vogt, H. H. Quick, F. Stock, T. Trarbach, M. E. Ladd, and J. Barkhausen Feasibility of Whole-Body MR with T2- and T1-weighted Real-time Steady-State Free Precession Sequences during Continuous Table Movement to Depict Metastases Radiology, March 1, 2008; 246(3): 910 - 916. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. C. Lauenstein, S. C. Goehde, C. U. Herborn, M. Goyen, C. Oberhoff, J. F. Debatin, S. G. Ruehm, and J. Barkhausen Whole-Body MR Imaging: Evaluation of Patients for Metastases Radiology, October 1, 2004; 233(1): 139 - 148. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Mazumdar, M. J. Siegel, V. Narra, and L. Luchtman-Jones Whole-Body Fast Inversion Recovery MR Imaging of Small Cell Neoplasms in Pediatric Patients: A Pilot Study Am. J. Roentgenol., November 1, 2002; 179(5): 1261 - 1266. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. C. Lauenstein, S. C. Goehde, C. U. Herborn, W. Treder, S. G. Ruehm, J. F. Debatin, and J. Barkhausen Three-Dimensional Volumetric Interpolated Breath-Hold MR Imaging for Whole-Body Tumor Staging in Less Than 15 Minutes: A Feasibility Study Am. J. Roentgenol., August 1, 2002; 179(2): 445 - 449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Siegel, H. Ishwaran, B. D. Fletcher, J. S. Meyer, F. A. Hoffer, D. Jaramillo, R. J. Hernandez, S. E. Roubal, B. A. Siegel, D. J. Caudry, et al. Staging of Neuroblastoma at Imaging: Report of the Radiology Diagnostic Oncology Group Radiology, April 1, 2002; 223(1): 168 - 175. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Schirrmeister, G. Glatting, J. Hetzel, K. Nussle, C. Arslandemir, A. K. Buck, K. Dziuk, A. Gabelmann, S. N. Reske, and M. Hetzel Prospective Evaluation of the Clinical Value of Planar Bone Scans, SPECT, and 18F-Labeled NaF PET in Newly Diagnosed Lung Cancer J. Nucl. Med., December 1, 2001; 42(12): 1800 - 1804. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. E. Daldrup-Link, C. Franzius, T. M. Link, D. Laukamp, J. Sciuk, H. Jurgens, O. Schober, and E. J. Rummeny Whole-Body MR Imaging for Detection of Bone Metastases in Children and Young Adults: Comparison with Skeletal Scintigraphy and FDG PET Am. J. Roentgenol., July 1, 2001; 177(1): 229 - 236. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Barkhausen, H. H. Quick, T. Lauenstein, M. Goyen, S. G. Ruehm, G. Laub, J. F. Debatin, and M. E. Ladd Whole-Body MR Imaging in 30 Seconds with Real-Time True FISP and a Continuously Rolling Table Platform: Feasibility Study Radiology, July 1, 2001; 220(1): 252 - 256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Taoka, N. A. Mayr, H. J. Lee, W. T. C. Yuh, T. M. Simonson, K. Rezai, and K. S. Berbaum Factors Influencing Visualization of Vertebral Metastases on MR Imaging Versus Bone Scintigraphy Am. J. Roentgenol., June 1, 2001; 176(6): 1525 - 1530. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
