| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Full paper |
Departments of 1 Diagnostic Radiology and 2 Nuclear Medicine, University of Ulm, Robert-Koch-Str. 8, 89081 Ulm and 3 Department of Obstetrics and Gynecology, University of Ulm, Prittwitzstr. 43, 89081 Ulm, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
Positron emission tomography (PET) is, like MRM, a cross-sectional imaging technique, although based on a different principle. Reports of the sensitivity and specificity of PET fall in the range of 68–94% and 83–97%, respectively, and like MRM, depend to a great degree on the criteria used for image interpretation [11–14]. On the whole there is a relative paucity of data regarding PET's role in the diagnosis of mammary carcinoma compared with MRM. Reliable findings regarding PET's pre-operative application have, to our knowledge, yet to be published. There are few data that would enable one to compare the efficacy of MRM vs PET in patients with mammary carcinoma.
The objective of the present study was to evaluate the efficacy of MRM and PET as pre-operative adjuncts to conventional diagnostic procedures.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
Assessment was finalized in Breast Imaging Reporting and Data System (BI-RADS) categories together with corresponding recommendations: 1, negative, routine screening; 2, benign findings, routine screening; 3, probably benign, short interval follow-up (6 months); 4, suspicious, biopsy; 5, highly suggestive of malignancy, biopsy; and 0, needs additional evaluation [15].
There was strong suspicion of mammary carcinoma in all patients, based on the results of at least one of the following examinations (BI-RADS category 5 [15]): palpation; diagnostic ultrasound; or mammography. Prior to histological confirmation, all patients underwent MRM and PET at an interval not longer than 3 days.
MRM and PET were interpreted independently of each other. The interpretation of MRM and PET findings was performed by one or two observers experienced with the respective method (MRM: KN, AR; PET: HS). The only clinical information provided was the fact that the patient had suspected mammary carcinoma. Observers were not told on which side the suspected carcinoma was located, nor were they informed regarding the findings of other diagnostic imaging methods.
The following questions were posed to the observers evaluating the MRM and PET images:
Tumour size was defined as the largest measureable diameter in one plane. In the case of bifocal, trifocal or multifocal carcinomas, or of invasive carcinomas with non-invasive portions, the overall tumour extent was determined. Thus, the sizes were summated. This summation was necessary especially in multifocal carcinomas with many small lesions. The definition of multifocality and the measurement technique were applied to both imaging and histological findings.
Because the phase coding direction produces significant horizontally running artefacts at MRM, evaluation of the axilla in most patients was unsatisfactory. In some patients, the axilla remained outside the field of view (FOV). Hence, we did not include MRM findings in our evaluation. Thus, only PET findings were available for comparison with the results of histology of the axillary region.
Following histological confirmation, a second evaluation of the images returned by both methods was performed by consensus. Also participating in this second evaluation was the surgeon performing operations on the breast (TK). The examiners were also able to refer to all available findings, including the results of diagnostic procedures, mammograms, histology and surgical reports. Besides the aforementioned points, this evaluation also discussed the degree to which the results of conventional clinical examinations, such as palpation, ultrasound and mammography, and MRM and/or PET had influenced surgical treatment in each patient. Finally, it was determined whether MRM and/or PET findings had overestimated or underestimated tumour extent. For comparison of PET and MRM findings with histological findings, a discrepancy margin of ±3 mm was accepted as tolerable owing to considerations of the methods' spatial resolution. For statistical analysis, sensitivities for the diagnosis of primary tumours, as well as sensitivities, positive predictive values and diagnostic accuracies in cases of multifocality, were calculated.
MRM
MRM was performed using a Magnetom Vision unit (Siemens, Erlangen, Germany) with a field strength of 1.5 Tesla.
After informed consent had been obtained and contraindications excluded, the patient was placed in the MRM unit in the prone position in order to minimize respiration artefacts. Examination of the breast was performed using commercially available bilateral breast surface coils.
First, fast T2 weighted images in axial projection were acquired (spin echo sequence: echo time (TE) 90 ms, repetition time (TR) 5376 ms, 2 acquisitions, field of view (FOV) 350 mm, matrix 252 x 256, slice thickness 4 mm, acquisition time 6 min 32 s). For the T1 weighted sequence and dynamic measurements, a gradient spin echo sequence (three-dimensional fast low angle shot, TE 5 ms, TR 11.8 ms, flip angle 30°, matrix 157 x 256) was utilized, first native, then at 1 min, 2 min, 3 min and 8 min after intravenous application of 0.15 mmol kg-1 body weight of gadolinium-DTPA (Magnevist®; Schering, Berlin, Germany). This three-dimensional sequence was performed with 32 partitions, corresponding to an effective slice thickness of 4 mm. A FOV of 350 mm was selected. Measurement time of this sequence was 1 min.
Injection of contrast medium took place via a cubital vein using a disposable 20 G catheter and the MR injector XD 7000 (Ulrich, Ulm, Germany), with a flow rate of 3 ml sec-1. Injection is automatically followed by flushing with physiological saline solution. Acquisition of data in the dynamic contrast medium series began immediately upon starting injection.
All 192 slice images acquired (32 individual T2 weighted images and 160 individual T1 weighted images) were documented. Qualitative contrast medium uptake was evaluated using subtraction images, produced by subtracting the individual images of the native sequence from the images acquired at the same respective slice positions 3 min after contrast medium application. For structures that appeared focal after contrast medium application, the increase in signal intensity over time ("mean curve") was calculated. For the evaluation of the dynamic contrast enhancement of a lesion, the region of interest was placed in the area with maximum contrast enhancement. Dynamic measurements were used that yielded reproducible results after repeated measurements. The computer software programs required for the subtraction images and signal intensity time curves are included in the standard software program of the Magnetom Vision unit. The investigational technique fulfills the recommendations of the German Roentgen Ray Society [16].
Morphological criteria suggestive of malignancy in T1 weighted images included the presence of skin thickening and contour irregularity (spiculated margins). In T2 weighted images the signal intensity of the tumour was compared with that of surrounding glandular tissue.
For comparison with data reported in the literature [1–8], the dynamic contrast medium behaviour of the lesions was investigated. The increase in signal intensity in per cent during the first and second minutes after contrast medium application and at the maximum achieved intensity were compared with the native value for the lesion. An increase in signal intensity of more than 100% during the first 2 min and a signal intensity similar to the peak signal in the third and eighth minute ("plateau") were determined to be suspicious for malignancy.
PET
All patients were examined using a whole body PET camera (ECAT EXACT HR+; Siemens/CTI, Knoxville, TN). This scanner has an axial FOV of 15.5 cm. The full width at half maximum at the centre FOV is 4.2 mm. Emission scanning was carried out 45–60 min after injection of an average 370 MBq (range 296–581 MBq) of fluorodeoxyglucose (FDG). Emission scans consisted of three bed positions covering a 45 cm FOV. These emission scans included the neck, thorax, breasts and liver. In order to enhance the practicability for routine surveys by shortening the examination time, attentuation correction was not performed. Patients fasted for at least 8 h prior to FDG-PET scanning. All patients were placed in the prone position with elevated arms. Two inserts in the table of the PET scanner prevented deformation of the breasts.
Lesions with clearly visible, focally increased tracer uptake located in the breast or axilla were defined as indicative of breast cancer or axillary lymph node metastases, respectively. Diffuse, homogeneous tracer uptake, mostly located in both breasts, was regarded as indicative for mastopathy and therefore not classified as malignant.
MRI and PET findings were correlated with histopathology. Histopathological evaluation was performed in knowledge of the findings of the imaging modalities. For this evaluation, the specimen orientation was similar to that of the tumour in pre-operative MRI.
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
T-stage
T-stage is defined by tumour size. In the case of multifocal carcinomas the T-stage is dependent on, among other factors, the size of individual lesions. The respective determinations of T-stage could be reliably compared in 39 patients. Patients who subsequently underwent neoadjuvant chemotherapy (n=5) or who had only been biopsied (n=2) were not included in this evaluation. For determination of patients' T-stage, we always used the absolute measurement returned by the method without considering the tolerated discrepancy margin of ±3 mm.
Results of histology showed the following T-stages: Tis (in situ) (n=1), and Tmic (microinvasive) (n=1); T1b (n=5), T1c (n=3), T2 (n=19), T3 (n=4), and T4 (n=6).
Conventional diagnostic techniques (palpation, mammography and/or ultrasound) identified all ipsilateral (sensitivity 100%; confidence interval (CI) 91.8–100%) and one of three contralateral (sensitivity 33.3%) carcinomas. In retrospect, the two remaining contralateral carcinomas were occult at conventional mammography (Table 1
).
|
). Contralateral lesions were extirpated following MR guided wire localization.
PET correctly recognized 40 of 43 ipsilateral carcinomas (sensitivity 93.0%; CI 80.9–98.4%). PET missed two 10 mm tumours and one 30 mm tumour. In two of these cases PET returned findings suggestive of axillary lymph node metastases. Histological types of the three undetected tumours included two invasive ductal carcinomas and one invasive lobular carcinoma with non-invasive lobular cancer portions (Figure 1). All contralateral carcinomas were correctly identified. In addition, PET returned false positive findings of a contralateral carcinoma (Table 1).
|
Unifocal, bifocal or multifocal lesions were suspected if the lesions showed dynamic contrast behaviour typical for malignancy in MRI, or glucose-uptake in PET.
Carcinomas were found, at histology, to be unifocal in 31 cases, bifocal in 4 cases, trifocal in 1 case and multifocal in 7 cases.
A total of 28 unifocal, three bifocal and five multifocal lesions were correctly identified as such on the basis of conventional diagnostic techniques. Three carcinomas (two unifocal and one trifocal) were incorrectly identified as multifocal. One bifocal and one multifocal carcinoma were incorrectly identified as unifocal lesions. Taking into consideration these and two carcinomas that escaped detection, conventional diagnostic techniques yield a sensitivity of 90.0% (CI 77.9–97.4%), a positive predictive value of 92.3% (CI 77.9–97.4%) and a diagnostic accuracy of 83.7% (CI 69.3–93.2%).
MRM correctly characterized 30 unifocal, 3 bifocal and 7 multifocal carcinomas. In one case a trifocal carcinoma was misclassified as bifocal, while in another case a bifocal carcinoma was misidentified as a unifocal carcinoma. Conversely, one unifocal carcinoma was misidentified as a bifocal carcinoma. These findings yield a sensitivity of 95.2% (CI 84.2–99.4%), a positive predictive value of 97.6% (87.7–99.9%) and a diagnostic accuracy of 93.0% (CI 80.9–98.6%).
PET correctly identified 28 unifocal, 3 bifocal and 6 multifocal carcinomas. Two unifocal carcinomas were incorrectly classified as bifocal or multifocal, respectively, while one trifocal carcinoma was misidentified as a multifocal lesion, giving a total of three false positive findings. Considering the three previously described false negative findings, there remains a sensitivity of 92.5% (CI 80.9–98.5%), a positive predictive value of 92.5% (CI 80.9–98.5%) and a diagnostic accuracy of 86.0% (CI 72.1–94.7%) for this method (Table 1).
Tumour size and non-invasive cancer portions
It was possible to make comparisons of tumour size in 39 carcinomas. Patients who underwent neoadjuvant chemotherapy (n=5) or who were biopsied only (n=2) could not be included in this comparison. At histology, tumour size ranged from 6 mm to 120 mm (average 35.3 mm). In 11 cases, invasive tumour was coupled with non-invasive cancer, while one tumour consisted exclusively of non-invasive ductal cancer. Of these 12 carcinomas, 3 (2 cases with non-invasive ductal cancer and 1 with non-invasive lobular cancer) were not correctly detected by any method.
A discrepancy margin of ±3 mm was accepted as tolerable owing to considerations of subjective measurement accuracy.
Two contralateral carcinomas escaped detection by conventional diagnostic techniques. In addition, the size of 12 carcinomas (25.6%) was underestimated. The measurable difference ranged from -4 mm to -40 mm (average -18.1 mm). At histology, these carcinomas included eight invasive ductal carcinomas (six unifocal and two multifocal) and five invasive lobular carcinomas, two of which also contained non-invasive lobular cancer segments. One other invasive ductal carcinoma included non-invasive ductal cancer segments. Of the 12 instances of non-invasive cancer portions demonstrated at histology, 4 (33.3%) were correctly identified, while in 8 cases (66.6%) findings were false negative (5 non-invasive ductal cancer, 3 non-invasive lobular cancer).
All carcinomas were correctly diagnosed by MRM. Tumour size was overestimated in three cases (7.7%; two invasive ductal carcinomas and one non-invasive ductal cancer) and underestimated in four cases (10.3%; three invasive ductal carcinomas and one invasive lobular carcinoma with non-invasive lobular cancer). The measureable difference ranged from +6 mm to +10 mm (average +8.6 mm) and from -10 mm to -20 mm (average -15 mm), respectively. Of 12 cases with non-invasive cancer, MRM findings were true positive in eight cases (66.6%). In four cases (33.3%; two non-invasive ductal cancer and two non-invasive lobular cancer), the method returned false negative findings.
PET returned three false negative findings. Overestimation of tumour size occurred in four cases (10.3%; three invasive ductal carcinomas and one tubular invasive carcinoma) and underestimation in three cases (7.7%; three invasive ductal carcinomas, of which one was trifocal and one multifocal). The measurable size difference ranged from +6 mm to +23 mm (average 13 mm) and from -11 mm to -20 mm (average -15.8 mm), respectively. Of the 12 cases of non-invasive cancer, PET returned correct findings in 8 patients (66.6%). In four cases (33.3%; three non-invasive ductal cancer and one non-invasive lobular cancer), the method returned false negative findings.
To summarize, both MRM and PET were superior to conventional diagnostic techniques in determining exact tumour diameter and in detecting non-invasive cancer portions. With regard to tumour size, the results of MRM and PET were equivalent (Table 1) (Figures 2 and 3).
|
|
Of 40 cases in which sufficient data were available for comparison, histology identified stage N0 or N1 in 20 cases. PET returned false negative findings in four cases and false positive findings in one case. These data yield a sensitivity of 80% (CI 64.6–90.6%), a specificity of 95% (CI 83.1–99.4%), a positive predictive value of 94.1% (CI 83.0–99.4%), a negative predictive value of 95% (CI 83.0–99.4%) and a diagnostic accuracy of 87.5% (CI 73.2–95.8%).
Additionally, PET visualized contralateral axillary lymph node metastases in one patient and mediastinal lymph node metastases in one other patient (Table 1).
Influence on patient management
In the retrospective analysis of the patient collective, therapy plans were constructed on the basis of either all clinical data and findings of conventional diagnostic techniques alone or a combination of all these data with the additional findings provided by MRM and/or PET. In no case did MRM findings negatively influence a patient's therapy plan based on conventional diagnostic techniques. Three ipsilateral carcinomas were not recognized by PET. In two of these patients, however, ipsilateral lymph node metastases had been detected by PET, which supported the patient's putative diagnosis. In the third patient, the clinical findings were so unequivocal that it would have been unlikely that the false negative PET findings would have called the putative diagnosis and the planned surgical therapy seriously into question. PET in combination with the findings of conventional diagnostic techniques resulted in two additional suspected contralateral carcinomas being recognized as such in both cases, while, in another case, false positive PET findings of a contralateral lesion would have resulted in unnecessary surgery. Because of the evidence of bifocality returned by the PET examination, one patient underwent a more generous resection than had originally been planned.
MRM recognized two additional contralateral carcinomas that had escaped conventional diagnostic techniques. Both patients underwent breast conserving contralateral surgery. Based on the more reliable size determination with MRM and PET, one patient underwent a more generous resection, thus sparing her a possible secondary resection. In two other patients, findings by MRM and/or PET resulted in their undergoing radical surgery including a latissimus-dorsi-plasty instead of the originally planned breast conserving treatment.
In summary, the additional data provided by MRM and/or PET examinations positively affected surgical management in five patients (12.5%) each. When all diagnostic methods are combined, surgical planning was optimized in a total of six patients (15%).
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
It would therefore seem reasonable to utilize MRM in the work-up of those entities in which this method's high sensitivity can be used to greatest advantage. This is the case, for example, in pre-operative staging, because of MRM's capabilities in visualizing both multifocal lesions and simultaneously developing contralateral carcinomas [8, 9]. In addition to overall tumour size, the relationship between tumour size and breast size, tumour localization, tumour grade and histological findings, the aforementioned factors help shape patients' surgical management. In addition, psychosocial and cosmetic considerations, not to mention the wishes of the patient herself, should be integrated into the therapeutic decision-making process [18].
In comparison to MRM, PET has not yet established a major role in routine clinical practice. Reports in the literature give the sensitivity and specificity of PET in the diagnosis of mammary carcinoma as 68–94% and 83–97%, respectively [11–14]. Reports of PET's capabilities within the framework of pre-operative staging have yet to appear in the literature. In general, data comparing MRM and PET and their respective roles in the diagnosis of mammary carcinoma remain rare [19, 20].
Detection of tumours and diagnosis of contralateral carcinomas
In patients in whom conventional diagnostic methods have returned a putative diagnosis of mammary carcinoma (BI-RADS category 5), it is the role of an adjunctive modality to reliably confirm this suspicion. The findings of the present study show that MRM is a reliable method. The high sensitivity of 100% confirms the findings of an earlier study but should not be taken out of context, since in our highly selected patient collective each diagnostic modality only serves to increase this high sensitivity [8].
With PET, three ipsilateral carcinomas were not recognized. Histology revealed invasive ductal carcinomas in two cases and an invasive lobular carcinoma with non-invasive lobular cancer portions in the other. Although it is known from the literature that the diagnosis of invasive lobular carcinomas by means of PET can be problematic (as is true for diagnosis of this type of tumour using MRM as well [21]), we could find no explanation for the false negative results in both cases of invasive ductal carcinomas.
On the basis of available reports on the results of MRM, approximately 4.7–6% of patients with mammary carcinoma may have a simultaneously developing carcinoma of the contralateral breast [8, 10]. In the present collective, the prevalence of contralateral carcinoma was 7%. At present, MRM is considered the most sensitive modality for detecting simultaneously developing contralateral carcinomas [8–10]. The most recent study by Fischer et al [10] to address this issue showed that of 19 contralateral tumours, 15 (78.9%) were detectable only by MRM. These results are confirmed in the findings of the present study. PET showed itself to be as sensitive as MRM. However, because of the small number of patients with contralateral carcinomas (n=3), the data were not statistically significant. To date there have been no reports in the literature describing the application of PET to the diagnosis of contralateral carcinomas. Only in one study is there mention of a member of the examined patient collective who suffered from bilateral mammary carcinoma [14]. In this case, the contralateral carcinoma was detected by MRM but not by PET [14]. A probable reason for PET's failure to detect this lesion lies in its small diameter (8 mm), because lesions under 1 cm in diameter often escape detection by PET [11, 12, 14, 21].
Based on the values for specificity given above, one must also expect a certain number of false positive findings for both MRM and PET [10–14]. In the study reported by Fischer et al [10], the rate of false positive findings was just under 3.5% for the diagnosis of contralateral carcinomas. In this study, PET returned one false positive finding. There are, however, no relevant references in the literature.
Number of lesions
The presence of multifocal lesions may often decisively alter a given patient's established management. In one recent paper by Fischer et al [10], the diagnosis of multifocal or multicentric lesions influenced the course of surgical therapy in 11% of patients. In 30 (71.4%) of 42 patients, however, these lesions were detected only by MRM. In our own study, MRM displayed a high sensitivity in the diagnosis of multifocal lesions, confirming our own results as well as reports in the literature [8, 22–26]. PET exhibited a sensitivity of 92.5%, which, though slightly lower than that of MRM (95.2%), was superior to those of conventional techniques. There are no reports in the literature to date describing PET's role in the diagnosis of multifocal mammary carcinoma. According to Avril et al [21], the present state of knowledge would suggest a more limited sensitivy for PET, because multifocal lesions are often small.
Of course, a false positive MRM or PET finding could affect the patient's surgical planning, resulting in the patient undergoing a more radical procedure than actually necessary [24]. Taking into consideration these possible mischaracterizations, MRM still shows a diagnostic accuracy of 93.0%. PET was slightly lower at 86.0%, but still higher than the results of conventional techniques. In our own series, these erroneous determinations had no negative impact on the eventual surgical treatment.
Tumour size
Only a few studies discuss the determination of tumour size by MRM and correlate it with conventional methods [24, 25]. Relevant findings for PET are lacking in the literature. According to Mumtaz et al [24], MRM provides more accurate findings regarding tumour size than either conventional mammography or ultrasound. In our own study we set a tolerance threshold of ±3 mm on the basis of subjective measurement accuracy. Both PET and MRM proved superior to conventional methods in determining exact tumour diameter, including detection of non-invasive cancer portions. Our data, however, also showed that, in general, the diagnosis of non-invasive cancer with both MRM and PET is problematic, which confirmed findings in the literature [4, 5, 10, 21].
N-stage
One of the advantages of PET is that the lymphatic system can be simultaneously visualized during an examination of the breast. Reports in the literature concerning the diagnosis of lymph node metastases of mammary carcinoma reports sensitivities of 79–95% and specificities of 66–96% [27–30]. Findings in the present study confirmed these data. Nevertheless, our study indicated a rather low sensitivity of 80%, but a high specifity of 95%. One explanation for this relatively low sensitivity may be the fairly high proportion of T1 carcinomas in the study collective (19.5%). In a study by Avril et al [29], the sensitivity of PET in the diagnosis of lymph node metastases of T1 mammary carcinomas was only 33%.
The detection or exclusion of axillary lymph node metastases with MRM is often not feasible because the images acquired by use of standard breast coils may not satisfactorily show the axillary region owing to significant artefact formation associated with the phase coding direction, or simply because it lies outside the FOV. For this reason we have omitted examination of this aspect with MRM. In those cases in which the axillary region is visualized on MRM images and is of sufficient quality for evaluation, findings by Mumtaz et al [24] suggest a sensitivity of 90% for the diagnosis of lymph node metastases, compared with 53% for palpation, and a specificity of 82%. In order to optimize the quality of the findings, it would be useful to perform a separate MR examination of the axilla.
Influence on patient management
As early as 1994, Fischer et al [9] reported that, based on the findings of pre-operative MRM, the surgical strategy in 18.5% of patients was modified, although it must be noted that additional pathological findings were returned by MRM in 20.9% of examined patients. These results were later confirmed in a more recent study by the same research group. According to Fischer et al [10], the planned surgical approach was corrected in 14.4% of 463 patients based on pre-operative staging using MRM. In the present study similar conclusions were reached. Findings obtained from pre-operative MRM or PET positively influenced the surgical strategy in 12.5% of cases. Considering the effect of both methods, there was an added synergistic effect resulting in modification of the surgical plan in up to 15% of cases. Therefore, it is our opinion that pre-operative MRM and/or PET in patients with suspicion of mammary carcinoma is clinically justified. These results have also been confirmed by Mumtaz et al [24]. According to these authors, 53% of patients with histological evidence of malignant tissue at the biopsy rims could have been spared a secondary excision had the resection volume been pre-operatively determined on the basis of an MRM examination.
|
Conclusion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
Received for publication August 1, 2001. Revision received March 21, 2002. Accepted for publication June 17, 2002.
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsDiscussionConclusionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. E. Brennan, N. Houssami, S. Lord, P. Macaskill, L. Irwig, J. M. Dixon, R. M.L. Warren, and S. Ciatto Magnetic Resonance Imaging Screening of the Contralateral Breast in Women With Newly Diagnosed Breast Cancer: Systematic Review and Meta-Analysis of Incremental Cancer Detection and Impact on Surgical Management J. Clin. Oncol., November 20, 2009; 27(33): 5640 - 5649. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P Veit-Haibach, G Antoch, T Beyer, H Stergar, R Schleucher, E A M Hauth, and A Bockisch FDG-PET/CT in restaging of patients with recurrent breast cancer: possible impact on staging and therapy Br. J. Radiol., July 1, 2007; 80(955): 508 - 515. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Shah, S. K. Shah, and K. V. Greatrex Current Role of Magnetic Resonance Imaging in Breast Imaging: A Primer for the Primary Care Physician J Am Board Fam Med, November 1, 2005; 18(6): 478 - 490. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. L. Rosen, T. G. Turkington, M. S. Soo, J. A. Baker, and R. E. Coleman Detection of Primary Breast Carcinoma with a Dedicated, Large-Field-of-View FDG PET Mammography Device: Initial Experience Radiology, February 1, 2005; 234(2): 527 - 534. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
