CASE 1 Stage IV presentation of breast cancer with bone metastases

A 60-year-old woman presented with a large, palpable right breast mass (same patient as in Case 19 in Chapter 4). Imaging workup, including mammography, ultrasound, breast MRI, and ultrasound-guided core needle biopsy and axillary lymph node aspiration, confirmed node-positive multifocal infiltrating ductal carcinoma (IDC). Based on physical examination, the patient was thought to have an advanced breast cancer, at least T3 by size, and she was started on neoadjuvant chemotherapy while her staging workup was completed. The staging evaluation included a bone scan and positron emission tomography (PET)/CT. The bone scan, obtained before starting chemotherapy, showed several foci of increased activity, including in the left anterior third rib, one focus each in the cervical and thoracic spine, and a focus at the left upper sacroiliac joint region (Figure 1). A breast MRI, obtained a few days after the first round of chemotherapy, included a coronal short tau inversion recovery (STIR) series of the thorax, obtained with the body coil. This showed evidence of widespread bone metastases, with hyperintense foci in the right medial clavicle and in a lower thoracic vertebral body (Figure 2). A PET/CT scan was performed about 1 week after the first cycle of chemotherapy. In addition to showing hypermetabolism of the multifocal breast carcinoma, hypermetabolic foci were identified in multiple bony sites, including the right clavicular head, right upper cervical spine, two adjacent lower thoracic vertebrae, the left medial iliac crest, and a left anterior rib (Figure 3). These sites of abnormal activity overlapped in distribution with the abnormalities noted previously on limited chest MRI and the prechemotherapy bone scan. No definite correlates were found at these sites on CT. Thus, as the staging was completed in this patient, it became apparent that what was initially thought to be T3N1 disease (stage IIIA) was actually stage IV disease.

FIGURE 1 A, Anterior image from a prechemotherapy bone scan shows a focus of increased activity at the left anterior third rib level. The focus is less punctate than usually seen with trauma, although fracture is in the differential diagnosis. Note the slight asymmetry in sternoclavicular activity, with the right-sided activity slightly proximal to the joint compared with the left. B, Posterior view from the same prechemotherapy bone scan shows an intense focus of activity at the right upper cervical spine, with smaller foci of activity on the right at the lower thoracic spine level and a small focus at the left upper sacroiliac joint level. Degenerative changes, metastases, or some combination are in the scintigraphic differential diagnosis.

FIGURE 2 Coronal STIR image of the thorax, obtained with the body coil as part of the breast MRI, shows a hyperintense focus in a lower thoracic vertebral body. Similar findings were seen elsewhere, including the right clavicular head (not shown), suggesting bone metastases.

FIGURE 3 A to D, Sagittal images through the spine from PET/CT, performed 1 week after one round of chemotherapy. Two adjacent foci of hypermetabolism are seen in the lower thoracic spine, at T11 and T12. The corresponding CT bone window showed no definite lesions. A smaller hypermetabolic metastasis is seen in C2, above a prior surgical level.

After one cycle of chemotherapy, CT scanning was repeated. The large breast mass and axillary adenopathy showed improvement, and new sclerosis was now evident at several levels in the spine (Figure 4).

FIGURE 4 A, Detail of T11 vertebral body from staging CT, obtained before neoadjuvant chemotherapy. No definite osseous abnormality is seen to correlate with the hypermetabolism subsequently demonstrated at this level 1 week later on PET/CT. B, Detail of the same level, 3 weeks later, after one cycle of chemotherapy, shows new sclerosis (arrows), at the level where PET/CT had shown hypermetabolism.

Before undergoing repeat bone scanning and spine MRI, a second cycle of chemotherapy was given. The repeat bone scan, obtained 1 month after the prechemotherapy study, showed an increase in number of foci of increased activity (Figure 5). Several preexisting foci of activity appeared larger and more intensely active. MRI of the lumbar spine showed multiple areas of abnormal signal intensity, at levels corresponding to the prior bone and PET scans (Figure 6).

FIGURE 5 Repeat bone scan, 1 month after the prechemotherapy study, after two cycles of chemotherapy. A, Anterior planar view shows increased activity of the right clavicular head. The activity is now seen more clearly to be bone centered (versus joint centered). Intense shine-through activity of a larger right upper cervical spine lesion is seen, with less intense shine-through noted of adjacent right lower thoracic foci. The left anterior rib activity persists without clear change from before. B, Posterior planar bone scintigraphy shows more numerous and intense foci of activity compared with the prior study, including multiple spinal foci, correlating with hypermetabolism on PET/CT.

FIGURE 6 Sagittal lumbar spine MRIs, obtained after two cycles of chemotherapy. A, Sagittal, T1-weighted image shows hypointense foci of altered marrow signal, contrasting with the normal bright signal of fatty marrow. Lesions are seen at T11, T12, and the inferior-anterior L4 level. The larger lesions at T11 and T12 correlate with bone and PET scan findings. B, Sagittal STIR image through the same level shows corresponding peripheral hyperintensity of the metastases.

After completion of three cycles of neoadjuvant chemotherapy, the breast mass and axillary adenopathy had shrunk. Three months after diagno-sis, the patient underwent bilateral mastectomies, right axillary dissection, left sentinel lymph node sampling, and removal of breast implants. The right mastectomy specimen contained a 3.5-cm IDC, with multiple satellite nodules and clear margins. Six of 16 axillary lymph nodes showed metastases.

Two weeks after surgery, CT-guided sampling of the sclerotic T11 lesion was performed and did not confirm metastatic disease. The needle biopsy was repeated 2 months later, confirming metastasis. A nearly concurrent repeat PET/CT for restaging had normalized (Figure 7).

FIGURE 7 A to D, Sagittal images through the spine on follow-up PET/CT, 5 months after the initial staging PET/CT scan. The PET activity has resolved. The lesions can now be seen on the CT bone windows as sclerotic foci. The focus at T11 was biopsied 2 days before this study, confirming metastatic carcinoma.

TEACHING POINTS

This case represents stage IV disease at presentation, with a locally advanced breast cancer, metastatic to bone. It provides an excellent take-off point for discussion of some of the imaging issues raised on initial identification of bone metastases, as well as the assessment of treatment responses.

Bone scanning is not routinely obtained in the initial staging of all newly diagnosed breast cancer patients, as it once was. The yield in early-stage disease is low. In general, bone scanning performed during initial staging is reserved for patients with stage III or greater disease, unless there is a symptom, physical examination finding, tumor marker, or other abnormal laboratory value to suggest advanced disease.

Bone metastases can have a variety of appearances on bone scintigraphy, from a solitary active lesion, to multiple focal areas of increased activity, to diffuse involvement. Whether a bone metastasis is visualized on a bone scan depends on the size of the lesion, the location, and the type of bone reaction elicited. Because breast cancer metastases can be lytic, sclerotic, and frequently are mixed, a variety of manifestations are possible. Adding to the difficulty is the potential for conversion of successfully treated lytic metastases to blastic lesions. In general, bone scans most reliably depict osteoblastic metastases and can miss purely lytic disease. The converse is true of PET, which is highly sensitive for lytic bone lesions, but less sensitive for blastic metastases.

How then to resolve and explain the apparently conflicting imaging study results obtained in this patient?

The initial bone scan, obtained before starting chemotherapy, was abnormal, although not definitively diagnostic of bone involvement. Without imaging correlation, provided subsequently by a limited chest MRI, a number of the bone scan findings could have been accounted for by other explanations. The rib lesion, for example, could have been post-traumatic, and a possible trauma history was elicited from the patient. The activity is less punctate than generally seen with fractures, but a long segment of involvement, which would more compellingly suggest neoplasia, is not seen either. The patient had had prior cervical spine surgery, raising the possibility of accelerated de-generative change at an adjacent level. At the same time, the activity in both the cervical and thoracic levels is disturbingly focal and relatively intense. Degenerative findings frequently involve multiple adjacent levels and, unless actively degenerating, are often less intense in activity than the findings seen here.

The correlation provided subsequently by the limited chest MRI and PET/CT resolves the questions and provides compelling evidence of bony metastatic disease.

What is the significance of the apparent increase in disease seen on the second bone scan, performed 1 month later, after two cycles of chemotherapy? Does this mean the metastases are progressing? Not necessarily. In fact, the changes we see between these two bone scans are an excellent example of the flare phenomenon, whereby a patient responding to hormonal therapy or chemotherapy may actually appear worse on bone scanning. Increased activity of previously identified lesions and even new lesions may be seen (as in this case). Recall that it is the osteoblastic response to a bone abnormality (neoplastic, infectious, or traumatic) that we see as an increase in activity on a bone scan. Metastases that are initially lytic may become sclerotic in response to treatment, and this is demonstrated in this case on the PET/CT. The hypermetabolic bone foci that were seen on the initial PET/CT were without a clear CT correlate (indicating they were lytic) but became sclerotic (within 3 weeks, after one cycle of chemotherapy!) and readily visible on the CT portion of the follow-up PET/CT, at which time the PET activity had resolved. This does not mean the patient is cured or that the disease is not active. This patient had biopsy confirmation of metastatic involvement of T11, concurrent with normalization of PET scan activity. This constellation of findings reflects the reality that initially lytic bone metastases, best seen on PET imaging, responded to treatment by developing sclerosis. In the process, the bone metastases became increasingly visible on CT and bone scan, and inactive on PET.

CASE 2 Stage IV presentation with bone metastases (breast cancer presenting with back pain)

A 35-year-old obese woman presented with severe low back, knee, and hip pain and 5 days of intermittent fever up to 102 degrees. She was evaluated with abdominal, pelvic, and lumbar spine CT scans for suspicion of a spinal infection. These studies showed multiple lytic lesions in bone, including the L5 vertebral body, and left sacrum, ilium, and proximal femur (Figures 1, 2, 3, and 4). Bony metastatic disease was suspected, and a lung primary was sought with a chest CT. This showed a 2-cm spiculated right breast mass, suspicious for a breast cancer primary (Figure 5).

FIGURE 1 Axial CT scan (bone window) through the sacrum shows lytic destruction of the left posterior sacrum, extending to the sacral foramina.

FIGURE 2 More inferior axial CT image (bone window) through the sacroiliac (SI) joints shows another lytic lesion in the left medial ilium, adjacent to the SI joint.

FIGURE 3 Sagittal lumbar spine reconstruction (bone window) shows another lytic lesion in L5, extending to the superior end plate.

FIGURE 4 Coronal reconstruction (bone window) shows the lytic change in L5 and another small lytic lesion in the left intertrochanteric proximal femur.

FIGURE 5 Axial unenhanced chest CT image shows a central right breast mass in largely fatty breasts.

Bone scan showed corresponding, highly suspicious, multifocal findings suggesting bone metastases (Figure 6).

FIGURE 6 Bone scan images (A, anterior and posterior whole-body; B, spot views of the ribs [oblique], knees [lateral], and pelvis [anterior and posterior]) show multiple foci of abnormally increased activity in a highly suspicious pattern for bony metastatic disease. In addition to activity that corresponds to the known lytic foci in spine and pelvis on CT, sites are seen in the right clavicular head, ribs, and left proximal tibia, among others. The long segments of rib activity (left posterior, right anterior) are a pattern of activity that is highly suspect for neoplasia.

Breast imaging workup with diagnostic mammography, breast ultrasound, and ultrasound-guided core needle biopsy confirmed right breast cancer. The mammogram showed a dominant, central, 5-cm spiculated mass (Figure 7). Ultrasound showed a corresponding irregular hypoechoic mass in the retroareolar region, with a separate 1.7-cm suspicious shadowing focus in the upper outer quadrant (UOQ) (Figures 8 and 9). A suspicious axillary lymph node was also noted, with mild cortical mantle thickening (Figure 10).

FIGURE 7 MLO (A) and CC (B) mammograms show a dominant central, spiculated right breast mass behind the nipple, with stranding extending anteriorly toward the nipple. This mass corresponds to the mass identified on CT. The breasts are predominantly fatty.

FIGURE 8 Ultrasound of the right central breast shows a large area of ill-defined hypoechoic shadowing, which corresponds in location with the dominant mass on CT and mammography.

FIGURE 9 Ultrasound of the right UOQ shows a separate hypoechoic shadowing mass at 10 o’clock.

FIGURE 10 Right axillary ultrasound shows two lymph nodes with hypoechoic, thickened cortices. Fine-needle aspiration confirmed metastatic tumor.

Histologic sampling of the breast masses confirmed infiltrating ductal carcinoma (IDC) from both sites, estrogen receptor and progesterone receptor positive, and HER-2/neu positive. Fine-needle aspiration of the axillary lymph node demonstrated metastatic carcinoma.

CT-guided sampling of a lucent sacral lesion showed metastatic adenocarcinoma, consistent with breast primary (Figure 11).

FIGURE 11 Sampling of the sacral lucent lesion was performed with CT guidance. Axial image with the patient in prone position shows the needle in the lesion. Histology confirmed metastatic adenocarcinoma, consistent with a breast primary.

Palliative radiation therapy to the lumbar spine (L5 and sacroiliac region) and left knee were given for pain relief, and the patient was started on tamoxifen and zoledronic acid (Zometa). Repeat CT images, 7 months later, show a variety of bone responses to these interventions (Figure 12).

FIGURE 12 Coronal CT reconstructions (A anterior to B), 7 months after initial CT in Figure 4, after lumbar spine radiation, and treatment with tamoxifen and Zometa. In A, the left intertrochanteric metastasis has grown. It now displays a dominant central sclerotic component, with a lucent rim. In B, the lytic focus at the superior end plate of L5 remains visible, but now shows a thick surrounding blastic response. At the superior end plate of L3, a blastic lesion is now seen. This may reflect healing of a subradiographic lytic metastasis.

TEACHING POINTS

This is a distinctly unusual presentation of breast cancer, with symptomatic bone metastases bringing this pre-screening-age patient with no notable family history to attention. Less than 10% of breast cancer patients present with stage IV disease. This patient’s primary tumor was also identified through an unusual route, being first seen on a chest CT. Breast cancers can certainly be picked up initially on chest CT, which usually is performed for other reasons. (See Case 15 in Chapter 1 for another example.) For this reason, CT interpreters should always scrutinize chest wall soft tissues and breasts as part of their routine search pattern.

The initial presentation of these bone metastases was lytic, manifested on CT as soft tissue replacement of bone. The lumbar spine was locally treated with radiation therapy for pain palliation. The follow-up study shows a new sclerotic metastasis at L3, which was not previously known. Sclerosis can be a manifestation of healing of a subradiographic lytic lesion. Sclerosis is now also seen surrounding the lytic end-plate change at L5. At the left hip level, the lytic lesion with central sclerosis grew between studies. On the later study, the lesion displays a larger and more sclerotic central component. Is this growth and progression of a mixed lytic and sclerotic metastasis, or growth despite a partial healing response manifested by the larger and more sclerotic central component? Unfortunately, this illustrates the difficulty of assessing by imaging the response of bone metastases to therapy.

CASE 3 Relapse with bone metastases

A 36-year-old woman sought orthopedic advice for low back and right leg pain that developed while training for a half marathon. She had been treated 4 years before for an estrogen receptor and progesterone receptor positive, T1cN1b(4) infiltrating ductal carcinoma with breast conservation and axillary dissection, with 4 of 20 lymph nodes positive with microscopic extracapsular extension. Additional treatment was with chemotherapy (four cycles of doxorubicin [Adriamycin] plus cyclophosphamide [Cytoxan] [AC] and four cycles of paclitaxel [Taxol]); right breast, supraclavicular, and axillary radiation therapy; and tamoxifen.

A pelvis x-ray was abnormal, with a moth-eaten appearance of the right ischium (Figure 1). Subsequent evaluations with pelvis and lumbar spine MRI, body CT scans, positron emission tomography (PET), and bone scan suggested extensive bony metastases (Figures 2, 3, 4, and 5). A bone marrow aspirate confirmed metastatic breast cancer. The patient was placed on goserelin acetate (Zoladex), zoledronic acid (Zometa), and anastrozole (Arimidex).

FIGURE 1 Anteroposterior view of the pelvis (cropped) shows expansion and abnormal bone mineral texture, with coarsened trabecula and mixed lucency and sclerosis, in the right ischium.

FIGURE 2 Sagittal MRIs of the lumbar spine show diffusely abnormal bone marrow signal. A, T1-weighted imaging shows complete loss of normal bright fatty marrow signal. B, T2-weighted, fat-saturated sequence shows heterogeneous patchy abnormal increased signal, involving portions of L1, L5, and S1 particularly. C, STIR sagittal view shows increased abnormal signal involving much of L1 and L2, and portions of L4, L5, and S1.

FIGURE 3 Axial pelvis CT image shows diffuse markedly abnormal bone mineral texture, with mixed lucency and osteosclerosis.

FIGURE 4 Bone scan is abnormal but under-represents the extent of the abnormalities apparent from MRI and CT. There are subtle abnormalities that correspond to known abnormalities (increased activity of the radiographically abnormal right ischium compared with the left), but overall the extent of disease suggested by bone scan is discrepant with other imaging studies. The most clear-cut evidence of bone metastases based on this study includes the patchy foci of increased activity in the shafts of both femurs and the increased activity in the right lower sternum. Activity is asymmetrically increased in the right proximal humerus, but also abnormal on the left. In fact, there is an overall “metabolic bone disease” pattern for much of the activity on this bone scan, with prominent periarticular activity of large joints.

FIGURE 5 A to D, Sagittal PET/CT images of the spine show diffusely abnormal bone mineral texture on the CT portion (left), with some areas predominantly lytic and others mostly sclerotic. Diffuse, but markedly inhomogeneous increased PET scan activity is seen in the sternum and spine. In a general way, the sites of greatest PET scan activity appear to correlate with levels with more lytic than blastic change.

A year later, while on Zoladex, Zometa, and exemestane (Aromasin), the patient was restaged with PET/CT for symptoms of increased aching, weight loss, depressed appetite, and rising tumor markers. PET/CT showed new liver metastases. Chemotherapy was recommended, but declined. The patient was started on fulvestrant (Faslodex). Restaging after 2 months, at which time the patient had fatigue, shortness of breath, and rising tumor markers, showed progression of liver metastases as well as evidence of lung involvement (see Case 13 in Chapter 10). She died a month later, 15 months after bony metastases were detected.

TEACHING POINTS

This case illustrates well the hierarchy of bone imaging studies in terms of how completely they depict osseous metastases. Lowest on the totem pole are x-rays. Still useful to evaluate a site of focal symptoms or to further evaluate a focal bone scan abnormality, radiographs are known to be limited in depiction of bone metastases because up to 50% bone destruction must occur before radiographs will suggest an abnormality.

Higher in the hierarchy is bone scintigraphy. In most cancers, bone scans are 50% to 80% more sensitive than radiographs for detection of bone metastases. However, as in this case, there are instances in which bone scans can severely underestimate the extent of skeletal involvement (see also Case 5). A possible explanation is the very diffuse distribution, with such extensive involvement that there is essentially no normal intervening bone between lesions to increase the conspicuity of metastatic lesions. Such extensive bony metastases could result in a “superscan” pattern of diffuse increased skeletal activity, with high target-to–soft tissue ratio and minimal renal and bladder activity. In such scans, tracer uptake is so increased in the skeleton that less tracer is available to excrete through the urinary system. This bone scan does not demonstrate the classic features of a superscan, but shows some features in the spectrum of a superscan.

The sensitivity of bone scintigraphy can be improved with the addition of single-photon emission computed tomography (SPECT), especially in the spine. This is most applicable if there is a target area of symptoms or if planar bone scan suggests a localized abnormality, but is impractical and time-consuming for routine application. Improved performance of bone scintigraphy can also be achieved with performance of fluorine-18 bone scanning, but at higher cost of the radiopharmaceutical and in imaging time.

MRI is the most sensitive anatomic imaging modality overall for the detection of bone metastases, whether lytic or blastic. CT is the most sensitive radiographic modality. PET imaging is very sensitive, particularly for lytic disease, but may underestimate blastic bone disease. Because the converse is true for bone scans, PET and bone scintigraphy may be considered complementary for comprehensive evaluation of the skeleton for metastases, whether lytic or blastic. Particularly if the PET is a PET/CT, the combination of PET, CT, and bone scan information is a very comprehensive approach for the detection and characterization of osseous metastases.

CASE 5 Recurrence with bone metastases: Disease extent discordant between imaging and bone scan; pathologic fractures

A 65-year-old woman with a history of treated, early-stage breast cancer 1.5 years before, presented with progressive severe hip pain, to the point of being unable to walk. Her breast cancer was stage I, T1bN0, infiltrating ductal carcinoma, which was treated with lumpectomy and brachytherapy. The tumor was estrogen receptor and progesterone receptor negative, and HER-2/neu positive. The patient declined chemotherapy during initial treatment.

Her laboratory data were abnormal at the time of presentation with hip pain, with elevated alkaline phosphatase and CA27.29. The patient was admitted with suspicion of metastatic disease. The workup included a head CT; brain and spine MRI; chest, abdomen, and pelvis CT; and bone scan. Head CT and MRI showed intracranial metastases. CT showed suspicious findings for liver and lung metastases. Bone metastases were shown on spine MRI (Figure 1), as well as on CT (Figure 2). Bone scan was abnormal, most notably for rib, verte-bral, and sacral fracture findings, but underrepresented the extent of diffuse osseous metastases (Figure 3).

FIGURE 1 Sagittal (A) T1-weighted and (B) fat-saturated, enhanced T1-weighted images through the thoracic spine show multiple foci of abnormal marrow signal and enhancement, consistent with metastases. There is also mild loss of height at the involved T12 level, suggesting pathologic fracture. C, Axial, fat-saturated, enhanced image through the sacrum shows diffuse, but heterogeneous enhancement of the sacrum and patchy enhancement of the adjacent ilia. A fracture line is visualized through the left sacral ala (arrow).

FIGURE 2 Images from CT show (A) a destroyed, expanded, soft tissue replaced right posterior rib, and (B) a fracture line through the left sacral ala, with patchy lucency and sclerosis of the upper sacrum.

FIGURE 3 Anterior and posterior views from a whole-body bone scan are disconcertingly disproportionate in representing the findings of osseous metastases. Multiple abnormalities are noted, most of which represent fractures by correlation with CT and MRI. At many levels of known bony metastases, without pathologic fracture, the bone scan markedly under-represents the disease. The transverse band of activity at T12 corresponds to the partially replaced vertebral body with compression on MRI. The typical scintigraphic findings of sacral alar insufficiency fractures correlate with pathologic fractures on CT and MRI. Adjacent rib fractures are seen on the right, laterally (best seen on the posterior view). Just above this level, the lateral rib is absent, corresponding to the destroyed lytic rib metastasis and soft tissue mass seen on CT (arrow). Other levels of vertebral body involvement by MRI in the thoracic spine are not visualized on this bone scan. There are some findings, such as the long segment of increased activity in the right humerus, which do suggest bone metastases.

TEACHING POINTS

The importance of imaging correlation when interpreting bone scans, PET scans, and other nuclear medicine studies cannot be overemphasized. Bone scans are highly variable in their depiction of metastases, depending on the size of the bone lesion, the location in the skeleton, and, importantly, the type of bone reaction incited (lytic, blastic, or mixed). There are patients with radiographic and imaging findings of extensive osseous metastases whose bone scans can appear deceptively benign. At first glance, the bone scan of this patient appears simply to be that of an elderly woman with typical scintigraphic findings of rib, vertebral compression, and sacral insufficiency fractures, with only subtle evidence on scintigraphy of metastatic disease as the etiology of the pathologic fractures. One of the most compelling findings on this bone scan for metastatic disease is the hardest to recognize: the cold lesion or absence of the destroyed right rib. With the roadmap provided by correlation with the CT, consequential oversights like this may be avoidable. At the same time, when presented with apparently discordant findings, the interpretation should reflect the fact that in individual patients, bone scintigraphy may not be the most accurate reflection of the full extent of osseous disease burden.

CASE 6 Diffuse bone metastases, underrepresented on bone scan

A 63-year-old woman was admitted from a rehabilitation center for evaluation of progressive anemia. She was recovering from an admission to another hospital for perforated ulcer, treated with gastrectomy. Her past medical history of right breast cancer was 7 years earlier, Stage 2, T1N2M0, treated with mastectomy, 6 months of adjuvant chemotherapy, and tamoxifen. A bone scan, performed for diffuse aches and bone pain, suggested metastatic disease as well as multiple fractures (Figure 1). Correlation with a concurrent body CT scan showed extensive bone mineral alterations of advanced and extensive osseous metastases, as well as multiple pathologic fractures (Figures 2 and 3). Metastatic adenocarcinoma, consistent with breast primary, was confirmed by CT-guided biopsy of a lytic focus in the left ilium (Figure 4).

FIGURE 1 Anterior and posterior whole-body bone scan images show multiple foci of increased activity in the skull and posterior ribs. The foci in the skull are randomly distributed and highly suggestive of metastases. The posterior rib activity shows multiple adjacent foci bilaterally, in a pattern suggesting fractures. Note subtle heterogeneity of osseous activity elsewhere, best seen in the extremities, as in the left humerus (posterior view) and the femora (anterior view). The right kidney is ptotic. Increased soft tissue activity in the pelvis suggests malignant ascites. A small, focal area of increased uptake is seen in the right inferior pubic ramus.

FIGURE 2 Bone windows from a concurrent body CT scan, at the level of the thorax, show a posterior rib fracture on the right (arrow). Additional fractures were seen at other levels (not shown). The bone mineral density is diffusely abnormal in the visualized sternum, spine, and ribs, indicating the rib fractures are most likely pathologic. There are moderate-sized bilateral pleural effusions as well.

FIGURE 3 The diffuse nature of the abnormal bone texture is even more apparent on a slice through the sacrum. All of the bones show mottled, mixed sclerotic and lytic changes.

FIGURE 4 Image from CT-guided biopsy, showing the planned needle trajectory into a lytic focus in the left ilium.

TEACHING POINTS

Bone scans should never be interpreted in a clinical or imaging vacuum, as this case clearly illustrates. Correlation should always be made with any available radiographs, CT, or MRI. This CT scan demonstrates that the patient has very extensive and essentially confluent bony metastases, involving the entire skeleton, with virtually no normal intervening bone. The very confluence of this osseous disease presumably accounts for the surprising subtlety of the extensive metastases on scintigraphy. In the skull, with scattered, randomly distributed foci of increased activity, the diagnosis of metastatic disease is clearly suggested. Elsewhere, these extensive bony metastases are manifested predominantly by slight heterogeneity of activity. Other than the skull lesions, the bone scan is most remarkable for the multilevel fractures, which presumably are pathologic. In addition to the ribs, CT showed a right inferior pubic ramus fracture, which correlates with the discrete focus of increased activity on scintigraphy.

Considerations in choosing a site for biopsy of a suspected bone metastasis include ease and safety of access (Can the patient lie prone? What structures might be injured?) and factors influencing the likelihood of obtaining adequate material (soft tissue replacing bone is easier to sample than bone forming tumor). In a case like this, in which there are diffuse changes, it makes sense to biopsy the pelvis, rather than risk a pneumothorax or neural injury sampling a rib or vertebral body. It is technically easier to obtain satisfactory material for analysis from a lytic bone metastasis than from sclerotic sites, when there is a choice.

CASE 7 Recurrence with mediastinal lymphadenopathy, vocal cord paralysis, and sclerotic bone metastases

A 59-year-old woman with a history of premenopausal breast cancer presented for medical oncologic evaluation for possible metastatic disease. Her breast cancer history was incomplete and obtained largely from the patient. It had been 10 years before, presenting as a large palpable left breast mass. She was treated with neoadjuvant chemotherapy with clinical response, followed by subsequent high-dose chemotherapy with stem cell rescue. A mastectomy and flap reconstruction was performed. The pathology specimen showed a 6-cm infiltrating lobular carcinoma (ILC). Six of 10 lymph nodes were involved. Chest wall radiation was performed. She declined recommended tamoxifen therapy.

A month before presentation, the patient developed left back pain, radiating laterally and to the front of the chest. A bone scan was abnormal at multiple levels, including T5, T11, the skull, and right humerus.

Oncologic evaluation included tumor markers, which were elevated (CEA and CA27.29). Positron emission tomography (PET) and CT imaging showed hypermetabolic mediastinal and hilar adenopathy and bone metastases (Figures 1, 2, 3, 4, 5, and 6). The patient had been noted to be hoarse on physical examination, and PET showed asymmetrical muscular activity of the vocal cords. Metastases in the spine were further evaluated with MRI to assess epidural tumor extension (Figures 7, 8, 9, and 10).

FIGURE 1 Three plane (left to right: coronal, sagittal, and axial), fused PET/CT with coronal projection volumetric image (right) shows multiple foci of hypermetabolism from soft tissue and bone metastases. Activity at the right proximal humeral (see coronal view) and L5 levels (seen on sagittal view) indicates bone metastases, whereas activity in the chest (seen on sagittal PET/CT and coronal projection image) shows aortopulmonary (AP) nodal disease, in a location to produce hoarseness. The coronal and axial views show asymmetry of vocal cord activity, with hypermetabolic muscular activity on the right at the level of the normally functioning cord, and absence on the left, paralyzed at the level of the left recurrent laryngeal nerve in the AP window.

FIGURE 2 Sagittal PET/CT through the spine (left to right: CT, PET, fused PET/CT, non-attenuation-corrected PET) shows two foci of hypermetabolism in the thoracic spine, which on CT appear to be sclerotic. Anterior to the upper thoracic bone lesion is seen hypermetabolism in mediastinal nodes. Activity in the neck is vocal cord.

FIGURE 3 Axial PET/CT images show a hypermetabolic metastatic focus in the right ilium.

FIGURE 4 CT images for correlation (A, soft tissue window; B, bone window). The blastic medullary bone change is visually much more apparent on soft tissue windows and comparatively subtle on the bone window.

FIGURE 5 Axial PET/CT images through the hips show a smaller focus of hypermetabolism in another metastasis in the right anterior acetabulum.

FIGURE 6 The corresponding CT images (A, soft tissue window; B, bone window) show subtle asymmetry, comparing the medullary density of one anterior acetabulum to the opposite side.

FIGURE 7 Sagittal MRI images of the thoracic spine, for correlation (A, T1-weighted; B, fat-saturated, enhanced, T1-weighted) show two thoracic metastases. A, On T1-weighted, unenhanced sequences, metastases show replacement of the normal bright fatty marrow of the vertebra with hypointense soft tissue intensity. The upper thoracic lesion shows evidence also of pathologic fracture, with loss of height. A small, hyperintense, fatty signal hemangioma is noted at the level below the lower thoracic metastasis. Increased soft tissue is seen in the visualized mediastinum anterior to the spine, corresponding to the nodal PET-positive disease. B, Fat-saturated, enhanced, T1-weighted image shows both vertebral lesions to be enhancing, as well as the mediastinal lymphadenopathy. Involvement of the lower thoracic posterior elements is easier to appreciate on this sequence. Enhancement is easier to visualize against a background of dark, suppressed fat signal. The hemangioma saturates, as would be expected on this sequence.

FIGURE 8 Axial CT images of T11 show lytic change of bone on the left, as well as the left transverse process. A subtle sclerotic margin is seen at the medial edge of the vertebral lysis.

FIGURE 9 Enhanced soft tissue window from the same level shows an enhancing rim of soft tissue extending into the epidural space on the left anterolaterally.

FIGURE 10 Corresponding axial fat-saturated, enhanced, T1-weighted image better shows the extent of enhancing tumor involvement of the left posterior elements, as well as the left anterolateral epidural tumor.

A fine-needle aspiration was performed of a mediastinal lymph node, confirming metastatic disease, which was estrogen receptor and progesterone receptor positive and HER-2/neu negative. Radiation therapy was begun to the spine for palliation of pain.

TEACHING POINTS

Development of musculoskeletal symptoms is a clear clinical indication for restaging of a breast cancer patient. Systemic metastases occur most commonly to bone and are seen in up to one third of patients with recurrent disease. Traditionally, new focal pain would be evaluated with a radiograph of the joint or extremity, with the entire skeletal system surveyed with bone scintigraphy. Today, bone scanning is still indicated as a cost-effective modality, which is particularly efficacious in identifying sclerotic metastases. If bone scanning does not provide an explanation for new symptoms, PET/CT may be indicated as the next imaging step for a highly sensitive comprehensive total-body evaluation for soft tissue and lytic bony metastases.

In this case, although bone scan showed evidence of bone metastases, PET/CT imaging showed soft tissue metastases to mediastinal lymph nodes. It also provided an explanation for the patient’s hoarseness, showing PET-positive nodal disease in the aortopulmonary window and absence of vocal cord activity on the left. The CT component of the study provided an imaging correlate for the PET-positive bone metastases, showing some of the lesions to be blastic and others predominantly lytic. As demonstrated in this case, at times subtle blastic bony change can be easier to see on soft tissue windows than on bone windows, so assessment of the bones for metastases optimally should include review with more than one window setting.

This case also illustrates the limitations of CT for the evaluation of intraspinal and epidural disease. Although enhancing epidural soft tissue from the expanding bone metastasis is fairly clearly demonstrated in this case, identification of more subtle intraspinal extension is generally best undertaken with enhanced MRI.

It is highly desirable to obtain histologic confirmation when metastases from breast cancer are initially identified. In this case, mediastinal and bone metastases were identified essentially concurrently, and the question arises as to which lesion is preferable to sample. In this case, why biopsy the soft tissue disease in the mediastinum, when there are potentially less risky sites to sample in bone?

Multiple factors come into play when making this decision. Although it may be safer in terms of not entering a body cavity to sample an osseous site of suspected metastasis, it may be more difficult to obtain satisfactory material from bone than from soft tissue sites of disease. Not only does sufficient material need to be obtained to confidently diagnose malignancy if present, but estrogen receptor, progesterone receptor, and HER-2/neu information needs to be derived from the analysis if the lesion is metastatic breast cancer. At times, the receptor status of metastases can differ significantly from that of the original primary tumor, which has a major influence on treatment decisions.

Another factor to be considered is that thoracic disease could signify a new, potentially treatable, primary lung carcinoma, rather than metastatic disease from breast cancer. Such a distinction could be more difficult to make by analysis of a bone metastasis and is an argument for sampling of thoracic soft tissue disease.

CASE 8 Significance of solitary rib activity on bone scan; chemotherapy-related marrow activation changes on PET

A 39-year-old woman developed abdominal distention and increased fatigue a year after being treated for a high-risk breast cancer. This was an estrogen receptor– and progesterone receptor–positive, HER-2/neu positive, 2.7-cm infiltrating ductal carcinoma (IDC) with an extensive intraductal component of the right breast. Ten of 12 lymph nodes showed metastatic carcinoma, including a 3-cm lymph node with focal extracapsular extension. She was treated with a mastectomy, four cycles of doxorubicin and cyclophosphamide, followed by four cycles of paclitaxel, as well as chest wall, supraclavicular, and axillary radiation.

When the patient developed constitutional symptoms, she was restaged with CT scans, which showed extensive hepatic metastases. Chemotherapy with carboplatin, docetaxel (Taxotere), and trastuzumab (Herceptin) was initiated, with good response. Later the same year, brain metastases were identified, and treated with whole brain radiation therapy, with MRI resolution of lesions.

Evaluations for persistent right rib pain showed a fracture, seen on bone scan (Figures 1, 2, 3, and 4), CT (Figure 5), and PET scan. The patient had been coughing, and the activity seen on PET associated with the fracture was demonstrated subsequently to resolve on repeat PET.

FIGURE 1 Anterior and posterior whole-body bone scan images show a solitary right anterolateral focus of increased activity.

FIGURE 2 Oblique views of the ribs were obtained to better delineate the pattern of increased rib activity. No long segment involvement is seen to particularly suggest this is a metastatic lesion, but the activity is also less focal than seen in many rib fractures. Multifocal activity involving multiple adjacent ribs is a pattern of activity that would allow one to more confidently diagnose fractures.

FIGURE 3 A follow-up bone scan (anterior and posterior projections), 2 months later, shows persistent activity at the same level.

FIGURE 4 Oblique rib views from this study show the activity to be focal and intense. This pattern of activity is more typical of a fracture.

FIGURE 5 Bone window from a concurrent chest CT scan shows a healing fracture. There is offset of the bone cortices, which may have contributed to the prolonged symptoms and bone scan activity. The patient is post–right mastectomy, and a left breast implant is noted.

PET scans obtained to monitor the patient’s response to the chemotherapy given for liver metastases show typical findings of the frequently encountered treatment phenomenon of marrow activation (Figures 6 and 7).

FIGURE 6 Sagittal PET/CT images (from left to right: CT, attenuation-corrected PET, fused PET/CT, non-attenuation-corrected PET) obtained while the patient was undergoing chemotherapy for liver metastases show a uniform increase in marrow activity, new from a prior study. Note the ease of visualizing the sternum on the PET images, which generally is difficult to see. CT images show uniform bone mineral density, without evidence of diffuse osseous metastases.

FIGURE 7 Comparable views from a repeat PET/CT, 2 months later, after cessation of chemotherapy, show resolution of the increased PET bone marrow activity. Bone mineral density on CT remains stable and normal.

TEACHING POINTS

A new, solitary bone scan finding in any patient with cancer poses a diagnostic dilemma. Unlike the more conclusive bone scan pattern of multifocal metastatic activity, such a bone scan result is often equivocal and will generally require correlation with imaging studies of the area with x-ray, CT, or MRI. Biopsy may ultimately be necessary to confirm or exclude metastatic bony disease. See a similar case, Case 9 in this chapter, for a discussion of bone scan rib findings.

In this case, in a patient with known liver and brain metastases, development of bone involvement would not be surprising.

How often is a solitary rib lesion on a bone scan in any cancer patient a metastasis? That depends on where the activity is located. Solitary spine foci are much more likely to be metastatic than a solitary rib focus: >40% versus <20%. This reflects the fact that vertebral bodies are highly vascularized and contain 75% of the body’s red marrow.

This case also illustrates typical PET findings of marrow activation.

Granulocyte colony-stimulating factor (G-CSF) is frequently given in conjunction with myelosuppressive chemotherapy to stimulate neutrophil precursors in bone marrow. On PET imaging, a diffuse increase in marrow activity is seen, which resolves after cessation of therapy. Generally, the increased marrow activity is diffuse and uniform. Rarely, this process may be less uniform and more difficult to differentiate from diffuse osseous metastases.

CASE 10 Solitary sclerotic rib metastasis (positive on bone scan, negative on PET)

A 57-year-old woman who had recently undergone bilateral mastectomies for newly diagnosed bilateral breast cancers was referred for medical oncology evaluation and therapy. Her breast cancer had been diagnosed the preceding month, when she presented with a palpable left-sided lump. The patient had never undergone mammographic screening.

On the left, her tumor stage was IIIA, T2N2, 3.2-cm infiltrating ductal carcinoma (IDC), with angiolymphatic invasion and negative margins. Eight of 16 lymph nodes were involved. The tumor was estrogen receptor and progesterone receptor positive and HER-2/neu positive.

Her right breast cancer was a T2NxM0, 2-cm IDC, estrogen receptor and progesterone receptor positive, and HER-2/neu negative, with negative margins and angiolymphatic invasion. One sentinel lymph node was involved.

The patient was staged with enhanced body CT scans, PET/CT, and bone scan. These evaluations showed a highly suspicious left anterior fourth rib focus of increased activity on bone scan (Figure 1), with subtle corresponding sclerosis at the same level on CT (Figure 2). There was no corresponding abnormal activity on PET. CT-guided biopsy showed metastatic adenocarcinoma, consistent with breast primary.

FIGURE 1 Whole-body bone scans (A, anterior and posterior whole-body views; B, oblique of the ribs) show a focus of intensely increased activity at the anterior left fourth rib level. Although the location brings trauma to mind, the activity is larger and less focal than generally seen with a rib fracture, especially when viewed in the left anterior oblique projection. The rib activity “shines through” on the posterior views. Increased chest wall soft tissue activity is seen, right greater than left, from recent mastectomies. Scoliosis is seen of the spine, but no other highly suspicious findings are noted (activity at L2 is indeterminate and seemed to correspond with degenerative spondylytic change on CT).

FIGURE 2 Detail views of the left anterior fourth rib from body CT scan show subtle blastic alteration of bone mineral density at the level in question, without expansion or fracture.

Because this seemed to be the only site of metastatic disease, aggressive chemotherapy was given, consisting of four cycles of doxorubicin (Adriamycin) and cyclophosphamide (Cytoxan) [AC], followed by 3 months of trastuzumab (Herceptin) and paclitaxel (Taxol). She then received Herceptin every 3 weeks for the remainder of the year, as well as anastrozole (Arimidex) and zoledronic acid (Zometa) every 6 months.

TEACHING POINTS

Systemic screening of this patient was indicated because of her high-risk tumor status (HER-2/neu positive and eight lymph nodes involved). The combination of studies obtained in this case to systemically stage this patient (PET/CT, enhanced body CT, and bone scan) is quite comprehensive, and arguably a very cost-effective combination. Even more cost-effective would be for the CT component of the PET/CT to be contrast enhanced, thereby eliminating one of the two CT scans obtained in the evaluation illustrated in this case. There are advantages and disadvantages to this approach.

This case illustrates well the complementary roles PET and bone scan play in the evaluation of breast cancer patients for osseous metastases. PET is a highly sensitive whole-body modality for the evaluation of soft tissue and visceral and lytic osseous metastases, but has known limitations in the demonstration of osteoblastic bone lesions. Conversely, bone scintigraphy is most reliable for identification of blastic bone metastases and may not show purely lytic lesions. As illustrated by this case, PET can miss osteosclerotic bone metastases that bone scan will find.

CASE 11 Assessing activity of sclerotic bone metastases

A 41-year-old woman with known sclerotic bone metastases taking letrozole (Femara) and monthly zoledronic acid (Zometa) developed generalized aches and pains, prompting imaging re-evaluation with bone scan, positron emission tomography (PET)/CT, and spine MRI. The patient had known sclerotic bone metastases for 3 years, which were inactive on prior bone and PET scans. Her initial diagnosis of breast cancer had been 7 years before at age 34, with T2N1M0, 3-cm, well-differentiated infiltrating ductal carcinoma. One of 11 lymph nodes was positive. She underwent right mastectomy with implant reconstruction, chemotherapy with four cycles each of doxorubicin (Adriamycin) plus cyclophosphamide (Cytoxan) (AC) and paclitaxel (Taxol), as well as chest wall and supraclavicular radiation.

When the development of symptoms prompted reimaging, a bone scan was obtained initially and showed new activity at T9 (Figures 1, 2, and 3), corresponding to a sclerotic metastasis known from CT (Figure 4). PET scan showed no associated hypermetabolism, and the size of the lesion on CT was unchanged. Spine MRI showed signal intensity characteristics of sclerosis at this and other levels. Subtle partial enhancement was noted of the T9 lesion, correlating with the new bone scan activity (Figure 5).

FIGURE 1 Whole-body bone scan in anterior and posterior projections (right side marked at skull level) shows a new focus of increased activity at the T9 vertebral level. Based on these views, it is difficult to say whether the activity is in a spinous process or the vertebral centrum.

FIGURE 2 Oblique views of the thoracic spine from the same bone scan localize the activity to the vertebral body.

FIGURE 3 Comparison whole-body bone scans for correlation, obtained 1 year (A) and 6 months (B) before. A, No abnormality is seen in the spine. B, No definite abnormality is seen. In retrospect, there may be very subtle change at the T9 level.

FIGURE 4 CT images for correlation (bone windows) from 3 years before (A), 1 year before (B), and concurrent with the bone scan in Figure 1 (C), show relatively stable findings of a sclerotic metastasis in the left side of the T9 vertebral body. No change in size is seen over this series of examinations. The lesion appears progressively more sclerotic, but the significance of this is uncertain.

FIGURE 5 Sagittal thoracic spine MRIs show foci of hypointensity on all sequences at T7, T9, and L1 (partially visualized). These signal intensity characteristics are consistent with sclerosis. Note subtle enhancement of the posteroinferior portion of the T9 lesion on D. This same component shows subtle higher signal on STIR than the rest of the sclerotic lesion. A, T2-weighted turbo spin echo (TSE). B, STIR. C, T1-weighted image. D, Gadolinium-enhanced, T1-weighted image.

TEACHING POINTS

Assessing the activity of sclerotic metastases is difficult. Generally, radiography and CT are not particularly helpful. Only in the rare instance of development of a new superimposed soft tissue mass or lytic components might x-ray or CT suggest clear progression or activation of a blastic metastasis. Assessing degrees of sclerosis of blastic me-tastases, either qualitatively or quantitatively, is problematic. PET and PET/CT also are frequently limited in their utility for assessment of the activity of sclerotic bone metastases, with many such lesions not associated with increased activity on PET. Development of new activity on PET associated with a sclerotic bone metastasis is helpful in suggesting disease reactivation (e.g., see Case 1 in Chapter 12). However, as this case illustrates, bone scan is very sensitive to changes in sclerotic metastases and may be the first or only modality to suggest disease reactivation. Typical imaging findings of sclerotic metastases are demonstrated by this case: focal increased activity on bone scan, increased bone density on CT, and hypointensity on all sequences on MRI. It is interesting to note the subtle enhancement seen on MRI at the level of the newly bone scan active T9 blastic metastasis. Other sclerotic bone metastases seen on the spine MRI of this patient did not show any enhancement and were scintigraphically inactive.

CASE 12 Diffuse bone metastases (superscan)

A 69-year-old woman with known extensive bony metastatic disease was evaluated with a bone scan (Figures 1, 2, and 3). Her bony metastatic disease had been diagnosed 6 years earlier, and she had undergone palliative radiation to the pelvis and hips, as well as systemic treatment with samarium. Her original diagnosis of left breast cancer was 18 years earlier. This was a T1cN1M0, left, 1.9-cm infiltrating ductal carcinoma (IDC) with 1 of 20 lymph nodes involved. She had been treated with mastectomy and with breast and regional lymphatic radiation.

FIGURE 1 A diffuse, heterogeneous pattern of increased bone scan uptake is seen, with only faint renal activity. The pelvis and hips had been previously radiated, and a corresponding relative decrease in bone scan activity is seen. Activity in the sternum, spine, and skull is increased diffusely, with more patchy foci of activity in the humeri.

FIGURE 2 Correlation with a chest CT bone window shows patchy sclerotic bone metastases, seen here in the sternum, right lateral rib, left posterior rib, and vertebral body and posterior elements.

FIGURE 3 Correlation with a sagittal, T1-weighted brain MRI shows diffuse hypointensity of bone marrow signal throughout the calvarial and upper cervical spine, correlating with the CT findings of diffuse blastic metastases and the diffuse increased skull uptake on bone scan.

TEACHING POINTS

The term superscan can be applied to bone scans with markedly increased skeletal uptake, resulting in less tracer available for renal excretion and visualization. Conditions resulting in a high bone-to–soft tissue activity ratio, with faint or little renal activity, include extensive bone metastases and metabolic bone diseases. The primary tumors most likely to produce the superscan pattern are carcinomas of the breast, lung, and prostate. Metabolic bone diseases that can produce a similar superscan pattern include renal osteodystrophy, osteomalacia, hyperparathyroidism, myelofibrosis, and systemic mastocytosis. As in this example, the diagnosis of diffuse osseous metastases is suggested by inhomogeneity of the activity, and correlation with x-rays, CT, or MRI often enables ready confirmation.