Metastatic Disease: Bone, Spinal Cord, Brain, Liver, and Lung

Published on 03/04/2015 by admin

Filed under Hematology, Oncology and Palliative Medicine

Last modified 03/04/2015

Print this page

This article have been viewed 2329 times

Chapter 23 Metastatic Disease

Bone, Spinal Cord, Brain, Liver, and Lung

Paul Okunieff, Michael T. Milano, Arthur T. Porter, Marc David

Local tumor control is now achieved with modern combined-modality therapy. Metastatic disease therefore dominates the survival outcomes of cancer patients. Metastases to the bone commonly cause pain, whereas those to the lung, liver, and brain cause organ dysfunction with substantial changes in quality of life. Metastases to all these organs can shorten life.

Although palliative care is a large part of the clinical practice of oncology, studies show that cancer pain is often inadequately managed.1,2 Among painful metastases, osseous metastases³ remain the most common cause of intractable pain in cancer patients. Bone is the third most common site of metastases after lung and liver.⁴ Metastases usually become apparent after the diagnosis of the primary tumor, but in up to 23% of patients they are the presenting problem. Bone pain results in immobility, anxiety, and depression and it severely affects a patient’s quality of life. Therefore treatment of bone pain is a high priority.

Brain metastases are a devastating complication of malignant disease. Most patients with brain metastases succumb to metastatic disease within a few months.5,6 Relatively few are candidates for open surgery, which can carry the risk of severe sequelae. The advent of radiosurgical techniques and high-quality magnetic resonance imaging (MRI) of the brain has greatly changed the outcomes of patients with brain metastases, especially when combined with whole-brain irradiation. Patients commonly have improved duration of cognitive performance and, in some, prolonged survival times.⁶ The success of this approach is limited to patients with tumors in favorable regions of the brain. Among patients with lung cancer, the actuarial incidence of brain metastases can exceed 70%.^7,⁸ Other cancers may also have high rates of metastasis, including 19% in female breast cancers.⁹

Liver metastases are common in many cancers and, like brain metastases, are usually associated with a limited median survival time. Liver metastases occur in 40% to 70% of patients with progressive colorectal cancer and in a similar range of patients with progressive cancer of the breast or lung.^10,¹¹ Treatment for liver metastases is often difficult and, therefore, inadequately managed.⁹ Recently there has been an effort to treat liver metastases using localized therapies such as stereotactic radiation techniques, radiofrequency ablation, hyperthermia, and embolic therapy in addition to resection. There has been satisfactory success with many of these treatments, which appear to be improving patient outcomes.

Bone Metastases

Table 23-1 shows the prevalence of skeletal metastases in several autopsy series.¹² The marked variation may be attributed to differences in the thoroughness of the pathologic examination of the skeleton. Bone scintigraphic surveys have, in general, reported higher rates of bone metastases. In a study by Tofe and colleagues,¹³ bone scans of 1143 patients with a nonosseous primary tumor were examined; 61% of the patients had an abnormal bone scan finding, and 33% had breast, lung, or prostate primary cancer.

TABLE 23-1 Prevalence of Skeletal Metastases at Primary Site

Primary Site	Prevalence (%)
Breast	47-85
Prostate	54-85
Thyroid	28-60
Kidney	33-40
Bronchus	32-40
Esophagus	5-7
Other gastrointestinal sites	3-11
Rectum	8-13
Bladder	42
Cervix	0
Ovaries	9
Liver	16

Data from Galasko CSB: Incidence and distribution of skeletal metastases. Clin Orthop Relat Res 210:14-22, 1986.

In a prospective series of hospital patients with bone metastases, the tumors carrying the highest risk of bone metastases were those originating in the prostate (32.4%), breast (21.9%), kidney (16.4%), thyroid (11.7%), lung (10.9%), and testes (10.2%).² Table 23-2 shows the incidence of patients developing bone metastases by primary site.

TABLE 23-2 Incidence of Bone Metastases According to Primary Site

Primary Site	No. Patients	Patients with Bone Metastases (%)
Breast	6423	17
Prostate	144	16
Esophagus	451	6
Lung	589	5
Bladder	172	5
Rectum	274	4
Thyroid	107	4
Uterine cervix	1981	3
Uterine corpus	509	3
Head and neck	2860	2
Ovaries	586	1
Colon	153	1
Stomach	118	1

Data from Tubiana-Hulin M: Incidence, prevalence and distribution of bone mets. Bone 12:S9-S10, 1991.

The distribution of skeletal metastases from breast cancer is shown in Table 23-3. Similar distributions have been noted from primary cancers in the prostate, lung, and breast.^14,15,16,17

TABLE 23-3 Distribution of Skeletal Metastases in 212 Breast Cancer Patients

Anatomic Site	At Presentation (%)*	At Any Time (%)*
Lumbar spine	52	59
Thoracic spine	35	57
Pelvis	31	49
Ribs	18	30
Femur	15	24
Skull	12	20
Cervical spine	11	17
Humerus	8	13
Other sites	3	3
Diffuse	1	12

* Of all patients.

Data from Tubiana-Hulin M: Incidence, prevalence and distribution of bone mets. Bone 12:S9-S10, 1991.

Pathophysiology

Numerous tumor cells gain access to the systemic circulation, primarily through the capillary system, but some gain access through the lymphatics; only a few of these cells are able to establish a metastatic focus successfully.¹⁸

Cancer cells metastasize to bone mostly via hematogenous spread. Skeletal blood flow accounts for only 4% to 10% of the cardiac output,¹⁹ and some authors think that the incidence of skeletal metastases is higher than expected from the relatively low perfusion of bone. A mechanism explaining the high incidence has been described by Weiss.²⁰ The microstructure of the hematopoietic marrow renders it particularly vulnerable to tumor cell accumulation and ultimate invasion. Nutrient arteries to the bone tend to subdivide into capillaries as they near the endosteal margin of the bone. These capillaries become continuous with a rich venous sinusoidal system, with a capacity six to eight times that of the osseous arterial system. More important, the circulation comes to a near standstill at this point, allowing tumor cells more time to invade the matrix.

To sustain growth, a colony of tumor cells needs to obtain its own vascular supply once it has been established. A hypothesis is that a tumor angiogenesis factor attracts vessels to a small tumor colony that would otherwise be dependent on local tissue circulation and incapable of further invasion.²¹ The production of such a tumor angiogenesis factor may be partly blocked by the immune responses, presumably mediated through lymphocytes. Therefore an established micrometastasis may attract vasculature required for growth several years later. This theory may explain the late appearance of metastases long after definitive treatment of the primary tumor.

Some tumors, notably of the breast, prostate, lung, kidney, and thyroid gland, produce and secrete humeral mediators that stimulate osteoclast activity. These include transforming growth factor, platelet-derived growth factor, tumor necrosis factor, prostaglandins, procathepsin D, interleukins, parathyroid hormone–related protein, and granulocyte-macrophage colony-stimulating factors.22,23

The distribution of metastases in the skeletal system is not uniform. Bone metastases tend to involve the axial skeleton more often than the appendicular skeleton. Considering the equally rich hematopoietic system in the appendicular skeleton, this higher predilection argues for specific bone marrow-derived growth factors that fertilize the “soil” of the bone for tumor growth.¹⁹

Diagnosis

Laboratory Studies

Biochemical parameters include alkaline phosphatase, urinary hydroxyproline, and the urinary hydroxyl-proline-creatinine ratio. These lack specificity, and are of no value in the diagnosis of skeletal metastases.^24,²⁵

Imaging Studies

Skeletal scintigraphy is usually the first-line imaging technique used for detecting skeletal metastases. A bone scan is more sensitive than plain radiographs and has the advantage of examining the entire skeleton. Most lesions evoke an osteoblastic response, which shows up as an increased tracer uptake.²⁶ Occasionally, metastases may show up as areas of decreased uptake. This may be observed in rapidly growing lesions, when bone destruction far exceeds new bone formation, or secondary to an infarction. Highly vascular metastases, such as those from a renal primary cancer, may be seen on the early vascular phase of the bone scan. Metastases not detected by a bone scan include tumors that do not evoke an osteoblastic response, such as myeloma, some lymphomas, and very small deposits.²⁷

Widespread metastatic disease may be misinterpreted as a normal scan with symmetric uptake. In these situations, a reduction in the urinary excretion of the isotope and faint or absent renal uptake with decreased bladder activity are clues of an abnormal scan.²⁸

Most skeletal metastases develop in the medulla and involve the cortex late; therefore plain radiographs are generally insensitive.²⁸ Within the spine, the vertebral body is affected first, although the radiologic findings of pedicle destruction are noted first.²⁹

Computed tomography (CT) scanning has been found to differentiate between metastases and degenerative joint disease, even though the two coexist, and the latter is a common cause of increased uptake on a bone scan. Muindi and colleagues ³⁰ reported that 50% of patients with breast cancer who have a positive bone scan and a normal radiograph had obvious skeletal metastases on a CT scan; 25% had a benign cause and 25% had a negative CT scan. None of the patients with a CT scan that was negative for metastases subsequently experienced metastases. CT scan is also valuable in evaluating soft tissue involvement and can be combined with myelography for detecting extradural tumor spread in patients unable to undergo MRI.

MRI has recently been described as the method of choice for examining the spine. It is more sensitive than a bone scan for detecting early metastases within the medulla, but both T1- and T2-weighted images are required.³¹ It is the procedure of choice when neural compression is suspected,³¹ because it is less invasive than CT myelography, and a small incidence of acute deterioration of neurologic function has been reported by CT myelography.³² When cord impingement is suspected, imaging of the entire spine should be considered because approximately 10% of patients have multiple levels of cord impingement.³³ It is also used in discriminating between benign and malignant vertebral collapse. In the future, whole body MRI could emerge for metastasis screening.³⁴ Disadvantages of an MRI include its high cost, exclusion of patients with metal implants, those with severe claustrophobia, and inferior visualization of the cortex compared with a CT scan.

Positron emission tomography (PET) with ¹⁸F-fluoride or 2-fluoro-deoxy-D-glucose (FDG) is used for the initial staging of many cancers and is helpful in the diagnosis of bony metastasis. ¹⁸F-fluoride is a bone imaging agent and forms fluoroapatite in osteoblastic cells. Uptake is higher than for the technetium-99m (^99mTc) used for bone scintigraphy.³⁵ FDG is a tumor imaging agent that exploits the higher glycolysis activity in the tumor cells.³⁶ FDG-PET scan compared with bone scintigraphy shows a similarly high sensitivity (range from 74% to 95%) but a higher specificity (range of 90% to 97%).³⁷^–⁴¹ Limitations include traumatic, infectious, and inflammatory processes that can also accumulate glucose. Accumulation of FDG requires the tumor to have an adequate metabolic rate, and neoplasms such as prostate adenocarcinoma are not consistently seen with PET scans.^42,⁴³ PET images provide poor anatomic imaging and are most useful when employed with a fused CT or MRI image.^35,44,45

Biopsy

Bone biopsy is not routinely necessary. It is helpful in patients with no history of cancer, in patients with a solitary lesion (in whom a more aggressive treatment approach may be indicated), and in patients with more than one suspected primary tumor. In the latter situation, it may make a difference in the systemic therapy.

Treatment

The primary goal of therapy is to improve quality of life. To achieve this goal, we need to decrease or eliminate pain and improve or maintain function for the duration or as a major component of the patient’s remaining life. The complexity, duration, and cost of therapy should be low, and complications should be avoided.

Treatment recommendations must be individualized. A key consideration is the patient’s overall prognosis. This assessment should be based on an understanding of the natural course of the specific disease. Although the survival of patients with bone metastases is generally poor, potential long-term survivors must be identified. Not only would they require more durable relief, but they are also more at risk for a late, treatment-related complication. In the Radiation Therapy Oncology Group (RTOG) trial,⁴⁶ median survival times in patients with solitary and multiple bone metastases were 36 weeks and 24 weeks, respectively. Patients with breast and prostate primary tumors survived significantly longer (30 to 73 weeks), whereas patients with lung cancer died within a median of 12 to 14 weeks. Patients with renal cell carcinoma with solitary metastasis are also likely to be long-term survivors. Mogens⁴⁷ monitored 25 such patients for 10 to 14 years. The median survival time was 4.3 years; the 5-year survival rate was 36%; and the 10-year survival rate was 16%.

Pharmacologic Treatment

Systemic therapy remains the mainstay of treatment for metastatic disease, including bone metastases. For patients with asymptomatic bone metastases who are not at immediate risk of fracture, disease-appropriate systemic therapy, including chemotherapy, hormonal therapy, and biologic therapy, is indicated. Patients may also benefit from bisphosphonate therapy.

Bisphosphonates

The discovery of compounds inhibiting calcium phosphate precipitation in plasma and urine led to an interest in the use of bisphosphonates as therapeutic agents. The inhibitory activity was attributed to inorganic pyrophosphate, but the use of this agent was limited because of its rapid hydrolysis when given parenterally. Subsequent research led to the development of pyrophosphate analogs resistant to endogenous phosphatases, now known as bisphosphonates.

Bisphosphonates inhibit osteoclast-mediated bone resorption. The exact mechanism is unclear, but postulated mechanisms include direct biochemical effects on the osteoclast, prevention of osteoclast attachment to the bone matrix, and inhibition of differentiation of osteoclast precursors and recruitment.

Four phase II trials of intravenous pamidronate every 2 to 4 weeks as the sole treatment of osteolytic bone metastases in patients with breast cancer reported similar results.⁴⁸^–⁵¹ Relief of pain was noted in approximately 50% of patients, and approximately 25% showed radiographic evidence of bone healing. Similar results for bone pain have also been reported in patients with prostate cancer.

In more recent studies, one phase II ⁵² and one phase III⁵³ trial showed equivalence between zoledronic acid and pamidronate. Rosen and colleagues⁵³ conducted a three-arm study in 1648 patients with bone lytic or mixed disease from either breast cancer or multiple myeloma. Patients received intravenous pamidronate (90 mg) or zoledronate (4 mg or 8 mg) every 3 weeks for 13 months. The primary endpoint was the incidence of a skeletal event and secondary endpoints were pain relief and performance status (Eastern Cooperative Oncology Group [ECOG] trial). All treatment groups showed equivalence with a skeletal event at 12 months and pain scores decreased by an average of 0.5 on a scale of 1 to 5. This randomized trial led to modification of the American Society of Clinical Oncology (ASCO) 2003 and the Cochrane Breast Cancer Review Group update recommendations on the use of bisphosphonates in breast cancer.^54,⁵⁵ Both boards now recommend either pamidronate 90 mg intravenously over 2 hours or zoledronic acid 4 mg intravenously over 15 minutes for patients with an abnormal bone scan and abnormal imaging by plain radiographs on CT scan or MRI. There is no current evidence to treat asymptomatic patients with an abnormal bone scan.

Zoledronic acid has also been used in patients with prostate cancer to treat blastic metastases.⁵⁶ Saad and associates⁵⁶ randomized 643 patients to placebo or to zoledronic acid 4 or 8 mg by intravenous infusion every 3 weeks for 15 months. Results show a reduction in skeletal-related events from 44% to 33%, with a significant p value of .021. Pathologic fractures were reduced from 22% to 13% (p = .015). Onset of the events occurred at a median time of longer than 420 days (median not reached) in the group receiving zoledronic acid and at 321 days in the placebo group. Time to disease progression or survival was similar in both groups. The need for local field irradiation was not significantly different in the two groups.

Zoledronic acid has also been evaluated for the treatment of bone metastases from other disease sites; 773 patients with lung, renal, head and neck, thyroid, and unknown primary tumors received either a placebo or zoledronic acid 4 or 8 mg.⁵⁷ Skeletal events, including hypercalcemia, were significantly reduced from 47% to 38% (p = .039), and the median time to the first event was slower in the zoledronic acid group (225 days vs. 155 days; p = .023).

Although zoledronic acid is well tolerated, the treating physician is advised to monitor the serum creatinine before each administration. Caution is also advised for patients receiving concomitant aminoglycoside or a loop diuretic because of an increased risk of hypocalcemia. Ruggiero and colleagues ⁵⁸ published a retrospective review of 63 patients taking bisphosphonates who suffered osteonecrosis of the jaw; 57% received pamidronate and 21% zoledronic acid. Surgical treatment was required. Oral agents such as clodronate have low bioavailability (2%) and produce gastrointestinal side effects.

Analgesics

The optimal management of pain begins with careful assessment of the degree of pain, the site, functional limitations, and concurrent neurologic symptoms.⁵⁹ The World Health Organization (WHO) analgesic ladder for cancer pain management provides a good guideline for analgesic use.⁶⁰

Step I Nonopioid with or without adjuvant therapy

Step II Opioid for mild to moderate pain, plus nonopioid, with or without adjuvant therapy

Step III Opioid for moderate to severe pain, with or without nonopioid, with or without adjuvant therapy

Step I nonopioid analgesics include acetaminophen, aspirin, and other nonsteroidal anti-inflammatory drugs (NSAIDs). The dosage of acetaminophen should not exceed 4 g/day. Step II opioids include codeine, dihydrocodeine, hydrocodone, oxycodone, and propoxyphene. Step III opioids include morphine, oxycodone, hydromorphone, and fentanyl.

Attention should be paid to selecting the appropriate analgesic, dose, route, and schedule. Continuous, slow-release medications are generally preferred over short-acting medications. The latter can be used effectively for breakthrough pain. Allowing pain to recur between doses causes unnecessary suffering and may allow tolerance to develop. When prescribing oral opioids, the dosage is about two times that of the subcutaneous dose and three times that of the intravenous dose. For patients unable to take oral medications, suppositories and transdermal patches are good options. When combining drugs, it is important to use drugs that act at different levels of the pain pathway (Fig. 23-1). The combined effect can be additive and at times synergistic.

Figure 23-1 Pain pathway and analgesia interventions. TENS, transcutaneous electrical nerve stimulation.

The pain of bone metastases is generally only partially responsive to opioids.⁶¹ Many osseous metastases produce prostaglandins that induce osteolysis. NSAIDs alleviate pain by inhibiting the synthesis of prostaglandins. Corticosteroids prevent the formation of arachidonic acid (the precursor of prostaglandins) from cell membrane phospholipids. The use of NSAIDs or corticosteroids combined with morphine is usually effective.

Corticosteroids can be used when pain is caused by nerve compression. They decrease edema and reduce the pressure on the nerve. Pain relief can be achieved within 48 hours. Corticosteroids can be used as a temporary measure, before a more definitive decompression is achieved with radiotherapy or surgery.

Side effects associated with the use of opioids include nausea, vomiting, constipation, urinary retention, dysphoria, mental clouding, tolerance, and addiction. Nausea and vomiting are usually self-limiting and resolve during the first week. Side effects should be anticipated, prevented, and managed aggressively.

Surgical Treatment

Surgery should be considered for patients with pathologic fractures or impending fractures (Fig. 23-2). In the former situation, fixation can reduce pain and expedite healing. In the latter, prophylactic fixation may prevent a fracture, thereby eliminating the functional loss and reducing the risk of nonunion of a fracture.

Figure 23-2 Treatment algorithm for bone metastases. HBI, hemibody irradiation; LF, local field; LF + ⁸⁹Sr, local field plus strontium-89; RT, radiation therapy.

To understand the role of surgery better, we need first to elaborate on the biomechanics of pathologic fractures. Cortical defects weaken bone, especially in the setting of torsional stress. The two general categories of cortical defects are (1) stress riser, a defect with dimensions less than the diameter of the bone, and (2) open-section defect, a discontinuity of dimensions greater than the diameter of the bone.⁶² By creating a nonuniform distribution of stresses in bone, stress risers can decrease bone strength by 60% to 70%.⁶³ An open-section defect has a greater impact on decreasing shear and torque-loading resistance. The volume of bone able to resist the load is significantly decreased compared with a closed section. A 90% reduction in load to failure and energy storage to failure is noted in torsion testing of the human adult tibia with open section.⁶⁴ Torsional or rotational forces occur in various daily movements such as getting out of a chair. Bone is weakest during torsion. A single quarter-inch hole made for a bone biopsy can decrease torsional strength by 50%.⁶⁵

The nature of the metastatic lesion affects the overall bone strength. Both lytic and blastic lesions dramatically alter bone elasticity; lytic lesions reduce bone strength more than blastic lesions. Irregular lesions are not necessarily more detrimental to the bone than smooth lesions, but elongated lesions drastically reduce bone strength.⁶⁶

Fidler ⁶⁶ described a technique for estimating the percentage of cortical bone involvement at any level of a long bone using anterior-posterior (AP) and lateral films. This technique is thought to be as accurate a prognostic indicator as cross-sectional CT images, except when cortical destruction progresses in a patchy or spiral form.⁶²

The distribution of pathologic fractures is shown in Table 23-4. Several series have examined various criteria predicting the risk of a pathologic fracture. Keene and associates⁶⁸ evaluated 2673 breast cancer patients in an attempt to predict pathologic fracture of the femur using clinical and radiologic criteria. Only 26 (13%) of 203 patients with evaluable proximal femur metastasis had pathologic fractures. They were unable to correlate lesion size and risk of pathologic fracture. No other risk factor was identified. The authors concluded that plain radiographs are insufficient diagnostic tools for identifying high-risk lesions. Of note is that this study was limited to single AP films.

TABLE 23-4 Distribution of Pathologic Fractures

Location	No.	%
Femur	258	65
Femoral neck	69	17
Peritrochanteric site	50	13
Subtrochanteric site	84	21
Femoral shaft	38	10
Supracondylar site	17	4
Acetabulum	34	8.5
Tibia	31	7.5
Humerus	68	17
Forearm	8	2
Total	399	100

From Mirels H: Metastatic disease in long bones. A proposed scoring system. Clin Orthop Relat Res 249:256, 1989.

Mirels ⁶⁹ designed a score system to predict the risk of a pathologic fracture (Tables 23-5 and 23-6). Of 78 patients, 51 experienced a fracture and 27 did not. The mean score for the nonfracture group was 7 versus 10 for the fracture group. This system provides a useful tool to evaluate patients for prophylactic fixation. Patients with a score of 10 to 12 should undergo surgery. Patients with scores of 7 or less are not likely to benefit from such therapy. In patients with a “gray zone” score, the status of surrounding bone and lifestyle (elderly osteoporotic woman vs. young athlete) should be considered.

TABLE 23-5 Scoring System by Mirels

TABLE 23-6 Pathologic Fracture Rate

Score	No. Patients	Fracture Rate (%)
0-6	11	0
7	19	5
8	12	33
9	7	57
10-12	18	100

From Mirels H: Metastatic disease in long bones. A proposed scoring system. Clin Orthop Relat Res 249:256, 1989.

The following guidelines may help make a decision regarding prophylactic fixation. Because each patient presents with unique circumstances, these guidelines cannot replace sound clinical judgment on the part of the attending physician.

1 Life expectancy is longer than 3 months.

2 The patient is medically fit to tolerate major surgery.

3 The procedure planned is expected to expedite mobilization.

4 The quality of bone both proximal and distal to the lesion is adequate to support any fixation device.

5 There is cortical bone destruction of 50% or more.

6 A lesion measuring 2.5 cm or larger is located in the proximal femur.

7 There is a pathologic avulsion fracture of the lesser trochanter.

8 Stress pain persists after irradiation.

The following principles govern the surgery of impending fractures:

1 Maximum effort is made to avoid disrupting the surrounding soft tissue to preserve the periosteal blood supply. This is of particular importance in these patients because the endosteal circulation has usually been disrupted by the metastatic deposits.

2 Highly vascular lesions (e.g., metastasis from renal cell carcinoma) should be considered for arteriographic evaluation and possible embolization before open curettage.

3 Defects that include the entire circumference of the cortex should be plugged by acrylic cement at fixation to reduce the biomechanical risks associated with stress risers or open-section defects.

4 For large, thin-walled lesions, the intramedullary nailing techniques should be augmented by direct reinforcement of the lesion using methyl methacrylate. This will enhance fixation of the distal long bone, particularly with regard to the torsional stability, and will also prevent shortening of the bone.

Pathologic fractures of the humerus commonly occur in the diaphysis followed by the proximal humerus. Fractures of the diaphysis can be fixed using an intramedullary interlocking device, such as a Brooker-Wills nail, which provides excellent strength and effective resistance against varus, torque, and distraction forces.⁷⁰ Proximal humerus fractures commonly require prosthesis. These patients usually achieve a limited flexion and abduction of about 90 to 100 degrees and enjoy good overall function, joint stability, and pain relief.⁶²

Fixation of the femoral neck-intertrochanteric area can be achieved with the use of a compression hip screw and side plate. Fractures involving the femoral neck may be better treated with prosthetic replacement because they are rarely amenable to internal fixation.⁷¹ Subtrochanteric, femoral shaft, and supracondylar femoral lesions are amenable to internal fixation, but large cortical lesions may benefit from an intramedullary acrylic cement filling.

The common problem encountered in acetabular lesions is the failure to appreciate the extent of bone lysis radiographically. Extensive destruction of bone may render efforts to reinforce such lesions with bone graft fruitless. Pathologic fractures of the acetabulum should be managed by total hip arthroplasty.

The spine is the most common site of skeletal metastases. The vertebral body is typically affected first, although pedicle destruction is noted first radiographically. In the absence of a blastic lesion, 30% to 50% of the vertebral body needs to be destroyed before any destruction can be noted on a radiograph. Vertebral metastases are often asymptomatic. Symptoms are usually a result of one of the following: (1) an enlarging mass within the vertebral body that breaks through the cortex and invades the paravertebral soft tissues, (2) a mass compressing or invading local nerve roots, (3) a pathologic fracture, (4) spinal instability secondary to a pathologic fracture (in particular when the posterior elements are involved), and (5) spinal cord compression.

An aggressive surgical approach to spinal metastases is usually not warranted.⁷² Spinal stabilization is a major operation involving multiple risks and prolonged recovery. Most patients with spinal metastases do not have progressive spinal instability or neurologic involvement and can be treated with irradiation, hormone therapy, chemotherapy, or temporary bracing. Even patients with vertebral body compression fractures can be treated with temporary bed rest and soft bracing. Indications for surgical intervention include (1) progressive spinal canal impingement and cord compression by a radioresistant tumor or a recurrence after the maximum tolerance dose to the intended area has been given; (2) bone or soft tissue detritus extruded into the canal as a result of progressive spinal deformity, with or without spinal instability; (3) progressive spinal deformity; (4) progressive kyphotic deformity associated with posterior disruption and shear deformity; and (5) solitary metastases of a histologic type that is unlikely to be controlled over the long term with tolerable doses of irradiation.

Vertebroplasty of bone metastasis was first described in 1987 and consists of direct injection of the affected vertebra with cement. Polymethylmethacrylate (PMMA) is active through several pathways and produces pain relief in 80% of patients.⁷³ The procedure is done under intravenous sedation or general anesthesia. Pain receptor destruction is achieved with exothermic reaction of the polymonomer and compressive effect on small nerves. Vertebroplasty effects are not modified by external beam radiation therapy, and PMMA conserves its properties despite irradiation.⁷³ Both treatments could become complementary measures in the future.

Radiotherapy

Local-Field Radiotherapy

The vast majority of patients can be managed successfully with external beam radiotherapy (EBRT). A large body of clinical evidence documents the effectiveness of such therapy.⁷⁴ The optimal dose and fractionation schedule are still not resolved. A summary of the major prospective clinical trials that addressed these issues is provided in Table 23-7. The results of these studies should be interpreted with caution because the inherent heterogeneity within the randomization groups may have precluded detection of significant differences, even when such differences could have existed. The use of different pain scoring systems (physician based vs. patient based) and different handling of concomitant use of analgesics, chemotherapy, or hormonal therapy preclude meaningful comparison of the results of these studies.

TABLE 23-7 Summary of Prospective Clinical Trials of Radiotherapy for Painful Osseous Metastases

Between 1974 and 1980, the RTOG conducted a large national study to determine the effectiveness of five different dose fractionation schedules.⁴⁶ A total of 1016 patients were entered, 266 into a solitary metastasis stratum and 750 into a multiple metastasis stratum. The former were randomly assigned to treatment with 40.5 Gy in 15 fractions or 20 Gy in 5 fractions. The latter were assigned to 30 Gy in 10 fractions, 15 Gy in 5 fractions, 20 Gy in 5 fractions, or 25 Gy in 5 fractions. A quantitative measure of pain, based on severity and frequency of pain and the type and frequency of pain medications used, was devised to evaluate response. Overall, 89% of patients experienced minimal relief, whereas 83% achieved partial relief and 54% obtained complete relief. There were no significant differences between the treatment arms in both strata. The initial pain score was found to be a useful predictor; patients with high scores were less likely to respond and less likely to experience a complete response. Patients with breast and prostate cancer were significantly more likely to respond than those with lung or other primary lesions. Patients completing their treatment as planned had a significantly higher rate of complete response than those who did not. Although some pain relief was experienced almost invariably within the first 4 weeks, complete relief was first reported later than 4 weeks after the start of treatment in about 50% of patients. The median duration of minimal and complete pain relief was 20 and 12 weeks, respectively. There were no significant differences in the duration of pain relief between the different arms. The authors concluded that all treatment dose schedules were equally effective.

Blitzer ⁷⁵ performed a reanalysis of the RTOG study. Using a stepwise logistic regression, he examined the effect of the number of fractions, the dose per fraction, and the presence of solitary versus multiple metastases on the probability of attaining complete pain relief and the need for retreatment. This multivariate technique allowed patients with solitary and multiple metastases to be analyzed together. By increasing the number of patients and events, the statistical power of the analysis was increased. The number of fractions was the only variable that was significantly associated with outcome. There was no correlation of the time-dose factor (TDF)⁷⁶ with outcome. It was concluded that the more protracted schedules resulted in improved pain relief.

The concept of TDF has long been replaced by the linear quadratic model. Using this model, and assuming an α/β of 10 for tumor, the biologically effective dose (BED) was calculated for the various schedules tested by the RTOG. Figures 23-3 and 23-4 depict pain response and freedom from retreatment as a function of BED, respectively. The solid lines are the regression functions, and the dotted lines represent the 95% confidence intervals. The results suggest that schedules with a higher BED resulted in better pain relief and reduced the need for retreatment.

Figure 23-3 Pain response as a function of biologically effective dose (BED) by means of linear regression analysis. Response = 39.1 + .38 × BED; r = .74; p = .15. The solid red line indicates the regression function, and the dotted blue lines indicate 95% confidence intervals.

Figure 23-4 Freedom from retreatment as a function of biologically effective dose (BED) by means of linear regression analysis. Response = 69.7 + .4 × BED; r = 0.95; p = .05. The solid red line indicates the regression function, and the dotted blue lines indicate 95% confidence intervals.

Price and colleagues ⁷⁷ randomized 288 patients to receive either 8 Gy in 1 fraction or 30 Gy in 10 daily fractions. Pain was assessed using a daily questionnaire completed by the patient at home. No differences were found in the probability of attaining pain relief, speed of onset, or duration of relief between the two arms.

Hoskin and colleagues ⁷⁸ randomized 270 patients to receive either 4 Gy or 8 Gy in one fraction. Pain (assessed by the patient) and analgesic usage were recorded before treatment and at 2, 4, 8, and 12 weeks. At 4 weeks, the response rates were 69% for 8 Gy and 44% for 4 Gy (p <.001). The duration of the effect was independent of dose.

Additional studies have evaluated single-fraction versus multiple-fraction regimens. A Danish randomized trial of 241 patients showed no significant difference with regard to pain relief or quality of life after receiving either 8 Gy in a single fraction or 20 Gy in 5 fractions.⁷⁹ Wu and colleagues⁸⁰ performed a meta-analysis of 16 trials including 5455 patients and proclaimed equivalence between single and multiple fractions. van der Linden and colleagues⁸¹ recently published a re-analysis of a Dutch Bone Metastasis Study that included 1171 patients. This study randomized patients to either a single dose of 8 Gy or to 24 Gy in 6 fractions. This new publication focuses on retreatment need. Mean time to retreatment was shorter (13 weeks vs. 21 weeks) with a single fraction. It was also more frequent (24% after a single fraction and 6% after 6 fractions; p = .001). An initial high pain score also influenced the need for retreatment.

Treatment Techniques

The target volumes for EBRT should be defined after review of all appropriate diagnostic studies. Attention should be paid to soft tissue masses, which are often associated with bone metastases and at times responsible for the observed symptoms. Such lesions are best assessed by CT or MRI. The target volumes are treated with appropriate margins. Depending on the treatment site and volume, suppression of the bone marrow should be anticipated. In patients for whom chemotherapy is planned, treatment volumes should be kept to a minimum to preserve marrow reserves. Because many patients have repeated courses of therapy, all previous ports and radiation records must be reviewed. To minimize late radiation damage, overlap of radiation fields should be avoided. Depending on the clinical circumstances, overlapping retreatment may be appropriate in patients with a short life expectancy.

Hemibody Irradiation

Most patients with bone metastases have multiple sites of involvement. As many as 76% of patients receiving therapy to a local field require additional treatment for pain at other sites within 1 year.⁸² Historically, wide-field radiotherapy was used to address this problem. Although large-field irradiation is less commonly used in the era of systemic therapy, it remains a treatment option. A summary of results of this form of therapy is provided in Table 23-8. Response rates are similar to those observed with local-field radiotherapy, but the onset of relief is more rapid, occurring often within 24 hours of treatment. Hemibody irradiation (HBI) may require hospitalization for hydration and premedication with steroids and antiemetics and tends to be associated with substantial morbidity. Most patients experience acute gastrointestinal toxicity, with nausea, vomiting, and diarrhea persisting for 24 to 48 hours. Myelosuppression is commonly observed but is rarely of clinical significance. Radiation pneumonitis is rare at doses below 7 Gy to the lungs.

TABLE 23-8 Wide-Field Radiotherapy for Painful Osseous Metastases

The RTOG conducted a phase III study to evaluate the efficacy of HBI in addition to local-field irradiation.⁸² A total of 499 patients were randomized to receive either HBI or no further therapy after completion of local-field irradiation to a symptomatic site. Entry was stratified by extent of metastatic disease (solitary or multiple) and the targeted hemibody area (upper, middle, or lower). Local-field irradiation consisted of 30 Gy in 10 fractions, and HBI consisted of 8 Gy in 1 fraction given within 7 days of completion of the local field. Partial transmission blocks were used to reduce the dose to the lungs to 7 Gy. Time to disease progression, time to new disease, and time to new course of therapy were significantly longer in the HBI arm. Progression of disease was faster in patients with involvement of the upper and middle hemibody (compared with the lower hemibody) and in patients with multiple metastases (compared with a solitary tumor). As expected, toxicity was significantly higher in the HBI arm, but there were no fatalities and no occurrences of radiation pneumonitis. Although an impact of HBI on clinically occult metastatic disease was demonstrated, the long-term benefit was relatively small. The ultimate progression rates were not significantly different between the arms, and at 1 year 60% of the HBI patients had to be retreated.

Stereotactic Radiosurgery of the Spine

Stereotactic radiosurgery (SRS) allows for the delivery of a single high-dose fraction of radiation, using conformal techniques as well as novel technologies to reduce the setup errors and, therefore, improve the targeting accuracy. Generally, the delivery of two to five fractions of radiation, using the same techniques, is referred to as stereotactic body radiation therapy (SBRT). SBRT and SRS offer several advantages, including one to five treatments versus 5 to 20 days of standard EBRT treatment for a patient with significant pain. It also helps reduce the dose delivery to neighboring structures such as the esophagus, lungs, kidney, bowel, and bone marrow of adjacent vertebral bodies. SBRT or SRS can be used as the sole initial treatment for symptomatic spinal metastases, as a boost treatment after the delivery of conventional radiation (generally prescribed to 20 to 40 Gy in 2.5- to 4-Gy fractions), or as salvage treatment in a patient previously irradiated (generally, >3 months prior) for spinal metastases.

A recent review offers a comprehensive analysis of outcomes and toxicity after spinal SBRT and SRS.⁵⁷ Generally, pain control and local tumor control is on the order of 85%.^83,84–88 There appears to be a dose response; greater palliation is seen with prescribed isocenter doses of 14 Gy or more.⁸⁹ Although traditional palliative spine radiation fields encompass the involved vertebral body plus a vertebral body above and below, SRS and SBRT generally treat only the involved vertebral body or a portion of the involved vertebral body. In a study from Henry Ford Hospital (Detroit), the development of subsequent metastases to adjoining vertebral bodies after SRS was rare (~5%) and associated with progression of disease elsewhere.⁸⁵ In a study from the M.D. Anderson Cancer Center (Houston), in-field failures occurred in approximately 25% of recurrences, whereas roughly half of the recurrences were in the epidural space, which was attributed to underdosing of this region to maintain spinal cord dose constraints. Patients also failed in other regions not included in the treatment volume (of any given patient), including the pedicles, posterior elements, and prevertebral and paravertebral regions.⁸³

Recent clinical data have shown that SRS and SBRT treatment of the spine is tolerable, albeit with limited patient follow-up ⁹⁰ because patients with spinal metastases generally have a poor survival rate, even those with a solitary spinal metastasis.⁹¹ It appears that myelopathy and radiculopathy rarely occur.^57,90 For single-fraction SRS, most institutions try to achieve a spinal cord maximum dose below 10 Gy.⁵⁷ In the University of Pittsburgh experience of spinal SRS in 393 patients with 500 lesions, for which the cord maximum was kept below 8 Gy, no acute or late neurotoxicity was observed, and no late toxicity was reported after a follow-up of 3 to 53 months (median, 21 months).⁸⁴ Several institutions have demonstrated that spinal cord maximum doses of 12 Gy⁸⁶ to 20 Gy^90,⁹² are tolerated in some patients, though from a recent multi-institutional pooled analysis, radiation myelopathy has only been documented to occur after a fractional dose maximum of 10 Gy to the spinal cord has been exceeded.⁹³ In the pooled-analysis study, dose-volume parameters such as maximal dose and mean and median dose to 0.1 to 5 mL of the spinal cord significantly correlated with the risk of radiation myelopathy. In the ongoing RTOG 0631 study (which is a randomization of spinal SRS delivered with 16 Gy versus conventional irradiation delivered in one fraction of 8 Gy), spinal cord dose constraints are 10% and 0.35 mL of spinal cord less than 10 Gy, and 0.035 mL less than 14 Gy, with the spinal cord volume defined as 5 to 6 mm above and below the target, based upon T2- and T1-weighted MRI.

Systemic Radionuclide Therapy

The first report on the use of systemic radionuclides for the treatment of bone metastases was published by Pecher ⁹⁴ more than 50 years ago. Using this modality, all involved osseous sites can be addressed simultaneously. Selective absorption into bone metastases limits irradiation of normal tissues and increases the therapeutic ratio. Administration as a single intravenous injection in the outpatient clinic is a further advantage for many patients.

Systemic radionuclides should be considered in the following circumstances:

1 In patients with widely metastatic disease, as an adjuvant to EBRT

2 In patients without a predominantly painful site, as a first-line therapy

3 When EBRT options have been exhausted and normal-tissue tolerance has been reached

4 When there is no evidence of imminent epidural cord compression, pathologic fracture, or mechanical instability

5 In patients with good marrow reserve with a white blood cell count of greater than 2400 cells/µL and a platelet count of more than 100,000 cells/µL

Historically, phosphorus-32 (³²P) was the first radionuclide to be widely used in the treatment of bone metastases,⁹⁵ although this isotope is now rarely used for bone disease due in part to myelosuppression with pancytopenia and an increased incidence of acute leukemia.⁸² Phosphorus-32 has since been replaced by newer, less toxic radionuclides (Tables 23-9 and 23-10).

TABLE 23-9 Physical Characteristics of Various Radionuclides