Prostate Cancer and Spinal Cord Compression
Prostate Cancer and Spinal Cord Compression
ABSTRACT: Prostate cancer metastasis to the spine is an extremely difficult clinical problem to treat. However, it occurs commonly, and all clinicians—not only oncologists—should undertake to understand its pathogenesis, diagnosis, clinical presentation, and current treatment options. This review emphasizes the surgical treatment of prostate cancer metastasis to the spine. The goals of this article are to (1) present an overview of the pathophysiology of this disease, with an emphasis on the mechanisms of metastasis and invasion, (2) provide a general overview of the clinical presentation and diagnosis of metastatic prostate carcinoma, and (3) discuss currently available treatment options. Such options include best medical management, nonsurgical treatments (radiation, chemotherapy), and surgical treatment of newly diagnosed and previously irradiated metastatic prostate carcinoma to the spine. Algorithms for the treatment of this disease are presented. [ONCOLOGY 15(7):841-861, 2001]
It is estimated that approximately 198,000 US men will be diagnosed with prostate cancer this year. Prostate cancer is the second leading cause of cancer death in men in the United States. The morbidity and mortality associated with prostate cancer can often be attributed to the consequences of bone metastases. The most common site of bone metastasis in prostate cancer patients is the spine, followed by the femur, pelvis, ribs, sternum, skull, and humerus. As a result, prostate cancer is second only to lung cancer as a cause of metastatic spinal cord compression in men. Symptomatic lumbar and cervical epidural metastases develop in 27% and 6% of prostate patients, respectively.[5,6]
Yamashita and coworkers found that, among responders to androgen deprivation, the absence of bone metastasis outside the pelvis and the lumbar spine was predictive of a longer survival. Because of the frequency of spinal cord compromise secondary to prostate carcinoma, the importance of early diagnosis and treatment of patients with spinal metastasis cannot be overemphasized.
Biology of Prostate Cancer Metastasis
The most common site of metastasis is the lumbar spine. Autopsy data reveal that spinal metastases precede lung and liver metastases in many patients with prostate cancer. Batson, in 1940, proposed that prostate carcinoma cells reach the lumbar vertebrae via the vertebral venous plexus—a network of longitudinal, valveless veins running parallel to the vertebral column, which comprises countless anastomoses to the sinusoidal structure of the vertebral marrow and epidural venous channels (now called Batson’s plexus, Figure 1). Under transient conditions of increased intra-abdominal pressure, the prostate cancer cells may reach the axial skeleton directly by retrograde hematogenous spread, without passing through the lungs. The cancer cells then invade through the sinusoidal endothelial cells of Batson’s plexus into the marrow space of the vertebral body; in the bone marrow, the prostate cancer cells are stimulated to proliferate.
Batson’s Plexus and the Metastatic Model
However, for this metastatic model to be true, several issues need to be addressed: (1) how valid is Batson’s plexus as a model for spine metastasis? (2) how do prostate carcinoma cells "home" into the sinusoidal endothelial cells of the vertebral body? (3) what is the invasion process that prostate carcinoma cells employ in order to get into the host bone marrow? (4) what is special about the vertebral body microenvironment to favor prostate cancer metastasis? (5) Why do prostate carcinoma cells induce an osteoblastic instead of an osteolytic response?
Batson originally injected the cadaveric dorsal vein of the penis with radio-opaque material and demonstrated the connection with the prostatic plexus and thereafter the pelvic vein, pelvic bones, and sacral canal, leading him to hypothesize the venous route of prostate metastasis. Subsequently, other investigators, working with animal models, validated Batson’s plexus as the preferred route of metastasis. Coman and DeLong were able to test this hypothesis directly using an animal model. In their experiment, Walker rat 256 carcinoma cells were injected into the femoral vein of rats while intra-abdominal pressure was exerted. The majority of animals developed vertebral metastasis while control animals (without increased abdominal pressure) only developed tumors in the lungs.
Other Routes of Metastatic Spread
However, other investigators have questioned whether prostate metastasis occurs via Batson’s plexus. Dodds et al in a review of various positive bone scintigrams from different cancers found little difference in overall distribution of the various cancers and prostate carcinoma, leading them to conclude that a systemic route of metastasis was the preferred method of spread for all cancers.
More recently, Nishijima et al reexamined the issue of bone metastasis by confining their examination of bone scintigrams to lung and prostate carcinoma patients with "early-stage" bony involvement (no more than two bony lesions). They concluded that there was a preponderance of metastases to the spine and pelvis in prostate cancer patients. Patients with "late-stage" bone involvement (more than three bony lesions) were indistinguishable in their metastatic distribution pattern. Moreover, a number of investigators have been able to reproduce metastatic tumor growth in the lumbar spine by injecting cancer cells into the tail vein of rats with temporary vena caval occlusion.[12,13]
If Batson’s plexus indeed offers a conduit for prostate carcinoma cells to travel up to the vertebral body, it does not explain how the prostate cancer cell is able to "home" into the sinusoidal endothelial cells of the vertebral body. Haq et al recently demonstrated that bone marrow-derived endothelial cells express adhesion ligands for prostatic cancer cells that are not expressed on either hepatic endothelial cells or nonendothelial cells of the marrow. The prostate cancer cells then cross the leaky endothelial cell barrier into the interstices of the marrow. Here, the cancer cells are surrounded and nurtured by the marrow.
Wu et al examined the interaction of human prostate cancer epithelial cells with bone stromal cells, and suggested that the bone stromal cells play a protective role in the development of metastatic cancer cells, inducing androgen independence in the prostate carcinoma cells. Moreover, Chackal-Roy et al have shown that marrow-conditioned medium is mitogenic for prostate carcinoma cells, suggesting that mitogenic factors produced by the marrow stromal cells may also account for the preferential growth of prostatic metastasis in bone.
Growth Factors That Enhance Metastatic Potential
In addition to the favorable host microenvironment of the vertebral body stroma, prostate carcinoma cells also secrete a variety of growth factors and proteases that enhance their metastatic potential and growth. In order to enhance their invasiveness, urokinase plasminogen activator may be secreted by prostate carcinoma cells. This very important protease converts the inactive zymogen plasminogen into the active serine protease plasmin, allowing for extravasation of cancer cells and breakdown of the skeletal matrix. Experimental evidence indicates that urokinase plasminogen activation is found at higher levels in highly aggressive prostate cancers as opposed to more well-differentiated lesions, and hyperplastic or normal tissues.
Urokinase plasminogen activator levels are also higher in metastatic prostate carcinoma vs the nonmetastatic variety. Chen et al have shown that rat adenocarcinoma PA III cells secrete bone morphogenetic protein 3 (BMP3)—a family of bone growth factors that have been linked to the transforming growth factor-beta family. The secretion of bone morphogenetic protein by prostate carcinoma cells may explain the "blastic" nature of prostate metastasis. In turn, the blastic response of prostate carcinoma may explain why new bone formation is found around the tumor cell deposits, often without prior osteoclastic resorption.
Ultimate Effects on Spinal Cord
Eventually, compression by the epidural tumor or bone fragment will result in spinal cord compression, leading to venous obstruction and development of vasogenic edema. At this stage, administration of dexamethasone would be extremely beneficial in decreasing cord edema and resolving many of the patient’s initial symptoms. Continued vasogenic edema, however, will lead to an objective decrease in the somatosensory-evoked potential, and then to a conduction block across the area of compression. This, in turn, will result in cord demyelination, decreased blood flow, and ischemia.
Ischemia induces a different type of edema (cytotoxic) with resultant spinal cord infarction. Figure 2 illustrates the sequence of events that occur during spinal cord compromise. There is a window of opportunity between the development of neurologic symptoms and complete loss of neurologic function that may range from a period of days to weeks. During this period, treatment options such as radiation, chemotherapy, and surgery may be used to reverse this process. It is the goal of this article to help the treating physician recognize the initial clinical presentation, order the appropriate diagnostic tests, and then propose a treatment plan for the patient with metastatic prostate carcinoma to the spine.
Metastasis to the spine initially results in pain that may subsequently progress to neurologic signs and symptoms if there is spinal cord or nerve root compromise. The initial consultation should consist of a well-taken history with evaluation of the patient’s symptoms, followed by a careful neurologic examination. The history is an important aspect of the evaluation. The timing of the onset of symptoms is crucial to an accurate determination of the acuity of the situation, and the patient will often be able to specify the exact time and location of onset of symptoms.
Osborn et al reviewed four large series of prostate metastases to the spine and concluded that patients had four main initial presentations: pain, weakness, autonomic dysfunction, and sensory loss. In most cases, pain is the initial presentation of spinal metastasis. A cancer patient with new onset of neck or back pain should be considered to have spinal metastasis until it is specifically ruled out. In prostate carcinoma patients, the lumbar spine is the most common site of initial metastasis. As a result, patients with lower back pain, but with a known history of prostate carcinoma, must be evaluated carefully for lumbar or sacral metastases.
Pain was the initial presentation in 75% to 100% of the patients reviewed by Osborn et al. The pain tends to be localized to the site of metastasis, and is usually secondary to periosteal stretching as the vertebral mass enlarges. However, if the tumor causes instability, it is important to determine whether the pain is mechanical in nature. It is extremely important when taking the history to ask the patient if the pain changes with position. For example, patients who receive high doses of narcotics to manage their pain will say that their pain is under control when they are lying in bed; however, if they attempt to sit up or stand, the pain becomes so intense that they feel as if they may "pass out." In those cases, a plain x-ray will often reveal extensive bony destruction and spinal compromise.
Obtaining information on other exacerbating factors besides position—such as laughing, coughing, sneezing, straining, or lifting—is important. It is also important to determine whether there are remitting factors for the pain, such as lying down or bending forward.
In addition, the patient’s motor strength needs to be determined. A patient will often notice subtle differences in strength that are not detectable on examination. For instance, a patient with a cervical metastasis and root involvement may notice some weakness while holding a cup, which may not have been apparent on initial neurologic assessment of strength. There may be changes in posture, gait, or balance that the physician did not notice when examining the patient in a seated or prone position.
Questions about sensory changes (ie, numbness, temperature sensitivity) should accompany those dealing with strength. Nerve root involvement may produce sensory changes in a dermatomal fashion and should be correlated with changes in associated motor strength noted by the patient. Finally, because prostate metastasis often involves the lumbar and sacral region, inquiries regarding autonomic dysfunction such as bowel and bladder compromise are extremely important. The latter is absolutely crucial because patients often do not associate changes in bowel/bladder function with problems noted in strength or sensation.
Physical and Neurologic Aspects
The initial examination should be divided into a physical and a neurologic exam. Much of the patient’s physical exam can be completed at the first meeting. The overall appearance of the patient provides clues to his general health status. The patient who is able to walk into the room without significant pain or weakness is probably not going to need a surgical procedure. A patient who is cachectic and appears ill may not be able to undergo a major operation. The examination must include an evaluation of posture, stance, and gait. This may be accomplished, in part, as the patient enters the examination room. It is important to examine the length of the spine visually, note any abnormal curvatures, and have the patient point to the location of his discomfort.
The neurologic examination consists of an evaluation of the patient’s general mental status, cranial nerve evaluation, motor examination, sensory examination, reflexes, and cerebellar function. In patients with nerve root compression, the neurologic exam may be specific for the dermatomal and myotomal distribution of the particular root involved, allowing for localization of the lesion to a particular vertebral segment. Patients with spinal cord compression may be more nonspecific, having complaints of bilateral weakness, sensory loss, increased deep tendon reflexes, and evidence of autonomic compromise such as bowel and bladder dysfunction.