Spine Pain and Metastatic Prostate Cancer: Defining the Contribution of Nonmalignant Etiologies

Abstract

PURPOSE:

In patients with metastatic prostate cancer (MPC), the contribution of nonmalignant etiologies to morbidity is often overlooked.

METHODS:

We retrospectively reviewed the documented specialist assessments of back pain in men with MPC in a joint medical oncology and physiatry clinic at our tertiary cancer care center. Data on cancer disease extent, hormonal status, sites of spread, pain characteristics, physiatric examination findings, imaging, and recommended management were reviewed, extracted, and codified. For those with back pain at a site of known disease, pain etiology was classified as malignant, nonmalignant, or mixed.

RESULTS:

Ninety-three men were collaboratively assessed for back pain, 24 (26%) with a biochemical recurrence and 69 (74%) with MPC of whom 53 (77%) reported pain in an area of known spinal metastases including 35 (66%) metastatic castration-resistant disease and 34 (64%) a precancer history of back pain. The presenting pain symptoms of the 53 patients were activity-related in 22 (42%), radicular in eight (15%), transitional movement–related in seven (13%), biologic in five (9%), and multifactorial in 11 (21%). Overall, pain was deemed malignant in 20 (38%; five castration-sensitive, 15 metastatic castration resistant prostate cancer), nonmalignant in 12 (23%; four castration-sensitive, eight CRPC), and of mixed etiology in 21 (40%; nine castration-sensitive, 12 CRPC).

CONCLUSION:

Nonmalignant etiologies contributed significantly to back pain at sites of metastatic spread for 33/53 (62%) patients with MPC assessed by medical oncology and physiatry. We recommend multidisciplinary care for patients with MPC and back pain to address nonmalignant etiologies that contribute to functional compromise.

INTRODUCTION

Back pain is one of the most common complaints managed by rehabilitation medicine physicians (physiatrists) but is rarely a harbinger of serious medical illness.^1,2 By contrast, bone, the spine in particular, is the most common site of metastatic prostate cancer (MPC) spread (90%) and recognized to produce some of the most feared and debilitating complications of the disease.^3-5 The result is that the evaluation and treatment of back pain and symptomatic skeletal-related events (SSREs) in men with metastatic disease may assume the etiology to be tumor-related with a lesser consideration to benign etiologies.^5,6

Sources of malignant pain in the context of skeletal metastases from prostate cancer are multiple, including the tumor itself, hypercalcemia, pathologic fractures, and neurologic deficits from tumor compression. Vertebral collapse secondary to bone loss because of the hypogonadal state that results from the hormone therapies used to treat the disease may also contribute. All have the potential to impair quality of life substantially.^4,7-11 Accordingly, patients, clinicians, and regulators recognize that reducing and preventing pain and SSREs are important therapeutic objectives for which clinically validated clinical trial end points have been established to support regulatory approvals for a drug indication.^12,13 Cancer-directed therapies to palliate pain in addition to analgesics include taxane-based chemotherapy, androgen receptor signaling inhibitors, bone-seeking radionuclides, glucocorticoids, and focal treatments such as radiation or surgery.^14-16 Zoledronic acid and denosumab are approved to reduce bone loss and the incidence of SSREs.¹⁷

Underconsidered, underappreciated, and at times incompletely addressed in oncologic practice and in the eligibility criteria for clinical trials are the nonmalignant causes of back pain that are frequently present in the age group at which prostate cancers are typically diagnosed.^18,19

Failure to accurately attribute causality of pain and/or neurologic symptoms may lead to overtreatment, exposure to unnecessary side effects from cancer-directed therapy and/or analgesics, and a failure to ameliorate symptoms, leading to further functional compromise. Here, we report the experience of a focused multidisciplinary outpatient clinic designed in January 2013 to improve the diagnosis, management, and treatment of back pain in men with prostate cancers who have known spinal metastatic disease. The care providers include medical oncologists with a prostate cancer focus and a physiatrist with clinical expertise in spinal disorders, including neurologic and musculoskeletal impairments, and in reducing pain using a patient-specific management plan focused on maximizing function.

METHODS

Patient Selection

Following approval from the Memorial Sloan Kettering Cancer Center (MSK) Institutional Review Board, the institutional database was queried to create a registry of patients evaluated at the joint medical oncology-physiatry clinic between January 2013 and December 2017. We included patients with a histologically confirmed diagnosis of MPC under the care of a prostate cancer–focused medical oncologist who were referred to the Rehabilitation Medicine Service for a physiatry assessment of back pain in the context of known metastases to the spine or pelvis. We excluded patients with localized prostate cancer or biochemical recurrence. Similarly excluded were men with back pain outside a region of known metastases (Fig 1).

FIG 1.

Men with prostate cancer evaluated for back pain in a multidisciplinary GU medical oncology/physiatry clinic. ARSi, androgen receptor signaling inhibitor; BCR, biochemical recurrence; CRPC, castration-resistant prostate cancer; CSPC, castration-sensitive prostate cancer; mCRPC, metastatic CRPC; mCSPC, metastatic CSPC.

Assessments

For all eligible patients, demographic data were collected from documented medical oncology assessments, including details on disease state,¹² site(s) of disease, presence or absence of bone and/or spinal metastases, disease and hormonal status (rising prostate-specific antigen with testosterone < 50 ng/dl), and prior cancer treatments at time of referral. Also captured were precancer history and current presence of non–cancer-related conditions that might have contributed to back pain, including a radiculopathy, degenerative disease, and prior spinal surgeries.

We extracted from the initial physiatry assessment (1) pain descriptors, including the quality, severity, duration, precipitating factors, location, and aggravating and alleviating factors; (2) associated symptoms; and (3) the timing of pain onset in relationship to current disease status, changes in oncologic therapies, medications, and interventions received. These descriptors were then used to categorize pain in one of four categories: activity-related, biologic, mechanical or transitional movement–related, and radicular. Our four pain classifications were chosen on the basis of the literature for nonspecific back pain (activity-related, mechanical, and radicular) and the three classic pain types described in spinal metastases (biologic, mechanical, and radicular).^15,20

Details of the documented physiatry examination were analyzed for binary categorizations on the presence or absence of tenderness to palpation of the spinal column and musculature, reduced spinal range of motion (ROM), muscle strength and sensation, and altered gait patterns. Classifications related to posture assessment were defined by the clinical observation of normal spinal alignment; presence of exaggerated or decreased cervical and lumbar lordosis or thoracic kyphosis; and abnormal head and shoulder positioning.

Spinal imaging completed before the joint assessment was reviewed to evaluate known metastatic lesions in areas of pain, along with the presence, location, and extent of degenerative spine changes. A physiatric diagnosis was then made on the basis of clinical findings contemporaneous to the initial assessment and imaging findings (Fig 2). On the basis of review of these records, a physiatrist retrospectively categorized the pain etiology as nonmalignant, malignant, or mixed (both cancer-related and non–cancer-related). Details of the initial management recommendations made at physiatry assessment were also recorded.

FIG 2.

(A) Nonmalignant pain etiologies: L5 radiculopathy in the setting of metastatic castration resistant disease progression. A 75-year-old with history of lumbar spine and pelvic metastases and chronic low back pain presents with escalating lower back and right calf pain, a rising PSA, and new S1 metastases. Physiatric examination recorded head forward posture, flattening of lumbar lordosis, Trendelenburg gait, and a severely impaired tandem gait. Power was 2/5 in the right long toe extensor and 4/5 in the right ankle dorsiflexor, with loss of balance when standing on the right leg and impaired proprioception at the right great toe. Examination findings were consistent with an L5 radiculopathy. An MRI spine confirmed right paracentral disc protrusion impinging the descending nerve roots at L4-5 and moderate facet arthrosis left greater than right at L4-5 with surrounding edema. The S1 metastasis sat anterior in the vertebrae. The patient was referred to physical therapy for pain management and neurosurgery for consideration of nerve decompression. (B) Mixed malignant and nonmalignant pain etiologies: progressive metastatic disease, kyphosis, and worsening scoliosis because of altered body mechanics. A 68-year-old man presents with worsening lumbar back pain and occasional upper thoracic pain with intermittent lower-extremity numbness and tingling. He has scoliosis, metastatic prostate cancer to bone, and prior palliative radiation to L3, complicated by compression fracture managed with L3 laminectomy and vertebroplasty. On examination, he had kyphotic posture, tenderness to the palpation of the thoracic spine, restricted lumbar spine motion, and paresthesias in an L3 distribution. MRI showed a new T5 metastatic lesion, stable compression fracture, and disc disease below his surgical intervention. MRI, magnetic resonance imaging; PSA, prostate-specific antigen.

Collected data were summarized by descriptive statistics and are presented for the cohort of patients with metastatic disease and back pain in a region of known cancer spread, and categorized by the assigned etiology of pain.

RESULTS

Our database search yielded 93 study-eligible men with histologically confirmed prostate cancer who were undergoing care in the Genitourinary Oncology clinic at MSK and referred for a physiatry back pain evaluation between January 2013 and December 2017. Twenty-four (26%) had localized or biochemically recurrent disease and were excluded. The remaining 69 patients had metastatic disease on imaging (23 with castration-sensitive and 46 with castration-resistant prostate cancer [CRPC]). Subsequently, 16 additional patients with disease outside of the spine or pelvis, or pain outside the sites of spread were excluded (Fig 1). The final analysis focused on the 53 men with back pain in an area of metastatic disease (Table 1).

TABLE 1.

Men With Spinal Metastases Who Presented With Pain in Region of Spinal Metastases

The median age was 72 years (range, 52-90 years). Eighteen (34%) had metastatic castration-sensitive prostate cancer (mCSPC) and 35 (66%) had metastatic castration resistant prostate cancer (mCRPC). Of these 53 men, the joint assessment of the physiatrist and medical oncologist determined 20 (38%) men to have pain because of the malignancy, 12 (23%) with a likely nonmalignant source, and 21 (39%) with pain of mixed malignant and nonmalignant etiology. Overall, 34 (64%) men reported back pain predating their cancer diagnosis, for which 14 had undergone surgical intervention. Men with nonmalignant and malignant pain were equally likely to have had a surgical intervention for pain before the physiatry evaluation (4/8 [50%] and 7/14 [50%], respectively).

Presenting pain symptoms were categorized as activity-related in 22/53 (42%), radicular in eight (15%), mechanical in seven (13%), and biologic in five (9%). Pain of nonmalignant and mixed etiology was more likely to be activity-related. Notable was that more than one pain type was reported in 11 (21%) patients. Three of the four assigned pain subgroups (radicular, mechanical, and biologic) had similar rates of precancer back pain. Forty-four (83%) men required analgesics for the management of their current presenting pain symptoms.

A single physiatrist completed all assessments and documentation relatively uniformly; there were no missing data for examination fields. Abnormal posture and reduced spinal ROM occurred with the highest frequencies (Table 1). In addition to spinal examination abnormalities, our cohort had significant rates of gait abnormalities (27/53; 51%) and weakness (16/53; 30%), both of which are known to contribute to falls.

Available imaging for review was completed within a median of 1 month before physiatry visit (range, 1 day-7 months [two cases]). Degenerative changes were noted on imaging in 32 (60%) patients including disc disease in 13 (41%), facet arthropathy in one (3%), stenosis in two (6%), multiple types in 11 (34%), and not otherwise specified in 5 (16%). Additional imaging with magnetic resonance imaging and/or fluorodeoxyglucose positron emission tomography was ordered for 12 of the 53 (23%) patients to further assess disease extent and determine the need to change or initiate a cancer-directed therapy.

Following discussion between the medical oncologist and physiatrist, oncologic management with focal radiation to a single site was recommended for six (11%) patients and radium-223 to two (4%) patients (data not shown). Neurosurgical and interventional radiology referrals were each recommended in two (4%) patients.

Therapeutically, physical therapy (PT) was recommended to 36/53 (68%) cases and postural bracing in addition to PT for 17/53 (32%). Additional management recommendations were discussed with 8/53 (15%) patients, which included topical anti-inflammatory gel, lidocaine patches, injections, nerve blocks, acupuncture/massage therapy, and glucocorticoids. No new narcotics were initiated for pain control. Management plans were individualized for each patient and often consisted of more than one recommendation.

Men With Malignant Pain

Twenty men, with a median age of 69 years, were determined to have back pain related to cancer. Five (25%) men had mCSPC and 15 (75%) had mCRPC. Fourteen (70%) reported a precancer history of back pain, with seven (50%) having required surgical interventions such as laminectomy, discectomy, and fusion for this pain.

Pain types classically related to malignancy were present in 11/20 (55%) of this group including radicular pain in five (25%), mechanical in three (15%), and biologic in three (15%). Interestingly, activity-related pain, not classic for malignancy, was seen in five (25%) men and multiple pain types were seen in four (20%) men. This cohort had not only the highest analgesic use overall (19/20; 95%), but the highest narcotic use (15/20; 75%). Abnormal posture and reduced spinal ROM were common in this group, similar to the others; however, these men were more likely to have gait abnormalities (13/20; 65%) and weakness (10/20; 50%).

On review of imaging, 10 (50%) had metastatic disease at the site of their pain and the other 10 (50%) had metastatic disease with a pathologic vertebral compression fracture. Degenerative changes were found on imaging in 12/20 (60%) patients but were designated as likely incidental findings and unlikely to be the source of pain in the clinical context.

Joint review of imaging and laboratory values supported clinical findings of malignant pain as the most likely source of pain in this group. Additional imaging was recommended for 5/20 (25%) patients to assess for progressive disease and in consideration of the need for change in the cancer-directed therapy. Palliative radiation to the metastatic pain source was ultimately recommended for 5/20 (25%) patients, radium-223 to 1/20 (5%), and neurosurgical evaluation to 2/20 (10%). In an effort to offload pressure on the spinal column, to improve strength, posture, and ROM, and to reduce fall risk, PT was recommended to 11/20 (55%) and bracing to 8/20 (40%).

Men With Mixed Pain Because of Both Cancer-Related and Non–Cancer-Related Causes

Twenty-one men, with a median age of 69 years, were categorized as having a mixed source of pain attributable to both cancer and noncancer sources. Nine (43%) men had castration-sensitive prostate cancer and 12 (57%) had CRPC.

Twelve (57%) men reported a precancer history of back pain, with 3/12 (25%) having required surgical intervention. Pain in the mixed-pain group was most commonly activity-related (10/21; 48%), although reports of multiple pain types were also common (5/21; 24%). In this group, 16/21 (76%) required analgesics to treat their active pain, with nonsteroidal anti-inflammatory drugs being most common (10/21; 48%). Narcotics and Tylenol were used with similar frequencies (6/21 [29%] and 5/21 [24%], respectively). Abnormal posture and reduced spinal ROM were the most common physiatric abnormalities, both of which were documented in 14/21 (67%). Imaging was completed within a median of 0.59 months (range, 4 days-6.87 months) before physiatry evaluation, and degenerative changes were found in 11/21 (52%) patients.

In the mixed-pain group (n = 21), oncologic management with radiation to a metastatic site was recommended to one patient (5%) and radium-223 treatment recommended to another patient (5%). Fifteen patients in this group (71%) were referred for PT and 5/21 (24%) were referred for bracing. Additional imaging with magnetic resonance imaging and/or positron emission tomography scan was recommended to 6/21 (29%) patients to better characterize the extent of disease and possible concurrent progression. Other interventions for pain management were discussed with 8/21 (38%) patients, which included pharmacologic management with topical anti-inflammatory gel, lidocaine patches and a steroid dose pack, injections including nerve blocks, kyphoplasty, and acupuncture/massage therapy.

Men With Nonmalignant Pain

This group comprised 12 men, with a higher median age than men with malignant or mixed pain. Four (33%) of the men had mCSPC and eight (67%) had mCRPC. Eight (67%) reported a precancer history of back pain, which had required surgical intervention in four (50%). Pain was most often categorized as activity-related (7/12; 58%). No radicular or biologic pain was found in this group. Nine (75%) men required analgesics for control of presenting pain symptoms. It is interesting to note that narcotic and nonsteroidal anti-inflammatory drug use occurred with equal frequency (both 5/12; 42%; Table 1). Decreased spinal ROM was the most common physiatric examination abnormality (9/12; 75%) in this group. Imaging was completed within a median of 1.3 months (range, 1 day-6.85 months) of physiatry evaluation, and degenerative changes were found in 9/12 (75%) patients.

Causes of nonmalignant pain in this group included lumbar stenosis, lumbar degenerative disc disease, and cervical spondylosis. No further imaging investigations were ordered for these patients following physiatry consultation. PT was recommended to 100% of these patients, and bracing was recommended to 4/12 (33%) for symptom management.

DISCUSSION

The morbidity of bone metastases is one of the most feared complications of MPC.²¹ Professional guidelines list the incidence of nonmalignant pain in prostate cancer to be as low as 3%, which can be misleading.²² Our findings challenge this view, showing that in a cohort of 53 men with progressing mCRPC in an area of pain in the spine, the malignant etiology fully accounted for the pain in only 38%. Features of patients' clinical history that suggested a noncancer etiology include the description of the pain as activity-related, found in 42% of cases but often associated with degenerative spine disease.^19,23 This was surprising, given that biologic pain, thought to be the result of periosteal stretching and inflammation from tumor growth,²⁴ was expected to be more frequent, yet was only reported in 9% (5/53) of the patients evaluated: eight (44%) and one (6%) in the 18 castration-sensitive and 15 (43%) and four (11%) in the 35 castration-resistant population for biologic versus activity-related, respectively.

Of particular importance was our finding of an abnormal posture observed in 72% and decreased spinal ROM found in 68%, which raised concerns about spinal kinematics and the competence of the vertebral bodies and intervertebral discs, and provided clues to nonmalignant etiologies or contributions to pain, fitness reduction, and functional compromise a patient was experiencing. From the radiologic point of view, degenerative spinal changes were noted on imaging in 60% of the cohort, which must be interpreted cautiously because the specific tests ordered were primarily intended to inform cancer-related pain, and similar changes are often found in asymptomatic individuals.²⁵ Similarly, lumbar spine imaging for low back pain is considered low value outside the context of a serious underlying diagnosis in health care as it does not improve clinical outcomes. The percentage of degenerative changes identified in asymptomatic individuals drives this thought process.^26,27 Our study supports imaging in evaluating back pain in the setting of MPC and highlights the importance of requesting degenerative change assessment in addition to the presence of degree of metastatic spread as it may affect imaging study protocols and radiologist interpretation.

Our findings support the importance of a shared model of care to ensure a broader evaluation and more directed management for back pain–specific symptoms experienced by patients with prostate cancer with metastatic disease and endorse the potential value of including a dedicated physiatry evaluation.^28,29 In this context, pain relief is critical to enable patient functionality and improve their overall outlook. Reports show that oncologists following analgesic guidelines will achieve pain control in up to 90% of patients. Rarely considered, however, is that referenced guidelines here are specific to cancer-related pain only.^30,31 Beyond applying spinal neurorehabilitation to malignant spinal cord compromise, the recommendations of dedicated physiatrists can add significantly to patient management by identifying and tailoring therapeutic recommendations to the nonmalignant etiologies of spinal pain.

Nonmalignant sources of pain in this study included, as expected in this population, spinal stenosis, facet joint arthropathy, sacroiliac dysfunction, and myofascial pain.^19,23,32-34 Additionally, in a group of men with a median age of 72 years, it is important to consider the impact of normal physiologic aging, which causes a loss of axial bone mass, extensor muscle strength, and resilience of the intervertebral discs that independently produce pain. Men with prostate cancer have an even greater risk for changes in bone density, vertebral compression fractures, and muscle strength as a result of their cancer and its treatment.^4,9-12,35

For all study patients, regardless of pain etiology, physiatry-prescribed rehabilitation efforts and bracing recommendations focused on pain prevention, along with preservation and improvement of functional status.^18,19,23 Programs were designed specifically around core muscle engagement, strengthening of spine extensor muscles, improving proprioception around joints, and correcting posture to offload the spinal column structures, most of which are not improved by medications alone and often obviate the need for invasive surgical procedures. Adherence results in a reduction of the risk of falls and pain, while promoting physical activity and exercise, reducing fatigue and the side effects related to androgen deprivation therapy, and reducing the dependence of the patient on others to carry out the activities of daily living, all of which can result in a marked improvement in a patient's outlook.^36,37

Limitations of the study include the relatively small sample size, and retrospective reporting of the experience of a single-practice pilot model of shared care, with the potential for an unrecognized referral bias. Further validation of our findings will require prospective trials that include patient-reported measures and standardized clinical assessments.

In conclusion, this retrospective review highlights the prevalence of non–cancer-related spine pain in men with MPC, which will not respond to tumor-directed therapies, but rather to rehabilitation-driven efforts guided by a shared care model.

AUTHOR CONTRIBUTIONS

Conception and design: Lisa Marie Ruppert, Erica Dayan Cohn, Howard I. Scher

Financial support: Howard I. Scher

Administrative support: Theresa Gillis, Howard I. Scher

Provision of study materials or patients: Howard I. Scher

Collection and assembly of data: Lisa Marie Ruppert, Erica Dayan Cohn, Abigail Bacharach

Data analysis and interpretation: Lisa Marie Ruppert, Erica Dayan Cohn, Niamh M. Keegan, Sungmin Woo, Theresa Gillis, Howard I. Scher

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Spine Pain and Metastatic Prostate Cancer: Defining the Contribution of Nonmalignant Etiologies

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).