 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1999 The Journal of Bone and Joint Surgery, Inc. Delay in the Diagnosis and Treatment of Primary Bone Sarcoma of the Pelvis*  , CHICAGO, ILLINOISInvestigation performed at the Section of Orthopaedic Surgery and Rehabilitation Medicine, Department of Surgery, University of Chicago Medical Center, Chicago
Background: Symptoms arising from primary bone sarcoma of the pelvic girdle are often insidious in onset and nonspecific in nature. To make the subtle initial signs and symptoms of these tumors more apparent to clinicians, we studied a cohort of patients who had a primary bone sarcoma of the pelvic girdle. Our purpose was to describe the initial clinical findings and to evaluate the duration, frequency, and implications of delays in the treatment of these tumors. Methods: We retrospectively analyzed the data on sixty-eight patients who had a primary bone sarcoma of the pelvic girdle. The data that we reviewed included demographic characteristics; histological diagnosis; anatomical location, size, and stage of the tumor; characteristics of the biopsy specimen; duration and description of symptoms before an accurate diagnosis was made; delay before recognition of the tumor on radiographs; results of diagnostic imaging; inaccurate diagnoses; type of intervention based on these inaccurate diagnoses; and outcome with regard to survival. There were forty high-grade sarcomas and twenty-eight low-grade sarcomas. Results: Excluding two asymptomatic patients in whom the sarcoma was noted incidentally, the average duration of symptoms before an accurate diagnosis was made was ten months (median, six months; range, one month to four years). Common symptoms and findings on physical examination included pain in the buttock (twenty-three patients; 35 percent), a mass (twenty patients; 30 percent), sciatica (nineteen patients; 29 percent), pain in the hip (seventeen patients; 26 percent), pain in the groin (thirteen patients; 20 percent), and low-back pain (fourteen patients; 21 percent). In thirty (44 percent) of the sixty-eight patients, the pelvic sarcoma was not recognized initially and an inaccurate diagnosis was made. The misdiagnoses included a herniated lumbar disc, spinal stenosis, spondylolisthesis, tendinitis, bursitis, an inguinal hernia, a stress fracture, a pilonidal cyst, a recurrent urinary-tract infection, and degenerative arthritis of the spine, hip, and knee. Inappropriate treatment for these misdiagnoses included seven operative procedures (two laminectomies, two débridements, one hip arthrotomy, one total knee replacement, and one inguinal herniorrhaphy), six courses of nonsteroidal anti-inflammatory medications, five chiropractic adjustments, four trials of physical therapy, and three local injections of steroids. It took an average of seven months for the clinicians to arrive at the diagnosis of primary pelvic sarcoma. With the numbers available, no significant association between the duration of symptoms before an accurate diagnosis was made and the grade or the stage of the tumor could be detected. In addition, no association between the duration of symptoms and the survival of the patient (p = 0.54) could be determined, with univariate analysis. The grade and the stage of the tumor were strongly associated with the outcome, with a low tumor grade proving to be a favorable prognostic indicator for survival (p = 0.006). Conclusions: Patients who have a primary bone sarcoma of the pelvis often have had symptoms for a long duration that mimic those of more commonly encountered non-neoplastic musculoskeletal conditions. When a patient has symptoms that appear to be out of the ordinary, particularly refractory pain or pain at rest, physicians should include the pelvic girdle in the evaluation and should carefully examine a high-quality radiograph of the entire pelvis.
Primary bone sarcomas of the pelvis present challenging problems with regard to management. Despite recent advances in adjuvant chemotherapy and radiation therapy, contemporary diagnostic imaging, and increased enthusiasm for limb-sparing resections of the pelvis instead of hemipelvectomies, the prognosis for patients who have a sarcoma (especially a high-grade tumor) of the pelvic girdle is worse than that for patients who have a tumor in an extremity13,15. The reason for this worse prognosis is that sarcomas of the pelvis generally remain undetected longer than do those of the extremities, are larger at the time of diagnosis, and are more difficult to treat. Early recognition of these sarcomas and early initiation of treatment should improve outcomes. Several reports have focused on the treatment and prognosis of primary bone sarcomas of the pelvis4,6,7,9,13. However, to our knowledge, there have been no studies of delays in the recognition of pelvic tumors with a focus on the implications of the duration of symptoms either before the correct diagnosis is established and the appropriate treatment is initiated or after the patients have been evaluated by a clinician. In the present study, we evaluated patients who had a primary bone sarcoma of the pelvis. Many of these patients had had a delay before the sarcoma was diagnosed. We analyzed the data on these patients to determine the duration of such delays in treatment and their effect with respect to survival.
From 1975 to 1995, seventy consecutive patients who had a primary bone sarcoma of the pelvis (the pubis, acetabulum, ischium, ilium, or sacrum) were referred to us. Patients who had been diagnosed with multiple myeloma or lymphoma of bone were not considered for the study. Of the seventy patients, one had a hemangioendothelioma of the acetabulum and was ultimately managed and followed elsewhere and another had a recurrent sacral chordoma that had been previously diagnosed and treated. These two patients were excluded. The remaining sixty-eight patients were followed from the time of diagnosis until the time of death or the time of the most recent follow-up. We retrospectively analyzed data from office and medical records, a tumor registry, a separate orthopaedic oncology file, and radiographs. The data included the age and gender of the patient, histological diagnosis, site of pelvic involvement, size of the tumor, stage of the tumor at the time of diagnosis, characteristics of the biopsy specimen, duration and description of the symptoms before an accurate diagnosis was made, duration of the delay before recognition of the tumor on radiographic imaging, results of diagnostic imaging, inaccurate diagnoses, type of operative treatment, and oncological outcome. All of the tumors were staged at the time of diagnosis according to the staging system for bone sarcomas described by the American Joint Committee on Cancer1 and by Enneking et al.3. We analyzed the data for an association between survival of the patient and duration of the symptoms, specific symptoms, size of the tumor, gender of the patient, anatomical site of the tumor, histological diagnosis, and grade and stage of the tumor. In addition, we specifically examined the time that it took for the evaluating clinician to recognize a pelvic sarcoma. We looked for an association between a delay in diagnosis (one month or longer) and the survival of the patient. The data were analyzed for significant associations with univariate and multivariate methods, and then a stepwise Cox proportional-hazards model2 was constructed for variables that had a significant prognostic value (p < 0.05).
The median age of this group of sixty-eight patients who had a primary sarcoma of the pelvis was forty-one years (range, eight to eighty-two years). There were thirty-four male and thirty-four female patients. Twenty-eight patients had a low-grade sarcoma, and forty had a high-grade sarcoma. Two (3 percent) of the tumors were stage IA, twenty-three (34 percent) were stage IB, thirty (44 percent) were stage IIB, and thirteen (19 percent) were stage IV. Thirteen (19 percent) of the patients had metastases; eleven were male and two were female. Of the sixty-eight patients, fifty-five (81 percent) had had a biopsy to establish a histological diagnosis. The remaining thirteen patients (19 percent), all of whom had a chondrosarcoma, were managed with an operative resection without a biopsy. There were thirty-eight incisional biopsies and seventeen needle biopsies. Thirty (55 percent) of the fifty-five patients had the biopsy before they were referred to our musculoskeletal tumor center, and twenty-five (45 percent) had it at our institution. According to the results of the histopathological studies, there were twenty-three (twenty-one low-grade and two high-grade) chondrosarcomas (34 percent), sixteen (two low-grade and fourteen high-grade) osteosarcomas (24 percent), ten Ewing sarcomas (15 percent), six high-grade malignant fibrous histiocytomas (9 percent), five chordomas (7 percent), three dedifferentiated chondrosarcomas (4 percent), two Paget sarcomas (3 percent), two undifferentiated sarcomas, and one low-grade hemangioendothelioma (1 percent). The anatomical locations of the epicenter of the tumor in the pelvis included the ilium (thirty patients), ischium (twelve patients), pubis (ten patients), acetabulum (eight patients), and sacrum (eight patients). The operative procedures included fifty-two pelvic resections and nine hemipelvectomies. Thirty-six patients had wide margins, seventeen had marginal margins, and eight had intralesional or contaminated margins3. Of the remaining seven patients, two who had a Ewing sarcoma received chemotherapy and radiation therapy, four had a high-grade sarcoma that could not be resected, and one had an undifferentiated sarcoma that progressed precipitously during preoperative chemotherapy and she died. The average duration of symptoms before an accurate diagnosis was made was ten months (range, one month to four years; median, six months), as reported by sixty-six patients. The other two patients were asymptomatic; one of them had a radiation-induced osteosarcoma of the acetabulum that was noted incidentally on a routine follow-up radiograph of the pelvis after operative treatment of a contralateral slipped capital femoral epiphysis, and the other was diagnosed during an evaluation for a pelvic injury after a fall. Common symptoms, which occurred alone or in combination, included pain in the buttock (twenty-three patients; 35 percent), pain with motion of the hip (seventeen patients; 26 percent), low-back pain (fourteen patients; 21 percent), pain in the groin (thirteen patients; 20 percent), and pain radiating down the lower extremity (nineteen patients; 29 percent). Thirty-seven (56 percent) of the patients reported a combination of these symptoms. Twenty (30 percent) of the patients had a mass that was found on physical examination at the time of diagnosis. One patient who had dyspareunia was found to have a large pelvic mass during a gynecological examination. Thirty (44 percent) of the pelvic sarcomas were not accurately diagnosed for at least one month from the time that the patient first visited a physician for an evaluation of symptoms. The incorrect diagnoses included a herniated lumbar disc, spinal stenosis, spondylolisthesis, tendinitis, bursitis, an inguinal hernia, a stress fracture, a pilonidal cyst, a recurrent urinary-tract infection, and degenerative arthritis of the spine, hip, and knee. The average time until these thirty tumors were accurately diagnosed was seven months (range, one month to three years). These misdiagnoses led to inappropriate observation or treatment, or both. The initial radiographs of eight of the thirty patients were misinterpreted by general practitioners, radiologists, or orthopaedic surgeons as showing normal findings, a stress fracture, or degenerative arthritis. In at least nine of the thirty patients, the symptoms were presumed to be the result of a disease of the lumbar spine; therefore, the initial diagnostic evaluation of these patients was focused on the lumbar spine rather than on the pelvis. Inappropriate treatment, due to the inaccurate diagnoses, included the administration of nonsteroidal anti-inflammatory drugs (six patients); local injections of steroids (three patients); physical therapy, including lumbar traction and electrical stimulation (four patients); and chiropractic adjustments (five patients). Inappropriate operative treatment included lumbar laminectomies (two patients), débridements (two patients), an inguinal herniorrhaphy (one patient), a hip arthrotomy (one patient), and a total knee arthroplasty (one patient). Various radiographic studies were ordered by clinicians during the initial evaluation of each of the sixty-eight patients. These tests included plain radiography, bone scintigraphy, computed tomography, and magnetic resonance imaging. Studies of the lumbar spine were common. The pelvic sarcoma was discovered as an abnormality on an initial anteroposterior radiograph of the pelvis in forty-nine (72 percent) of the patients before any other radiographic images were made. The following three reports illustrate the typical findings in the patients in the present study. CASE 1. A fifty-six-year-old woman had increasing pain in the right hip that radiated to the right leg in a radicular fashion. The pain had persisted for four months despite rest, the administration of nonsteroidal anti-inflammatory drugs, and physical therapy. Suspecting that the symptoms had a lumbar discogenic etiology, the treating physician referred the patient to a neurosurgeon, who ordered magnetic resonance imaging of the lumbar spine and diagnosed the condition as spinal stenosis. Radiographs of the pelvis were not made. The patient was managed with a laminectomy and a partial discectomy at the fourth and fifth lumbar levels. However, the symptoms persisted postoperatively and progressed over the next seven months. A radiograph of the pelvis then revealed an abnormality in the right ilium (Fig. 1-A). A computed tomographic scan confirmed the presence of a bone lesion in the inferior portion of the ilium, with a soft-tissue mass at the sciatic notch (Fig. 1-B). A computed tomography-guided needle biopsy was performed, and chondrosarcoma was diagnosed. The patient was managed with a limb-salvage resection of the right hemipelvis and a reconstruction with a hemipelvic allograft and a femoral endoprosthesis.
CASE 2. A seventeen-year-old girl had a gradual onset of vague pain in the right buttock that radiated to the right leg. After these symptoms had persisted for six months, the general practitioner had a radiograph of the pelvis made (Fig. 2-A). The patient was told that she had arthritis, and she was managed with aspirin. The pain increased, and eight months later the patient sought another opinion. Another pelvic radiograph was made (Fig. 2-B), and, in retrospect, it was realized that the first study had also showed an abnormality in the right ilium, but this was not appreciated at the time. The patient was referred to an orthopaedic surgeon, who performed an incisional biopsy; the findings were interpreted as demonstrating a small-cell osteosarcoma. Clinical staging studies (Fig. 2-C) and a pathological review confirmed the presence of a stage-IIB small-cell osteosarcoma. The patient subsequently was managed with an extended hemipelvectomy of the right lower extremity followed by chemotherapy.
CASE 3. A forty-five-year-old man had had low-back pain and dysesthesias in the anterior aspect of the left thigh for eight months. He had had a series of chiropractic spinal manipulations that had temporarily relieved the symptoms. The pain became progressively severe, and the patient went to a local emergency room, where a radiograph of the pelvis showed a lesion in the wing of the left ilium (Fig. 3-A). A needle biopsy was performed under computed tomographic guidance (Fig. 3-B). A diagnosis of chondrosarcoma was made, and the patient was referred to our institution. On physical examination, a mass was noted in the left lower quadrant of the abdomen. The patient had weakness of the left quadriceps muscle and a slight decrease in the left patellar reflex. A magnetic resonance image of the pelvis showed a huge intrapelvic and lower abdominal mass (Fig. 3-C). We performed a resection of the ilium and the mass (Fig. 3-D) from the supra-acetabular region to just lateral to the sacral foramen, almost to the left kidney, and reconstructed the pelvis with an allograft. The patient had evidence of metastasis to the paraspinal lymph nodes.
The median duration of follow-up of the thirty-nine surviving patients was four years and one month (average, four years and nine months). Thirty-three (49 percent) of the sixty-eight patients were alive and had been continuously disease-free at the time of follow-up; five (7 percent) had no evidence of disease after the treatment of a relapse; and one (1 percent) was alive with disease. Twenty-nine of the sixty-eight patients died: twenty-eight (41 percent) died of the disease and one (1 percent), of chemotherapy-related sepsis. Of the twenty-eight patients who had a low-grade tumor, twenty-five (89 percent) survived, whereas only fourteen (35 percent) of the forty patients who had a high-grade sarcoma survived. Univariate analysis revealed an association between female gender (p = 0.001), histological diagnosis (p = 0.0002), and grade of the tumor (p = 0.0001) and continuously disease-free survival. On the basis of the Cox proportional-hazards model2, female gender (p = 0.01) and a low-grade tumor (p = 0.006) were associated with longer continuously disease-free survival. With the numbers available, no significant association between the duration of symptoms before an accurate diagnosis was made and the grade or the stage of the tumor could be detected. With the numbers available, we detected no significant association between the duration of symptoms reported by these patients and survival (p = 0.54). We also detected no association between any specific symptom, the location of the epicenter of the tumor, or the size of the tumor and survival. With the numbers available, we detected no significant difference between the survival of the thirty patients who had a delay in diagnosis of at least one month and that of the remaining patients (p = 0.62). In fact, we could detect no significant association between the delay in diagnosis and survival even when this group of thirty patients was incrementally divided into smaller groups with successively increasing delays.
The present study shows that delays in the diagnosis of primary bone sarcomas of the pelvis are common. Forty-four percent (thirty) of the sixty-eight lesions were initially misdiagnosed, resulting in a considerable delay (average, seven months) before an accurate diagnosis was made. There are several reasons for the lack of early recognition of a primary sarcoma of the pelvis. First, primary bone sarcomas are rare. A recent report from the American Cancer Society revealed an estimate of 2500 bone sarcomas annually in the United States10. This relatively small number includes bone sarcomas in all anatomical locations. The pelvis, including the sacrum, accounts for only a very small fraction. Consequently, because pelvic sarcomas are encountered so infrequently, physicians undoubtedly have a low index of suspicion for this diagnosis. Understandably, commonly encountered conditions, such as arthritis of the hip, trochanteric bursitis, and disease of the lumbar spine are suspected first, and symptoms that arise in this region are apt to be attributed to a cause other than cancer. Second, our patients had had symptoms for a surprisingly long duration (average, ten months) before an accurate diagnosis was made. A long period of symptoms is commonly viewed by physicians as less consistent with a diagnosis of cancer because malignant tumors tend to grow rapidly locally, thus drawing attention to their existence early in the course of the disease. This view cannot be applied to the pelvic region, where tumors can grow and extend from relatively thin, flat bones into both intrapelvic and extrapelvic soft tissues. Patients with this type of tumor often have an extraordinarily large soft-tissue mass without symptoms before the lesion becomes evident on clinical examination. As we discovered, symptoms secondary to bone sarcomas of the pelvis are often vague and misleading. Common symptoms included pain radiating down the lower extremity in a radicular fashion; pain with motion of the hip; and pain in the buttock, groin, and low back. With the numbers available, we were unable to detect an association between these common symptoms and the anatomical site of the tumor within the pelvis. Finally, the pelvis is a difficult location for early radiographic recognition of bone abnormalities. Other authors have discussed the inherent difficulties in recognizing tumors, both benign and malignant, in the pelvic girdle8,17. Osseous changes secondary to pelvic bone sarcomas are often undetected on early plain radiographs because of subtle features, an inconspicuous location, or superimposition of other bones and soft-tissue shadows. Orthogonal radiographs, which are made routinely for bones in the extremities, are lacking for the pelvis. A detailed evaluation of the pelvis to search for bone tumors often requires more sophisticated radiographic imaging. Our patients often had symptoms that mimicked those commonly associated with diseases of the lumbar spine. This finding is in agreement with those of other reports demonstrating that benign, malignant, and metastatic bone tumors in the pelvis frequently cause symptoms that are typical and suggestive of tumors originating from discogenic disease of the lumbar spine8,11,14,16. Findings of sciatica or radiculopathy may be related to the proximity of the tumor to, or involvement of, elements of the lumbosacral plexus. These lesions are often associated with a soft-tissue mass that extends into the sciatic notch (producing sciatica) or into the psoas muscle (causing quadriceps weakness). Recent studies have shown that the most important prognostic factor for patients who have a primary bone sarcoma of the pelvis is the histological grade of the tumor7,13. Our data confirm that high-grade tumors are associated with a poor prognosis for survival (p = 0.0001) and low-grade tumors are associated with longer continuously disease-free survival (p = 0.006). Female gender was also associated with a favorable prognosis (p = 0.001). However, this association might be due to the large number of male patients (eleven [32 percent] of thirty-four) who had metastatic disease at the time of diagnosis; metastatic disease conferred a worse prognosis for this group. We cannot explain the larger number of male patients who had metastatic disease. We detected no significant difference (with the numbers available) in the delays in diagnosis or the grades of the tumors between male and female patients who had metastatic disease at the time of clinical staging. Symptoms may be present for a long period of time before the patient seeks medical attention. By that time, the pelvic tumor may have progressed considerably, becoming large locally and possibly metastasizing to distant sites. With the numbers available, however, we detected no significant association between the duration of symptoms and survival. Furthermore, we detected no association between a long (at least one-month) delay in diagnosis, which was the case for thirty (44 percent) of the patients, and overall survival. The prognosis, regardless of a delay in diagnosis, was most strongly associated with the grade of the tumor. Simply stated, patients who had a high-grade bone sarcoma of the pelvis had a poor prognosis regardless of any delay before an accurate diagnosis was made. We found no association between the size of the tumor and the survival of the patient; however, previous studies4,15, including one in which one of us (M. A. S.) participated13, showed that adequate operative margins are technically more difficult to achieve with larger tumors. Furthermore, a recent report by Kawai et al.7 showed inadequate operative margins to be an independent, adverse prognostic factor in local recurrence of sarcoma of the pelvis. Thus, a delay in treatment may compromise the opportunity for a successful limb-salvage resection to achieve adequate local control of the tumor. We recommend evaluation with plain radiography and, in ambiguous clinical situations, computed tomography of the pelvis5,12. The rationale for this recommendation and our preference for computed tomography rather than bone scintigraphy rely heavily on our findings in the present study and on the clinical experience of most musculoskeletal oncologists. First, in the present study, forty-nine (72 percent) of the diagnoses of pelvic sarcoma were based on plain radiographic findings. In addition, some patients (for example, Case 2) had subtle abnormalities on the plain radiographs that were not detected by the clinician or the radiologist. A bone scintiscan is often used by clinicians when patients have low-back pain or other painful musculoskeletal conditions of unknown cause. Often, under these circumstances, another benign tumor (such as enchondroma) or a benign non-neoplastic condition is diagnosed, a biopsy is performed, and the tumor in the appendicular skeleton is treated, to the detriment of the patient. Thus, if the goal is to diagnose a pelvic sarcoma, a radiograph and a computed tomographic scan of the pelvis are probably more specific and more sensitive than a bone scintiscan5,12. Furthermore, a bone scintiscan is of little value for the detection and evaluation of a soft-tissue mass or another soft-tissue abnormality that may be responsible for symptoms in some patients. Compared with magnetic resonance imaging, computed tomography is relatively quick and inexpensive. Physicians must remember to include the pelvic girdle in the evaluation of patients who have symptoms that appear to be out of the ordinary, particularly when the patients have refractory pain or pain at rest.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |