Dr. Henry J. Mankin published this classic article nearly 25 years ago in an effort to warn orthopaedic surgeons of the potential hazards of performing biopsies on malignant musculoskeletal tumors. Subsequent papers by Dr. Mankin and other authors, however, have revealed that concerns about the biopsy persist.
Dr. Mankin was born in Pittsburgh and received his MD from the University of Pittsburgh in 1953. He interned at the University of Chicago and, after military service in the US Navy, he took a residency in Orthopaedics at the Hospital for Joint Diseases in New York. In 1972 he became Chief of Orthopaedic Surgery at the Massachusetts General Hospital in Boston, following service at the University of Pittsburgh and the Hospital for Joint Diseases, where he served as Chief of Orthopaedics, and Mount Sinai School of Medicine, where he held an appointment as Professor. From 1962 to 2003 Dr. Mankin was continuously funded by the National Institutes of Health for research in the fields of cartilage, osteoarthritis, allografting, Gaucher's disease and, in particular, for many aspects of musculoskeletal tumors. He has published over 600 articles on these subjects and has received numerous awards and honors for his contributions to Orthopaedics. He has been elected president of the Musculoskeletal Tumor Society, the American Orthopaedic Association, the Orthopaedic Research Society, and the American Board of Orthopaedic Surgery, and has a Chair in Orthopaedics named after him at the University of Pittsburgh. Dr. Mankin is one of the most influential orthopaedic surgeons in North America, respected as a leader in education, research, and clinical practice.
Henry H. Sherk, MD
Most protocols for the staging of primary malignant bone and soft-tissue tumors require accurate histological diagnosis and grading after biopsy1,4. Although this essential operative procedure has been demonstrated to involve certain hazards, the discussion of the biopsy in standard textbooks and in the literature usually has been too general and non-directive2-11.
Surgeons who deal frequently with malignant tumors of bone and soft tissue recognize the possible complications that may result from a biopsy. Their experience has also led them to believe that such complications are more frequently encountered than is commonly thought. The Musculoskeletal Tumor Society is a group of forty individuals, mostly orthopaedic surgeons, whose primary interest is the surgical management of patients with malignant tumors of the musculoskeletal system. In 1979, the Society initiated a study to determine the frequency with which problems with biopsy occur and the impact of these problems on the patient's course and ultimate outcome. The purpose of this report is to describe the results of that study and to offer some suggestions to physicians who care for such patients, to aid them in avoiding the hazards inherent in the biopsy procedure.
Materials and Methods
The Society assigned us the task of gathering data that could answer the following questions.
- How accurate is the diagnosis made after biopsy when compared with the so-called definitive diagnosis?
- What are the incidence and nature of the complications of the biopsy procedure?
- Do inaccuracies or technical errors in the biopsy procedure materially affect either the treatment plan or the patient's outcome, or both?
- Is there a significant variation in the answers to these questions depending on whether the biopsy is performed in a referring institution or in the center where the patient subsequently receives definitive treatment?
A questionnaire was devised for the purpose of answering these questions (Table I). Each member of the Musculoskeletal Tumor Society was asked to complete this questionnaire on each of twenty sequentially seen, unselected, newly discovered patients who were treated prior to January 1, 1979. All of the patients had to have had a definitive diagnosis of malignant primary bone or soft-tissue tumor, for which an incisional or needle biopsy was performed, followed at an interval by definitive surgery (at which procedure tissue was obtained), and all patients had to have been followed for long enough to determine the early outcome of the treatment. The members were asked to append pathology reports of the results of the original biopsy and of the subsequent definitive surgery. The questionnaire also required the surgeon to explain specifically how the outcome was affected.
Twenty Society members in sixteen separate treating centers* responded and submitted data on a total of 329 patients who fulfilled the criteria. One hundred and seventy-three (52.6 per cent) of the patients were male and 156 (47.4 per cent) were female. The mean age was 36.5 years (range, two weeks to eighty-three years). Two hundred and twenty-two of the tumors were primary in bone and 107 arose in soft tissue.
The distribution of tumor types is shown in Table II. It should be noted that the slightly greater incidence of malignant soft- tissue tumors in the general population compared with primary malignant bone tumors is not reflected by the data reported in this study (only 32.5 per cent of the tumors arose in the soft tissues). This discrepancy is partly explained by the composition of the Musculoskeletal Tumor Society, which has only one member who is a general surgeon. The remainder are orthopaedic surgeons, who sometimes share the management of patients with tumors of the soft tissues with other services but have more or less exclusive access to patients with primary tumors of bone. Another conjecture is that many more tumors of soft tissues were marginally excised (shelled out) prior to referral and thus did not meet the criteria for incisional or needle biopsy, which was required for inclusion in the study.
Despite this variation from the reported relative incidence of bone and soft-tissue tumors, the distribution of the histological diagnoses approximated the incidence of these tumors in the general population3,5. Almost 50 per cent of the bone tumors were osteosarcomas, 22 per cent were chondrosarcomas, and 6.7 per cent were Ewing's tumors. The relative infrequency of the other tumors was reflected by the few cases that were reported. The data for the soft-tissue tumors similarly reflected the frequency with which surgeons encounter synovial sarcoma and malignant fibrous histiocytoma (22 per cent of the series), liposarcoma (17 per cent), and the now less commonly made diagnosis of fibrosarcoma5 (11 per cent of the series). The other soft-tissue tumors were encountered less frequently, and some were quite rare.
One hundred and forty-three patients (43.5 per cent) had the initial biopsy done in a hospital (the referring institution) and were subsequently transferred to a second hospital (the treating center) for definitive care. One hundred and eighty-six patients (56.5 per cent) had the initial biopsy procedure done at the treating center.
Accuracy of the Biopsy
Variation between the diagnosis made on the basis of the initial biopsy specimen and that made from examination of the tissue obtained at the subsequent definitive procedure occurred in eighty-two (25 per cent) of the 329 patients (Table III). We reviewed these differences in diagnosis and arbitrarily divided them into major and minor variances, depending on the likelihood that the discrepancy would materially affect the outcome. A variance in diagnosis that would be likely to mislead the surgeon, alter the treatment plan, change the indication for adjunctive therapy, or in some way interfere with timely and appropriate treatment was considered a major variance. A variation in diagnosis that would be unlikely to have an effect on the patient's course was considered a minor variance. Of the eighty-two variations in diagnosis, sixty (73 per cent) were major, of which forty-three (71 per cent) occurred in referring institutions and seventeen (19 per cent) occured in treating centers. Examples of major variances in the diagnosis of osteosarcoma included diagnoses of aneurysmal bone cyst, benign bone neoplasm, fibrosarcoma, and non-ossifying fibroma. For chondrosarcoma, the errors included diagnoses of osteocartilaginous exostosis, enchondroma, and chondromyxoid fibroma. A biopsy specimen from a patient with chordoma was interpreted as showing fat necrosis; that of a patient with an epithelioid sarcoma, nodular fasciitis; and that of one with a clear-cell sarcoma, benign leioblastoma. Examples of minor variances included the interpretation of histological material from patients with fibrosarcoma as showing mixed sarcoma or malignant fibrous histiocytoma and the identification of a liposarcoma as a fibromyxosarcoma. It is of interest that variances in the diagnoses of frequently encountered tumors occured far less frequently than for less common tumors (implying greater familiarity with the radiographic and histological presentation on the part of the surgeon, radiologist, and pathologist) (Table III). It should be noted, however, that if the 110 osteosarcomas and fifteen Ewing's sarcomas (both of which commonly have very distinctive patterns) are excluded from analysis, the percentage of variances in diagnosis for the entire series would rise considerably. Errors in diagnosis were made in only 10 per cent of the patients with osteosarcoma and in 6.7 per cent of those with Ewing's sarcoma, but were much more frequent for other relatively common lesions such as parosteal osteosarcoma, malignant fibrous histiocytoma, synovial sarcoma, fibrosarcoma, and so on. It is interesting to note that there were only seven lesions for which there were no variances in diagnosis, and that some of these tumors are not easily diagnosed. These included neurofibrosarcoma (four lesions); Paget's sarcoma (two); lymphangiosarcoma (two); and adamantinoma, plasmacytoma, leiomyosarcoma, and hemangiopericytoma (one each).
We created a special category for initial biopsy material that was considered to be non-representative or technically unsatisfactory prepared (that is, stained poorly, cells squashed, only necrotic area included, and so on). The biopsies of thirty-four patients (10.3 per cent of the entire series) were included in this group, and could account for the problems in diagnosis of some of the lesions. It should be noted that only five of these thirty-four biopsies were needle or trochar biopsies (compared with fourteen needle biopsies in the entire series). The needle biopsies in this small series were as accurate (or inaccurate) as open biopsy. However, patients are often preselected for needle biopsy because the suspected lesion is considered to be easy to diagnose from a limited amount of tissue7-9. Twenty-eight (82.4 per cent) of the thirty-four non- representative or technically poor biopsies were performed in referring centers, while six (17.6 per cent) were done in treating centers. In seven patients (30.4 per cent) the outcome was adversely affected.
Complications of the Biopsy Procedure
The questionnaire asked the respondent to categorize surgical complications of the biopsy according to the tissue involved (skin, soft tissue, bone, combinations of these, or “other”) (Table IV). In this group of fifty-seven patients(17.3 per cent of the series), complications involving skin and soft tissue included wound breakdown, hemorrhage, and infection, and fracture was the most common complication involving bone. In thirty (53 per cent) of these patients, the complications had an effect on either treatment or outcome. Forty-four (77.2 per cent) of these patients were initially treated in a referring institution and thirteen (22.8 per cent), in a treating center.
Effect of Biopsy on Treatment and Outcome
In sixty (18.2 per cent) of the patients in the series, there was a significant alteration in the treatment or outcome as a result of some difficulty with the initial biopsy (Table V). The biopsy was performed in the referring institution in forty-five (75 per cent) of these patients and in the treating center in fifteen patients (25 per cent). The alterations in the treatment plan ranged from having to perform a radical resection instead of a wide or marginal one to amputation in situations in which local resection would have been possible. Fifteen (4.5 per cent) of the 329 patients had what might be considered an unnecessary amputation (based on analysis of the type of lesion and its location, which in the opinion of members of the Society would ordinarily have been treated by local resection).
The treating surgeons were asked whether the biopsy adversely affected the patients' outcome and, if so, to provide specific information. This is obviously a difficult assessment to make and some of the reports were, in our opinion, less than straightforward or lacked sufficient detail. For twenty-eight patients (8.5 per cent), however, sufficient documentation was included to allow us to conclude that the patient's prognosis and outcome were adversely affected by some aberration in the biopsy procedure.
Comparison of Results of Biopsies Performed in the Referring Institution with Those Performed in the Treating Center
Patients whose staging biopsy was performed in the referring institution fared less well than those whose biopsy was done in the treating center (Table VI). The latter group of patients had a more than threefold increase in the likelihood of a major error in diagnosis; a sixfold increase in non-representative or poorly prepared biopsies; a more than fourfold greater frequency of problems with skin, soft tissue, or bone; an almost fourfold increase in the likelihood of being treated in some manner other than the one that the surgeon would have chosen if the biopsy had been performed or interpreted appropriately; and an almost threefold increase in the risk of the outcome being adversely affected. It should be clearly noted, however, that although these data are well defined by this study they do not take into account the possibility that some patients may have been transferred to the treating center from the referring institution because of problems with the biopsy. The possible bias introduced by that factor should be carefully considered in interpreting these results.
The data from this study show that in this group of 329 sequentially seen, unselected patients with malignant primary bone and soft-tissue tumors who were treated by experienced surgeons in major centers, errors in diagnosis, non-representative or poorly stained biopsy material, and wound-healing problems at the biopsy site occurred in a startlingly high percentage and frequently affected the patients' treatment and prognosis. Almost 20 per cent of these patients had to be managed by a less than optimum treatment plan and at least 8 per cent had a distinctly adverse prognosis because of problems with the biopsy. Fully 4.5 per cent of the patients who might have had a limb-sparing procedure required an amputation as a result of a problem with the biopsy. Furthermore, analysis of the data strongly suggests that the risk of all of these problems increases markedly if the initial biopsy is performed in a referring institution rather than in a treating center (Table VI).
Before fully accepting these statistics, however, several issues must be clarified. The first is the validity of the sampling technique. Although each member of the Musculoskeletal Tumor Society was asked to contribute data on twenty patients, only twenty of the forty members of the Society complied. One must ask whether the findings would have been altered if data had been obtained from the non-contributors as well. In reviewing the individual submissions, variability both in diagnoses and in aberrations of the biopsy were noted, suggesting that the responding members were not a homogeneous group. Proof of the likelihood that the sample was unselected was the distribution of malignant bone and soft-tissue tumors (Table II), which was generally consonant with the relative incidence of these uncommon lesions in the general population. The fact that the incidence of primary bone tumors exceeded that of soft-tissue tumors (in contrast with generally accepted incidence figures) is probably a reflection of the composition of the Society and of the members who responded to the questionnaire.
A second factor that may have affected the results of the study is related to the method by which the study was performed. The questionnaire requested data only on those patients in whom the definitive diagnosis was of a malignant primary bone or soft-tissue tumor, and was designed to detect errors in diagnosis or in the performance of the biopsy that had resulted in increased problems for the patient. This eliminated patients in whom the initial diagnosis made on the basis of a biopsy was of a malignant lesion, but in whom the final definitive diagnosis was of a benign lesion. One must logically conclude that problems with biopsy affecting the patient's treatment and outcome undoubtedly occurred in these patients as well. The existence of this second, unstudied group certainly suggests that the data obtained in this study are probably on the conservative side, and that the actual values for problems arising in association with biopsy may be considerably higher.
The third consideration is related to the finding in this series that the incidence of problems appeared to be markedly increased for the initial biopsies that were performed in a referring institution. This apparent bias against the referring institution, however, may be because only those patients in whom the diagnosis from the initial biopsy was obscure or in whom some serious problems developed with the biopsy site were referred to the treating center. To test this hypothesis, the diagnoses in the 143 patients in whom the biopsy was performed in a referring institution were compared with those in the 186 patients in whom the initial biopsy was performed in the treating center. If this hypothesis were correct, the percentage of rare lesions would be higher in patients whose biopsy was done in a referring institution rather than in the treating center, but this was not the case (Table VII). This evaluation of the data, however, does not preclude the fact that some of the patients who were sent to the treating institution were referred for help following a technically poor biopsy or one from which a diagnosis could not be made, and that this form of preselection contributed to the poor record for biopsies done in referring institutions. Our findings strongly support the concept, however, that unless the surgeon and the hospital facilities are equipped to carry out the definitive therapy required by the diagnosis, the biopsy should not be done before the patient is sent to a treating center.
Although it is tempting to single out the tumors that caused the most problems, such conclusions can only be tentative because of the numbers of diagnoses included in the study and the variety of problems that occurred. It is evident that the rarer lesions, particularly among the soft- tissue tumors, caused the most difficulty in diagnosis. Analysis of the cases of those patients in whom a problem with the biopsy caused an alteration in the course showed that the most frequent diagnoses were malignant fibrous histiocytoma, chondrosarcoma, unclassified sarcoma, synovial sarcoma, and parosteal osteosarcoma. Of some significance was the fact that the course of only three patients with osteosarcoma was adversely affected by the biopsy, and in the patients with liposarcoma or Ewing's sarcoma the course was altered only once.
We suggest that the following important principles be observed in order to reduce problems with biopsy to a minimum and to provide optimum care for the patient.
- Plan the biopsy procedure as carefully as the definitive surgery. It is not a simple procedure.
- Pay as close attention to asepsis, skin preparation, hemostasis, wound closure, and so on as with any other operation.
- Place the skin incision in such a manner so as not to compromise a subsequent definitive surgical procedure. (Avoid transverse incisions!)
- Be certain that an adequate amount of representative tissue is obtained, and that the pathologist prepares the slides in a manner that will allow a definitive diagnosis.
- If the pathologist cannot make a diagnosis because of unfamiliarity with bone and soft-tissue tumors, urge him or her to seek consultation promptly.
- If the orthopaedist or the institution is not equipped to perform accurate diagnostic studies or definitive surgery and adjunctive treatment, the patient should be referred to a treating center prior to performance of the biopsy.
1. AMERICAN JOINT COMMITTEE FOR CANCER STAGING AND END RESULTS REPORTING: Manual for Staging of Cancer, pp. 153-158. Chicago, American Joint Committee, 1977.
2. BROSTRÖM, L.-A.; HARRIS, M.A.; SIMON, M.A.; COOPERMAN, D. R.; and NILSONNE, ULF: The Effect of Biopsy on Survival of Patients with Osteosarcoma. J. Bone and Joint Surg., 61-B(2):
3. DAHLIN, D.C.: Bone Tumors. General Aspects and Data on 6,221 Cases. Ed. 3. Springfield, Illinois, Charles C Thomas, 1978.
4. ENNEKING, W. F.; SPANIER, S. S.; and GOODMAN, M. A.: Current Concepts Review. The Surgical Staging of Musculoskeletal Sarcoma. J. Bone and Joint Surg., 62-A:
1027-1030, Sept. 1980.
5. HAJDU, S. I.: Pathology of Soft Tissue Tumors. Philadelphia, Lea and Febiger, 1979.
6. JAFFE, H. L.: Tumors and Tumorous Conditions of Bone and Joints, pp. 14-17. Philadelphia, Lea and Febiger, 1958.
7. MOORE, T. M.; MEYERS, M. H.; PATZAKIS, M. J.; TERRY, ROGER; and HARVEY, J. P., JR.: Closed Biopsy of Musculoskeletal Lesions. J. Bone and Joint Surg., 61-A
: 375-380, April 1979.
8. SCHAJOWICZ, F.: Tumors and Tumor-like Lesions of Bone and Joints, pp. 5-16. New York, Springer, 1981.
9. SCHAJOWICZ, FRITZ, and DERQUI, J. C.: Puncture Biopsy in Lesions of the Locomotor System. Review of Results in 4050 Cases Including 941 Vertebral Punctures. Cancer, 21:
10. SPJUT, H. J.; DORFMAN, H. D.; FECHNER, R. E.; and ACKERMAN, L. V.: Tumors of Bone and Cartilage, pp. 26-28. Atlas of Tumor Pathology, series 2, fasc. 5. Washington, D. C., Armed Forces Institute of Pathology, 1971.
11. WOOD, W. D., and BINDER, S. C.: Biopsy Principles. In
Cancer: A Manual for Practitioners, pp. 18-21. Boston, American Cancer Society, Massachusetts Division, 1978.
* The participating programs included: Beaumont Hospital, Royal Oak, Michigan; Case Western Reserve University; Massachusetts General Hospital; Mayo Clinic; M. D. Anderson Hospital; Ohio State University; Rizzoli Institute, Bologna; University of California, San Francisco; State University of New York at Buffalo; University of Chicago; University of Florida; University of Iowa; University of Kansas; University of Miami; University of Minnesota; and University of Wisconsin.