Pigmented villonodular synovitis and synovial cysts of the spine [ comment].

The most commonly diagnosed tumor in the spine is metastatic carcinoma and it is often stated that primary tumors are rare in this location. In Dahlin 's series of 1,447 surgically diagnosed benign skeletal tumors, 5.0% were in the vertebral column excluding the sacrum and of 4,774 surgically diagnosed malignant tumors, 8 % (1). However, it must be borne in mind that tumors in the spine may have been underdiagnosed because of their location, and this was especially likely before the introduction of modern imaging techniques. On the other hand, perhaps we have a tendency to overestimate the mass of the vertebral column with respect to total skeletal mass and the incidence of primary tumors is, after all, proportional to the mass of bone. Be that as it may, it is certain that some tumors predilect the spine. For example, osteoblastoma is twice as common in the spine as in the appendicular skeleton, myeloma is more likely to be diagnosed in the spine than elsewhere, and chondroma is unique to the vertebral column. Furthermore, lesions that are very common in the appendicular skeleton, such as nonossifying fibroma , are not seen in the spine. In this issue of the American Journal of Neuroradiology, there are two case reports (2, 3) of benign lesions that, though not that uncommon in the peripheral skeleton, have rarely been reported in the spine. The first of these two reports (2) deals with a case of pigmented villonodular synovitis (PVNS), a condition more likely to be diagnosed in the fingers than elsewhere, possibly because a small lump in the finger is much more obvious and likely to cause problems than when it involves the synovium elsewhere, such as in the knee joint. (In this latter location, at least in some cases, the lesion may only be discovered accidentally.) In our department, which specializes in orthopaedic pathology, we see approximately 60 cases of PVNS a year, mostly of the fingers . We have seen no cases originating in the spine, thus supporting the contention that this lesion is rare in the vertebral column. Because this is a synovial lesion, it can only arise in structures lined by synovium, such as tendon sheath or a diathrodial joint. Considering the number of synovial-lined joints in the vertebral column, it is perhaps surprising that the condition is so rare in this location. In the case reported by Titelbaum et al (2) , the patient was initially thought to have a disk herniation and, in our experience as well as that of others (4), a number of cases that were subsequently diagnosed histologically as tumors had been diagnosed initially both clinically and radiographically as disk herniation. With a lesion in the neural arch, the radiologic differential diagnosis may also have included osteoblastoma. As is pointed out in this paper, it is likely that MR imaging, had it been done in this case, would not only have provided excellent localization but also that the iron present in the lesion would have given a characteristic T2-dephasing quality, producing a low signal on long TR/TE images. Lesions of PVNS may on occasion be seen to erode the underlying bone, as was the case in the lesion reported; this is more likely to occur with a tight joint such as a facet joint than it is with a looser structure such as the knee. In a hip, also a tight joint, which has been involved with PVNS, we have observed not only erosion of bone but joint-space narrowing, suggesting radiographically an inflammatory arthritis (5). In the discussion of the case, it is suggested that PVNS is closely related to nodular tenosynovitis. I believe many pathologists would go further and say that they are identical lesions. In our opinion, PVNS is best regarded as a neoplasm, albeit a benign neoplasm of histiocytic origin (6, 7). A few cases have been reported with malignant degeneration and metastasis (8) . Another benign lesion that also may erode bone and lead to problems of differential diag-