Case 5 -

Spindle Cell Tumor of Prostate Lawrence D. True University of Washington Medical Center Seattle, WA

Click on each slide thumbnail image for an enlarged view

Case History This 65 year old male presented with a bulky pelvic tumor. Nine years prior to admission a routine physical exam revealed an asymmetrically enlarged prostate. Prostate biopsies showed a "prostate adenocarcinoma, poorly differentiated." The tumor was treated with 6,800 Gy of radiation to the pelvis. The patient remained asymptomatic for 9 years, at which time he developed bladder outlet obstruction. A biopsy revealed tumor, which was treated by radical cystoprostatectomy. Based on evidence of residual tumor, additional radiation therapy, followed by en-bloc resection of residual tumor and rectum, was undertaken. The slides are from this latter specimen. Case 5 - Figure 1 - The lesion cells have a predominantly spindle shape and haphazard growth pattern. Markedly atypical cells are present in this relatively hypercellular field. Case 5 - Figure 2 - Spindle shaped lesion cells exhibit only moderate atypia and cellularity in this field. Case 5 - Figure 3 - Lesion cells are admixed with an osteoid stroma in this field. Case 5 - Figure 4 - Lesion cells form palisading aggregates of Verocay bodies in this field. Histology of the case In the section is a multifocally necrotic tumor that has regions of histologically intact, cytologically malignant spindle cells. The architecture of these spindle cells is undifferentiated. Specifically, there is no formation of fascicles or of specific stromal elements and no epithelioid features. The main differential diagnostic considerations are sarcoma of the prostate, sarcomatoid prostate carcinoma, and pseudotumor that exhibits marked atypia. Immunostains of the spindle cells in this and additional sections were positive only for vimentin. No immunoreactivity for antibodies to keratin, smooth muscle actin, desmin, S100 protein, alk, HMB45, or PSA was exhibited by the spindle cell population. However, keratin immunostains of other sections revealed rare, PSA-positive nests and cribriform aggregates of epithelial cells that exhibited cytologic atypia, characterized by mildly enlarged nuclei and somewhat prominent nucleoli. Based upon these findings and the history of a previously diagnosed prostate carcinoma, a diagnosis of sarcomatoid carcinoma of the prostate was made. History (M181): 27 year-old male who presented with hematuria. Radiological exam showed a single, large, midline, multicystic pelvic mass that was situated between and did not appear to involve the prostate/bladder neck region or the rectum. Cystoscopy and colonoscopy failed to reveal a mucosal lesion. An exploratory laparotomy showed a localized mass, which was biopsied. After treatment with chemotherapy, the tumor was resected as a radical cystoprostatectomy/tumor mass resection. Lesions to discuss The differential pathologic diagnosis of the more common spindle cell tumors of the prostate are: Pseudotumor Post-operative spindle cell nodule Inflammatory pseudotumor Stromal "tumors" Stromal hyperplasia/nodule/leiomyoma STUMP (stromal proliferation of uncertain malignant potential) Miscellaneous Sarcoma Sarcomatoid carcinoma Sarcomatoid prostate adenocarcinoma Sarcomatoid transitional cell carcinoma Epithelial-stromal tumors Fibroadenoma Phyllodes tumor Pseudotumor A number of terms apply to what is apparently the same benign biologic process, which consists of a proliferation of predominantly fibroblastic cells and small numbers of admixed smooth muscle cells and vessels. These terms include "Inflammatory pseudotumor", "Pseudosarcomatous" (fibromyxoid) pseudotumor", and "Post-operative spindle cell nodule". Post-operative spindle cell nodule presents as a mass at the site of a previous surgical procedure, most frequently a bladder neck mass after a TURP or TURB. Typically, the procedure was done within the 3 months prior to appearance of the mass. Occurring at any age, this rare lesion is characterized by a proliferation of spindle cells that focally may form fascicles and/or storiform aggregates. Often they have a high mitotic rate. Atypia and hypercellularity may be marked. Immunohistochemically the only antigen diffusely expressed is vimentin. A minority of lesional cells express desmin and smooth muscle actin; in up to 40% of cases keratin-immunoreactive cells are found. In some post-operative spindle cell nodules cells are cytologically so atypical that malignancy should be considered. You can avoid the pitfall of diagnosing sarcomatoid carcinoma or sarcoma by remembering that a nodule that appears within 3 months of a local surgical procedure is most likely a benign, reactive process. Localized excision is curative. Inflammatory pseudotumor is, basically, indistinguishable from the post-operative spindle cell nodule with respect to histology and benign outcome. Without a history of prior local trauma/surgery, these are idiopathic masses that are self-limited and without malignant potential. The range of cellularity, inflammation, and edematous/myxoid stroma exhibited by these lesions has provided the basis for different terms [1, 2]. This lesion affects both prostate and bladder [3]. Our challenge is to distinguish a malignant stromal tumor from an inflammatory pseudotumor that arises without a history of prior surgery and is composed of an atypical, hypercellular stroma. An immunostain for smooth muscle actin that demonstrates no more than rare to infrequent smooth muscle cells argues strongly against a leiomyosarcomas, which is the most frequent prostate sarcoma in adults. Stromal Tumors of the prostate are predominantly of smooth muscle phenotype. Stromal hyperplasia/nodule/leiomyoma. The earliest stage of nodular hyperplasia of the prostate is thought to be microscopic, periurethral nodules of stromal cells composed of fibroblasts and smooth muscle cells and free of glands. Since the distinction between a stromal cell-only nodule of a prostate with BPH and leiomyoma appears to be arbitrary, we can question whether a large smooth muscle cell nodule warrants being termed "leiomyoma". These lesions are similar in a number of ways to uterine smooth muscle tumors. Histologically atypical and "bizarre" stromal tumors of the prostate have been reported [7]. However, to my knowledge, the biology of these is no different than that of stromal tumors free of atypia and/or of smaller size, specifically, they are benign. Distinction of a stromal nodule/leiomyoma from leiomyosarcoma is based upon features used as criteria in many organs – small size, circumscription and localization to the organ, lack of necrosis, absence of cell/nuclear pleomorphism and low mitotic activity. Atypical stromal tumors . Atypical spindle cell tumors (recently termed STUMPs, for, "stromal proliferations of uncertain malignant potential") exhibit a wide range of histological appearances, ranging from atypical stromal cells interposed between histologically benign prostate glands to phyllodes tumors in which both epithelial and mesenchymal components are proliferative (see below for more discussion of phyllodes tumors) [4-6]. There is evidence, mostly anecdotal, that these tumors represent a biologic continuum, and, consequently, that there may be progression between stages characterized by overgrowth of the stromal component. Antigens variably expressed by STUMPs include vimentin, smooth muscle actin, desmin, CD34 and estrogen and progesterone receptor [5]. Miscellaneous benign prostate spindle cell tumors have been reported, e.g. nerve sheath tumor and solitary fibrous tumor. We have had the experience of a GIST (GI stromal tumor) being regarded clinically to be of prostate origin. When resected, this was found to be a rectal GIST that abutted the prostate gland [8]. Sarcomas of the prostate are rare, representing < 0.1% of primary prostate neoplasms [9]. The most frequent types are leiomyosarcomas (in adults in whom the median age is 60, with a range of 20 to 80 years) or rhabdomyosarcomas (in children in whom the median age is 4, with a range from newborn to 6 years of age, though cases have been reported in adults). Other histologic types of sarcomas have been reported, including angiosarcoma, osteosarcoma, chondrosarcoma, malignant peripheral nerve sheath tumor, synovial sarcoma, and malignant mesenchymoma. Leiomyosarcomas are typically hypercellular, composed of cytologically high-grade spindle cells that tend to form fascicles, and are often necrotic. Other histologic patterns include pleomorphic and epithelioid. Based on the Broders' grading system (a scale of 1 to 4) the majority (>67%) are high grade. Mitoses and necrosis are typically frequent. A definitive diagnosis can be made by demonstrating diffuse smooth muscle actin immunoreactivity. Leiomyosarcoma can usually be distinguished from sarcomatoid carcinoma by absence of an epithelial tumor cell population and by demonstrating widespread expression of smooth muscle-actin (66% of cases), desmin (20% of cases), and vimentin (virtually all cases). Keratin expression does not rule out leiomyosarcoma since 25% of leiomyosarcomas express keratin [10]. The median survival after diagnosis is 2 to 4 years [11]. The only known curative therapy is radical surgery. No systemic therapy has produced a tumor response. Rhabdomyosarcoma of the prostate present at a median age of 4 to 5 (range: newborn to 19 years of age, though cases have been reported in adults). Histologically the majority are embryonal type. Expression of nuclear factors myogenin and myo-D and of cytoplasmic myoglobin confirm a diagnosis of rhabdomyosarcoma. Multimodal therapy (tumor resection, chemotherapy, radiation therapy) has resulted in cures of many patients with bladder preservation; 85% of patients survive at least 5 years. Prognostic parameters of rhabdomyosarcoma include tumor stage, histologic type, and patient age [12]. However, that said, virtually all patients with metastatic disease have died of tumor. Sarcomatoid carcinoma (or, "carcinosarcoma") As with carcinomas of other organs, i.e. breast, lung, pancreas, and urinary bladder, adenocarcinomas of the prostate rarely have a malignant spindle cell component. This may be the predominant component. These tumors are extremely rare in the prostate; fewer than 50 cases have been published. Patients tend to be older (median age 70; range 50 to 90 y.o.) than the median age of patients with pure prostate adenocarcinoma. About half of the cases are associated with a previously diagnosed, pure acinar prostate adenocarcinoma, which is often of high grade. In these patients there is a correlation, albeit weak, with prior radiation or hormonal therapy that is followed by recurrence of tumor as sarcomatoid carcinoma an average of 3 years after initial diagnosis. The other cases co-occur with an acinar prostate adenocarcinoma. Histology: The proportion of the tumor that is sarcomatoid is highly variable. Where the dominant element is sarcomatoid, extensive sampling to find histologic evidence of epithelial differentiation may be necessary. Immunostaining sections for epithelial cell antigens (for transitional/urothelial cell carcinoma, though I am not aware of only sarcomatoid transitional/urothelial cell carcinomas that clinically presented as a prostate tumor: low- and high-MW keratin, keratin 7/20 and p63; for prostate adenocarcinoma, low-MW keratin or PSA) can either highlight rare epithelial tumor cell nests, or can provide indirect evidence of epithelial differentiation in a tumor that is virtually exclusively composed of spindle cells. When the only evidence of epithelial differentiation is the expression of keratin by tumor cells, the caveat that sarcomas have been reported to, infrequently, express keratin should be recollected. The sarcomatoid component may be composed of heterologous elements. In decreasing frequency of occurrence these are osteosarcoma (40% of cases), chondrosarcoma (20%), leiomyosarcoma (5%), rhabdomyosarcoma (5%), and angiosarcoma (1%). The presence of these elements is of no apparent clinical significance. Regarding immunophenotype the sarcomatoid component expresses, in decreasing frequency, vimentin (100%), smooth muscle actin (45%), prostatic acid phosphatase (35%), desmin (5%), keratin (5%), and PSA (2%). A case in which the sarcomatoid element expressed progesterone receptor has been reported [13, 14]. Although the epithelial element in most cases is adenocarcinoma, there are case publications that report a squamous and urothelial epithelial component [15]. In general, carcinomas that exhibit extensive sarcomatous differentiation are associated with a particularly rapid course. The median survival in one series of patients with sarcomatoid carcinomas was 7 months (range: 1 to 48 mos.) following the appearance of sarcomatoid foci, and 36 mos. (range: 8 to 108 mos.) following the diagnosis of adenocarcinoma. The SEER (an NCI program: Surveillance, Epidemiology and End Results) data provides further documentation of the malignant nature of sarcomatoid prostate carcinomas. The histologic elements of metastases are, with equal frequency, pure epithelial, pure sarcomatoid, and sarcomatoid carcinoma. These tumors appear to be non-responsive to systemic therapy, including androgen blockade and chemotherapy. With respect to etiology, the suggestion has been made that radiation, either external beam or brachytherapy, confers an increased risk of prostate carcinosarcoma and/or prostate sarcoma [16]. Given the relatively short duration after radiation (< 5 years after radiotherapy) and rarity of carcinosarcoma, and the even greater rarity of prostate sarcoma, this etiology seems unlikely, unless there are yet to be identified host-specific predisposing factors [17]. Regarding molecular pathogenesis, little is known. That 2 patients exhibited increasing p53 provides a basis for arguing that "the progressive accumulation of p53 suggests increasing clonal dominance of dedifferentiated tumor cells with p53 mutations" [18]. A final point concerns semantics. "Sarcomatoid carcinoma", which has been strictly defined as a spindle cell carcinoma, has the same clinical behavior as "carcinosarcoma", which has been strictly defined as a malignant tumor composed of both carcinoma and sarcoma histologic elements. In the latest WHO Classification of Tumors of the Prostate (2004), the authors state that "given their otherwise similar clinico-pathologic features and identically poor prognosis, these two lesions are best considered as one entity." Finally, the term "collision tumor" [19] has sometimes been applied to sarcomatoid carcinomas of different organs. Since it is now clear that the sarcomatoid component appears to be a histologically metaplastic manifestation of the respective carcinoma, and not a de novo tumor, the term collision tumor is best reserved for co-occurent tumors of adjacent, but different organs, such as concurrent rectal and prostate adenocarcinomas. A number of reviews of sarcomatoid prostate adenocarcinoma have been published [20, 21, 22, 23, 24]. Pertinent discussions are also available in the recently published WHO monograph Tumours of the Urinary System and Male Genital Organs [25]. Epithelial-stromal tumors of prostate Primary epithelial-stromal tumors of the prostate are also extremely rare, consisting for the most part of case reports. These tumors occur in adults (median age of 49 years; range 22 to 78 years of age). Although a variety of different names have been used for these tumors, "phyllodes tumor" and "fibroadenoma" seem the most appropriate since they have gross and microscopic similarities to the corresponding tumors of the breast: When large, they may be cystic; the cystic spaces are lined by the same epithelium that lines the glands [26]. The size of the lesion ranges from incidental, microscopic, proliferative changes to bulky tumors. The stromal component has malignant potential [27]. These tumors are composed of cytologically benign epithelial cells and a stroma that varies widely in cellularity in different cases. Similar to phyllodes tumor of the breast, the histology of the stromal cells contributes to predicting tumor behavior. A proliferative, cellular stroma leading to stromal overgrowth and cellular atypia with increased (>2 to 9 per 10 hpf) mitotic rate predicts likely recurrence of tumor with a significant probability of death due to the neoplasm. These tumors (termed "malignant phyllodes tumor" or "cystosarcoma) are lethal in 75 % of patients due to either bulky local tumor or metastases. Conversely, in the absence of these features, a diagnosis of "low-grade phyllodes tumor (or, cystadenoma)" is appropriate, with the caveat that the course of the tumor in individual cases is unpredictable. The risk of local recurrence of incompletely excised tumor is high. 80% of tumors that were surgically treated by transurethral excision recurred locally. The risk of local invasion and of distant metastases can be predicted base on the histologic features as discussed above. Responsiveness to either hormonal therapy or to chemotherapy is unknown. These tumors are apparently insensitive to radiation therapy [28]. Additional comments Several intriguing observations can be made about sarcomatoid carcinomas: The sarcomatoid element virtually never has functional phenotypic features of the specific epithelium and, when free of heterologous elements (such as skeletal muscle, bone, cartilage), are phenotypically undifferentiated. Sarcomatoid carcinomas exhibit a more malignant course than the carcinoma of the respective organ, growing more rapidly and disseminating as metastases earlier in the course of disease. Sarcomatoid carcinomas are resistant to systemic therapy. Cases such as the present one raise yet-to-be answered questions: Why is a sarcomatoid element so rare? What induces the sarcomatoid phenotype? Why does the sarcomatoid phenotype behave in a particularly malignant manner? Basic science findings based on intermediate filament cell biology provide intriguing clues to some answers. There are > 65 intermediate filaments (originally defined as long, 10 nm diameter cytoplasmic filaments that are distinct from muscle filaments). The broad classes of intermediate filaments include keratins (types I and II), vimentin, desmin, neurofilament, glial fibrillary acidic protein, (nuclear) lamins, etc. The functional complexity of these proteins is exemplified by such functions as cell and nuclear structure/scaffolding, intracellular transport, cell signaling, and cell motility [29]. Known for years is the loss (in general} of keratin immunoreactivity and gain of vimentin expression by the spindle cell element of sarcomatoid carcinomas. Functional features of these respective intermediate filaments that provide an explanation for the more aggressive course of carcinomas with a sarcomatoid component are the following: Cells that express vimentin as their predominant intermediate filament are more rapidly motile than comparable cells expressing keratin as the predominant intermediate filament [30]. Transition of epithelial cells to a mesenchymal phenotype has been correlated with resistance to EGFR inhibitor Erlotinib (shown in non-small cell lung carcinoma] [31]. Although the reason for why some carcinomas undergo this switch in intermediate filament type is unknown and whether the functions reported above for vimentin and some keratins are relevant to the behavior of sarcomatoid carcinomas is unproven, characterizing molecular events begins to provide us with specific tools to better understand the complex biology of tumors, and, in our present case, of such phenotypically unusual tumors as sarcomatoid carcinomas. Summary points (sarcomatoid prostate carcinoma) If it look like sarcoma, rule out carcinoma and post-op spindle cell nodule Look for epithelial nests to confirm the diagnosis of sarcomatoid carcinoma Immunostains may be less helpful than expected in confirming the diagnosis since stromal cells and some smooth muscle cell tumors can (infrequently and, then, only focally) express keratin Sarcomatoid prostate carcinoma is rapidly progressive and resistant to systemic therapy References Huang, W. L., Ro, J. Y., Grignon, D. J., Swanson, D., Ordonez, N. G. & Ayala, A. G. (1990) J Urol 143, 824-6. Proppe, K. H., Scully, R. E. & Rosai, J. (1984) Am J Surg Pathol 8, 101-8. Iczkowski, K. A., Shanks, J. H., Gadaleanu, V., Cheng, L., Jones, E. C., Neumann, R., Nascimento, A. G. & Bostwick, D. G. (2001) Mod Pathol 14, 1043-51. Wee, H. M., Ho, S. H. & Tan, P. H. (2005) Ann Acad Med Singapore 34, 441-2. Gaudin, P. B., Rosai, J. & Epstein, J. I. (1998) Am J Surg Pathol 22, 148-62. Leong, S. S., Vogt, P. J. & Yu, G. S. (1988) Urology 31, 163-7. Karolyi, P., Endes, P., Krasznai, G. & Tonkol, I. (1988) Virchows Arch A Pathol Anat Histopathol 412, 383-6. Ishii, T., Kuroda, K., Nakamura, K. & Sugiura, H. (2004) Hinyokika Kiyo 50, 405-7. Sexton, W. J., Lance, R. E., Reyes, A. O., Pisters, P. W., Tu, S. M. & Pisters, L. L. (2001) J Urol 166, 521-5. Mondaini, N., Palli, D., Saieva, C., Nesi, G., Franchi, A., Ponchietti, R., Tripodi, S., Miracco, C., Meliani, E., Carini, M., Livi, L., Zanna, I., Trovarelli, S., Marino, V., Vignolini, G., Pomara, G., Orlando, V., Giubilei, G., Selli, C. & Rizzo, M. (2005) Eur Urol 47, 468-73. Cheville, J. C., Dundore, P. A., Nascimento, A. G., Meneses, M., Kleer, E., Farrow, G. M. & Bostwick, D. G. (1995) Cancer 76, 1422-7. Lobe, T. E., Wiener, E., Andrassy, R. J., Bagwell, C. E., Hays, D., Crist, W. M., Webber, B., Breneman, J. C., Reed, M. M., Tefft, M. C. & Heyn, R. (1996) J Pediatr Surg 31, 1084-7. Sak, S. D., Orhan, D., Yaman, O., Tulunay, O. & Ozdiler, E. (1997) Urol Int 59, 50-2. Ordonez, N. G., Ayala, A. G., von Eschenbach, A. C., Mackay, B. & Hanssen, G. (1982) Urology 19, 210-4. Rogers, C. G., Parwani, A., Tekes, A., Schoenberg, M. P. & Epstein, J. I. (2005) J Urol 173, 439-40. Moreira, S. G., Jr., Seigne, J. D., Ordorica, R. C., Marcet, J., Pow-Sang, J. M. & Lockhart, J. L. (2004) BJU Int 93, 31-5. Canfield, S. E., Gans, T. H., Unger, P. & Hall, S. J. (2001) Tech Urol 7, 294-5. Delahunt, B., Eble, J. N., Nacey, J. N. & Grebe, S. K. (1999) Anticancer Res 19, 4279-83. Pavelic, J., Lamovec, J., Novak, J., Gall-Troselj, K., Kapitanovic, S. & Pavelic, K. (2000) J Cancer Res Clin Oncol 126, 95-100. Randolph, T. L., Amin, M. B., Ro, J. Y. & Ayala, A. G. (1997) Mod Pathol 10, 612-29. Dundore, P. A., Cheville, J. C., Nascimento, A. G., Farrow, G. M. & Bostwick, D. G. (1995) Cancer 76, 1035-42. Lauwers, G. Y., Schevchuk, M., Armenakas, N. & Reuter, V. E. (1993) Am J Surg Pathol 17, 342-9. Wick, M. R., Young, R. H., Malvesta, R., Beebe, D. S., Hansen, J. J. & Dehner, L. P. (1989) Am J Clin Pathol 92, 131-9. Shannon, R. L., Ro, J. Y., Grignon, D. J., Ordonez, N. G., Johnson, D. E., Mackay, B., Tetu, B. & Ayala, A. G. (1992) Cancer 69, 2676-82. Epstein JI, H. B., Algaba F, Humphrey PA, et al. (2004) Acinar adenocarcinoma (WHO, Lyon). Maluf, H. M., King, M. E., DeLuca, F. R., Navarro, J., Talerman, A. & Young, R. H. (1991) Am J Surg Pathol 15, 131-5. Young, J. F., Jensen, P. E. & Wiley, C. A. (1992) Arch Pathol Lab Med 116, 296-9. Young, R. H. (1995) Am J Clin Pathol 104, 607-10. DePianto, D. & Coulombe, P. A. (2004) Exp Cell Res 301, 68-76. Helfand, B. T., Chang, L. & Goldman, R. D. (2004) J Cell Sci 117, 133-41. Thomson, S., Buck, E., Petti, F., Griffin, G., Brown, E., Ramnarine, N., Iwata, K. K., Gibson, N. & Haley, J. D. (2005) Cancer Res 65, 9455-62.