Low-Grade Myofibroblastic Proliferations of the Urinary Bladder

In human pathology, mesenchymal fibroblastic/myofibroblastic spindle cell proliferations constitute an exceedingly broad and heterogeneous category of disorders, ranging from indolent lesions to highly aggressive neoplasms. In this setting, and especially when dealing with reactive and low-grade lesions, the differential diagnosis is often challenging.¹ 

In 1980, Roth²  reported a particular spindle cell proliferation arising in the urinary bladder of a 32-year-old woman. This lesion shared several similarities with nodular fasciitis, so it was classified as a reactive proliferation because of the absence of malignant epithelial elements, its benign clinical course, and the patient's well-known history of chronic cystitis.²  Four years after the Roth²  initial description, Proppe et al³  reported on 8 cases of similar spindle cell lesions that occurred 5 weeks to 3 months after genitourinary tract interventions. These benign reactive lesions resembling a sarcoma were called postoperative spindle cell nodules (PSCN).³  Later on, several authors reported cases of PSCN that had occurred up to 15 months (usually about 3 months) after urinary bladder surgery or instrumentation.^4–6 

Over time, these myofibroblastic proliferations have had a number of designations, such as inflammatory pseudotumor, inflammatory pseudosarcomatous fibromyxoid tumor, pseudosarcomatous myofibroblastic tumor, fibromyxoid pseudotumor, plasma cell granuloma, plasma cell pseudotumor, and xanthomatous pseudotumour.^6,7  Although most of the above-mentioned terms are now considered outdated, the nosology and definition of these lesions are still a matter of debate. Inflammatory myofibroblastic tumor (IMT) of the urinary bladder is probably still the most frequent designation^8–19 ; however, some authors prefer the term pseudosarcomatous myofibroblastic proliferation (PMP) for a fraction of these myofibroblastic proliferations. These authors believe that, both morphologically and biologically, there are sufficient reasons to distinguish IMT and PMP. In addition to histopathologic differences (see below) and even though PMPs seem to be clonal proliferations with a potential for aggressive, local growth and recurrence, they are characterized by an excellent outcome and do not have the same malignant potential as IMT. These same scientists also group PSCNs with PMPs.^20,21 

This review will focus on what is currently known about these low-grade myofibroblastic proliferations (LGMPs) of the urinary bladder—which is the term we have chosen to embrace all of the above designations and entities.

A PubMed (US National Library of Medicine, Bethesda, Maryland) search was carried out to identify original articles focusing on this subject. We have based our review of the literature on selected study articles and textbooks, recently published review articles, editorials from peer-reviewed journals, and the reference list of each searched publication. All of these manuscripts were full-text, English-language, and published between 1980 and 2012. When necessary, we contacted by e-mail the corresponding authors of the article to obtain clarifications and/or details about the cases on which they had reported.

Because of the rarity of LGMPs affecting this anatomic site, we obtained most of the clinicopathologic features from an overall patient population composed of the most-relevant cases reported in the English-language literature (Table 1).^{2–6,8,12,15,17–50}  Because of the relative paucity of specific articles dealing with PMP and/or PSCN, fewer data were available regarding these entities, compared with those dealing with IMT. Information regarding tumor designation, patients' characteristics, primary symptoms, tumor size, urinary bladder infiltration, treatment, and follow-up was obtained from each selected publication. The LGMPs shown in the various figures are unpublished and are selected from files belonging to our Section of Pathology and the Department of Pathology, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS), Arcispedale Santa Maria Nuova (Reggio Emilia, Italy).

Among the LGMPs, IMT, which was first described in the lung, is the most-frequently reported entity.⁵¹  In the genitourinary tract, the urinary bladder is the most common anatomic site of origin of this lesion; additional locations include the kidney, urethra, prostate, ureter, and rete testis.⁵² 

Our review of the most-salient literature concerning cases of LGMP occurring in the urinary bladder (Table 1) showed that most of those cases occur during the fourth and fifth decades of life, with men more frequently affected than women (1.33:1 ratio), and without any perceptible difference between IMT and PMP. Overall, this male predominance seems to be less evident when considering patients younger than 18 years. The case distribution during the various decades of life is shown graphically in Figure 1.

In LGMPs, painless hematuria was the most-common symptom, whereas dysuria, pelvic pain, and symptoms related to obstruction or infections occurred more rarely (Table 1). Unlike extragenitourinary tract IMT, systemic symptoms—such as fever and weight loss—are uncommon in urinary bladder locations.⁵³  In only one case were myalgia, fatigue, and fever reported; however, at least in part, these signs and symptoms could have been secondary to an underlying systemic lupus erythematosus.⁴⁸ 

Grossly, LGMPs of the urinary bladder may be either polypoid masses or, less frequently, rather flat nodules, with or without surface ulceration. On cut section, their color is gray-white to yellow with a macroscopic appearance from fascicular to myxoid.^7,20,52  In the cases reported (Table 1), at first diagnosis, the tumor's greatest diameter was between 1 and 11 cm (median, 4 cm), with no difference between patients aged older than or younger than 18 years, and no site predilection.

Three main histopathologic patterns of IMT have been described: (1) a myxoid/vascular pattern (Figure 2, A) characterized by spindle/stellate myofibroblasts with abundant eosinophilic cytoplasm and vesicular nuclei; (2) a compact, spindle cell proliferation (Figure 2, B)—sometimes with some ganglion-like cells—arranged in fascicular or storiform patterns, with varying degrees of cellularity; and (3) a hypocellular, fibrous pattern (Figure 2, C), displaying abundant collagen, plasma cells, lymphocytes, and eosinophils (similar to desmoid fibromatosis).⁵² 

These patterns are also typically characterized by varying amounts of acute (with neutrophils and/or eosinophils), chronic (lymphoplasmacellular), or mixed inflammatory infiltrates (Figure 2, D). Myofibroblasts show elongated-to-oval nuclei, often with prominent nuclei, and do not exhibit clear-cut atypia. The reported mitotic index ranges from 0 to 20 mitoses/10 high-power fields with most IMTs having a mitotic count of less than 5.²⁰  Based on Table 1, muscularis propria infiltration is common (about 60% of the whole patient population), whereas perivesical soft tissue invasion is not frequent (about 5% of the cases, without any apparent difference between IMT and PMP).

Necrosis is usually quite rare to absent in LGMPs of the urinary bladder (Figure 2, E). Iczkowski et al⁵  found no necrosis in any of their cases and, in addition, suggested it as key to a diagnosis of sarcoma. Montgomery et al⁶  detected necrosis in 30% of their patient population with IMT, whereas Harik et al²⁰  found it in 22 of 42 lesions (52%), even though it was mainly focal, minimal, and confined to the surface. Seven of those 22 cases (32%) showed necrosis (extensive in 2 of 7; 29%) in the deep urinary bladder wall and were associated with muscularis propria invasion.²⁰  Overall, when present, necrosis in LGMPs is rarely extensive and/or is located at the tumor-detrusor interface. Lesions with these findings should be carefully evaluated to rule out any malignancy.

Some authors^20,21,46,54  claim that, in the urinary bladder, IMT and PMP should be considered separate entities with sufficiently different clinical behavior to warrant their distinction on morphologic grounds. According to these scientists, the distinguishing histologic features of PMPs are haphazard-to-loose, fascicular arrangements of spindle cells, many of which have elongated, bipolar cytoplasmic processes that are often eosinophilic (rhabdomyoblast-like) (Figure 2, F); edematous-to-myxoid stroma with prominent vascularity and a variably dense, acute, and/or chronic inflammatory infiltrate; occasionally, focal, more-cellular fascicular areas, particularly in the deeper portions; rare occurrences of storiform and hypocellular, fibrous patterns; and inflammatory infiltrates that are generally less dense than they are in IMT.⁵⁴  The malignant potential of IMT of the genitourinary tract is seen in anecdotal cases of transformation into true inflammatory fibrosarcoma, which is characterized by increasing atypia and cellularity in a background superimposable on that of classic IMT (see index case in Figure 3, A through C).⁶ 

Moreover, the synchronous occurrence of a transitional-cell carcinoma, as either a separate nodule or admixed, has been reported sporadically (Table 1).⁷  Interestingly, metastatic urothelial carcinoma may exhibit a similar myofibroblastic proliferation, whereas the corresponding primary tumor may not.⁵⁵ 

In the LGMPs of the urinary bladder, there is significant immunohistochemical overlap between myofibroblastic proliferations diagnosed as IMT or PMP. In fact, they express vimentin (95%–100% of cases), smooth-muscle actin (48%–100%; Figure 4, A), AE1/3 (73%–78%; Figure 4, B), desmin (5%–80%; Figure 4, C), muscle-specific actin (62%), clone OSCAR keratin (70%), and CAM 5.2 (70%), whereas CK5/6 and 34βE12 seem not to be expressed.^1,6,10,52,54  Bladder IMT is more prone to express keratins than is IMT occurring in other genitourinary tract locations, and, rarely, findings are positive for myogenin or MyoD1.⁵²  Unlike bladder sarcomas, in IMT, p53 nuclear positivity is weak or even negative.⁵ 

Immunohistochemical analysis with an antianaplastic lymphoma kinase-1 (ALK-1) antibody is specific but only slightly sensitive for IMT (provided that myogenin and MyoD1 findings are negative because rhabdomyosarcoma may seldom express these markers).¹  Results for ALK-1 were found to be positive in 20% to 89% of urinary bladder IMTs (Figure 4D).^{1,6,10,20,42,43,46} 

The pathologic differential diagnosis of bladder LGMPs is summarized in Table 2.

Radiologic imaging plays a key role in the preoperative management and follow-up of patients with LGMP. Ultrasound, contrast-enhanced computed tomography, and magnetic resonance images provide valuable information on tumor size and infiltration of the urinary bladder wall.⁵⁶  On ultrasound, according to the particular case, LGMP can appear as either a well-defined, nonspecific mass or as simple bladder-wall thickening with variable echogenicity, and the lesion being described as hypoechoic or hyperechoic. On ecocolor Doppler (or ecopower Doppler), and especially on contrast-enhanced ultrasound examinations, LGMPs may appear as hypervascular nodules.^57–59  Computed tomography (Figure 5) may show either a focal or diffuse bladder-wall thickening and, more rarely, a polypoid mass; the lesion may be hypodense, isodense, or hyperdense and well enhanced after intravenous contrast medium administration. On magnetic resonance imaging, LGMP is characterized by low T1 signal intensity and nonhomogeneous high T2 signal intensity. In some cases, on T2 imaging, LGMP may show high signal intensity centrally, reflecting central necrosis, and low signal peripherally. Postcontrast, T1-weighted magnetic resonance imaging demonstrates heterogeneous enhancement with nonenhanced areas of necrosis.⁵⁹ 

Clonal genetic aberrations in the short arm of the chromosome 2 (region 21–23 of the long arm) have been described in approximately 50% to 60% of IMTs. This rearrangement involves the 2p23 region that contains the ALK gene, which encodes a tyrosine kinase receptor expressed predominantly in the central nervous system and is a member of the superfamily of insulin growth-factor receptors.⁶⁰  Although fluorescence in situ hybridization (FISH) in many IMTs has shown ALK gene rearrangement, in most cases, the fusion partner has not been identified. The ALK gene rearrangements identified that have led to fusion with other genes include ATIC,⁴¹  CARS,^61,62  TPM3,^63–65  TPM4,^63,65,66  TPM3 and TPM4,⁶⁷  CLTC,^60,65,68  RANBP2,^60,69  and SEC31L1 (SEC31A).⁷⁰  The ALK-ATIC rearrangement has been reported in bladder IMTs.⁴¹ 

In normal ALK signaling, ligand-induced homodimerization of the extracellular domains brings the tyrosine kinase domains into sufficient proximity to enable transphosphorylation and kinase activity, whereas translocations resulting in pathogenic fusion partners provide dimerization domains that are ligand-independent, leading to unregulated constitutive kinase activity.⁷¹ 

Fluorescence in situ hybridization analysis can be used to detect ALK gene break-apart rearrangements in IMT (Figure 6). The FISH kits are a mixture that consists of 2 fluorophore-labeled DNA probes in hybridization buffer containing ALK-upstream and ALK-downstream probes characterized by 2 different colors. The hybridization targets of these 2 probes are situated on opposite sides flanking the breakpoint of the ALK gene. Nuclei with ALK gene rearrangement display the normal ALK region as a 2-color fusion signal, whereas, if a breakapart has occurred, the 2-color signals split, separating them.⁵²  Using FISH analysis, rearrangements of the ALK gene have been detected in 67% to 72% of IMTs of the urinary bladder.^6,20,43,46  Some PMPs and PSCNs show positive immunohistochemistry for ALK and/or express ALK gene rearrangements.^20,52  Detection of ALK gene rearrangements by FISH is fundamental to distinguishing IMTs from other spindle cell proliferations of the urinary bladder, especially when the immunohistochemical analysis is not conclusive.⁴⁶ 

Some studies on extragenitourinary tract IMT have investigated its etiology; however, its pathogenesis remains obscure. In addition, no studies have clarified the pathogenesis of genitourinary IMT. An underlying infection has long been thought to be the primum movens of this spindle cell proliferation. Various infectious agents have been detected in IMTs, such as bacteria (Bacteroides corrodens, Klebsiella pneumoniae, Mycobacterium avium, Corynebacterium jejuni, Bacillus sphaericus, Escherichia coli, and Coxiella burnettii) and viruses, the most-frequently reported being Epstein-Barr virus and human herpesvirus 8.^52,72,73 

An autoimmune etiology of IMT has been postulated based on single reports. In one case, an IMT of the submandibular gland was found in a patient with polyclonal hypergammaglobulinemia, high titers of antinuclear antibody, but without any signs of known autoimmune disease.⁷⁴  Additional reports described cases of IMT of the spleen with thrombocytopenic purpura and IMT with Riedel thyroiditis.⁵² 

In 20 LGMPs listed in Table 1, previous urinary bladder instrumentation was ascertained, so that they were labeled as PSCN.

As shown in Table 1, 65% (185 of 285) of the reported cases were treated with transurethral resection (TUR), 27% (77 of 285) with partial cystectomy, 6% (16 of 285) with total cystectomy, and 3% (7 of 285) with other surgical procedures. In fact, surgical management is a sign of the controversy surrounding the nature of LGMPs, the consequent limited workload of such lesions for a single surgeon, and the anecdotal and subjective judgments.^6,12,20,45 

Based on available data from the literature, even though the malignant potential of LGMP of the genitourinary tract is still uncertain and is sometimes difficult to evaluate, the indolent and often benign clinical course that characterizes most of the cases does not warrant aggressive therapy (radical cystectomy, radiation, or chemotherapy).⁵²  On the other hand, follow-up cystoscopy and biopsies are mandatory to detect potential recurrences early on, especially in large, multinodular, and/or incompletely resected lesions.⁷⁵ 

The World Health Organization classification of soft tissue tumors classifies IMTs as neoplasms of intermediate malignancy, subtype rarely metastasizing.⁷⁶  According to this classification, extrapulmonary IMTs have a recurrence rate of approximately 25% (related to location, resectability, and multinodularity), whereas only a few cases (<5%) metastasize.⁵³  In most cases, it is difficult to predict the recurrence and/or malignant potential on the basis of histopathologic findings alone.⁷⁶ 

In our review of the literature (Table 1), data on patients' follow-up were available in 218 cases. Only 7% (15 of 218) of patients with urinary bladder myofibroblastic proliferations experienced local recurrence: 5% (10 of 218) with IMT, or similar designations, and 2% (25 of 218) with myofibroblastic lesions clearly labeled as PMP. None of them developed metastases or died from the disease.

In 5 of 218 cases (2.3%) there was a synchronous carcinoma (4 urinary bladder transitional-cell carcinomas and 1 renal clear-cell carcinoma). Of 9 deaths, 1 patient (of 9; 11%) died from urosepsis, 1 (of 9; 11%) from postoperative infection, 3 (of 9; 33%) from unrelated causes, 1 (of 9; 11%) of inflammatory fibrosarcoma, 2 (of 9; 22%) from synchronous bladder transitional-cell carcinoma, 1 (of 9; 11%) from neurofibromatosis complications, and 1 (of 9; 11%) from unknown causes.

We need more data to explore ALK-1 as a prognostic factor in IMT and related myofibroblastic lesions of the urinary bladder.

Although more than 30 years have passed since its first description, several questions remain unanswered regarding LGMP of the urinary bladder. The most urgent problem seems to be both the nomenclature and the real nature of these lesions.

Some authors feel that PMP and PSCN represent the same entity, which should be separated from true IMT. Grounds for this classification are a fairly distinct histopathologic picture and the quite different clinical behavior, with IMT representing proliferations with some malignant potential.^20,21,46,54  At the opposite end of that spectrum, the inflammatory fibrosarcoma shares a similar background with clearly malignant spindle cells and, above all, a much worse prognosis.⁶  A widespread unifying theory hypothesizes that such lesions would be part of a single continuum, ranging from benign pseudosarcomatous lesions to low-grade sarcomas.⁵²  Molecular genetics does not yet help us to solve the arcana of these LGMPs. In fact, only a fraction of these cases (but a fraction that includes IMT, PMP, and PSCN cases) harbor ALK gene rearrangements with evidence of clonal proliferation, whereas the remaining LGMPs are either reactive or due to an unknown genetic abnormalities.^20,52  Unfortunately, so far, ALK expression and ALK gene rearrangements are unable to predict prognosis. In addition, in most cases, the histopathologic appearance alone does not help us forecast the clinical behavior of a given myofibroblastic proliferation of the urinary bladder.

The LGMPs of the urinary bladder include different lesions extending from benign, pseudosarcomatous entities to low-grade, inflammatory fibrosarcomas, and they mainly affect patients in their middle decades. With the exception of frank sarcomas, these lesions should be considered of uncertain clinical behavior, even though no metastatic cases in this anatomic location have been reported in the English-language literature. However, in view of the indolent clinical course, in most cases, aggressive treatments (such as radical cystectomy, chemotherapy, and radiotherapy) are probably unjustified.⁵² 

Further investigations dealing with the molecular genetics in LGMPs of the urinary bladder will certainly help us to clarify their nature and to establish more-sound classifications and prognostic criteria.

We are indebted to Gabriella Becchi, B. Tech, for her competent technical assistance. In addition, we would like to express our gratitude to the various authors who provided additional information on their published cases. Lastly, a special thank to William R. Sukov, MD, Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota, for providing Figure 6; to Alberto Cavazza, MD, Department of Pathology, IRCCS, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy, for his precious help in retrieving adequate histologic material and to Augusto Vaglio, MD, PhD, Department of Nephrology, University Hospital, Parma, Italy, for his critical review of this manuscript.