Primary Neuroendocrine Tumors of the Ureter: A Short Review

Urothelial carcinoma is relatively frequent in the United States, with 54 000 new patients diagnosed with the disease each year.¹  Most of these tumors arise in the urinary bladder, while primary upper urinary tract urothelial carcinoma represents only 5% of urothelial neoplasms.²  On the other hand, neuroendocrine tumors (NETs) of the urinary system are extremely rare, accounting for less than 0.05% of primary urinary tract (UT) malignancies. Primary urinary NETs are most commonly located in the bladder, where they account for less than 1% of all bladder malignancies.^3,4  To the best of our knowledge, the first case of ureteral NET was published in 1986, and fewer than 30 cases have been reported since, including tumors with a mixed urothelial/neuroendocrine histology.⁵  Even though primary ureteral NETs are believed to show a male predominance, several cases have occurred in women, most often after the sixth decade of life.^6–10  Interestingly, ureteral neuroendocrine neoplasms seem to occur more frequently in Asians than in any other ethnic group (Table). Although some of the reported cases include a history of tobacco consumption, a relationship between smoking and urinary tract NETs has not been well established.^7,9 

The pathogenesis of ureteral NETs is debated, because the normal ureters are thought to be devoid of neuroendocrine cells. Interestingly, a thorough systematic immunohistochemical and histochemical search for neuroendocrine cells in the genitourinary system of the female pig demonstrated that the neuroendocrine cell population density shows a concentration gradient. The highest concentration was found in the vulvar/distal urethral region and decreased steeply toward the proximal genitourinary tract. In fact, this study¹¹  found only occasional neuroendocrine cells in the bladder, and none in the ureters. Therefore, 4 different hypotheses have been formulated to explain the origin of ureteral NETs. The first one posits that the original lesion is a urothelial carcinoma, which undergoes neuroendocrine metaplasia. The next 2 hypotheses suggest that ureteral NETs are derived either from neuroendocrine cells normally residing in the ureter (which has not been demonstrated yet) or from neural crest cells abnormally entrapped during embryonic migration. Finally, similar to NETs arising in other organs that lack neuroendocrine cells, some authors^6,10,12,13  support a cancer stem cell theory. A recent study documented that Notch1 and its target Hes1 show significantly diminished immunohistochemical positivity in the neuroendocrine phase of biliary mixed adenoneuroendocrine carcinomas when compared to both the adenocarcinoma component and the normal biliary epithelium. Moreover, Notch1 mRNA knockdown in a cholangiocarcinoma cell line induced an increase in the concentration of chromogranin A and Ascl1 (a neuroendocrine inducer normally inhibited by Notch1 target Hes1) mRNA.¹⁴  This suggests that, at least in some epithelia, neuroendocrine differentiation might be actively suppressed. Consequently, the loss of normal suppression mechanisms during urothelial carcinogenesis might allow transdifferentiation to a NET. While it is uncertain whether this mechanism operates in vivo, or if it can be extrapolated from biliary to urinary NET pathogenesis, it could potentially explain why many reported ureteral NETs show a mixed histologic profile.^6,15,16 

Clinical presentation is widely variable, from asymptomatic microscopic hematuria to complete ureteral obstruction and hydronephrosis. The most common initial manifestations are flank pain and hematuria.¹⁰  Locally advanced disease at presentation, as well as lymph node metastasis, are not uncommon.^6,17  Contrary to large masses in the urinary bladder, which can cause bilateral ureteral obstruction, upper UT NETs are usually unilateral and do not cause oligoanuria in binephric patients.

Besides, since unilateral obstruction in ureteral NETs ensues slowly, serum creatinine concentration would not be expected to increase dramatically if renal function was previously normal. Tumor markers are usually not helpful. In contrast, urine sediment frequently reveals different degrees of hematuria and leukocyturia.^6,7,17–19  If obstruction is complete, urine cultures may show growth of bacteria that are not the usual uropathogens.⁷  Cytology specimens can demonstrate malignant neuroendocrine cells, especially if obtained during an endoscopic study.^6,7 

Ultrasonography is usually the first imaging study performed and, although it is effective at detecting ureteropelvic distention, it is insensitive for the demonstration of ureteral tumors. Intravenous and retrograde urography can indirectly detect the presence of a luminal mass, by showing a filling defect of the ureteral lumen. However, most upper UT masses are initially identified with helical computed tomography or magnetic resonance imaging.^6,10,17,19  Finally, histologic diagnosis is usually made on tissue biopsy specimens obtained through endoscopic ureteroscopy.

Ureteral NETs should be distinguished from other causes of upper UT obstruction and hematuria. Differentiation between ureteral NETs and urolithiasis is quite straightforward thanks to helical CT. Moreover, lithiasis usually causes sudden obstruction and acute pain, while tumoral obstruction tends to ensue slowly, producing subacute or chronic pain. More problematic is the differentiation of NETs from other, much more common ureteral malignancies. Although the morphology of NETs is very characteristic on hematoxylin-eosin–stained sections, immunohistochemical studies are an essential tool for differential diagnosis. It is particularly important to rule out the presence of a NET when urothelial and/or squamous carcinoma are diagnosed, since ureteral NETs can have a mixed histologic profile, and the prognosis seems to be determined by the neuroendocrine component.^6,15–17  Although a NET metastatic to the ureter would be exceptionally rare, it has been reported at least once in the literature.²⁰  Therefore, before making a diagnosis of primary ureteral NET, it is necessary to rule out metastatic disease from primary sites where NETs are more commonly found, such as the digestive system and the respiratory tract.²¹ 

Macroscopically, ureteral NETs are solid sessile tumors with a firm consistency, gray to white smooth cut surface, and peritumoral mural thickening.^6,7  Their microscopic appearance is similar to that of neuroendocrine tumors from other locations, and they are invariably malignant (Figure 1, A through C). Cells are usually small- to medium-sized with speckled chromatin, scant to moderate amounts of cytoplasm, and high nuclear to cytoplasmic ratio (Figure 1, D). Most of these neoplasms show typical small cell neuroendocrine carcinoma histology, with only 1 large cell neuroendocrine carcinoma reported to date, to the best of our knowledge.⁷  The architecture shows patterns that are similar to NETs from other primary locations: rosettes, diffuse sheets, and solid tumor nests. A desmoplastic stromal reaction is commonly observed, and additional nonneuroendocrine tumoral elements can be identified in more than 50% of the cases (Table).^6,7  Necrosis, lymphovascular invasion, and direct invasion of adjacent structures have been reported.^6,7,10,19  Furthermore, mixed ureteral neoplasms containing urothelial carcinoma and/or squamous carcinoma in addition to a NET have been well documented.^6,15–17  Ureteral NETs are almost invariably aggressive neoplasms, with an elevated number of mitotic figures per high-power field.

Immunohistochemical stains for synaptophysin (Figure 2, A), chromogranin A (Figure 2, B), neuron-specific enolase, and/or CD56 (Figure 2, C) show positivity in ureteral NETs. At least 1 ureteral NET has demonstrated positivity for cytokeratin 20, cytokeratin 7, pan-keratin, and epithelial membrane antigen, but uroplakin III invariably shows negativity.⁷  Therefore, uroplakin III seems to be an additional tool for differentiating ureteral NETs from urothelial carcinoma, with the caveat that this marker shows positivity in 57% to 81% of urothelial neoplasms.^22,23  Nuclear Ki-67 labeling index is frequently greater than 50% to 60% (Figure 2, D), which is consistent with the high mitotic rate normally found in these tumors.^7,10 

Electron microscopy studies often reveal the presence of membrane-bound electron-dense cytoplasmic granules measuring 100 to 200 nm in diameter, which are characteristic of neuroendocrine cells.⁷ 

The treatment of ureteral NETs is far from well established. Oncologists seem to favor a multimodal therapeutic approach, including surgery and adjuvant chemotherapy with or without radiation.²⁴  A review of 39 patients with upper UT NETs (ureter-renal pelvis) showed that patients treated with surgery alone had a median survival of 12 months, while patients receiving adjuvant chemotherapy survived for a median of 24 months. However, statistical significance was not reached. Platinum-based regimens seem to confer an important survival advantage, compared with schemes lacking platinum-derived drugs.^25,26  Among the 25 cases of ureteral NETs found in the literature, only 9 patients received chemotherapy, and 4 were treated with platinum-based regimens. On the basis of results from 1 case, another group¹⁰  has suggested that an adjuvant regimen combining irinotecan and platinum would add to the therapeutic benefit without increasing toxicity. Furthermore, neoadjuvant administration of alternating ifosfamide/doxorubicin and etoposide/cisplatin to 30 patients with urinary tract NETs yielded promising results, especially for those patients with resectable disease at initial presentation.²⁷ 

Although sufficient data are lacking to accurately estimate the prognosis of ureteral NETs, reports thus far indicate that patients have fared poorly. Most present with locally advanced disease (T3), and up to one-fifth show lymph node involvement (Table). Clinical course and overall prognosis seem to be similar for pure and mixed ureteral NETs, hence the same treatment strategies have been used for both types of neoplasms.²⁵  Review series of renal and upper UT NETs show a median survival of 8 to 15 months, a high rate of metastatic dissemination (32%–53.8%, synchronous or metachronous), and frequent relapses.^25,26  For practical purposes, the most important prognostic factor seems to be the pathologic stage at diagnosis.²⁵