Incidence of downstaging and complete remission after neoadjuvant chemotherapy for high-risk upper tract transitional cell carcinoma




The authors evaluated the incidence of pathologic downstaging and complete remission (CR) in patients with high-grade ureteral and renal pelvic transitional cell carcinoma (TCC) (upper tract TCC) who received neoadjuvant chemotherapy followed by surgery.


The study group comprised patients with biopsy-demonstrated, high-grade disease who received neoadjuvant chemotherapy followed by nephrouterectomy from 2004 to 2008, during which time patients uniformly were considered for neoadjuvant chemotherapy. The control group comprised patients with biopsy-demonstrated, high-grade disease who underwent initial nephroureterectomy from 1993 to 2004, when patients uniformly underwent initial surgery. Multiple clinical and pathologic features were evaluated, and the primary endpoint was pathologic tumor classification.


One hundred seven patients in the control group underwent initial surgery, and 43 patients in the study group received neoadjuvant chemotherapy. Baseline demographics were similar between the groups except for a higher rate of sessile tumor architecture in the study group (72.1% vs 49.5%; P = .018). There was significant downstaging in study group patients compared with the historic control group (P = .004). The incidence of tumors classified as pathologic T2 (pT2) or as pT3 or higher was significantly lower in the study group (pT2, 65.4% vs 48.8%; P = .043; pT3 or higher, 47.7% vs 27.9%; P = .029). Fourteen percent of patients who received neoadjuvant chemotherapy had a pathologic CR.


Neoadjuvant chemotherapy was associated with a 14% CR rate and a significant rate of downstaging. While longer follow-up is awaited for survival data to mature, the current data provide justification for the sustained support of trials using this strategy. Cancer 2010. © 2010 American Cancer Society.

Survival rates for patients with high-grade and high-stage upper tract transitional cell carcinoma (UTCC) are poor; patients with locally advanced disease experience a 50% or greater risk of progression and less than a 2-year median survival duration.1-4 Contemporary analyses indicate that there has been no improvement in survival rates in the past several decades for patients with high-grade and high-stage disease.1 Some have advocated a paradigm shift in the treatment of such high-risk patients toward multimodality management that adds neoadjuvant chemotherapy to the traditional surgical consolidation; however, there are minimal data, randomized or otherwise, supporting this more aggressive approach.1, 5 Any interim data that demonstrate significant pathologic downstaging and complete remission (CR) rates (as evidenced by pathologic T0 [pT0] disease) could provide surrogate data to justify the continuation of current trials that incorporate the neoadjuvant approach. With this objective in mind, we evaluated the possibility of pathologic downstaging in patients with high-risk UTCC who received neoadjuvant chemotherapy followed by surgical consolidation by comparing their pathologic outcomes with those of a historic control group who had similar clinical high-risk features but underwent initial surgery.


We performed a retrospective review of patients who underwent nephrouterectomy at The University of Texas M. D. Anderson Cancer Center from 2004 to 2008. Patients with biopsy-demonstrated, high-grade disease who received neoadjuvant chemotherapy were selected as the study group. During this period, patients uniformly were considered for neoadjuvant chemotherapy followed by surgery if high-risk features were present. High-risk features primarily entail the finding of high-grade disease on biopsy, sessile architecture, and a large tumor burden (defined subjectively as being observed on computed tomography). The control population was obtained from a database of patients who were characterized previously; those with biopsy-demonstrated, high-grade disease who underwent initial nephroureterectomy during 1993 to 2004 were selected.1 During this previous era, neoadjuvant chemotherapy was considered infrequently (<9% of patients); it was given in extreme circumstances, such as patients who had overt evidence of clinical T4 disease or grossly involved lymph nodes.1 Thus, most patients uniformly had undergone initial nephroureterectomy. The following clinical and pathologic features were evaluated: patient age, sex, laterality, Eastern Cooperative Oncology Group performance status, prior history of bladder cancer, location of tumor, tumor architecture, type of neoadjuvant chemotherapy, radiographic tumor size, surgical technique, pathologic classification, pathologic lymph node classification, number of lymph nodes removed, and the presence of extranodal extension, lymphovascular invasion, carcinoma in situ (CIS), and multifocality. A pathologic CR was defined as the absence of any malignant elements in the resected surgical specimen. This study was approved by the institutional review board.

Initial statistical power calculations were made on the basis of previous data indicating an approximate 66% probability that disease will be classified as pT2 or higher when high-grade disease is encountered on biopsy.6, 7 We hypothesized that the incidence of disease classified as pT2 or higher would be reduced in the setting of neoadjuvant chemotherapy. By using a 2-sided Fisher exact test with a significance level of .05, to have 80% power to detect a difference, a reduction of high-stage disease from 67% to 40% would require 44 patients, a reduction to 35% would require 28 patients, and a reduction to 33% would require 24 patients. The Fisher exact test and the chi-square test were used to evaluate the association between clinical and pathologic variables. Differences in variables that had a continuous distribution across dichotomous or ranked categories were assessed using the Mann-Whitney U test or the Kruskal-Wallis nonparametric analysis of variance, respectively. All reported P values are 2-sided, and significance was set at P ≤ .05. All statistical tests were performed with SPSS software (version 13.0; SPSS Inc., Chicago, Ill).


Among patients with biopsy-demonstrated, high-grade disease, 107 patients in the historic control group underwent initial surgery, and 43 patients in the contemporary study group received neoadjuvant chemotherapy before they underwent surgery. The association of clinical and pathologic variables between the 2 groups is shown in Table 1. Baseline demographics were similar between the groups, except that patients in the study group underwent laparoscopic excision more frequently, reflecting current trends in the surgical management of this disease. With regard to tumor features, there was a higher incidence of sessile architecture in the study group (72.1% vs 49.5%; P = .018), reflecting enrichment of patients with high-risk features.

Table 1. Comparison of Clinical and Pathologic Characteristics of 150 Patients Who Underwent Radical Nephroureterectomy With or Without Neoadjuvant Systemic Chemotherapy for High-Grade Urothelial Carcinoma of the Ureter and Renal Pelvis
CharacteristicNo. of Patients (%)P
All PatientsNeoadjuvant Systemic Chemotherapy
  • ECOG indicates Eastern Cooperative Oncology Group; SD, standard deviation; UC, urothelial carcinoma; Tis, tumor in situ; N+, positive lymph node status; CIS, carcinoma in situ.

  • a

    Unknown in 1 patient.

  • b

    Measurable radiographic tumor size was available in 80 patients.

  • c

    Patients with Nx and N0 were combined for statistical analysis.

No. of patients150 (100)107 (71.3)43 (28.7) 
 Women55 (36.7)39 (36.4)16 (37.2) 
 Men95 (63.3)68 (63.6)27 (62.8).930
ECOG performance statusa
 087 (58)61 (57)26 (61.9) 
 163 (42.3)47 (43.9)16 (38.1).583
Age: Median ± SD, y69.0 ± 10.968.5 ± 10.970.5 ± 10.8.529
Previous bladder UC diagnosis
 Yes80 (53.3)58 (54.2)22 (51.2) 
 No70 (46.7)49 (45.8)21 (48.8).857
Surgical technique
 Open120 (80)93 (86.9)27 (62.8) 
 Laparoscopic30 (20)14 (13.1)16 (37.2).001
Tumor size: Median ± SD, cmb3.6 ± 1.93.8 ± 2.13.6 ± 1.4.957
Pathologic tumor classification
 T06 (4)0 (0)6 (14) 
 Tis10 (6.7)7 (6.5)3 (7) 
 Ta17 (11.3)11 (10.3)6 (14) 
 T127 (18)19 (17.8)8 (18.6) 
 T232 (21.3)22 (20.6)10 (23.3) 
 T349 (32.7)40 (37.4)9 (20.9) 
 T49 (6)8 (7.5)1 (2.3).004
Pathologic lymph node statusc
 Nx46 (30.7)42 (39.3)4 (9.3) 
 N081 (54)50 (46.7)31 (72.1) 
 N113 (8.7)8 (7.5)5 (11.6) 
 N210 (6.8)7 (6.5)3 (7).707
No. of lymph nodes removed: Median ± SD7 ± 95 ± 6.712 ± 10.4<.001
No. of positive lymph nodes: Median ± SD1 ± 4.41 ± 1.81 ± 6.9.227
Extranodal extension in patients with lymph node metastases
 Absent14 (60.9)7 (46.7)7 (87.5) 
 Present9 (39.1)8 (53.3)1 (12.5).086
Stage grouping with pT2 threshold
 Organ confined: <pT2, any N60 (40)37 (34.6)23 (53.5) 
 Locally advanced: ≥pT2, any N90 (60)70 (65.4)20 (46.5).043
Stage grouping with pT3 threshold
 Organ confined: <pT3, N087 (58)56 (52.3)31 (72.1) 
 Locally advanced: ≥pT3 and/or N+63 (42)51 (47.7)12 (27.9).029
Index tumor location
 Renal pelvis94 (62.7)63 (58.9)31 (72.1) 
 Ureter43 (28.7)33 (30.8)10 (23.3) 
 Ureteroenteric anastomosis13 (8.7)11 (10.3)2 (4.7).274
Tumor architecture
 Papillary66 (44)54 (50.5)12 (27.9) 
 Sessile84 (56)53 (49.5)31 (72.1).018
Lymphovascular invasion
 Absent88 (58.7)60 (56.1)28 (65.1) 
 Present62 (41.3)47 (43.9)15 (34.9).361
Concomitant CIS
 Absent77 (51.3)51 (47.7)26 (60.5) 
 Present73 (48.7)56 (52.3)17 (39.5).206
 Absent86 (57.3)58 (54.2)28 (65.1) 
 Present64 (42.7)49 (45.8)15 (34.9).274

The distribution of pathologic stage is represented in Figure 1. There was a significant reduction of pathologic stage in patients who received neoadjuvant chemotherapy compared with those who underwent initial surgery (P = .004). The overall incidence of patients who had pT2 or higher disease was reduced significantly in the study group (46.5% vs 65.4%; P = .043). When grouping was based on pT3 or higher disease, the reduction was greater, and the association was stronger (27.9% vs 47.7%; P = .029). In addition, 14% of patients who received neoadjuvant chemotherapy had a CR, whereas no patient in the control group did.

Figure 1.

Pathologic tumor classification is illustrated for patients who underwent initial surgery and patients who received neoadjuvant chemotherapy before surgery. A shift toward the left was observed in those who received chemotherapy. CIS/A indicates carcinoma in situ and stage pTa grouped together.

Patients in the study group were more likely to have undergone lymphadenectomy and to have more lymph nodes removed than patients in the historic control group (median number of lymph nodes removed, 12 vs 5; P < .001), again reflecting current trends in the surgical management of this disease and paralleling the management of bladder cancer, in which greater import is given to lymphadenectomy and improving lymph node harvest in the current era.8-12 There also was a trend toward lower rates of extranodal extension in the study group. Although no statistical difference was observed in the rate of lymph node-positive disease between the 2 groups, we presume that the number of events was low and that the analysis potentially was underpowered to demonstrate any difference. Dramatic responses were observed in individual patients (Figs. 2, 3).

Figure 2.

(A,B) These are computed tomography images from a patient aged 78 years who had a biopsy-proven retrocaval lymph node metastasis (A, arrow) associated with a right renal pelvic high-grade transitional cell carcinoma (B, arrow). The estimated glomerular filtration rate (GFR) was 52 mL per minute per 1.73 m2. (C,D) After the patient received gemcitabine and cisplatin chemotherapy, there was a complete radiographic response in the primary tumor and in the lymph node metastasis. Pathologic evaluation of the resected specimen revealed no evidence of cancer in the renal pelvis; there was evidence of an effect of therapy in the stroma of the lamina propria and in 4 lymph nodes but no evidence of any disease in 16 other lymph nodes. After 2.5 years of follow-up, this patient remained without evidence of disease and had an estimated GFR of 28 mL per minute per 1.73 m2.

Figure 3.

(A,B) These computed tomography images from a patient who had a history of nonsteroidal anti-inflammatory drug overuse and exposure to phenacetin reveal a large right renal pelvic transitional cell carcinoma (TCC) with apparent parenchymal invasion and segmental hypoperfusion. The estimated pretreatment glomerular filtration rate (GFR) was 35 mL per minute per 1.73 m2 and prompted initial chemotherapy with gemcitabine, paclitaxel, and doxorubicin. (C,D) After 3 cycles of chemotherapy, a significant reduction in tumor volume and improvement in the GFR to 50 mL per minute per 1.73 m2 were evident, allowing conversion of chemotherapy to a dense-dose regimen of methotrexate, vinblastine, doxorubicin, and cisplatin. Pathologic evaluation of the resected specimen revealed a 3-cm grade 3 TCC that invaded the renal parenchyma with no involvement of lymph nodes. At 3 years of follow-up, the patient had a GFR of 35 mL per minute per 1.73 m2 and remained without evidence of disease.

The types of neoadjuvant chemotherapy regimens given, the median number of cycles, and tumor radiographic responses are summarized in Table 2. Nearly half of the 43 patients received dose-dense, combined methotrexate, vinblastine, doxorubicin, and cisplatin13 chemotherapy for a median of 4 cycles, resulting in a median 50% decrease in radiographic tumor size. Seventy-seven percent of patients received either a cisplatin or high-dose ifosfamide combination preoperatively.14, 15

Table 2. Tumor Response to Neoadjuvant Systemic Chemotherapy Administered to 43 Patients With High-Grade Urothelial Carcinoma of the Upper Urinary Tract Before Radical Nephroureterectomy
VariableNo. of Patients (%)
  • MVAC indicates methotrexate, vinblastine, doxorubicin (Adriamycin), and cisplatin; CGI, cisplatin, gemcitabine, and ifosfamide; GTA, gemcitabine, paclitaxel, and doxorubicin; GC, cisplatin and gemcitabine; SD, standard deviation; CR, complete response; PR, partial response; SD, stable disease.

  • a

    Radiographic tumor response could be evaluated in 29 patients. CR, PR, and SD were defined according to the Response Evaluation Criteria in Solid Tumors.

Neoadjuvant chemotherapy regimen, no. of patients (%)
 MVAC19 (44.2)
 CGI9 (20.9)
 GTA6 (14)
 GC5 (11.6)
 Other4 (9.3)
No. of chemotherapy cycles
 Median ± SD4 ± 1.8
Response of index tumora
 Tumor size decrease: Median ± SD, cm2.0 ± 1.3
 Tumor size decrease: Median ± SD, %50.0 ± 3
 CR9 (20.9)
 PR14 (32.6)
 SD6 (14)


We observed a significant reduction in pathologic classification for patients who received neoadjuvant chemotherapy, including a 25.4% reduction in the incidence of pT2 or higher disease and a 41.5% reduction in the incidence of pT3 or higher disease. This reduction occurred despite the presence of higher risk factors in the study group (ie, sessile tumors) that predicted a worse pathologic stage. In addition, we observed a CR rate of 14% in patients who received neoadjuvant chemotherapy. To our knowledge, this study represents the largest reported group of patients with UTCC who received neoadjuvant chemotherapy. Igawa et al5 reported their experience with 15 patients with locally advanced UTCC who received neoadjuvant chemotherapy and demonstrated a 13% CR rate and a 53% overall response rate. The patients in that study who had responses had a significant improvement in survival duration. Although the validity of downstaging has not been confirmed in upper tract urothelial tumors, it has been considered a valid surrogate that correlates with outcomes in urothelial tumors of the bladder, and it seems reasonable to extend it to this population.15-17 Moreover, pathologic confirmation of CR is more meaningful than radiographic evaluation because of the limitations of imaging with this disease, as discussed further below.

The motivation to pursue neoadjuvant, as opposed to adjuvant, chemotherapy for patients with high-risk UTCC is based on several important observations. First, data from the treatment of high-risk bladder cancer indicate improved survival in patients who receive neoadjuvant chemotherapy.16, 18 Second, most patients with high-risk UTCC suffer from underlying renal insufficiency, and significant loss of renal reserve occurs after nephrectomy.19, 20 This loss precludes effective dosing of chemotherapy in the adjuvant setting, providing 1 of the most compelling reasons for pursuing the neoadjuvant approach. Third, data indicate that the survival rates of patients with UTCC have not improved in the past 2 decades, during a time when initial nephroureterectomy was considered the standard of care and advances in imaging and endoscopy presumably may have allowed earlier detection of disease.1, 7, 21-23

Limitations of the neoadjuvant approach include the probability that some patients may not tolerate both chemotherapy and surgery. Given the association of UTCC with elderly age and tobacco-associated comorbidities, many patients are unable to tolerate the toxic effects of multimodality treatment. In such patients, surgery is chosen as the therapy that may offer the best chance of cure. Another obvious limitation of the neoadjuvant approach is the limited means of preoperative risk stratification and the possibility of over treatment in some patients.

Particularly difficult with UTCC are the potential inaccuracies in preoperative staging based on imaging studies and ureteroscopy. Abdominal imaging with computed tomography and magnetic resonance imaging is prone to under staging, and excretory urography or retrograde pyelography provide limited staging information.24 During ureteroscopy, the small, delicate instrumentation used for biopsy, coupled with the risk of perforation of the thin, muscular layers of the renal pelvis and ureteral walls, may preclude adequate tissue sampling, unlike during transurethral resection of bladder tumors, when muscle invasion can be determined readily and an examination under anesthesia can be performed. Because the recurrence of UTCC is associated with a high likelihood of death within 2 years, our position is that the potential for over treatment may be justified, especially with the associated loss of renal function after nephroureterectomy.

Our current strategy is to offer neoadjuvant chemotherapy to patients who have high-risk disease. This group includes patients with high-grade tumors demonstrated on biopsy; with sessile architecture, which almost always is observed with high-grade tumors; and with large-volume disease, which commonly is associated with higher stage upper tract tumors.3, 25 The results from a large, retrospective, multi-institutional study with central pathologic review were published recently, and multivariate analyses from that study indicated that these factors were highly predictive of pathologic stage.26 Because of the difficulty in accurate clinical staging, these features may be used as surrogates for estimating high-risk pathologic disease. Although using biopsy-demonstrated, high-grade disease alone has an approximate 66% positive predictive value for predicting high-stage disease,6, 7 the addition of the factors described above may improve diagnostic accuracy. Survival outcomes of patients with UTCC are highly dependent on the stage and grade of disease; high stage or grade and lymph node metastasis confer a significant increase in mortality. Patients with pT3 or higher disease and those with lymph node disease have a ≤50% chance of surviving for 5 years, and the median survival duration for patients with pT4 disease is <6 months.2 The median survival duration for patients who have retroperitoneal or distant metastasis is <3 years.2, 26

The limitations of the current study include its retrospective nature and the absence of survival data. Another potential criticism is the possibility that some patients may not have undergone surgery because of progression on chemotherapy. Although we cannot be certain that this did not happen to any patient, our experience from clinical trials suggests that this type of event is very unlikely. In a recent study that evaluated neoadjuvant chemotherapy for locally advanced urothelial cancer, all patients with upper tract tumors completed chemotherapy and surgery.15 By comparing the contemporary study group with an historic control group controlled for similar biopsy features, we hope that we minimized the selection bias inherent to retrospective studies. During the corresponding time periods, the treatment algorithm at our institution largely was standardized; before 2004, >90% of patients underwent initial surgery; and, after 2004, the neoadjuvant approach was considered uniformly for all patients who had high-risk features and were deemed capable of undergoing multimodality therapy. Ideally, randomized studies would indicate definitively the relative benefit of the neoadjuvant approach, but the low incidence of this disease renders the consideration of randomization prohibitive. Given these limitations, our goal was to evaluate the next most reliable, readily available level of evidence that could provide justification for continuing the neoadjuvant approach. Although survival data will be needed to recommend the neoadjuvant strategy as a standard of care, the data described here suggest a significant benefit of this treatment approach for high-risk UTCC and provide justification for the continuation and support of current trials using this strategy.


Funded in part by a grant from the National Institutes of Health (CA091846-08).