Features and outcomes of renal cell carcinoma of native kidneys in renal transplant recipients

Authors


  • T.K and C.S. contributed equally

Tobias Klatte, Department of Urology, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria.
e-mail: tobias.klatte@gmx.de

Abstract

Study Type – Therapy (case series)
Level of Evidence 4

OBJECTIVE

To outline the features and outcomes of renal cell carcinoma (RCC) in native kidneys of renal transplant recipients, who are at increased risk of developing this disease.

PATIENTS AND METHODS

We retrospectively studied the clinicopathological features and survival of 28 surgically treated RCCs, which developed in 24 renal transplant recipients. Features and outcomes were compared with 671 patients with RCC who had no renal transplant.

RESULTS

The median interval between renal transplantation and the occurrence of RCC was 5.6 years. Acquired cystic kidney disease was present in 83% of the transplanted patients. Compared with the patients with RCC and no renal transplant, RCCs of native kidneys in transplant recipients were more frequently incidental findings (92% vs 77%, P = 0.092), multifocal (39% vs 15%, P < 0.001), bilateral (17% vs 4%, P = 0.006), had lower T stages (P = 0.040), were smaller (P = 0.027), of lower grades (P = 0.010), were more frequently papillary (43% vs 19%, P = 0.019) and occurred at a significantly younger age (P = 0.022). After a median follow-up of 6.7 years, eight renal transplant recipients had died (33%), but only two deaths were due to RCC. Survival with metastatic RCC was only 4 months, if a full resection of all metastatic sites was not achieved. In multivariate analysis the presence of a renal transplant had no effect on survival.

CONCLUSIONS

Most RCCs in renal transplant recipients are incidental low-stage, low-grade tumours with a favourable prognosis. The outstanding pathological findings are bilateral occurrence, papillary subtype and multifocality. Prognosis of metastatic RCC is poor but might be favourable if all metastases are resected. Screening for early detection of asymptomatic RCC is advocated.

Abbreviations
AKCD

acquired cystic kidney disease

OS

overall survival

CSS

cancer-specific survival

US

ultrasonography.

INTRODUCTION

After successful renal transplantation for end-stage kidney disease the risk of developing RCC in native kidneys is about 15 times greater than in the normal population [1]. A major risk factor for this tumour is acquired cystic kidney disease (ACKD) [2], which is found in 70–90% of the patients [3,4]. For the pathogenesis of ACKD, it has been speculated that nephron loss leads to compensatory tubular hyperplasia and cyst formation, driven by up-regulation of growth factors and activation of proto-oncogenes [5,6]. Due to the diffuse nature of ACKD, high incidences of multifocal and bilateral tumours have been reported [4,7–9]. On a pathological and genetic level these tumours appear to be different from sporadic forms, e.g. more frequently papillary and with different genetic pathways affected [10,11].

It has been suggested that most of these RCCs are low-stage tumours with a favourable prognosis [7,8]. However, studies indicate a high proportion of high-grade tumours [8,9] and mortality rates of up to 25% despite aggressive surgical treatment [4]. Due to the relative rarity of the disease, there have been only a few larger series comprising few patients [4,7–9]. In an attempt to better understand the characteristics and outcomes of surgical therapy we present the largest series to date, derived from one institution over a 20-year period.

PATIENTS AND METHODS

Between 1989 and 2008, 24 renal transplant recipients had a nephrectomy for a total of 28 renal tumours, at the Department of Urology, Medical University of Vienna, Vienna, Austria. After approval by the institutional review board, a retrospective study was undertaken based on a chart review, slide review and assessment of outcome.

Clinical, pathological and follow-up data were reviewed for each patient. Clinical data included age, gender, symptoms at presentation, duration of dialysis, the date of renal transplantation, and history of ACKD. Incidental diagnosis was defined as the RCC being completely asymptomatic at the time of diagnosis. Radiographic, operative and pathology reports were used to assess the extent of disease. Preoperatively, all patients had cross-sectional imaging of the abdomen, pelvis (CT or MRI) and chest (CT or X-ray), serum electrolytes, liver function and coagulation tests. Tumour stage was determined according to the 2002 TNM classification [12]. For a diagnosis of ACKD the presence of at least three cysts with involvement of both kidneys on imaging was required [13]. Histological slides of all tumours were reviewed by one dedicated uro-pathologist (A.H.) for histological subtype and grade. The subtype of tumour was assigned according to the Heidelberg classification of renal tumours [14]. Tumours were graded according to the Fuhrman scheme [15].

The clinical and pathological features were compared with a group of 671 patients with RCC but no renal transplant, using chi-square tests and Student’s t-tests. The endpoints of this study were overall (OS) and cancer-specific survival (CSS) time, which were both calculated from the date of nephrectomy to the date of death and death from RCC, respectively. The Kaplan-Meier method was used to estimate survival functions. A multivariate Cox proportional hazards regression model was used to identify independent prognostic factors of OS and CSS. As there were only two RCC-related deaths in renal transplant recipients, a separate multivariate analysis for this group was not meaningful. Instead, the presence of a renal transplant was entered as a dichotomized variable in the Cox model.

RESULTS

RCC was diagnosed in 18 men (75%) and six women (25%), with a mean (range) age at nephrectomy of 58.1 (32.0–81.3) years. Twenty patients (83%) had unilateral disease and four (17%) presented with synchronous bilateral RCC. The median interval between renal transplantation (start of dialysis) and diagnosis of RCC was 5.6 (9.6) and the range was 0.5–15.5 (3.6–18.3) years. ACKD was present in 20 of the 24 patients (83%); the initial nephropathies are outlined in Table 1.

Table 1.  The initial nephropathies leading to end-stage renal disease
Diseasen (%)
Chronic glomerulonephritis10 (42)
Hypertensive nephrosclerosis6 (25)
Chronic interstitial nephritis4 (17)
Haemolytic uraemic syndrome1 (4)
Renal tuberculosis1 (4)
Unknown2 (8)

All patients had received standard immunosuppressive treatment at the time of RCC diagnosis. Of the 24 patients, 15 received a conventional triple immunosuppression regimen consisting of calcineurin inhibitors (cyclosporin A in 13; tacrolimus in two), mycophenolate mofetil in seven or azathioprine in eight and prednisolone in 15. Six patients received dual therapy with cyclosporin A and prednisolone, one patient received only azathioprine and prednisolone, one highly sensitized patient received quintuple immunosuppression with immuno-adsorption, antithymocyte globulin, rapamycine, mycophenolate mofetil and prednisolone, while another highly sensitized patient received tacrolimus, azathioprine, antithymocyte globulin and prednisolone. Immunosuppressive treatment was only modified if there were metastases not amenable to surgery.

Only two patients presented with symptoms due to RCC, both reporting gross haematuria and one reporting flank pain. Both were the only tumours that were metastasized at diagnosis. In incidental tumours, the primary diagnostic procedure was ultrasonography (US) in 10 (45%), CT in 11 (50%) and MRI in one (5%), all either as part of the annual routine screening for ACKD or its follow-up.

All patients had an open radical nephrectomy through a flank incision (16) or retroperitoneoscopic radical nephrectomy (12) for the tumour-bearing kidney. In the four patients with a synchronous bilateral tumour, two had a single procedure and the remaining two a two-stage procedure. A lymph node dissection was performed in three patients, where enlarged or suspicious lymph nodes were noted on preoperative imaging. There were no perioperative deaths or major complications.

All of the removed lesions were malignant RCCs. On final pathology, 21 tumours (75%) were staged pT1N0M0 and 96% were Fuhrman grade 1 or 2. The mean (range) tumour size was 4.0 (0.7–10) cm. In terms of subtype, 43% were papillary and only 54% were clear cell; 39% were multifocal. Only two tumours were metastasized at diagnosis (nos 7 and 12, Table 2).

Table 2.  The patient and tumour characteristics
N/gender ageInterval NTX to RCC, yearsSymptomsSideACKDSize, cmTNM stage, Fuhrman gradeSite of metastasesMultifocalSubtypeFollow-up years
  1. M, male; F, female; NED, no evidence of disease; DOD, died of disease; AWD, alive with disease; DOC, died of other causes, NTX, renal transplantation; FP, flank pain; H, haematuria.

1/F, 32.012.4norightNo0.8T1a N0 M0 G1 Nochromophobe7.5, NED
2/M, 63.93.8norightYes3.1T1a N0 M0 G2 Yespapillary14.9, NED
3/M, 63.50.5norightNo4T1b N0 M0 G1 Noclear cell13.5, NED
4/M, 68.83.1norightYes6.5T3b N0 M0 G2 Noclear cell1.8, DOC
5/M, 64.79.6norightYes3.7T1a N0 M0 G2 Noclear cell3.9, DOD
6/M, 66.413.7norightYes5.5T3b N0 M0 G2 Yesclear cell9.4, DOC
7/M, 54.33.9HleftYes6T3b N0 M1 G3LiverNoclear cell0.4, DOD
8/F, 50.32.3norightNo4T1b N0 M0 G2 Noclear cell1.8, DOC
9/M, 70.29.6norightYes4.6T3a N0 M0 G2 Noclear cell3.2, AWD
10/F, 61.915.5noleftYes5T1b N0 M0 G2 Nopapillary3.4, NED
11/F, 51.23.1norightYes1.7T1a N0 M0 G1 Noclear cell6.7, NED
12/M, 66.25.4FP, HleftYes10T3c N1 M1 G2Adrenal, nodeYespapillary9.9, NED
13/M, 53.05.6noleftYes3.5T1a N0 M0 G2 Noclear cell14.8, NED
14/M, 53.63.3noleftNo4.5T1b N0 M0 G2 Nopapillary5.1, NED
15/M, 81.412.6noleftYes3T3a N0 M0 G2 Noclear cell0.4, NED
16/F, 56.53.6noleftyes5T1b N0 M0 G2 Noclear cell9, DOC
17/M, 72.613.0norightyes4.5T1b N0 M0 G2 NoPapillary5, NED
18/M, 65.55.6norightyes5T1b N0 M0 G2 Noclear cell4.3, DOC
19/M, 59.211.8norightyes3.4T1a N0 M0 G2 YesPapillary6.2, NED
20/F, 50.55.3norightyes5T1b N0 M0 G2 Noclear cell13.4, NED
21/M, 58.73.3nobilateralyes2.5T1a N0 M0 G2 YesPapillary5.1, DOC
1.6T1a N0 M0 G2YesPapillary
22/M, 38.615.3nobilateralyes1.5T1a N0 M0 G2 YesPapillary4.6, NED
1.4T1a N0 M0 G1YesPapillary
23/M, 46.711.6nobilateralyes2T1a N0 M0 G2 YesPapillary6.7, NED
6.5T1b N0 M0 G2YesPapillary
24/M, 45.914.1nobilateralyes0.7T1a N0 M0 G1 Noclear cell8.3, NED
7.5T2 N0 M0 G2Yesclear cell

After a median (range) follow-up of 6.7  (0.3–14.9) years, eight patients had died (33%) but only two deaths were due to RCC. The remaining causes of death included coronary heart disease in three, septicaemia in two and metastatic lung cancer in one. The 5- and 10-year survival rates (sem) were 72 (9)% and 51 (14)% for OS, and 91 (6)% and 91 (6)% for CSS (Fig. 1). During the follow-up seven patients (29%) developed other malignancies, including lung cancer in two, thyroid cancer in two, melanoma in one, prostate cancer in one and gastric cancer in one.

Figure 1.

CSS and OS for renal transplant recipients with RCC in native kidneys. Survival rates after 5- and 10-years were 91% and 91% for CSS, and 72% and 51% for OS, respectively.

Of the four patients with metastatic RCC, two presented with synchronous metastatic disease, while two developed recurrence after surgery for N0M0 RCC (at 38 and 42 months, respectively). Immunosuppressive treatment was modified in all but one patient (nos 5, 7 and 9). Among the four patients, two died from the disease 4 months after the diagnosis of metastases without receiving systemic therapy (nos 5 and 7). Of note, both tumours were of the clear cell subtype. One patient (no. 12) presented with papillary RCC metastatic to retroperitoneal lymph nodes and the ipsilateral adrenal gland. He had a full resection of all disease sites and is currently free of disease after almost 10 years of follow-up. The remaining patient was diagnosed with lung metastases 38 months after nephrectomy and is awaiting sunitinib treatment (no. 9).

Compared with the patients with RCC and no renal transplant (Table 3), RCCs of native kidneys of renal transplant recipients were more frequently incidental findings, multifocal, bilateral, had lower T stages, were smaller, with lower Fuhrman grades, more frequently papillary and occurred at a significantly younger age. Interestingly, the incidence of multifocal tumours was only higher in papillary, but not in clear cell RCC.

Table 3.  Clinical and pathological characteristics of patients with RCC, stratified by presence (24) or absence (671) of a renal transplant
Mean (sd) or
n (%) variable
No renal transplantRenal transplantP
Age63.3 (14.0)58.1 (11.2)0.022
Incidental517 (77)22 (92)0.092
Men422 (63)18 (75)0.227
T stage  0.040
 T1-2397 (59)22 (79) 
 T3-4274 (41)6 (21) 
Lymph node metastasis22 (3)1 (4)0.932
Distant metastasis85 (13)2 (7)0.386
Fuhrman grade  0.010
 G1-2505 (75)27 (96) 
 G3-4166 (25)1 (4) 
Tumour size5.0 (2.9)4.0 (2.2)0.027
Bilateral30 (4)4 (17)0.006
Multifocality99 (15)11 (39)<0.001
Histological subtype  0.019
 Clear cell499 (74)15 (54) 
 Papillary126 (19)12 (43) 
 Chromophobe40 (6)1 (3) 
 Collecting duct6 (1)0 

Finally, multivariate Cox proportional hazards regression models showed that the presence of a renal transplant did not affect the CSS and OS; T stage, M stage, grade and presence of symptoms (for CSS only) were retained as independent prognostic factors (Table 4).

Table 4.  Multivariate Cox proportional hazards models for prediction of CSS and OS. In both models the presence of a renal transplant had no effect on the prognosis
Covariate categoriesHR (95% CI), P
CSSOS
Symptoms, yes vs no1.88 (1.14–3.12), 0.0141.37 (0.92–2.03), 0.120
T stage, T3/4 vs T1/22.13 (1.17–3.86), 0.0131.52 (1.01–2.27), 0.045
N stage, N + vs Nx/01.34 (0.95–2.04), 0.0941.30 (0.89–1.91), 0.177
M stage, M1 vs M010.14 (5.78–17.78), <0.0014.58 (2.95–7.10), <0.001
Grade, G3/4 vs G1/22.34 (1.37–3.99), 0.0021.70 (1.13–2.56), 0.011
Multifocality, yes vs no0.92 (0.46–1.85), 0.8150.94 (0.58–1.52), 0.809
Subtype, non-clear cell vs clear cell1.28 (0.71–2.31), 0.4101.11 (0.74–1.69), 0.601
Renal transplant, yes vs no0.33 (0.07–1.58), 0.1650.91 (0.41–2.01), 0.818

DISCUSSION

RCC of the native kidneys affects 1–3% of all renal transplant recipients [7–9]. In a study of >35 000 first-time renal transplant recipients the risk of RCC was estimated to be 15 times greater than in the general population [1]. RCC of the native kidneys is one of the most common tumours among patients after renal transplantation, and accounts for ≈8% of all malignant tumours in this group [4]. In the present study, the actual incidence of RCC among renal transplant recipients could not be calculated, as we analysed only patients who were referred to our institution for surgical treatment of suspected RCC.

The main risk factor for RCC in renal transplant recipients is ACKD [2], which is present in 70–90% of the patients [3,4]. The diagnosis of ACKD is established via US, CT or MRI [16]. Cysts in ACKD are usually <0.5 cm in diameter but can be up to 2–3 cm. For the diagnosis of ACKD the presence of at least three cysts with involvement of both kidneys is mandatory [13]. The probability of RCC among patients with ACKD is 4–7% over a 7–10-year period [17,18] with a yearly incidence of 0.18%[19]. In a recently published prospective ultrasonographic study on 561 renal transplant recipients, the prevalence of RCC among patients with ACKD was as high as 19.4%[3]. The pathogenesis of ACKD and carcinogenesis of RCC in ACKD are poorly understood. It has been speculated that nephron loss leads to compensatory tubular hyperplasia and cyst formation, driven by epidermal growth factor, hepatocyte growth factor, Bcl-2 and activation of proto-oncogenes such as jun[5,6].

The current study supports the concept that pathological features of RCC in renal transplant recipients differ from sporadic forms. While papillary RCCs comprised 19% of the sporadic tumours, 43% of RCCs in renal transplant recipients were of this subtype, which is in agreement with previous reports [7–9]. Moreover, the incidence of multifocal tumours and bilateral cases was significantly higher than in sporadic forms. All these features might be explained by the diffuse nature of the underlying disease, ACKD. In our study the proportion of incidentally detected RCCs in renal transplant patients was >90%, which was higher than in those with no renal transplant. We speculate that this is the result of annual US screening of patients who had received a renal transplant. This might also explain why the detected tumours were smaller and of lower TNM stage. Our findings are supported by recent data [8,9]. Interestingly, two studies reported that the incidence of high-grade lesions (Fuhrman grade ≥3) might be as high as 50%[8,9]. This finding is in stark contrast to the current study, in which only 4% of the tumours were high-grade. We suggest that methodological differences, such as absence of slide review, account for this discrepancy.

We carried out classical radical nephrectomy through a flank incision in 16 cases, but managed the last 12 cases by retroperitoneoscopic nephrectomy. Compared with the transperitoneal approach the working space is less, but this is not an issue in patients with cirrhotic end-stage kidneys [20]. Given the high incidence of bilateral tumours, prophylactic bilateral nephrectomy might be considered even in patients with unilateral tumours. Currently there are no published reports or guidelines but this is definitely a direction for further discussion [8]. In the opinion of the authors, bilateral nephrectomy is not justified, given the relatively low incidence of bilateral disease of ≈20%, and that no tumours were found in the contralateral native kidney during the follow-up.

The prognosis of RCC in renal transplant recipients appears to be favourable. In the study of Ianhez et al.[7], all nine patients treated by radical nephrectomy survived. There was only one patient presenting with metastases, who died 4 months after diagnosis with no surgical intervention. In the study of Moudouni et al.[8] on 10 patients with 12 tumours, only one patient died from progressive RCC at 6 years after diagnosis. Neuzillet et al.[9] reviewed data of 933 transplant recipients, of whom 11 developed a total of 12 tumours. After a median follow-up of 39 months, one patient had died from RCC. By contrast, one study showed that four of 12 patients died from RCC during the follow-up [4]. However, most data including the present study indicate that RCC in renal transplant recipients has an excellent prognosis. Furthermore, our study showed that the presence of a renal transplant does not diminish OS or CSS.

Interestingly, the prognosis of metastatic RCC in renal transplant recipients was very poor but might be favourable if all metastases are resected. There were two deaths due to RCC, both occurring 4 months after the diagnosis of metastatic disease. Importantly there was one patient who had a full resection of all metastatic sites and is currently alive with no evidence of disease after almost 10 years of follow-up. Therefore, metastasectomy should be as complete as possible in renal transplant patients with metastatic RCC. If this is not possible, systemic antitumour treatment should be instituted. In addition, it is generally accepted that immunosuppressive therapy should be reduced in the presence of metastases [9]. Immunotherapy is generally contraindicated in patients receiving immunosuppressive therapy but might be considered in selected patients [8]. A new attractive treatment option for metastatic RCC in transplant patients might be to switch from standard immunosuppressive drugs to the rapamycin analogue temsirolimus, which interacts through mTOR inhibition [9]. This drug showed prolongation of OS in poor-risk metastatic RCC [21], and has both immunosuppressive and antitumoral properties. Also, tyrosine kinase inhibitors such as sunitinib and sorafenib might represent treatment options, as they are not contraindicated in patients with a transplant under immunosuppressive therapy.

Given the high incidence of incidental tumours among patients with RCC [7,9] and that the incidence of ACKD increases with increasing duration of dialysis, periodic screening for ACKD and RCC has been suggested for the early detection of premalignant and malignant lesions in patients who have been on dialysis for 3–5 years [22]. US appears to be less sensitive than CT for characterizing cysts but is similar for detecting solid tumours [23]. For RCC, early contrast-enhanced CT might show higher sensitivity and specificity than delayed contrast CT [24]. At our institution, annual US screening for RCC is carried out in renal transplant recipients. Schwarz et al.[3,25] proposed the following screening approach, according to the Bosniak renal cyst classification: (i) all transplant patients should undergo yearly US of their native kidneys, independent of ACKD; (ii) patients with ACKD and Bosniak I or II cysts should have US twice a year and CT if progression is evident; (iii) patients with ACKD and Bosniak IIF cysts should have US four times a year and CT or MRI annually; nephrectomy should be performed if there is progression, even if category III or IV is not reached; (iv) patients with ACKD plus Bosniak III or IV cysts should have a nephrectomy. However, there are no prospective studies on the efficacy and cost-effectiveness of this approach. As a result, routine screening is currently not recommended in the USA. By contrast, we advocate routine screening for RCC in renal transplant recipients. Most tumours were incidental findings in asymptomatic patients. Further, RCC can be only cured surgically, even in the presence of metastases. Therefore routine screening for surgically amenable RCC should be mandatory in all renal transplant recipients.

In conclusion, most RCCs in renal transplant recipients are incidental low-stage, low-grade tumours with a favourable outcome. The outstanding clinical and pathological findings are bilateral occurrence, papillary subtype and multifocality. The prognosis of metastatic RCC is poor but might be favourable if all metastases are resected. Screening for early detection of asymptomatic RCC is advocated.

CONFLICT OF INTEREST

None declared.

REFERENCES

EDITORIAL COMMENT

This important paper provides some data on a rare but significant topic. Urologists and nephrologists who manage transplant patients daily, as well as those clinicians who sporadically encounter tumour masses in native renal units, will find in it interesting and practical information.

The current study suggests that the pathological features of RCC in renal transplant recipients differ from the conventional sporadic form of RCC, and are most probably associated with concomitant ACKD, i.e. frequently incidental (92% vs 77%), multifocal in papillary (75% vs 29%), but not in clear cell RCC (≈13%), bilateral tumours (17% vs 4%) with lower T stages (T1 or 2, 79% vs 59%), smaller tumour diameter (4.0 vs 5.0 cm) and lower Fuhrman grade (grade 1 or 2, 96% vs 75%). Tumours are more frequently papillary (43% vs 19%), as was reported earlier [1] and tend to occur at a younger age. However, when entered into a multivariate Cox proportional hazards regression model, renal transplantation was not found to be an independent variable associated with worse prognosis. While the 5- and 10-year CSS rates were 91%, the respective OS rates were only 71% and 51%, respectively, an alarming reflection of other more fatal threats to this particular group of patients, amongst whom were 29% cases of second primary malignancy, coronary heart disease and sepsis.

ACKD is present in 83% of cases and the median interval between the start of dialysis and diagnosis of RCC is ≈10 years. This naturally provokes the imagination to speculate on the unknown pathogenesis of RCC in native non-functioning renal units, especially in those cases that are both multifocal and with papillary histology, comprising a third of these patients. Two issues that are not addressed in this report are the prevalence of RCC in native kidneys of transplanted patients and the prevalence of renal benign lesions. Moudouni et al.[1] followed 373 patients and discovered 10 (2.7%) cases of RCC during 10 years of follow-up. Neuzillet et al.[2] found 11 cases of 933 patients that were followed (1.2%), whereas Denton et al.[3] found RCC in 4.2% of native kidney nephrectomy specimens and oncocytoma in only 0.6%. Given that these are non-functioning renal units, the reported low incidence of benign histology, and the multifocal nature of this particular entity, a straight-forward total nephrectomy is definitely indicated. However, due to the cystic nature of these kidneys, I now find it rather rewarding to base the final diagnosis of a solid mass on MRI rather then CT or US.

Laparoscopy allows one-session bilateral transperitoneal native kidney nephrectomy with a relatively low morbidity. It is an unanswered question whether such an approach is warranted when a solid mass is diagnosed in one native kidney. The pros and cons are obvious, but the lifetime low likelihood of contralateral malignant transformation might not justify such a direct aggressive approach unless special indications co-exist.

Klatte et al. suggest, based on one metastatic case of four, that complete surgical removal of the primary and the metastatic outgrowths is warranted to prevent a catastrophic disease progression. Although it makes sense to do so in immunocompromised patients with non-clear cell tumour histology, one should be very careful in extrapolating this one case of a durable complete response, as this study is not sufficiently powered to make such a suggestion, and second, there are currently alternative medical treatments for such patients.

Amnon Zisman

Department of Urology, Assaf Harofeh MC, Tel Aviv University, Tel Aviv, Israel

REFERENCES

Ancillary