Follow-up guidelines for nonmetastatic renal cell carcinoma based on the occurrence of metastases after radical nephrectomy


Dr Ljungberg Department of Urology & Andrology, Umeå University, S-901 85 Umeå, Sweden. e-mail:



To define guidelines for the follow-up management of nonmetastatic renal cell carcinoma (RCC), by assessing tumour recurrences and the clinical course in patients who had undergone radical nephrectomy.

Patients and methods

The records of 187 patients with pT1–3, N0–X, M0 RCC who underwent radical nephrectomy between 1982 and 1997 were reviewed prospectively. Clinicopathological variables were compared with the time of first recurrence, site of metastasis and reason for diagnosis.


Metastases were diagnosed in 98 sites in 56 of the 187 patients (30%). The risk for developing metastases increased with stage; 80% of the patients had their metastases diagnosed within 3 years (median 14.5 months) after nephrectomy. The time to first diagnosis was longer for patients with pT1 tumours and for those with skeletal metastases. The cause-specific 5-year survival rate for pT1 tumours was 95%, for pT2 87% and for pT3 tumours 37%. All patients with diploid pT1–2 RCC survived, having a survival advantage over those with aneuploid pT1–2 tumours (P=0.018). Also, pT1–2 tumours of < 5 cm were associated with better survival rates. Among 74 patients with pT3 tumours, 45 got metastases; DNA ploidy in these tumours did not influence survival. Of 30 patients with lung metastases, 28 were diagnosed during follow-up, while 25 of 26 other metastatic sites were diagnosed because of symptoms.


The risk for tumour progression depends mainly on stage; these results indicate no need for follow-up in patients with diploid pT1–2 tumours or with aneuploid pT1 tumours of < 5 cm. For patients with aneuploid pT1–2 tumours of > 5 cm and pT3 tumours, follow-up is indicated.


The only curative treatment for RCC remains surgery; at the time of diagnosis, about a third of patients already have metastatic disease and of the remaining patients, ≈40% will have distant metastases during follow-up. In 1963, Robson et al. [ 1] showed that radical nephrectomy, with complete removal of the perinephric tissue, including the ipsilateral adrenal gland and the lymphatic field with paracaval or para-aortic lymph nodes, was associated with a better survival rate than was simple nephrectomy. Radical nephrectomy remains the standard surgical treatment of resectable RCC. Currently, for small localized tumours, there is a trend towards nephron-sparing surgery. The prognosis for patients with RCC mainly depends on the local tumour extent and metastatic spread [ 1, 2].

There is a clinical need to identify those patients who are at risk of tumour progression. The detection of metastatic disease can be facilitated in scheduled follow-up regimens. However, increasing healthcare costs will lead to economic decisions to avoid unnecessary follow-up procedures or those with little effect. To enhance efficiency, follow-up efforts should be directed at patients with a high probability for recurrent disease. Thus, individualized follow-up protocols are desirable and necessary to optimize the use, and justification for, follow-up management.

In the present prospective study of RCC, we analysed the results of follow-up after radical nephrectomy. Based on these data, an individual patient follow-up protocol was designed, with the aim of identifying those patients with a high or a low risk for disease recurrence.

Patients and methods

Between January 1982 and December 1997, 187 patients with pT1–3N0(X)M0 RCC (mean age 65.9 years, sd 11.1, range 28–87) underwent radical nephrectomy. Before surgery, the patients were examined using a chest X-ray, ultrasonography and CT, and patients with symptoms or signs underwent supplementary examinations by bone scintigraphy, MRI and cavography. Most patients were followed according to a scheduled follow-up programme and the remaining patients were followed at other hospitals. Over the last few years the follow-up schedule has changed, insofar as it is now based on the preference of the surgeon. The follow-up programme consisted of a clinical examination, laboratory tests, a pulmonary X-ray at 2 months, then every 6 months during the first 3 years and thereafter yearly throughout the 10 years of follow-up. Abdominal CT was carried out at 6 and 24 months after nephrectomy and additional radiological examinations used whenever indicated. In August 1998, the median (range) follow-up for the surviving patients was 66 (5–179) months. Of the 187 patients, 98 died during the follow-up, 52 from RCC and 46 from intercurrent diseases with no evidence of RCC.


Radical nephrectomy was routinely performed through a transverse abdominal incision dividing both recti abdominis muscles. This incision provides an excellent exposure of the retroperitoneal space and enables the early ligation of the renal vessels. The nephrectomy was performed en bloc, removing the kidney, perirenal fat and Gerota’s fascia, including the ipsilateral adrenal gland and the surrounding tissue from the abdominal aorta in left-sided tumours and from the middle of the vena cava in right-sided tumours. If there were enlarged or palpable lymph nodes between the aorta and vena cava, or in other sites, lymph nodes were dissected in that area, but no extensive radical retroperitoneal lymph node dissection was carried out. As prophylaxis for thrombosis, low-dose heparin (5000 IU subcutaneously) was given twice daily for patients treated between 1984 and 1990; during recent years, low molecular weight heparin (5000 IU subcutaneously, Fragmin® , Pharmacia, Uppsala, Sweden) was used daily for at least 7 days or until discharge.

Tumour stage was defined according to 1997 TNM classification system [ 3]. From the tumour-bearing kidney, normal kidney cortex and 3–8 tumour specimens were obtained by schematically taken biopsies [ 4]. Each sample was divided; one part was processed for flow cytometric DNA analysis and the other for conventional histopathological examination. Histopathological nuclear grade was determined using a four-grade scale according to Skinner et al. [ 2]. The method for flow cytometric DNA analysis was described previously [ 4]. The kidney cortex tissue samples were used as the standards for diploidy. The tumour samples were defined as diploid (DNA index of 1.0) when only one peak was detected, and aneuploid when there were at least two separate peaks, as it was assumed that all tumour samples contained normal and tumour cells. A tumour sample was regarded as tetraploid when > 15% of the cells had a tetraploid DNA index (1.95–2.05). In most samples, trout and chicken erythrocytes were added and used as reference cells. A tumour was then defined as diploid when all analysed tumour samples had a diploid DNA index (0.95–1.05), and aneuploid when at least one aneuploid tumour cell clone was found.

The values were expressed as the mean (sd) and analysed statistically using the chi2-test and the Mann–Whitney U-test, with P < 0.05 considered to indicate significance. The survival time curves were calculated according to the method of Kaplan–Meier; survival time was compared using the log-rank test.


During the follow-up, 56 of the 187 (30%) patients had their first metastases diagnosed as shown in Table 1; in all, 98 metastatic sites were diagnosed. There were 37 lung metastases, 24 bone, 11 liver, seven local/retroperitoneal, three abdominal, four brain, three skin and nine others. Table 1 also shows the occurrence and time to diagnosis of the first metastasis in relation to tumour stage. The median (range) time to the diagnosis of metastatic disease after the nephrectomy was 14.5 (2–144) months. In 43% of the 56 patients the metastases were observed within 1 year, in 70% within 2 years, in 80% within 3 years and in 93% during 5 years of follow-up. The risk of metachronous metastases was significantly higher in pT3 tumours than in both pT1 and pT2 tumours (P<0.001). There was no significant difference in time to occurrence of first metastasis between pT1 and pT2 RCC or between pT2 and pT3 tumours ( Table 1). Patients with pT1 tumours had significantly longer times (P=0.001) to the diagnosis of recurrence than had patients with pT3 tumours. Also, for patients with skeletal metastases, the time to diagnosis of the first metastases was significantly longer than for other metastatic sites (P=0.041).

Table 1.  The time to the diagnosis of first metastases in 56 patients with RCC, with metastatic site and with tumour stage Thumbnail image of

All primary pT1 tumours with subsequent occurrence of metachronous metastases were ≥5 cm in diameter. Table 2 shows the location of the first metastases and the reason for detection; of six patients with pulmonary metastases, five were diagnosed at 3–60 months through scheduled follow-up examinations. The remaining five patients with other metastatic lesions were diagnosed from symptoms. All the pT1 and pT2 tumours with metastases were aneuploid; none of 41 patients with homogeneously diploid primary pT1–2 tumours had metachronous metastases during the follow-up, differing significantly from the 11 of 64 patients with aneuploid tumours (P=0.005). Among the 64 aneuploid tumours, 15 were <5 cm in diameter and 49 tumours were ≥5 cm.

Table 2.  Reason for and interval to diagnosis of the first metastases in 11 patients with pT1–2 disease, and in 45 patients with pT3 disease, divided according to metastatic site (tumour stage according to 1997 TNM classification [3]) Thumbnail image of

Metastases occurred in 45 of 74 (61%) patients with locally advanced tumours (pT3N0–1,M0) after a mean time of 17.8 months; Table 2 shows the sites of the first metastases and the reason for their detection. The lung was the first site of metastatic disease in 24 patients with pT3 tumour. Only one of these 24 relapses in the lung was discovered through symptoms, the other 23 being detected in follow-up assessments. The earliest pulmonary lesion was detected 2 months after nephrectomy by routine pulmonary X-ray. All 17 patients with bone liver and cerebral metastases were diagnosed from symptoms. One of four local or retroperitoneal recurrences was diagnosed on routine follow-up CT. In all, metachronous metastases were diagnosed in 53% of the relapsed patients because of follow-up procedures. Five of 12 patients with homogeneously diploid pT3 tumours developed metachronous metastases, an incidence not significantly different from the 34 of 51 aneuploid pT3 tumours.


The 5-year survival rate was 95% for patients with pT1 tumours, 87% for pT2 and 37% for patients with pT3 RCC, as shown in Fig. 1. There was no significant difference in survival between patients with pT1 and pT2 tumours, but patients with pT3 tumours had significantly worse survival (P<0.001). There was no survival difference between the different pT3 tumour groups or depending on whether lymph node metastases were diagnosed. Figure 2 shows Kaplan–Meier survival curves of patients with pT1 and pT2 RCC, and pT3 tumours, grouped according to tumour DNA ploidy. Patients with diploid pT1–2 tumours survived significantly longer than those with aneuploid tumours (P=0.0179, log-rank test). There was no such survival difference based on ploidy in patients with pT3 tumours. In 41 patients with homogeneously diploid pT1–2 tumours, and in 37 patients with pT1 RCC of < 5 cm in diameter, there was no tumour progression.

Figure 1 Kaplan–Meier survival curves of patients with RCC subdivided according to stage. Seventy patients had pT1N0M0 tumours (green), 43 had pT2N0M0 tumours (red), 11 pT3aN0M0 tumours (light green), 36 pT3b–cN0M0 (light red) and 27 had pT1–3N1.

Figure 1 Kaplan–Meier survival curves of patients with RCC subdivided according to stage. Seventy patients had pT1N0M0 tumours (green), 43 had pT2N0M0 tumours (red), 11 pT3aN0M0 tumours (light green), 36 pT3b–cN0M0 (light red) and 27 had pT1–3N1.

M0 tumours (black).

Figure 2 Kaplan–Meier survival curves of patients with RCC subdivided according to tumour stage and DNA ploidy; 41 patients had homogeneously diploid pT1–2N0M0 tumours (green), 65 had aneuploid pT1–2N0M0 tumours (red), 12.

Figure 2 Kaplan–Meier survival curves of patients with RCC subdivided according to tumour stage and DNA ploidy; 41 patients had homogeneously diploid pT1–2N0M0 tumours (green), 65 had aneuploid pT1–2N0M0 tumours (red), 12.

had diploid pT3N0-1M0 tumours (light green) and 51 had aneuploid pT3N0–1M0 tumours (light red).


Tumour stage is the best predictor of prognosis in RCC and the incidence of tumour progression is correlated with the stage of the primary tumour. Other variables predicting tumour progression are tumour grade and DNA ploidy [ 2, 4, 5]. In patients with metastatic disease, the best treatment results reported are in those with minimal tumour spread and predominantly lung metastases [ 6[7]–8]. Such results have been shown after surgical treatment and long-term survival was reported in selected patients after surgery combined with immunostimulation [ 9[10]–11]. Furthermore, remission rates of 15–35% after treatment with IFN and IL-2 alone or in combination have been reported, most frequently for metastases in the lung and in those with limited metastatic spread [ 12]. Thus, early detection of metastases is important, as it increases the possibility for prolonged survival. The early detection of metastases is facilitated by a scheduled follow-up regimen, but most protocols have been directed towards all patients, regardless of the individual risk of metastasis. The increasing total costs of healthcare will lead to economic decisions that minimize unnecessary follow-up procedures that are not cost-effective. To better justify follow-up management, individualized follow-up protocols are needed.

In the present study, 30% of patients treated with radical nephrectomy had metachronous metastases, a proportion slightly higher than the ≈24% reported in previous studies, excluding lymph node metastases [ 13, 14]. In the present patients, there was a significant stage-dependent risk for metastases, as also reported by Hafez et al. [ 15], after nephron-sparing surgery. Tumours progressed in 10% of the present pT1–2 tumours, a frequency similar to the 14% reported by Levy et al. [ 14]. All patients with tumours of < 5 cm in diameter survived at 5 years, illustrating the less malignant character of small tumours. However, RCCs of < 5 cm can follow an unfavourable clinical course and the better prognosis reported might be caused by early detection rather than a lower malignant potential [ 16]. Supporting the present results after radical nephrectomy, Sandock et al. [ 13]found no occurrence of metastases in patients with tumours of < 2.5 cm. After nephron-sparing surgery in patients with RCCs of < 2.5 cm, Hafez et al. [ 15] found no local tumour recurrences but in 4.4% of the patients distant metastases occurred. Patients undergoing nephron-sparing surgery generally have a higher incidence of local tumour recurrence than after radical nephrectomy, because the affected kidney remains in place, with possible incomplete surgical excision of the original tumour, tumour spillage, multifocal existing tumours or growth of a new tumour. Patients treated by nephron-sparing surgery need more intense surveillance than those treated with radical nephrectomy, as do patients with hereditary RCC. The present patients with pT3 tumours had a significantly shorter survival time than patients with pT1–2 tumours; most patients with pT3 tumours had progressive disease, with a slightly higher proportion than in previous surveillance studies [ 13, 14].

Those with homogeneously diploid pT1–2 tumours had an excellent survival rate, significantly better than those with aneuploid pT1–2 tumours, confirming previous results [ 4, 17]. In those patients with homogeneously diploid pT1 and pT2 RCCs, no metachronous metastases occurred, while those with aneuploid pT1–2 tumours had a 17% incidence of metastases. Because tumours are frequently heterogeneous, several samples must be obtained from each tumour to verify DNA ploidy [ 4]. At least four samples should be analysed to have a 90% probability of finding a possible aneuploid cell population. For pT3 tumours, DNA ploidy had no influence on tumour progression.

Previous follow-up protocols were mostly based on empirical clinical data, with similar follow-up imaging and time intervals for all patients [ 18]. Based on the present results, we recommend a follow-up protocol that considers tumour stage, ploidy and tumour size (Fig. 3). For patients with aneuploid pT1–2 tumours of ≥5 cm and pT3 RCCs, follow-up is advisable, as patients with such tumours are at significant risk for metastases. Levy et al. [ 14] also proposed stage-specific surveillance after radical nephrectomy, but in their guidelines all patients were followed. In our protocol after radical nephrectomy, follow-up is not recommended for those with no or minimal risk of recurrent disease, i.e. about a third of the patients. This proportion could be greater, because more RCCs are being diagnosed incidentally and they are generally smaller and of higher stage [ 19]. Also, according to the surveillance guidelines of Hafez et al. [ 15], tumours of < 2.5 cm in diameter should not require follow-up after nephron-sparing surgery. The current costs for the routine follow-up procedures of a patient is US$ 5000–10 000, an expense which can be saved by using our proposed follow-up schedule (Table 3).

Figure 3.

The follow-up scheme according to tumour stage, tumour size and DNA ploidy of the primary tumour.

Table 3.  The follow-up protocol for patients with RCC and who undergo radical nephrectomy, according to TNM stage, tumour size and DNA ploidy of the primary tumour Thumbnail image of

No general consensus about the time schedule for the follow-up of patients with RCC has been established. In the present study, 80% of the patients had their metastases diagnosed within 3 years, similar to the 85% in the study of Sandock et al. [ 13]. The metastases were diagnosed after a median time of 15 months, with no statistical difference between pT2 and pT3 tumours. The mean time to the first diagnosis was in accordance with 29.5 and 22 months in pT2 and pT3 tumours, respectively, previously reported [ 13]. In contrast, in the study of Levy et al. [ 14], metastases were diagnosed after a longer interval (median 32 and 17 months in the corresponding groups of patients). Based on the present results, the follow-up should be most intense during the first 3 years. As most metastases were diagnosed early, we recommend follow-up studies at 3 and 6 months, thereafter every 6 months up to 3 years, and then yearly for at least 5 years. As only 7% of metastases occurred after 5 years and were mostly symptomatic, there is less need for follow-up after 5 years.

A follow-up protocol should be focused on the most common metastatic sites. RCC may metastasize to any site in the body, but lung and bone are the predominant sites, as found in the present study. In 52% of the present patients the first metastasis was diagnosed by the follow-up procedures, compared with 28–68% in other studies [ 13[14]–15]. Almost all of the present lung metastases were diagnosed by a routine chest X-ray, compared with only 26% in the study of Sandock et al. [ 13]. Despite the low diagnostic value of their chest X-ray, they recommended it for routine follow-up, but found no indications for any other imaging procedure. Because the routine chest X-ray diagnosed a high proportion of asymptomatic lung metastases in the present patients, its use is accordingly recommended in our routine follow-up protocol. Most patients with intra-abdominal/retroperitoneal metastases showed symptoms, as reported in other studies [ 19, 20]. The value of CT in the follow-up has not been determined, although it is recommended by some [ 14, 15, 19, 20], and only when symptoms appear by others [ 13, 21]. Parienty et al. [ 21] suggested that it should be reserved for high-risk patients during routine follow-up. As one of the nine intra-abdominal/retroperitoneal metastases was detected by scheduled CT, we also recommend it routinely in high-risk patients (e.g. pT3 tumours) at 6 and 12 months, and as an optional examination ( Table 3). All patients with skeletal metastases were diagnosed by bone scintigraphy, indicated by symptoms. Whether routine bone scans should be used for surveillance and to facilitate early detection has not yet been assessed.

In conclusion, the present study clearly shows the value of a stage-specific follow-up protocol including DNA-ploidy and tumour size. Individualized surveillance allows a better selection of follow-up procedures. For a third of patients with RCC, no follow-up is needed and the subsequent costs can be saved, thus enhancing the efficiency.


This study was supported by the Swedish Cancer Society, Lions Cancer Research Foundation of Umeå University, and The Medical Faculty, Umeå University, Umeå.