Here we report the hitherto largest clinical series of patients with nephroblastomatosis treated in two consecutive nephroblastoma trials in Switzerland, Austria and Germany over a timespan of more than twenty years with a median follow-up of more than nine years. Even though a uniform treatment was recommended (Supplemental file A) diagnostic approach and treatment varied. This individualization was caused by difficulties in making a correct diagnosis based solely on imaging studies for such polymorphous renal lesions (11, 14). In our series, overall diagnostic accuracy was low with only 48.7% of all nephrogenic rests being recognized on initial imaging, while 41% were mistaken for nephroblastoma. Furthermore, both open as well as cutting-needle-biopsies were not able to reliably discern between HNR and nephroblastoma, leading to changes in treatment decision in only 14%. The small sample size of a cutting-needle biopsy does not give a true overview and impression of the pseudo-capsule, rendering it impossible for pathologists to discern NR from blastema reliably. While this is usually possible with open “wedge” biopsies, it would lead to upstaging (stage III) in the case a nephroblastoma is found. Furthermore, even when taking this risk, only 62% were diagnosed of having a NR and it rarely led to change in decision. Thus, in line with current recommendations, biopsy must be strongly discouraged (15). The GPOH offered real-time radiology review to all participating sites during SIOP93-01 and SIOP2001/GPOH studies. This increased accuracy from 30 to 62% underlining the importance of radiologic experience in discriminating these rare lesions from other renal lesions (16). However, criteria to discriminate NR from WT are limited, mainly based on the nephroblastoma suggesting presence of a pseudo-capsule and contrast enhancement inhomogeneity in solid parenchyma at diagnosis. Size of a lesion had not been a defining criterion in SIOP93-01 and SIOP2001. But smaller single or oligo-focal lesions suggest NR. Sandberg et al. found a <1.75 cm diameter cut-off being highly suggestive for HNR, which however does not help with larger lesions (14 ). We found a significantly higher rate of single or oligo-focal nephrogenic rests in predisposed patients (Table 1b), most likely due to successful routine surveillance, as compared to non-predisposed patients having more diffuse and multifocal lesions (Table1b) (17, 18). Decision-making, especially in predisposed patients, should therefore include a thorough review of imaging and multidisciplinary board discussion, including physicians with specific experience in treatment and diagnosis of nephrogenic rests and other childhood kidney tumors to avoid unnecessary invasive procedures.
The prevention of nephroblastoma is the main aim of treatment directed at hyperplastic nephrogenic rests (HNR). A gold standard is lacking so far. Acute side effects and long-term effects of different treatment modalities have to be balanced against their efficacy in this non-metastasizing neoplastic disease (9, 19-21). Thus, clinical approach spans from “watch and wait” (W&W) to long-term maintenance treatment and tumornephrectomy. Similar to infant-neuroblastoma, W&W has been suggested as standard approach for HNR (22). This was not successful in our seven patients with six PD (Table 1b). However, all seven had diffuse or multifocal tumors, we can hence only conclude that clinicians need to be aware of the high probability of progression under W&W in this cohort. In contrast, W&W was successfully used in 25 of 34 patients who had undergone early surgery, achieved a 1st CR and were followed up after no or only short treatment (<12 weeks) without maintenance treatment for nephroblastomatosis. Despite 76% nephron-sparing surgeries, only four events (16%) occurred, including three nephroblastoma. Though single or oligo-focal HNR predominated, also a small group of multifocal and diffuse HNR patients (n=4) achieved and maintained an early CR (Figure 3, Table 1a). Contrary, the current SIOP-RTSG Umbrella recommends maintenance of monthly AV for a total duration of one year in case of nephron-sparing surgery if on histology any histologic NR are present, even if the remaining kidneys are in CR (SIOP-RTSG Umbrella, summarized in (23)). This recommendation is based on reported association of histologic NR with metachronous relapse and hyperplastic NR’s chemosensitivity (24, 25). However, 42-45% of all WT are accompanied by NR on histology, whether pre-treated or not, suggesting a limited responsiveness of microscopic & histologic NR to treatment (3, 4). One reason might be their wide histologic and biologic spectrum from benign, completely quiescent NR to active, highly proliferative, clinically apparent HNR (6, 7). We provide evidence that even in the latter HNR group, in case of CR, a short response-adapted treatment in combination with NSS followed by W&W is successful, thus challenging the current SIOP-RTSG Umbrella approach of one-year maintenance therapy.
On the other hand, already small residual lesions have a risk of relapse of more than 60% (Supplemental Figure C). In these cases, a maintenance treatment might be justified. Treatment duration varied significantly, ranging from about one (SIOP2001) to two years (SIOP93-01), with no difference in the rate of events or WT. Mean and median time to nephroblastoma was 3.8-4 years, independently whether patients had received a short treatment, a long or a short maintenance therapy. Thus, we could not find evidence of efficacy of the current maintenance treatment. Considering the long time to WT, its duration of up to two years might not be sufficient. A treatment inducing renal cell differentiation such as cis-retinoic acid could be an alternative. Friesenbichler et al. reported on two patients with residual lesions who had extended maintenance treatment with cis-retinoic acid (biweekly on-off) and achieved sustained CR (26).
Diffuse and multifocal hyperplastic NR (DHNR & MHNR) occur more frequently bilaterally and are usually not amenable to early surgical CR (Table 1). The COG suggested 6-week doxorubicin-intensified AV to induce resectability in bilateral nephroblastomatosis (27). In our cohort, prolonged treatment of ≥12 weeks AV facilitated the achievement of a first-line CR in 60% (32/54) of all upfront CTx patients, including 10 MHNR and DHNR. Interestingly, an early reduction of >70% of the initial tumor volume correlates significantly with superior event- and nephroblastoma-free survival (Figure 4 & 5d). Hence, patients not responding convincingly to initial AV within 6-12 weeks are candidates for treatment escalation. Data on treatment escalation in HNR are lacking. However, data from bilateral nephroblastoma series, which frequently include HNR, suggest that doxorubicin induces a better response. Coppes et al. also found a lower rate of metachronous WT in patients having NR if treated with doxorubicin (24). Alternatively, SIOP recommends carboplatin and etoposide in unresponsive bilateral cases (8, 23, 27). Beside standard cytotoxic drugs, add-on cis-retinoic acid as a differentiation-inducing drug has recently been used successfully in two HNR case series (26, 28). Persistent and thus treatment-resistant nephrogenic rests seem to harbor the origin of nephroblastoma clones. In case of an event, the probability of a nephroblastoma was relevant in our cohort with 29% for the first event. Furthermore, 65% of all patients experiencing an event eventually developed a WT. Similarly, Perlman et al. reported a WT development rate of 39% in their series of 53 DHNR patients (29). Failure to achieve a 1st line CR was significantly and independently correlated with an inferior event-free and nephroblastoma-free survival (Table 2, Figures 3&4b). Furthermore, patients suffering from nephroblastoma had a remarkably higher proportion of 45% high-risk nephroblastoma (blastemal subtype and DA) as compared to 5-7% in the only shortly pre-treated SIOP nephroblastoma population (12, 30, 31). Similarly, Perlman et al. and a report from GPOH indicated a significantly increased risk of 32% and 38% for diffuse anaplasia after long-term pre-treatment for nephroblastomatosis, which is even higher than the 20% in our cohort (8, 29). DA is a late event in nephroblastoma evolution and also in our cohort one more patient developed DA at a later relapse of a nephroblastoma. Prolonged observation of suspicious lesions comes with a significant risk of development of a WT, that has a 45% probability of high-risk histology. It is thus not surprising that three of our patients had pulmonary metastases at first diagnosis of a WT. Early oncologic resection to rule out WT, especially high-risk WT, should hence always be considered for stable or progressive suspicious lesions.
The relapse and progression pattern of NR patients differs significantly from WT patients. While 85% of first events after diagnosis of nephroblastomatosis occur within three years, 95% of nephroblastoma occurred 2.7 to 6.2 years after diagnosis of nephroblastomatosis and up to the age of 11. Similarly, Perlman et al. reported a case developing his first WT at the age of 11 years (29). Bergeron and Beckwith et al. stressed the risk of very late metachronous relapse in case of NR or NR-predisposing syndromes (7, 25). Coppes et al. mentioned a late relapse 13.1 years after initial diagnosis even in contralateral NR (24). Similarly, Scalabre et al. reported one metachronous WT after nephron-sparing surgery occurring at the age of 12 years (32).This is in contrast to unilateral nephroblastoma, where typically 95% of relapses occur within two years of initial diagnosis (33). This suggests to continue follow-up at least up to early adolescence. Table 3 summarizes treatment recommendations with respect to the clinical situation based on published data and our experience (3, 4, 6, 8, 11, 25, 26, 29, 32, 34-37).
Based on different phylogenetic background, intralobar (ILNR) and perilobar (PLNR) nephrogenic rests can be distinguished (6, 34). Hyperplastic ILNR are more common in patients with WT1 germline mutation, typically in Denys-Drash syndrome or WAGR patients (3, 6). This was the case in five of 13 patients in our series. While hyperplastic ILNR (HILNR) had a significantly lower risk of an event than hyperplastic PLNR, the risk of nephroblastoma is similar in both groups. Despite of the well-documented association of ILNR with stromal predominant WT, both WT were blastemal predominant (3, 17). In summary, there is no evidence to support a different surveillance or treatment approach for HPLNR and HILNR so far.
Overall, patients having HNR in our cohort had 40% risk of relapse or progression, which is dramatically lower if a CR can be achieved (Figures 3&4b). Although HNR are considered non-metastasizing, benign lesions, the overall survival of the entire cohort of 93.5% is not superior to the 94% for stage I-III WT treated in the GPOH (12). Interpretation of our data could be limited by under-registration of uneventful focal lesions. However, compared to 12% focal HPLNR in the NWTSG cohort, our higher rate of 46% single or oligo-focal HNR supports the representativeness of our dataset (table 1b).
Smaller case series reported homogeneously favorable outcomes and supported prolonged maintenance treatment. Taken together, only two out of 18 patients developed WT (9, 26, 38-40). Contrary, Perlman et al. reported that 46% of their 52 patients developed WT, most likely due to a higher rate of 88% DHPLNR. In our cohort prevention of nephroblastoma as the main goal achieved a 5-y-NFS of 78.9%. Risk factor analysis yielded two main risk factors: gender and non-CR after 1st-line treatment (Figures 3&4, Table 1). While it is self-explaining that residual HNR tissue harbors the origin of further lesions, it is surprising that females had a higher risk of progression and WT-development. It is known that WT is slightly more common in females, and females were more likely to have bilateral and DHNR, both factors showing a significantly higher risk in univariate analysis. However, when all three were adjusted in a COX-model, female gender remained independently associated with higher risk of event (HR 3.715; p=0.005; Table 2) and had a strong trend towards a higher risk of WT (HR 2.575; p=0.092). Recently, Vujanic et al. demonstrated a bi-partite peak in females with histologic NR at one and three to four years (3). Interestingly, our cohort with HNR had a single incidence peak in the first year of life. It is furthermore intriguing that all but one patient who later developed WT were younger than 24 months at diagnosis of HNR (Supplemental File E). Coppes et al. found an increased risk of metachronous contralateral WT development in case of histologic NR and age at diagnosis ≤12 months compared to older patients (p<0.001) (24). In a recent Children’s Oncology Group study of 8 patients treated for DHPLNR, 25% developed favorable histology WT, which is in line with our results (41). Any DHNR and bilateral tumors correlated significantly with inferior event-free survival. DHNR, especially when occurring in both kidneys, limit the surgical options and hence the likelihood to achieve a CR during first line. In contrast, single and oligo-focal HNR are usually amenable to radical surgery and have less frequent events. However, the risk of developing a nephroblastoma was similar in both groups, and underlines the same malignant potential. This underlines the importance of meticulous follow-up at least in the first decade and a half of life, irrespectively of the clinical presentation of HNR.