CORRESPONDENCE

 

Response to letter regarding 'Paediatric nephroblastoma at a South African tertiary hospital: A 21-year retrospective analysis'

 

 

To the Editor: Thank you for allowing us to respond to the letter regarding our article 'Paediatric nephroblastoma at a South African tertiary hospital: A 21-year retrospective analysis.' We appreciate the reader's thoughtful questions and insights, contributing to a more comprehensive discussion of our findings.

We acknowledge the reader's appreciation of our collaboration with undergraduate medical students, and extend our gratitude to Profs David Stones and Jan du Plessis for their support in accessing the database.

Regarding the histological assessments and preoperative use of fine needle aspiration (FNA) and cutting needle biopsy, we clarify that FNAs were routinely performed before initiating chemotherapy. When FNAs were inconclusive, a cutting needle biopsy was performed to aid in diagnosis. This approach ensured accurate preoperative cytological or histological assessment.

The discrepancy between the number of patients undergoing surgery (n=180) and those with histopathological data (n=207) is explained by including preoperative FNAs and cutting needle biopsies. Additionally, some patients defaulted on treatment (n=5) or passed away (n=22) before reaching surgery. We also acknowledge that one patient with anaplastic histology was inadvertently omitted from Table 3, and this should be corrected to n=8 (4.4%).

Concerning syndromic patients, one patient had Denys-Drash syndrome, and another had Beckwith-Wiedemann syndrome. Unfortunately, both patients passed away before completing treatment. We had no patients with a horseshoe or single kidney in this study cohort.

We emphasise that lactate dehydrogenase levels were included in our study as part of our routine biochemical workup for paediatric patients presenting with abdominal distension, rather than as a specific marker for Wilms' tumour.

Concerning the use of isotope studies and bone marrow aspiration, we clarify that these investigations were conducted as part of a broader diagnostic approach for abdominal masses, which includes differential diagnoses such as neuroblastoma and B-cell lymphoma. While not routine for Wilms' tumour, these investigations were valuable in identifying metastases in cases where Wilms' tumour was ultimately diagnosed.

Concerning the site of metastases (single or multiple sites), 67 patients had metastasis, 53 (79.1%) of whom had metastasis at a single site and 14 (20.9%) at multiple sites. The metastasis at single sites included lung (n=32), liver (n=10), marrow (n=3), skeletal (n=1), pleural effusion (n=3), inferior vena cava (IVC) wall infiltration (n=3) and thoracic nodes (n=1).

The metastases at multiple sites included lung and liver (n=7), lung and IVC wall (n=1), lung, liver and IVC wall infiltration (n=2), lung and spleen (n=1), lung and bone marrow (n=1), lung, liver and skeletal (n=1) and lung, liver and lymph nodes (n=1). Non-pulmonary metastases were present in 20 (30.3%) patients, and non-hepatic metastases were present in 45 (67.2%) patients.

We want to clarify the IVC involvement in six patients, as depicted in Table 1. The metastatic breakdown has been provided, and we confirm that these patients had IVC wall infiltration, and the lesions were not IVC thrombi.

We acknowledge the reader's suggestion regarding the impact of diagnostic delay and the need to determine the optimal cut-off for it. Using Youden's index, we found that the cut-off time for experiencing symptoms leading to lower survival was 22 days.

We confirm that patients with metastatic disease were included in the preoperative staging data, and that radiotherapy was primarily administered to patients with local stage III disease or metastases. Unfortunately, specific data on the exact anatomical sites of radiation were not collected.

Our definition of a 'good response' to neoadjuvant chemotherapy was based on clinical and imaging assessments of tumour volume reduction by at least 30 - 50%. However, we acknowledge that further refinement in defining response categories, such as distinguishing between fast and slow responders, would enhance the robustness of future analyses.

Concerning relapsed patients, we confirm that of the 18 who relapsed during treatment, 3 completed treatments, while 1 abandoned therapy. Of those who relapsed post treatment, only 1 abandoned further care, with subsequent management decisions based on individual patient needs such as second-line chemotherapy, radiation, re-resection, or palliation. We are in the process of conducting a dedicated study on relapse patterns, which will provide more detailed insights.

We recognise the need to calculate and present the median age (with interquartile range (IQR)) for unilateral and bilateral tumours separately. In our comparison, we looked at the age at diagnosis (measured in months) for patients with bilateral nephroblastoma (n=12) v. those with unilateral tumours (either left or right, n=195). The median (IQR) age at diagnosis for bilateral tumours was 31 (22 - 39) months, while for unilateral tumours, it was 39 (24 - 64) months. The difference in ages was not statistically significant (p=0.13, based on the Wilcoxon rank sum test), indicating that there was no significant difference in age at diagnosis between bilateral and unilateral cases.

Regarding survival analysis, our study primarily focused on 2-year overall survival (OS), given the increasing loss to follow-up over more extended periods. However, we recognise the value of 5- and 10-year OS rates, and will attempt to extract these in future studies. We also acknowledge the inconsistency in reporting treatment default impact, and confirm that treatment default significantly affects survival (p=0.019).

The sites of relapse beyond the lungs and liver included the peritoneum, mediastinum, lymph nodes, pelvis, facial bones, intracranial regions and residual kidney, in a case of bilateral disease.

We recognise that our hazard ratio (HR) for lung metastases seems higher than expected, and may have been affected by our approach of including lung metastases in both the single and multiple metastatic site categories. The initial analysis indicated that patients with lung metastases had an 85% higher risk of mortality (HR 1.85, 95% confidence interval (CI) 1.15 - 2.95, p=0.01) compared with those with metastases in other locations, which contradicts global observations. Therefore, we conducted a re-evaluation to analyse lung metastases separately, both as a single site and as part of multiple metastatic sites, to determine if the original estimate was influenced by classification bias. The updated analysis showed that lung metastases as a single site were not significantly associated with increased mortality risk (HR 0.58, 95% CI 0.31 - 1.08, p=0.088). In contrast, patients with multiple metastatic sites had a significantly higher hazard of mortality (HR 2.29, 95% CI 1.14 - 4.58, p=0.019), confirming that the overall metastatic burden, rather than lung involvement specifically, was the key determinant of prognosis. Additionally, when lung metastases were compared against all other metastases in the dataset, the HR remained non-significant (HR 1.48, 95% CI 0.86 - 2.56, p=0.16). These findings indicate that the previously high HR for lung metastases was likely overestimated owing to the inclusion of multiple-site cases, and that the actual driver of poor survival outcomes was the presence of multiple metastatic sites rather than lung metastases alone.

The reason for choosing 2 weeks for a delayed presentation is that it indicates that the symptoms were not acute but have become chronic, which should be investigated for their persistence. Symptom-to-diagnosis time does not mean that the patient arrived on time. The patient was still being worked up or held up (e.g. transport and passport problems) in the periphery with symptoms, or was wrongly diagnosed elsewhere, and we were only able to diagnose it a month after the onset of symptoms. Survival times are determined from the time of diagnosis, and the median survival rate in this article refers to the OS.

Regarding Kaplan-Meier survival curves, we agree that additional subgroup analyses (e.g. unilateral v. bilateral disease, localised/single site v. metastatic disease, stage-specific survival) would be valuable. These results are shown in Fig. 1A-C, and discussed as follows.

 

Tumour side

Survival probability was assessed based on tumour laterality, categorised as left (purple), right (brown), or bilateral (green). The log-rank test showed a statistically significant difference in survival between these groups (p=0.0367). Patients with bilateral tumours exhibited the worst survival outcomes, with a more rapid decline in survival probability over time compared with those with unilateral tumours. The curves for left- and right-sided tumours were similar, suggesting that bilateral disease was associated with a poorer prognosis, likely owing to increased tumour burden and treatment challenges.

 

Metastasis status

The presence of metastatic disease (purple) was associated with significantly reduced survival probability compared with non-metastatic cases (red), with a highly significant log-rank p-value <0.001. Patients without metastasis showed substantially better survival rates over time, whereas those with metastases had a steep and continuous decline in survival probability. These findings highlight the negative prognostic impact of metastatic spread in nephroblastoma, and reinforce the importance of early detection and aggressive management in metastatic cases.

 

Tumour stage

Survival probability was compared across tumour stages (I - V) and unstaged cases, with a log-rank p-value of <0.001, indicating a significant difference in survival among the different stages. Advanced-stage tumours (likely stages IV and V) showed markedly worse survival, with a steep decline in survival probability, whereas early-stage tumours (I - II) had significantly better outcomes. The unstaged group had variable survival patterns, possibly owing to heterogeneity in disease severity or incomplete staging data. This analysis underscores the tumour stage as the most critical predictor of survival in nephroblastoma, emphasising the importance of early diagnosis and treatment initiation.

These findings confirm that bilateral disease, metastatic spread and advanced tumour stage are all associated with poorer survival outcomes in nephroblastoma. Metastasis and tumour stage have the most pronounced impact on survival, while tumour laterality, though significant, plays a relatively lesser role than metastatic status and stage. These insights highlight the need for aggressive management in advanced and metastatic cases, and support the continued use of staging systems to guide treatment decisions.

Finally, regarding the request for raw data, any request for data access is welcome. It should be made through the editor with an indication of the specific data or analysis required, and for what purpose it will be used.

We appreciate the opportunity to clarify these points, and welcome further constructive discussion.

E Brits

Department of Surgery, School of Clinical Medicine, Faculty of Health Sciences, University of the Free State, Bloemfontein, South Africa britse@ufs.ac.za

On behalf of the authors