Clinical Features of Children with Retinoblastoma and Neuroblastoma

Purpose Retinoblastoma and neuroblastoma are the most common malignant extracranial solid tumors in children. This study aimed to summarize the clinical features, especially the delayed diagnosis in children with retinoblastoma and neuroblastoma. Methods In a single hospital-based case-control study, a retrospective cohort of 175 children with retinoblastoma and neuroblastoma diagnosed from January 2016 to January 2018 were reviewed. The state of enucleation in retinoblastomas and pathological prognosis in neuroblastomas were outcome indicators. Hereby, the patients were divided into two groups, and clinical features including age at presentation and delayed diagnosis were compared. Results A total of 112 patients with retinoblastoma and 63 with neuroblastoma were included. In the retinoblastoma cohort, the median age at presentation was 17.2 months (0.3–110 months). The mean delay of diagnosis was 1.6 ± 2.3 months, and the rate of enucleation was 61.6%. Unilateral disease, the International Classification of Intraocular Retinoblastoma (IIRC) stage E, and delay of diagnosis over 2.5 months were independent risk factors of ocular outcomes. Notably, the risk of enucleation was increased by 474% when the delay was longer than 2.5 months. In the neuroblastoma cohort, the delay of diagnosis of the unfavorable histology (UH) group was longer than that of the favorable histology (FH) group (1.9 months vs. 1.4 months, P=.487). The levels of serum ferritin and neuron-specific enolase were higher in the UH group than in the FH group (P < .05). Conclusions This study summarized the clinical features and diagnosis biomarkers of retinoblastoma and neuroblastoma patients in China. These results might help to focus on early detection and treatment in children with retinoblastoma and neuroblastoma.


Introduction
Retinoblastoma and neuroblastoma, which are more common in children than in adults, are different from adultonset cancers in tumorigenesis and transformation [1,2]. ey are both embryonal tumors that originate from developing organs and demonstrate a close biological connection with those organs [3][4][5]. Moreover, these tumors have an early age of onset and rapid progression [6]. According to the International Classification of Childhood Cancer (ICCC), neuroblastoma is classified as group IVa and retinoblastoma as group V [7]. In terms of molecular mechanisms of pathogenesis, mutation or amplification in specific genes, including retinoblastoma 1 (RB1) [8,9], MYCN [10], TP53 [11], and DEAD box 1 [12], are found in both diseases.
Despite unprecedented progresses in multiple therapeutics in the past decades, enucleation and unfavorable prognosis are still challenges in patients with retinoblastoma and neuroblastoma [1]. ere is no consensus on the relationship between the delayed diagnosis and adverse outcomes of these two diseases. Moreover, diagnostic biomarkers of retinoblastoma and neuroblastoma may contribute to early diagnosis and improve prognosis [13]. erefore, we aimed to summarize the clinical characteristics and diagnostic biomarkers of patients with retinoblastoma and neuroblastoma from a tertiary hospital, the National Center of Children's Health.

Study Design.
is retrospective, single-institution study was reviewed and approved by Beijing Children's Hospital, Capital Medical University Review Board (IRM2020-k-4). All children with retinoblastoma and neuroblastoma that presented to Beijing Children's Hospital from January 2016 to January 2018 were reviewed. Only newly diagnosed patients who had not received previous treatment were included. e age range was 0 to 18 years. All retinoblastoma and neuroblastoma patients were followed up for at least 2 years after diagnosis.

Data Collection.
e medical records of patients with retinoblastoma and neuroblastoma were reviewed to identify clinical characteristics and gather a complete clinical history. Data concerning demographic characteristics, family history, onset of symptoms, time from presenting symptoms to diagnosis, age at presentation, laterality, treatment, outcomes, IIRC/INSS classification, and the levels of lactate dehydrogenase (LDH) in retinoblastoma patients and the levels of LDH, serum ferritin (SF), and neuronspecific enolase (NSE) in neuroblastoma patients were collected from medical records.
Enucleation is a major concern for retinoblastoma patients; therefore, we used enucleation as an outcome indicator. e patients were divided into the enucleation (ENU) group and the no enucleation (NENU) group. In the neuroblastoma cohort, the patients were divided into favorable histology (FH) and unfavorable histology (UH) groups based on pathological classification, onset age, tumor pathology type, degree of differentiation, mitosis-karyorrhexis index (MKI), and other factors [14]. e prognosis outcomes of the FH and UH groups were used as outcome indicators among neuroblastoma patients. Moreover, the expression of serum LDH in retinoblastoma was counted, and 110-295 U/ L was the normal range in children.

Statistical
Analysis. Data were analyzed using IBM SPSS software (version 20.0) and JMP software (version 14.0). Categorical data were described as numbers and percentages. Quantitative data were described as means ± standard deviations or medians with ranges. Comparisons of categorical variables between different groups were performed using the chi-square test. When more than 20% of the cells had an expected count less than 5, Monte Carlo correction was used. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated from logistic regression models. Parametric tests were used for data with a normal distribution, and nonparametric tests were used for data that were abnormally distributed. P < .05 was considered significant unless otherwise specified.

Patient Characteristics.
e medical records of 208 patients were reviewed. One hundred and seventy-five patients were eligible: 112 patients with retinoblastoma and 63 patients with neuroblastoma ( Figure 1). Patients were scattered geographically, representing 23 provinces and three municipalities, four economic regions, and seven ethnic groups; patients were from both rural and urban areas ( Figure 2). In the retinoblastoma cohort, the median age at presentation was 17.2 months (range: 0.3-110 months). Approximately half of the retinoblastoma patients were males (57/112, 50.9%), and the Han ethnicity accounted for 92.9% (104/112) of patients. In the neuroblastoma cohort, the median age at presentation was 30.5 months (range: 1.0-102.0 months). e male : female ratio was 1.25, and there were 35 male and 28 female patients. e Han ethnicity accounted for 96.8% (61/63) of patients ( Figure 3 and Table 1).

Clinical Features of the Retinoblastoma Cohort.
Overall, 88 (78.6%) of 112 retinoblastoma patients presented at younger than 2 years of age, and 109 patients (97.3%) were younger than 5 years of age at presentation. e most common symptoms of retinoblastoma patients were leukocoria (64/112, 57.1%) and strabismus (18/112, 16.1%). Over half of the cases were unilateral retinoblastoma (72/ 112, 64.3%). e mean time from presenting symptoms to diagnosis was 1.6 ± 2.3 months (median: 0.8 months; range: 0.12 to 12.0 months). e delay of diagnosis was 1.7 months in patients with unilateral disease and 1.3 months in patients with bilateral disease (P � .502). A delay longer than 2 months occurred in 18 patients (16.1%), and a delay over 6 months occurred in 5 (4.5%) patients.
Of 112 patients with retinoblastoma, 69 (61.6%) underwent enucleation before or after chemotherapy. Univariate analysis showed that unilateral disease (P � .004), delayed diagnosis longer than 2.5 months (P � .039), stage E (P � .003), and red and inflamed eyes (P � .015) were prognostic factors of ocular outcomes ( Table 2). Multivariate analyses showed that unilateral disease, IIRC stage E, and delay of diagnosis over 2.5 months were independent risk factors of ocular outcomes when factoring in other covariates. Notably, the risk of enucleation was increased by 474% when the delay of diagnosis was longer than 2.5 months (Table 3) e median delay of diagnosis was 1.0 months (range: 0.2 to 24.0 months) in patients with neuroblastoma ( Figure 4).

Discussion
Retinoblastoma and neuroblastoma are the most common malignant extracranial solid tumors in children. ese cancers increased the burden on the state and caused pain in children and anxiety in parents [6,15]. Our findings suggested that delayed diagnosis may affect the prognosis of pediatric retinoblastoma and neuroblastoma patients, inevitably leading to heavier economic and psychological burdens. Although the survivals of retinoblastoma and neuroblastoma have improved significantly, early diagnosis is still closely correlated with better prognosis, earlier disease stage, and smaller tumor size [16]. Pediatric solid tumors have unique anatomical locations, cellular origins, and clinical presentations [2]. As pediatric embryonal tumors, retinoblastoma and neuroblastoma share similar histogenesis and histopathological features, despite having different  anatomical locations [4,17]. No study has examined the clinical features or delayed diagnosis in both diseases. erefore, this study aimed to describe the clinical features and diagnostic biomarkers of these two diseases to identify clinical features that affect the prognosis.
Because of the retrospective nature of the study and the short follow-up time, we selected ocular outcome and histology as short-term outcomes. Although these are not as accurate as survival analysis, they are still representative. Eye preservation is a concerning indicator for retinoblastoma patients and their parents, and the FH and UH classification can correctly predict prognosis in neuroblastomas according to the International Neuroblastoma Pathology Classification System [14,18].
In the retinoblastoma cohort, delayed diagnosis longer than 2.5 months was an adverse prognostic factor, which is in agreement with the results of other studies [16,19]. Xiao et al. [19] found that high-risk retinoblastoma had a longer interval from symptom presentation to enucleation than standard-risk retinoblastoma. According to Soliman et al. [16], delayed diagnosis is significantly associated with advanced disease in both unilateral and bilateral retinoblastomas. Moreover, children with a delay of diagnosis longer than 30 days are at a significantly higher risk of extraocular invasion [19]. Early diagnosis in retinoblastoma is a key concern. In the neuroblastoma cohort, delayed diagnosis   interval of the UH group was longer than that of the FH group, but it was not statistically significant. However, it is controversial that the real revenue for early screening in neuroblastoma patients considering the clinical heterogeneity of the disease. In 1985, a nationwide large-scale screening program for NB was started in Japan, but it was stopped in 2003 because of suspected overdiagnosis [20,21]. erefore, more randomized controlled trials and careful consideration are needed when implementing populationwide strategies.
Our study also found that patients with unilateral retinoblastoma typically presented later than patients with bilateral disease, and they had a worse prognosis. Patients with bilateral retinoblastoma are more likely to have a family history and experience early onset, multifocal tumors, and more severe symptoms [5].
is might be related to the timely detection of bilateral retinoblastomas by standard surveillance [22]. is also highlights the importance of early diagnosis of retinoblastomas. e causes of delayed diagnosis of retinoblastoma patients are mainly patient delay and healthcare system reasons [23,24]. For patients, ignoring symptoms and delaying seeking medical care are the most common reasons for delayed diagnosis [25]. Delayed diagnosis after seeing a general physician or nonspecific presentation is also common [26]. Moreover, socioeconomic factors, forcing some parents to refuse to face the problems after enucleation, also lead to poor compliance [17]. An example is that the age at diagnosis was significantly older in children from lower socioeconomic provinces in Argentina [27]. e recent development of body fluid-based biomarker analysis has drawn great interest. is technique overcomes the limitations of traditional tissue-based tumor analysis [13]. For neuroblastoma, the classic urine catecholamine or catecholamine metabolites, LDH, SF, and NSE are widely used in clinical disease diagnosis and prognosis [15,28]. In our study, the levels of LDH, NSE, and SF were higher in the UH group than in the FH group, which is consistent with the result of previous studies [29][30][31][32][33][34]. LDH is an important enzyme in the glycolytic pathway in tumor tissues and can be used to evaluate tumor burden and prognosis [29]. NSE, which is closely related to the advanced disease stage and poor prognosis, is a cell-specific isozyme of glycolytic enolase found in neurons, peripheral nerve tissues, and neurosecretory tissues [30]. Massaron et al. [35] found that repeated NSE measurements during follow-up could predict recurrence in patients without any clinical symptoms. ere are two forms of SF: the glycosylated form that the cell actively secretes and the nonglycosylated form, which is directly released by damaged cells [31]. Moreover, the level of ferritin returns to normal with clinical remission, suggesting that the tumor is the source of elevated ferritin [33,34]. Additionally, several genetic alterations including MEG3 polymorphisms [36], ERCC1/XPF [37], and NRAS [38] were associated with the occurrence of neuroblastoma and might serve as important diagnosis biomarkers in the future [13]. However, few clinical biomarkers are used in retinoblastoma, and we explored the value of LDH in disease diagnosis but without significance. More accurate biomarkers for therapeutic stratification and prognostication are needed for retinoblastoma.
Our findings indicated an adverse effect of delayed diagnosis on the outcomes of patients with retinoblastoma and neuroblastoma. is should be considered when formulating or revising national health policies. e current situation of delayed diagnosis of retinoblastoma and neuroblastoma emphasizes the necessity of popularization, dissemination, and promotion of medical science. Additionally, multidisciplinary management and algorithms must be widely used to reduce diagnosis and treatment delays by general practitioners.
ere are several limitations in our study because of its retrospective nature. is was a single-center study, and the reliability of our conclusions would increase by cooperation with other centers. Additionally, we could not obtain specific detailed information on the delay of diagnosis. erefore, we did not analyze the differences between parents' delays and doctor's delays. Finally, only short-term follow-up was performed, which reduces the reliability of the results. We will continue to follow our study cohort and provide results in larger cohorts. More detailed information and reliable evidence are needed.

Conclusion
is study summarized the clinical features and diagnostic biomarkers of retinoblastoma and neuroblastoma in China.
ese results might help to focus on the early detection and treatment of children with retinoblastoma and neuroblastoma.
Data Availability e datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. e data are not publicly available due to privacy and ethical restrictions.