Association of image-defined risk factors with clinical features, tumor biology, and outcomes in neuroblastoma: a single-center retrospective study

Image-defined risk factors (IDRF) in neuroblastoma have been developed to predict tumor resectability and surgical complications; however, the potential prognostic value of IDRF in neuroblastoma has been variably reported. Previous studies did not report the IDRF status separately from the International Neuroblastoma Risk Group (INRG) stage. Moreover, the association between IDRF and clinical and pathological factors has not been discussed further. In this retrospective study, we investigated the clinical and biological features of neuroblastoma at different INRG stages based on IDRF. Event-free survival (EFS) and overall survival (OS) related to the INRG stage were analyzed using log-rank tests, and the prognostic value of the IDRF number and type was also evaluated. Among 72 patients, 182 IDRF at diagnosis were found in 79.2%. The distribution of the INRG stages was 10 L1 (13.9.0%), 25 L2 (34.7%), and 37 M/MS (51.4%). Patients with stage M/Ms had a larger tumor volume, a higher percentage of age ≥ 18 months, elevated lactate dehydrogenase (LDH) level, elevated ferritin level, and a higher percentage of COG high-risk compared with stage L1 and L2 patients. EFS and OS were similar for stage L1 and L2 tumors but were significantly poorer for metastatic disease. However, EFS (P = 0.06) and OS (P = 0.07) were similar for IDRF-negative and positive neuroblastomas. Patients with stage M/Ms with IDRF-positive had poorer EFS (P = 0.001) and OS (P < 0.001) compared with patients in stage L2. An IDRF ≥ 4, vascular IDRF, and infiltrative IDRF of the tumor were significant indicators of poor prognosis. Conclusion: Our study indicates that increasing the INRG stages based on IDRF is associated with various unfavorable clinical features of neuroblastoma. The principal determinant of survival in neuroblastoma is the presence of metastatic disease more than IDRF alone at diagnosis. Both the number and type of IDRF have important clinical significance in the protocol planning of neuroblastoma, rather than just considering the absence or presence of IDRF. What is Known: • The International Neuroblastoma Risk Group Staging System (INRGSS) now employs image-defined risk factors (IDRFs) to stratify and stage disease. • The presence of IDRF at diagnosis are associated with higher rates of operative complications and incomplete surgical resection. What is New: • The principal determinant of survival from neuroblastoma is the presence of metastatic disease at diagnosis, more than IDRF alone. • IDRF number and type should also be considered during the diagnosis and treatment planning of neuroblastoma, rather than just considering the absence or presence of IDRF. What is Known: • The International Neuroblastoma Risk Group Staging System (INRGSS) now employs image-defined risk factors (IDRFs) to stratify and stage disease. • The presence of IDRF at diagnosis are associated with higher rates of operative complications and incomplete surgical resection. What is New: • The principal determinant of survival from neuroblastoma is the presence of metastatic disease at diagnosis, more than IDRF alone. • IDRF number and type should also be considered during the diagnosis and treatment planning of neuroblastoma, rather than just considering the absence or presence of IDRF.


Introduction
Neuroblastoma is the most frequent extracranial solid tumor in childhood, accounting for more than 7% of malignancies in children and approximately 15% of all pediatric cancer deaths [1]. Clinical heterogeneity is a hallmark of neuroblastoma. A subset of tumors can undergo spontaneous regression, whereas others can show rapid progression and death, Communicated by Peter de Winter.
Dan-dan Yang and Chuan Liu are equal contributions to this work. depending on multiple clinical risk factors and tumor biology [2,3]. Since 2009, the International Neuroblastoma Risk Group (INRG) staging system has been developed to stratify patients and stage disease, with image-defined risk factors (IDRF) as an integral part of this system [4,5].
To date, numerous studies have focused primarily on individual risk factors in the management of neuroblastoma and their prognostic implication [6][7][8][9][10][11][12][13][14][15]. To the best of our knowledge, tumor resectability and the risk of surgical complications have been proven to be associated with the presence of IDRF [6][7][8][9]. The impact of IDRF on neoadjuvant chemotherapy has also been studied in detail for neuroblastoma at different primary sites [10][11][12]. The role of IDRF in predicting outcomes has been studied in heterogeneous populations with varying results [9,[13][14][15]. Some studies have not reported the IDRF status separately from the INRG stage. In such studies, all patients were categorized as having IDRF absent or present, but due to the unknown IDRF status of patients with metastatic disease (INRG M/MS). Therefore, it is necessary to generalize the results to patients with metastatic disease. Moreover, the associations between IDRF and other risk factors, such as age, stage, tumor biology, and pathology, have not been discussed further.
Therefore, we investigated our patient cohort regarding the association of IDRF status at diagnosis with clinical features, tumor biology, and outcomes in neuroblastoma based on INRG stages. We also focused on the impact of the number and type of IDRF because we recognized that not all IDRF at diagnosis were equally important.

Patient selection
Patients with a confirmed diagnosis of neuroblastoma or ganglioneuroblastoma (intermixed or nodular) between January 2010 and May 2020 were reviewed retrospectively. This study was approved by the Institutional Review Board. The electronic medical records of 126 patients were reviewed. All patients provided informed consent for data collection and treatment. Twenty-two patients without available staging computed tomography or magnetic resonance imaging scans at diagnosis and 32 with insufficient clinical data for the analyses were excluded. Finally, 72 patients were included in the study.

IDRF analysis
CT images at diagnosis were reviewed by two senior radiologists who were blinded to the patient's clinical information. IDRF was identified according to INRG guidelines [5]. The number and type of IDRF [6] were analyzed in all patients simultaneously. IDRF was divided into five types (Supplementary Table 1): vascular IDRF (vessel encasement, regardless of the location), infiltrative IDRF (infiltration of adjacent organs or structures), neurological IDRF (encasement of brachial plexus roots, crossing of the sciatic notch, or intraspinal extension), extensive IDRF (involvement of two adjacent anatomical compartments), and airway compression. Tumor volume was assessed by imaging as follows: length × width × height × 0.52 [16]. The location and calcification of primary tumors were also recorded.

Outcome variables
Data on clinical characteristics and tumor biology were collected from electronic medical records. Clinical variables of interest included sex, age, primary tumor site, tumor calcification, International Neuroblastoma Staging system (INSS) [17], COG risk group [18], lactate dehydrogenase (LDH) level, and ferritin level. The biological variables of interest included ploidy, MYCN amplification [19], International Neuroblastoma Pathology Classification (INPC), histologic classification [20], and grade of differentiation. Event-free survival (EFS) time was calculated as the time from diagnosis to the last examination if the patient had no event. An event was defined as a relapse, disease progression, death, or a second neoplasm. Overall survival (OS) was defined as the time interval from diagnosis to death or the last follow-up.

Statistical analysis
Continuous variables were described using the median and interquartile range or mean and standard error of the mean, and categorical variables were described using absolute numbers with percentages. Continuous variables were analyzed using the Kruskal-Wallis H test and Nemenyi test (for pairwise comparisons). Categorical variables were analyzed using the chi-square test and partition of χ 2 . The EFS and OS were estimated using the Kaplan-Meier method, and survival curves were compared between each category using log-rank tests. All statistical tests were two-sided, and a P < 0.05 was considered statistically significant. Statistical analysis was performed using the SPSS software (version 22.0).

Patient characteristics and distribution of IDRF
Among the 72 patients included in the analysis, 37 were female and 35 were male before treatment. Patient characteristics are shown in Table 1. The distribution of the INRG stages was 10 L1 (13.9%), 25 L2 (34.7%), 7 MS (9.7%) and 30 M (41.7%). A total of 182 IDRF at diagnosis were found in 57/72 (79.2%) patients, and the maximum number of IDRF in a single patient was 9 (n = 1) and the median number of IDRF was 2 [1][2][3][4]. Of the 25 patients in stage L2, 17 (68.0%) had IDRF numbers ranged one to three and 17 (32.0%) had greater than four number of IDRF. Only one (1/7) patients had no IDRF in patients with stage MS. Among the 30 patients with stage M, 4 (13.3%) patients had no IDRF, 11 (36.7%) had an IDRF number ranged one to three, and 15 (50.0%) had greater than four IDRF (Table 2).

Clinical and biological features related to INRG stages
Clinical features were found to be significantly associated with INRG stages in the full cohort (

Clinical outcomes related to IDRF number and type
In the full cohort, we observed that patients with IDRF ≥ 4 had poorer EFS (X 2 = 5.743, P = 0.017) and OS (X 2 = 6.478, P = 0.011) than patients with IDRF < 4 (Fig. 3). However, in both stage L2 and M/Ms patients, no significant differences in EFS and OS were identified between IDRF ≥ 4 and IDRF < 4 (all P > 0.05).

Discussion
IDRF is of utmost importance in the INRG staging system because its presence or absence heavily influences the treatment strategy. The presence of one or more IDRF indicates the need for neoadjuvant chemotherapy. The results of our study confirm this treatment strategy. In our study, direct resection of neuroblastoma was 90.0% in stage L1, 44.0% in stage L2, and 24.3% in stage M/Ms. A robust body of literature had concluded that the presence of IDRF was associated with an increase in intraoperative complications and/or incomplete resection [7,8]. This study aimed to explore whether there is an association between IDRF and other important biological aspects and outcomes.
We found that increasing the stage of tumors based on IDRF was associated with various unfavorable clinical features of neuroblastoma. Furthermore, the principal determinant of survival from neuroblastoma is the presence of metastatic disease at diagnosis, more than IDRF alone. The amount of IDRF appears to affect oncological outcomes rather than the presence or absence of IDRF. Additionally, the tumors were found to encase important blood vessels, and infiltrating adjacent organs or structures may also be a poor prognostic factor for neuroblastoma.
Although surgery is not required for INRG staging system, the clinical biological characteristics of the tumor must be known to stratify the patients according to the INRG pretreatment classification system. Previous studies have reported the associations between IDRF and tumor clinical features and biology based on the presence or absence of IDRF [14,15,21]. Phelps et al. [15] showed that a larger tumor diameter or volume was associated with the presence of IDRF. van Heerden et al. [22] found neither an association between MYCN amplification nor an association between LDH and ferritin levels and IDRF. In our study, we found that patients with stage M/Ms had larger tumor volume, a higher percentage of patients aged ≥ 18 months, elevated LDH levels, and elevated ferritin levels compared with stage L1 and L2 patients. This also indicated that patients with a higher tumor stage had more adverse clinicopathological features, implying great clinical heterogeneity.
MYCN amplification is strongly correlated with aggressive tumor behavior and poor prognosis and is detected in 20-30% of neuroblastomas [23,24]. Patients with MYCN amplification at any age and stage were classified as "high-risk". About the correlation between MYCN amplification and IDRF, previous studies have reported different results [14,15,21,22,25]. Laura et al. [14] found that the presence of IDRF was associated with MYCN amplification, and this result was different from that of Heerden et al. [22]. Temple et al. [21] found that tumor infiltration into adjacent organs or structures is associated with MYCN amplification, and the presence of IDRF is associated with the grade of neuroblastic differentiation. Of note, one recent study found that MYCN amplification may be associated with the number of IDRF and not the absolute presence of IDRF at diagnosis [15]. Brisse et al. [25] found that MYCN-amplified neuroblastomas also had a greater number of IDRF compared to tumors with numerical-only or segmental chromosome alterations. In our study, we observed an increase in the percentage of MYCN amplification in patients with metastatic disease than in those with non-metastatic disease. Our study suggests that MYCN amplification in neuroblastoma correlates with advanced disease stages. Whether IDRF obtained from radiological images is associated with MYCN amplification also requires further investigation. The emergence of radiogenomics may provide a promising method for the noninvasive characterization and prediction of molecular biomarkers [26,27].
Previous studies have revealed different results regarding the association between IDRF status at diagnosis and long-term outcomes [28,29]. The European LNESG1 study reported that the presence of IDRF at diagnosis was associated with worse survival in localized neuroblastoma [28]. In contrast, the NB97 trial reported that the presence of IDRF at diagnosis was not significant for predicting EFS or OS, with 3-year survival rates of 99% (no IDRF) versus 98% (any IDRF) [29]. These conflicting results may be attributed to the underpowered sample size. Although IDRF-positive tumors tended to have more aggressive tumor biology, there was no significant survival advantage in patients with or without IDRF in our study. We found that patients with metastatic neuroblastoma had poorer survival than those with localized neuroblastoma (stages L1 and L2). In addition, we observed that patients with metastatic disease with IDRFpositive had poorer EFS and OS compared with patients with stage L2 disease. Therefore, the principal determinant of survival from neuroblastoma is the presence of metastatic disease rather than IDRF alone at diagnosis.
The amount of IDRF appears to affect oncological outcomes rather than the presence or absence of IDRF. In our study, ≥ 4 IDRF were predictive of poor EFS (P = 0.017) and OS (P = 0.011). Neuroblastoma with more IDRF achieves more limited resection and more surgical complications [12]. Neoadjuvant chemotherapy reduced the number of IDRF and affected the extent of delayed surgical resection. Avanzini et al. [11] showed that prolongation of chemotherapy over five courses was a risk factor for EFS and OS. These tumors may have poor chemosensitivity, and thus, the highest aggressiveness. A longer duration of chemotherapy may lead to increased toxicity and a worse prognosis. Zhang et al. [30] found that tumors with ≥ 4 IDRF and intraspinal tumor extension were significant indicators of poor prognosis in neuroblastoma by univariate analysis. Our study showed that tumors with vascular and infiltrative IDRF had poor survival outcomes. Therefore, not all IDRF are considered to be equally important. Temple et al. [21] also found that tumor infiltration into adjacent organs and structures is associated with decreased survival. Among the multiple vascular IDRF, tumor encasement of the celiac axis and/or mesenteric artery origin unfavorably affect EFS and OS [11]. Therefore, IDRF number and type should also be considered during the diagnosis and treatment planning of neuroblastoma, rather than just considering the absence or presence of IDRF.
This study has several limitations, including its retrospective nature and the small number of patients. In this study, there may also have been inherent biases in the use of medical records to collect related data. Moreover, precise histological and tumor biology data were lacking for some patients owing to the lack of standardized reporting criteria. This may have limited the interpretation of results. In addition, medical technology in different periods may have an impact on prognosis, and our study included some recent patients with a short follow-up period. Future studies with larger patient populations and longer follow-up periods are necessary.
In conclusion, the results of the current study add to our overall understanding of IDRF and its association with key clinical, biological, and outcome variables. The current study's findings demonstrate that tumors with metastatic disease have more adverse clinicopathological features and poorer survival. The presence or absence of IDRF at diagnosis was not significantly correlated with EFS or OS in neuroblastoma. Tumors with an IDRF ≥ 4 and those with vascular and infiltrative IDRF may be predictive of poor survival outcomes. Analysis of IDRF at diagnosis and extensive biological studies may assist in decision-making, surgical planning, and treatment of neuroblastoma.
Authors' contributions YDD, LC and GJ designed the study; YDD, LC and HQJ collected the data; YDD and LC reviewed the CT data; LY, LJ,GJ and YDD analyzed and interpreted the data; YDD and LC drafted the manuscript; LC, LJ and GJ contributed to critical revision of the manuscript. All authors have read and approved the final manuscript.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable.

Declarations
Ethical approval This was a retrospective study. Consent was obtained from patients' parent. No finical or nonfinancial benefits have been received or will be received from any party related directly or indirectly to the subject of this article.