Bone Marrow Metastases In Neuroblastoma Research Articles

Radiomics models to predict bone marrow metastasis of neuroblastoma using CT.

Bone marrow is the leading site for metastasis from neuroblastoma and affects the prognosis of patients with neuroblastoma. However, the accurate diagnosis of bone marrow metastasis is limited by the high spatial and temporal heterogeneity of neuroblastoma. Radiomics analysis has been applied in various cancers to build accurate diagnostic models but has not yet been applied to bone marrow metastasis of neuroblastoma. We retrospectively collected information from 187 patients pathologically diagnosed with neuroblastoma and divided them into training and validation sets in a ratio of 7:3. A total of 2632 radiomics features were retrieved from venous and arterial phases of contrast-enhanced computed tomography (CT), and nine machine learning approaches were used to build radiomics models, including multilayer perceptron (MLP), extreme gradient boosting, and random forest. We also constructed radiomics-clinical models that combined radiomics features with clinical predictors such as age, gender, ascites, and lymph gland metastasis. The performance of the models was evaluated with receiver operating characteristics (ROC) curves, calibration curves, and risk decile plots. The MLP radiomics model yielded an area under the ROC curve (AUC) of 0.97 (95% confidence interval [CI]: 0.95-0.99) on the training set and 0.90 (95% CI: 0.82-0.95) on the validation set. The radiomics-clinical model using an MLP yielded an AUC of 0.93 (95% CI: 0.89-0.96) on the training set and 0.91 (95% CI: 0.85-0.97) on the validation set. MLP-based radiomics and radiomics-clinical models can precisely predict bone marrow metastasis in patients with neuroblastoma.

Predictive value of 18F-FDG PET/CT versus bone marrow biopsy and aspiration in pediatric neuroblastoma.

Neuroblastoma (NB) is the most prevalent solid extracranial malignancy in children, often with bone marrow metastases (BMM) are present. The conventional approach for detecting BMM is bone marrow biopsy and aspiration (BMBA). 18F-fluorodeoxyglucose-positron emission tomography/computed tomography (18F-FDG PET/CT) has become a staple for staging and is also capable of evaluating marrow infiltration. The consensus on the utility of 18F-FDG PET/CT for assessing BMM in NB patients is still under deliberation. This retrospective study enrolled 266 pediatric patients with pathologically proven NB. All patients had pretherapy FDG PET/CT. BMBA, clinical, radiological, and follow-up data were also collected. The diagnostic accuracy of BMBA and 18F-FDG PET/CT was assessed. BMBAs identified BMM in 96 cases (36.1%), while 18F-FDG PET/CT detected BMI in 106 cases (39.8%) within the cohort. The initial sensitivity, positive predictive value (PPV), specificity, and negative predictive value (NPV) of 18F-FDG PET/CT were 93.8%, 84.9%, 90.6%, and 96.3%, respectively. After treatment, these values were 92.3%, 70.6%, 97.3%, and 99.4%, respectively. The kappa statistic, which measures agreement between BMBA and 18F-FDG PET/CT, was 0.825 before treatment and 0.784 after treatment, with both values indicating a substantial agreement (P = 0.000). Additionally, the amplification of MYCN and a positive initial PET/CT scan were identified as independent prognostic factors for overall survival (OS). 18F-FDG-PET/CT is a valuable method for evaluating BMM in NB. The routine practice of performing a BMBA without discrimination may need to be reassessed. Negative result from 18F-FDG-PET/CT could potentially spare children with invasive bone marrow biopsies.

Machine-learning radiomics to predict bone marrow metastasis of neuroblastoma using magnetic resonance imaging.

Neuroblastoma is one common pediatric malignancy notorious for high temporal and spatial heterogeneities. More than half of its patients develop distant metastases involving vascularized organs, especially the bone marrow. It is thus necessary to have an economical, noninvasive method without much radiation for follow-ups. Radiomics has been used in many cancers to assist accurate diagnosis but not yet in bone marrow metastasis in neuroblastoma. A total of 182 patients with neuroblastoma were retrospectively collected and randomly divided into the training and validation sets. Five-hundred and seventy-two radiomics features were extracted from magnetic resonance imaging, among which 41 significant ones were selected via T-test for model development. We attempted 13 machine-learning algorithms and eventually chose three best-performed models. The integrative performance evaluations are based on the area under the curves (AUCs), calibration curves, risk deciles plots, and other indexes. Extreme gradient boosting, random forest (RF), and adaptive boosting were the top three to predict bone marrow metastases in neuroblastoma while RF was the most accurate one. Its AUC was 0.90 (0.86-0.93), F1 score was 0.82, sensitivity was 0.76, and negative predictive value was 0.79 in the training set. The values were 0.82 (0.71-0.93), 0.80, 0.75, and 0.92 in the validation set, respectively. Radiomics models are likely to contribute more to metastatic diagnoses and the formulation of personalized healthcare strategies in clinics. It has great potential of being a revolutionary method to replace traditional interventions in the future.

Abstract 41: 47 kda CXCR4: A marker for highly invasive neuroblastoma

Abstract Neuroblastoma; a pediatric cancer of the sympathetic nervous system, which accounts for 15% of childhood cancer deaths, presents with metastatic disease of the bone and the bone marrow at diagnosis in 70% of the cases. Although studies have shown that the CXCL12 chemokine receptor CXCR4 mediates the process of metastasis in many cancer types, its role in the bone marrow metastasis of neuroblastoma is unclear. In an effort to better understand the role of the CXCR4- CXCL12 axis in bone marrow metastasis of neuroblsatoma, we screened 20 neuroblastoma cell lines for their invasive potential, and expression of CXCR4, and sought to identify a correlation between tumor invasiveness and CXCR4 expression. Western blot analyses of cell lysates revealed the presence of multiple CXCR4 isoforms with 3 prominent species at about 35 kda, 47 kda and 62 kda. Cell lines expressing the 47 kda isoform showed higher secretion of MMP9 in gelatin zymography and higher migration and invasion towards mesenchymal stem cells (MSCs) and MSC-conditioned-RPMI, in transwell migration/invasion studies. Far western blot analysis with immuno-precipitated CXCR4, showed that the 47kda protein bound with CXCL12 at a higher rate. Blocking CXCR4 with AMD3100 or a neutralizing antibody against the 47 Kda isoform decreased the secretion of MMP9, the expression of integrin α3 and integrin β1 and the invasive potential of the cell. These data suggest that the 47 kda CXCR4 plays a prominent role in the bone marrow metastasis of neuroblastoma and could possibly be used as a diagnostic marker for highly invasive neuroblastoma and the development of minimal residual disease. Citation Format: Vipin Shankar, Lei Qi, Kentaro Kihira, Yoshihiro Komada, Hiroki Hori. 47 kda CXCR4: A marker for highly invasive neuroblastoma. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 41. doi:10.1158/1538-7445.AM2014-41

The involvement of the fractalkine receptor in the transmigration of neuroblastoma cells through bone-marrow endothelial cells

Transendothelial migration (TEM) of tumor cells is a crucial step in metastasis formation. The prevailing paradigm is that the mechanism underlying TEM of tumor cells is similar to that of leukocytes involving adhesion molecules and chemokines. Fractalkine (CX3CL1) is a unique membrane-bound chemokine that functions also as an adhesion molecule. CX3CL1 can be cleaved to a soluble fragment, capable of attracting fractalkine receptor (CX3CR1)-expressing cells. In the present study, we asked if CX3CR1 is involved in the TEM of neuroblastoma cells. We demonstrated that biologically functional CX3CR1 is expressed by several neuroblastoma cell lines. Most importantly, CX3CR1-expressing neuroblastoma cells were stimulated by CX3CL1 to transmigrate through human bone-marrow endothelial cells. A dose dependent phosphorylation of ERK1/2 and AKT was induced in CX3CR1-expressing neuroblastoma cells by soluble CX3CL1. In addition to CX3CR1, neuroblastoma cells also express the CX3CL1 ligand. Membrane CX3CL1 expression was downregulated and the shedding of soluble CX3CL1 was upregulated by PKC activation. Taken together, the results of this study indicate that CX3CR1 plays a functional role in transmigration of neuroblastoma cells through bone-marrow endothelium. These results led us to hypothesize that the CX3CR1–CX3CL1 axis takes part in bone-marrow metastasis of neuroblastoma.

Which MRI sequence of the spine best reveals bone-marrow metastases of neuroblastoma?

MRI is an effective tool in evaluating bone marrow metastases. However, no study has defined which MRI sequences or image characteristics best correlate with bone-marrow metastases in neuroblastoma. To identify and refine MRI criteria and sequence selection for the diagnosis of bone-marrow metastases in children with neuroblastoma. Ninety-one children (mean age: 3.2 years; standard deviation: 2.8 years) enrolled in the RDOG IV study participated in our study. Forty-five children had bone metastases determined by bone-marrow aspiration or biopsy (n=4), radionuclide imaging (n=2), or both (n=39). Spine lesions were characterized using coronal T1-weighted (T1W) sagittal short tau inversion recovery (STIR) and coronal gadolinium-enhanced T1-weighted (GAD) MR sequences. Contingency table analysis was performed to determine which MRI sequences and characteristics were associated with metastases. The MRI criteria for metastatic disease were then developed for each imaging sequence. The sensitivity, specificity, predictive values, and accuracy of these criteria were determined for the whole group, children younger than 12 months old, and children 12 months and older. The MR characteristics that had significant (P<or=0.05) associations with metastases were homogeneous low T1-signal intensity, homogeneous high STIR-signal intensity, and heterogeneous pattern on T1, STIR, or GAD. Homogeneous low T1-signal had the highest sensitivity (88%), but a specificity of 62% for detecting metastases. A heterogeneous pattern on GAD was highly specific (97%), but relatively insensitive (65%) for detecting metastases. These MR characteristics were most accurate in children 12 months and older. The combination of non-contrast-enhanced T1W and GAD sequences can be used to determine the presence of spinal metastases in children with neuroblastoma, particularly those children who are 1 year and older.

A possible role for CXCR4 and its ligand, the CXC chemokine stromal cell-derived factor-1, in the development of bone marrow metastases in neuroblastoma.

The homing of hemopoietic stem cells to the bone marrow is mediated by specific interactions occurring between CXCR4, which is expressed on hemopoietic stem cells, and its ligand, stromal cell-derived factor-1 (SDF-1), a CXC chemokine secreted by bone marrow stromal cells. In the present study we evaluated the possibility that neuroblastoma cells use a mechanism similar to that used by hemopoietic stem cells to home to the bone marrow and adhere to bone marrow stromal cells. Our study suggests that CXCR4 expression may be a general characteristic of neuroblastoma cells. SH-SY5Y neuroblastoma cells express not only CXCR4, but also its ligand, SDF-1. CXCR4 expression on SH-SY5Y neuroblastoma cells is tightly regulated by tumor cell-derived SDF-1, as demonstrated by the ability of neutralizing Abs against human SDF-1alpha to up-regulate CXCR4 expression on the tumor cells. The reduction in CXCR4 expression following short term exposure to recombinant human SDF-1alpha can be recovered as a result of de novo receptor synthesis. Recombinant human SDF-1alpha induces the migration of CXCR4-expressing SH-SY5Y neuroblastoma cells in CXCR4- and heterotrimeric G protein-dependent manners. Furthermore, SH-SY5Y cells interact at multiple levels with bone marrow components, as evidenced by the fact that bone marrow-derived constituents promote SH-SY5Y cell migration, adhesion to bone marrow stromal cells, and proliferation. These results suggest that SH-SY5Y neuroblastoma cells are equipped with adequate machinery to support their homing to the bone marrow. Therefore, the ability of neuroblastoma tumors to preferentially form metastases in the bone marrow may be influenced by a set of complex CXCR4-SDF-1 interactions.

Bone marrow metastasis of neuroblastoma analyzed by MRI and its influence on prognosis.

Bone marrow metastases (BM mets) of neuroblastoma were analyzed by magnetic resonance imaging (MRI). BM mets were demonstrated in 45 femurs of 23 patients with neuroblastoma by MRI. They were classified into two patterns, nodular (N) type and diffuse (D) type, on MRI, with an incidence of 16% (7/45) and 84% (38/45), respectively. No bone metastases (B mets) were accompanied by N type BM mets, whereas all of them were accompanied by D type BM mets. The rate of disappearance of N type BM mets after chemotherapy was 100% (7/7) and was significantly higher than that (26%, 10/38) of D type BM mets. The rate of disappearance (58%, 11/19) of BM mets in bones without associated B mets was also significantly higher than that (23%, 6/26) of BM mets in bones with associated B mets. BM mets present in bones with B mets tended to persist after chemotherapy. The survival rate of patients with residual BM mets after chemotherapy was significantly lower than that of patients without residual BM mets. This suggests that powerful systematic chemotherapy and additional targeted therapy to lesions remaining after chemotherapy are needed to improve the cure rate of patients with advanced neuroblastoma.

Effects of stem cell factor and other bone marrow-derived growth factors on the expression of adhesion molecules and proliferation of human neuroblastoma cells.

Metastasis in children with neuroblastoma (NB) is a poor prognostic factor despite intensive therapy. In the near future, stem cell factor (SCF) is likely to be used clinically to accelerate bone marrow (BM) recovery after high-dose chemotherapy in patients with advanced NB. The high frequency of BM metastases in NB could be secondary to BM-derived human growth factors (HGF) modulating the adhesion, secondary growth (or both) of circulating metastatic NB cells. To test this hypothesis, we studied the in vitro effects on NB cell lines grown in chemically defined medium of SCF, interleukin (IL)-1 beta, IL-3, IL-6, basic fibroblast growth factor (bFGF), transforming growth factor-beta (TGF-beta) used alone or in combination. The antigenic expression of NB-associated cell adhesion molecules (CAM) HLA class 1, intercellular CAM-1, neural-CAM and CD44 were assayed by monoclonal antibodies and flow cytometry, and DNA synthesis by 3H-thymidine uptake. The expression of CAM was not modulated by SCF or other HGFs. An increase in thymidine uptake was induced by bFGF alone in IMR-32 cells, while SCF and other HGFs had no notable effect. Our results indicate that SCF and other BM-derived HGFs are unlikely to have a generalised effect on the expression of adhesion molecules by NB cells or proliferation. The clinical administration of recombinant human SCF to children with NB should be safe.

Evaluation of bone marrow metastasis of neuroblastoma and changes after chemotherapy by MRI.

To evaluate the usefulness of MRI for diagnosing bone marrow metastasis of neuroblastoma, we compared MRI findings with histological findings. MRI was performed 26 times in 20 patients with neuroblastoma to detect metastasis to the bone marrow of the femur and tibia. Abnormal areas observed by MRI were histologically examined. The lesion visualized by MRI as a low-intensity area on T1-weighted images and as a high-intensity area on T2-weighted images was histologically confirmed to be neuroblastoma in 81% (17/21). The percentage varied according to the treatment state: 89% (8/9) by MRI imaging performed before the initiation of chemotherapy, 67% (6/9) within 3 weeks after cessation of chemotherapy (during chemotherapy), and 100% (3/3) in recurrent cases 1 year or more after chemotherapy. During the follow-up period after chemotherapy, tissue with signal intensities similar to that of bone marrow was observed in a speckled pattern in the intramedullary space on T1- and T2-weighted images. This tissue was histologically demonstrated to be normal bone marrow and was considered to be bone marrow remaining after chemotherapy. In this small series, histological findings supported the results of MRI, confirming the usefulness of MRI for diagnosing bone marrow metastasis of neuroblastoma. However, bone marrow metastasis after chemotherapy was difficult to evaluate by comparing signal intensities alone.

Detection of tyrosine hydroxylase mRNA and minimal neuroblastoma cells by the reverse transcription-polymerase chain reaction

To facilitate the diagnosis of bone marrow metastasis in neuroblastoma, we have developed a method of amplifying and detecting the tyrosine hydroxylase (TH) mRNA sequence in bone marrow cells using a combination of reverse transcription and the polymerase chain reaction (RT/PCR). By this method, the sequence of TH was detected clearly in the neuroblastoma tissues of all 6 patients and not detected in the bone marrow cells of any of the 9 negative control children. In a reconstitution experiment, 1 neuroblastoma cell per 100 000 normal bone marrow cells could be detected, thus indicating the great sensitivity of this method. Based on these results, this technique may be of value in the diagnosis and treatment follow-up of bone marrow metastasis of neuroblastoma.

Immunocytochemical detection of neuroblastoma cells infiltrating clinical bone marrow samples.

To evaluate the feasibility and clinical usefulness of immunocytochemical detection of bone marrow metastases in neuroblastoma, we studied bone marrow samples from patients undergoing intensive therapy, followed in the majority of cases by autologous bone marrow rescue. Two monoclonal antibodies were used in an indirect immunoenzymatic assay to test 384 samples collected from multiple bone marrow sites during 79 staging procedures in 48 patients. Of 578 immunocytochemical tests, 59 (10%) yielded non-evaluable results. Analysis by individual bone marrow sites showed an agreement between cytological and immunocytochemical examinations in 276 of 309 (89%) evaluable tests with 5 A7 and in 179 of 210 (85%) with UJ 13 A. Infiltration by neuroblastoma cells was reported in 9% of samples by cytology, in 6% by immunochemistry with 5 A7 and in 16% with 13 A. Analysis of results by staging demonstrated agreement between cytological examination and immunocytochemical detection with both monoclonal antibodies in 60 of 75 (80%) evaluable stagings. Bone marrow metastasis was detected by cytology in 22% of stagings, by immunochemistry with 5 A7 in 23%, with UJ 13 A in 25%. Detailed analysis of discordant results revealed that they were related partly to bone marrow sampling variability associated with focal and minimal metastasis of neuroblastoma cells. These data suggest the clinical usefulness of immunocytochemical detection as a complementary test to cytological examination for accurate evaluation of bone marrow infiltration in patients with disseminated neuroblastoma.

Chromosome Studies in Malignant Diseases in Children with Special Reference to Bone-marrow Metastasis of Neuroblastoma

Chromosome Studies in Malignant Diseases in Children with Special Reference to Bone-marrow Metastasis of Neuroblastoma