Malignant bone tumors constitute 3-5% of all cancers in children under the age of 15 years [1]. Osteosarcoma (OS) and Ewing’s sarcoma family tumors (ESFT) are the most common subtypes, accounting for 94% of all bone tumors in children. In comparison, the remaining 6% include chondrosarcomas, malignant fibrous histiocytomas, and adamantinomas [2]. Both OS and ESFT are more common in adolescents and young adults but can also present in younger children. Metastatic disease, large tumor size, elevated lactate dehydrogenase (LDH), poor response to chemotherapy, axial tumor location, and inability to perform a complete surgical resection are associated with poor prognosis. With the best available the-rapies, current overall survival for non-metastatic disease is around 80% for both OS and ESFT, while it reduces to less than 30% for metastatic disease [2].

A recent study based on the Saudi bone tumor registry described the characteristics of bone tumors, primarily focusing on incidence, type of bone tumor, site of disease, and survival in the Saudi population across all ages. However, there is no published study from this region on the outcome and disease-associated risk factors in the pediatric age group [3]. Lebanese bone tumor collaborative program has highlighted the challenges associated with multimodality care needed for bone tumor patients in developing countries and successfully demonstrated the importance of uniformity of treatment approaches for this group of patients in the Middle East region [4]. Saudi Arabia, although a high-income country that offers free access to health care for its citizens with health spending of over 1,300 USD per capita in 2019, according to World Bank data, still lags behind the developed world in its health indices [5,6]. Looking at pediatric malignant bone tumors’ outcome should help to highlight the areas for improvement. We have analyzed our cohort of pediatric bone tumors treated at a major tertiary care center over twenty years with regard to event-free survival, overall survival, and the factors impacting the survival figures in this small but significant subgroup of pediatric malignancies.

We retrospectively reviewed both paper and electronically held data on all patients diagnosed with a malignant bone tumor in King Abdullah Specialized Children’s Hos-pital (KASCH) between Jan 2000 and Dec 2020 who were under the age of 18 years at the time of diagnosis. KASCH is a tertiary care facility that provides specialist services for all pediatric sub-specialties. Bone tumors are managed by a multidisciplinary team consisting of orthopedic surgery, bone pathology, musculoskeletal radiology, radiation oncology, and pediatric oncology services. We had limited pediatric orthopedic and radiation oncology services for the first half of the study, but from 2008 onwards, these facilities were available on-site. Variables of interest included age, gender, diagnosis, site of the tumor, symptoms at presentation including pain, swelling, weight loss, fever, and pathological fracture, radiological investigations including plain x-rays, computerized tomography (CT) scans, magnetic resonance imaging (MRI), bone scan and positive emission tomography (PET) scan, metastatic status and sites, treatment details including chemotherapy agents, surgery and radiotherapy, response assessment on imaging and treatment-related percentage of necrosis on histological specimens, outcome and treat-ment-related late effects in surviving patients.

Osteosarcoma patients were treated according to the standard arm of Children’s Oncology Group (COG) AOST0331 protocol with Methotrexate, Doxorubicin, and Cisplatin (MAP) based chemotherapy. Local control was offered to all patients in the form of surgical resection, including limb salvage where feasible. All of the ESFT patients were treated with Vincristine, Doxorubicin, Cyclo-phosphamide, Ifosfamide, and Etoposide (VDC/IE) regimen, according to recommendations of COG AEWS0031 protocol, initially at three weekly intervals before the pu-blication of results of the above study and subsequently according to the interval-compressed regimen at two weekly intervals. Local control included surgical resec-tion, including limb salvage where possible, and radiotherapy if clinically indicated because of incomplete resection and/or less than 90% necrosis on resected tumor specimens post-induction chemotherapy.

Most of the patients’ data was stored in hospital electronic records called Bestcare, which also included archived copies of patients’ historical records, while for some patients, we had to explore paper-based medical records for missing information. Patients with benign bone pathologies were excluded from this study. Means and proportions of the entire cohort were calculated to characterize the subjects. The primary outcome variable was having an event of relapse or death. To determine the factors associated with having an event, the subjects in both disease cohorts were divided into two groups based on whether a subject had had an event. The two groups of the event (relapse or death) and without event were compared using Chi-square or Fisher exact test for categorical variables and t-test or Kruskal Wallis Test for continuous variables as appropriate. We generated Kaplan– Meier survival curves for survival analyses, and comparison was done using the log-rank test. The level of significance was declared at a=0.05. Statistical analysis was conducted using SAS 9.4 (SAS Institute Inc., Cary, NC, USA).

31 patients under 18 were treated for malignant bone tumors during the study period. Six patients were excluded from the study as either they opted to be treated at other facilities or we could not find reliable records for them in our hospital medical records, leaving behind 25 evaluable patients, out of which 13 had OS, and 12 had ESFT. The male-to-female ratio was 0.8:1, while the median age at diagnosis was 8.75 years, with a range of 2-17 years for the entire cohort. Disease characteristics at presentation, including age, site, and symptoms, have been summarized in Table 1. Pain and swelling were the predominant symptoms, while a minority of patients presented with pathological fracture and systemic symptoms like fever and weight loss. Variables of prognostic significance have been summarized in Table 2 for the Osteosarcoma cohort and Table 3 for the ESFT cohort. The median follow-up time was 3.6 years, with a range of 1.1 to 10.4 years.

Table 1 . Presentation characteristics of entire cohort.

Variable	Osteosarcoma (n=13)	Ewing’s sarcoma family tumors (n=12)
Age at diagnosis in years (median)	10	7.5
Site	Femur 6 (46%) Tibia 3 (23%) Mandible 2 (15%) Humerus 1 (8%) Skull 1 (8%)	Femur 5 (42%) Humerus 2 (17%) Chest wall 2 (17%) Spine 2 (17%) Scapula 1 (8%)
Pain	7 (54%)	8 (67%)
Swelling	4 (31%)	6 (50%)
Fever	5 (39%)	2 (17%)
Pathological fracture	1 (8%)	2 (17%)
Weight loss	1 (8%)	3 (25%)

Table 2 . Treatment and prognostic variables of Osteosarcoma cohort.

Variable	Total (n=13)	Patients with event (n=6)	Patients without event (n=7)	P-value
Metastatic disease	4 (30.8%)	3 (50.0%)	1 (14.3%)	0.2657
Lung mets	4 (30.8%)	3 (50%)	1 (14.3%)
Chemotherapy	13 (100%)	6 (100%)	7 (100.0%)	0.1923
Limb salvage surgery	8 (61.5%)	3 (50.0%)	5 (71.4%)	0.7552
Negative microscopic margins	11 (84.6%)	6 (100.0%)	5 (71.4%)	0.4615
Tumor necrosis >90%	3 (23.1%)	0	3 (42.9%)	0.1923
Functional disability at follow-up	4 (30.8%)	0	4 (57.1%)	0.0699
Death	4 (30.8%)	4 (66.6%)	0	0.0210
Disease progression as cause of death	4 (30.8%)	4 (66.6%)	0	0.0210

Table 3 . Treatment and prognostic variables of Ewing’s sarcoma family tumors cohort.

Variable	Total (n=12)	Patients with event (n=6)	Patients without event (n=6)	P-value
Metastatic disease	6 (50.0%)	6 (100.0%)	0	0.0022
Lung mets	5 (42.0%)	5 (83.0%)
Chemotherapy	12 (100.0%)	6 (100.0%)	6 (100.0%)	1.0000
Limb salvage surgery	8 (66.7%)	4 (66.6%)	4 (66.6%)	0.4545
Negative microscopic margins	5 (42.0%)	3 (50.0%)	2 (33.3%)	0.6104
Tumor necrosis >90%	7 (58.3%)	4 (66.6%)	3 (50.0%)	1.0000
Radiotherapy	12 (100.0%)	6 (100.0%)	6 (100.0%)
Functional disability at follow-up	2 (16.7%)	1 (16.7%)	1 (16.7%)	1.0000
Death	4 (25.0%)	4 (66.6%)	0	0.0606
Disease progression as cause of death	4 (25.0%)	4 (66.6%)	0	0.066

In the Osteosarcoma cohort, four (31%) patients had metastatic disease at presentation in the OS group, and the lung was the most common metastatic site, present in all four patients. Out of 4 patients with metastatic disease, 3 (75%) developed relapse, and 2 (50%) of them subsequently died of disease progression. The remaining three relapses were in the non-metastatic disease group, out of which 2 (66%) died of disease progression—no significant association of metastatic disease with developing an event (P=0.26). Tumor necrosis was found to be <90% in 9 (69%) patients and >90% in 3 (23%) patients, while it was not relevant in one patient with skull-based primary who had an upfront surgical resection. Five (83%) out of 6 patients with relapsed disease had evidence of <90% tumor necrosis in resected specimens. Out of the remaining seven patients who did not develop relapsed disease, 3 (42.9%) had tumor necrosis >90%, while the remaining 4 (57.1%) had tumor necrosis <90%. The Chi-square test failed to show a statistically significant association of tumor necrosis with the risk of developing an event (P=0.20). Two (15%) patients out of 13 had evidence of positive histological margins on histopathology evaluation of resected tumor specimens, but none of them developed disease relapse and remained alive on follow-up. Of the 11 (85%) remaining patients without positive microscopic margins, 6 (55%) developed an event, while the remaining 5 (45%) did not. Again, positive histological margins were not associated with developing an event (P=0.46). Amongst six patients with relapsed disease, there were 2 (33%) patients, each with local, distant, and combined relapse. Event-free survival of the entire group was 54%, while overall survival was 69% (Fig. 1). Relapsed disease was significantly associated with poor long-term survival compared to non-relapsed disease, with a P=0.012 (Fig. 2).

Figure 1. Overall survival of Osteosarcoma and ESFT cohort.

Figure 2. Association of relapsed disease with mortality in Osteo-sarcoma cohort.

In the ESFT cohort, 6 (50%) patients had metastatic disease, with the lung being the most common site in 5 (83%) patients with metastatic disease. All six of them developed relapsed disease subsequently, evident from a P=0.002 on the risk of association with developing an event, and 4 (66%) of them subsequently died of progressive disease. We had information on tumor necrosis in 8 patients, of which 7 (88%) had it >90%. Four patients with relapsed disease, where this information on tumor necrosis was available, all had >90% necrosis. The remaining four patients, for whom we did not have information on tumor necrosis, were either operated upfront or refused surgical resection. Four patients had evidence of positive histological margins at the time of surgical resection, and only one of them went on to develop relapsed disease. In 5 patients where microscopic margins were negative, 3 (60%) developed relapse and 2 (40%) did not. For the remaining three patients for whom we had no information on histological margins, 2 of them never had any surgery, and this information was missing in one patient. Both tumor necrosis and microscopic marginal resection had no association with the risk of developing an event (P=1.0 and 0.61, respectively). Amongst six patients with relapsed disease, there were three patients, each with distant and combined relapse, but none with isolated local relapse. EFS of the entire group was 50%, while overall survival was 66% (Fig. 1). Although relapsed disease in ESFT was not associated with a poor outcome in our cohort, there was a trend toward this association with a P=0.09 (Fig. 3).

Figure 3. Association of relapsed disease with mortality in ESFT cohort.

The median age of 8.75 years for the entire study cohort was younger than the one reported by multiple authors, both in the developed and developing world, where the median age was 11-12 [1,7,8]. Male to female ratio of 0.8:1 is mostly in line with published studies on single-center experiences [7,9]. Localized pain, along with swelling or mass, has been reported as the commonest symptom for malignant bone tumors, as was the case in our study [9]. Pathological fractures and systemic symptoms like fever and weight loss are less common but important features of bone tumors, presenting in a minority of patients and should not be ignored. Osteosarcoma and Ewing sarcoma were our study cohort’s two most common malignant bone tumors, which corresponds well with published literature [1,7-9].

Non-metastatic disease, good response to preopera-tive chemotherapy by achieving >90% tumor necrosis in resected tumor specimens of malignant bone tumors, and clear microscopic margins are well-known good prognostic factors as evidenced by multiple international co-operative clinical trials [10,11]. Apart from a subgroup of ESFT patients, where there was a statistically significant association of metastatic disease with the risk of developing an event, none of the other risk factors, both in the ESFT and OS subgroups, showed any statistically significant association with worse outcomes in our cohort. Although most relapses in the OS subgroup had evidence of <90% tumor necrosis, it did not show statistical significance because of similar findings in the non-relapse group. The dose intensity was maintained during the preoperative chemotherapy phase, and no dose adjustments were required because of toxicity. We do not have any obvious explanation for this finding, but our results may have been skewed because of the small sample size. PET scan was available in hospital since 2009 but has not been consistently used in bone tumor patients for staging or response assessment purposes. Hence, we are not in a position to comment on its correlation with histological response to preoperative chemo-therapy.

The relapsed disease was associated with a worse outcome in the Osteosarcoma subgroup, where 4 out of 6 patients with relapsed disease died of disease progression while two were still alive. Although relapsed disease did not return a statistically significant association in the ESFT group, presumably because of the small sample size and late events, there was a trend toward significant association. Looking at the pattern of relapsed disease of the entire cohort, there were five distant relapses, five combined relapses, and only two isolated local relapses. This does not point toward surgical failure as a potential reason. In the osteosarcoma subgroup, one of the two survivors has had multiple relapses with local and metastatic disease, was treated with multiple resections, chemotherapy, and radiotherapy, and is recently being treated for a frontal lobe Glioblastoma multiforme, presumably secondary to radiotherapy for skull-ased recurrence, with a background of Li Fraumeni syn-drome. Second survivor of relapsed Osteosarcoma developed recurrent disease in lungs within two years from first diagnosis, was treated with Ifosfamide and Etoposide based chemotherapy along with pulmonary metastatec-tomy and remains alive four years from her relapse. In the EFST subgroup, out of the two survivors with relapsed disease, one was treated with salvage topotecan and cyclophos-phamide-based chemotherapy along with radiotherapy and remains alive three years after completing her treatment, while the second survivor has had multiple relapses and is still being treated with different combinations of second-line chemotherapy along with radiotherapy to symptomatic sites with palliative intent.

Five-year overall survival for both ESFT and OS subgroups was around 70%, while the event-free survival was 54% for the OS group and 50% for the ESFT group in our cohort. Kaatsch et al. reported a similar overall survival for combined bone tumors in the pediatric age group in Germany observed over 24 years from 1987 to 2011 [1]. The European and American Osteosarcoma Study (EURAMOS), which is a consortium of four large study groups across North America and Europe, reported a five-year overall survival of 71% and five-year event-free survival of 54% in 2,260 registered patients under the age of 40 years with a median follow up of 54 months [10]. European Pediatric Soft tissue Sarcoma group (EpSSG), in their international multicenter phase III Euro Ewing 2012 trial, reported a 3-year event-free survival of 67% in dose-ompressed VCD/IE regimen and a 3-year overall survival of 82%, which is significantly better than our cohort [12].

Despite multiple collaborations within the pediatric oncology networks, the survival figures for bone tumors have not improved significantly worldwide. Our experience adds to existing literature on pediatric bone tumors, particularly from the developing world, and has shown mixed results compared to the developed world. There is room for improvement, particularly for our ESFT cohort, where we had more events than internationally published literature. We have had access to dedicated pediatric orthopedic oncology services since 2008, supported by diagnostic radiology, pathology, and pediatric oncology services in a multidisciplinary team structure where all issues related to patient management are dis-cussed.

A critical look at our pediatric bone tumor service did not highlight any significant deficiencies compared to the standard of care worldwide. We have not used Mifamur-tide in our non-metastatic osteosarcoma patients, which is licensed for use in Europe based on the original study by Meyers et al. but is yet to be approved by regulatory authorities in North America [13,14]. There does seem to be a slight advantage in published literature by using Mifamurtide, but this drug has yet to find its place in the standard management of Osteosar-coma. We do not have access to proton radiotherapy within the Kingdom, but we send our ESFT patients to the USA for this purpose if the patient’s general condition allows traveling. Protons may not be more effective, but they offer a better long-term late effects profile in the pediatric population than conventional photon radiotherapy [15]. There were no other significant differences in our practice, and all patients received internationally accepted standard-of-care treatment for their bone tumors.

Small sample size, retrospective review over a relatively long period, and single-center experience are major limitations of our study. However, this study is the first on the Saudi pediatric population. We need to engage with other service providers within the country, ideally under the leadership of the Saudi Association of Pediatric Hematology Oncology (SAPHOS), to better understand service structures across the board and look at a larger sample size to come up with more representative conclusions. Our work will generate this dialogue among the major stakeholders in the country to improve the outcome of pediatric bone tumors.

The authors have no conflict of interest to declare.