Background

Ewing’s sarcoma (ES)/primitive neuroectodermal tumors (PNETs) are a well-known mesenchymal tumor belonging to a group of small round cell tumors (SRCTs) with simple sarcoma-specific genetic alterations between genes of TET/FET family and Erythroblast Transformation Specific (ETS) family with translocation of EWSR1 gene on chromosome 22q12 t(11;22)(q24;q12) occurring in 90% of cases [1]. ES/PNET expresses CD99 on its membrane that differentiates it from other SRCTs [2]. ES is commonly seen as a primary bone tumor, although it often arises from soft tissues (extra-osseous ES [EES]) [3]. ES/PNETs belong to the same spectrum of neoplasms known as Ewing sarcoma family of tumors (EFT) which also includes malignant small cell tumor of the chest wall (Askin tumor) and atypical ESS [4,5,6]. ES represents the least differentiated and PNET represents the most differentiated tumors [7]. Abdominal ES/PNETs are rare with few case reports in literature [8,9,10,11,12,13,14,15,16,17,18,19,20]. Abdominal ES/PNETs are often confused with other SRCTs such as embryonal rhabdomyosarcomas, neuroblastomas, and lymphomas. This makes it even more important for accurate diagnosis to provide evidence-based multimodality management for optimal outcomes.

The primary objective was to study the management of abdominal ES along with a review of literature.

Methods

Data Source and Patient Population

The demographic, clinical, radiological, treatment, histopathological, and survival data of patients were obtained retrospectively from hospital’s electronic medical records. The inclusion criteria were:

  1. 1.

    Biopsy proven abdominal ES/PNET disease

  2. 2.

    Age ≥ 14 years

  3. 3.

    Eastern Cooperative Oncology Group (ECOG) performance status 0–1

  4. 4.

    Surgical treatment in the institute

The exclusion criteria were:

  1. 1.

    Poor performance status (ECOG ≥ 2)

  2. 2.

    Presence of metastatic disease

Ten patients identified and registered between June 2005 and November 2019 were included in the study. The diagnosis was established by histopathological examination, and immunohistochemical analysis. Molecular study was not routinely performed for diagnosis in the institution. Disease was staged using 18-Fluorodeoxyglucose Positron Emission Tomography with Contrast Enhanced Computed Tomography (FDG-PETCECT) scan (except for two patients for whom CECT scan was done) and bone marrow biopsy was obtained in patients treated till 2010 (four patients). As a part of Institutional policy, PETCECT scan was selectively used from 2005 to 2010. After 2010, PETCECT scan was incorporated as a staging tool and subsequently bone marrow biopsies were omitted. Clinical tumor (cT) size (defined by its largest dimension) and maximum standardized uptake value (SUV max) were recorded and compared before and after chemotherapy. All patients were discussed in multidisciplinary joint clinics (MDJCs) involving surgical oncologist, medical oncologist, radiation oncologist, pathologist, nuclear medicine consultant, and radiologist. The chemotherapy regimen followed was EFT 2001 protocol [21, 22]. Post-chemotherapy, surgery was performed after clinical and radiological assessment through MDJC. Post-surgical resection, patients were continued on maintenance chemotherapy as per protocol. Complications were graded and recorded as per Clavien-Dindo classification of surgical complications [23]. Adjuvant radiation was given to patients based on histopathological findings, tumor volume, margin status, and response to chemotherapy. All patients were followed-up at three-monthly intervals for the first three years, six-monthly intervals for the next two years, and annually thereafter. In view of COVID-19 pandemic, all follow-ups after February 2020 were done via telephonic calls.

Statistical Analysis

The date of registration at the hospital was considered as date of diagnosis for statistical purposes. RFS was defined as interval between dates of diagnosis and appearance of first recurrence. RFS was calculated using the Kaplan–Meier method and compared using the log rank test. A univariate Cox regression hazard model analysis was used to analyze any prognostic factors of outcome and the hazard ratio (HR), relevant 95% confidence intervals (CIs) were calculated. P < 0.05 was considered to indicate a statistically significant difference. Wilcoxon signed rank test was used to test whether there was any significant change between pre-chemotherapy FDG-PETCECT scan (cT) size and SUV max due to chemotherapy. Spearman rho coefficient test was used to derive correlation between pathological tumor size (pT) with post-chemotherapy cT size and post-chemotherapy SUV max. The database of all patients in the review was compiled and analyzed using SPSS v.21 software (IBM Corp.).

Results

Baseline Patient Characteristics and Staging

Ten patients were analyzed in this study and the baseline patient characteristics, staging investigations, baseline cT size, and SUV max and post-chemotherapy cT size and SUV max are shown in Table 1. There were six males and four females with a median age of diagnosis of 36.5 years (range 19 to 46 years). The site of origin was retroperitoneum (six patients), small bowel (two patients), and abdominal wall (two patients).

Table 1 Demography, staging, FDG-PETCECT response
Full size table

Surgical Aspects

All patients (except one — patient E) underwent R0 resection of which five patients required multivisceral resections [Table 2 (A)]. The median blood loss was 450 ml (range: 100–1900 ml) and only three patients (A, D, and E) had Clavien-Dindo grade-1 complications. Table 2 (A) shows the histological features of these patients. The median pathological tumor size (pT) was 13.50 cm (range 2.20 cm to 20 cm). Five patients had disease involving the resected organs (B, E, F, G, and I). Two patients (E and I) had peritoneal metastasis, patient I underwent R0 resection while patient E underwent R2 resection (95% tumor was grossly debulked, residual multiple small deposits were present on parietal peritoneum and sigmoid colon mesentery).

Table 2 Surgical, histopathological, chemotherapy, radiotherapy and recurrence details
Full size table

Table 2 (B) provides the details of induction, maintenance chemotherapy, and adjuvant radiotherapy received by these patients. Patient B underwent emergency surgery in view of intra-luminal bleeding. The patient presented with anemia (hemoglobin 5.1 gm/dL) and 12 × 10 cm mass. Two patients (E and I) had recurrences [Table 2 (B)], patient I was re-challenged with chemotherapy and underwent cytoreduction to achieve R0 resection whereas patient E was deemed unresectable and continued palliative chemotherapy.

FDG-PETCECT Scan

Eight patients underwent baseline FDG-PETCECT scan, and post-chemotherapy FDG-PETCECT scan of one patient was not available, hence, seven patients’ data were available for analysis (Table 1). Chemotherapy did not cause any statistical significant difference between pre-chemotherapy and post-chemotherapy median SUV max (p = 0.31); however, there was significant reduction in cT size with chemotherapy (p = 0.034). On univariate analysis, post-chemotherapy SUV max and cT size were not prognostic or predictive of recurrence. There was a weak positive correlation between pT size and cT size (rho = 0.252; p = 0.585) and between pT size and post-chemotherapy SUV max (rho = 0.500; p = 0.253).

Outcomes and Prognostic Factors

The median follow-up was 25 months (range: 3–178 months). The 1- and 2-year RFS of the cohort was 88.90% (95% CI: 70.60 to 100%) and 76.20% (95% CI: 52.1 to 100%), respectively (Fig. 1). Overall survival (OS) could not be calculated, as there were no events in the series. The only prognostic factors associated with recurrence were presence of pathological peritoneal disease/metastasis (p = 0.0026) and ability to achieve R0 resection (p = 0.11) (Table 3). Two patients developed recurrences, patient E had recurrences as retroperitoneal and mesenteric disease with subcutaneous umbilical nodule, and patient I developed recurrence in the form of peritoneal deposits in diaphragmatic peritoneum, pelvis, and transverse mesocolon.

Fig. 1
figure 1

Recurrence-free survival

Full size image
Table 3 Prognostic factors
Full size table

Discussion

There are subtle biological differences between ES/PNET arising from bones and EES that lead to varying presentation in terms of age, tumor characteristics, clinical features, treatment strategies and outcomes [24,25,26]. ES/PNETs most often present as primary skeletal tumor in the trunk or axial skeleton in adolescents and young adults. Due to the rarity of EES the available literature is scarce and limited to few case reports. However, in recent years, there has been an increase in case reports which might be because of the recognition of EES and abdominal ES/PNET as a separate entity and improved cytogenetic and molecular diagnostic methods. Hence, it becomes necessary to identify this rare group of patients early in the course of disease.

Patient characteristics in EES differ than in those with skeletal ES/PNET. In our study of 10 patients (six males, four females), the median age at diagnosis was 36.5 years. The median age across different case reports of abdominal ES/PNET have been in similar age-groups (15–35 years) [8, 10, 12,13,14,15, 20, 21]. National Cancer Database (NCDB) and Surveillance, Epidemiology, and End Results (SEER) database studies show that the median age of patients at diagnosis of EES is older than its skeletal counterparts and has female predilection. EES patients are more likely to be < 5 or ≥ 35 years at diagnosis [25, 26]. The average tumor size did not differ between skeletal ES/PNET and EES [25].

ES/PNET tumors have increased rate of glycolysis and hence the cells retain avidity. Meta-analysis of 23 studies by Huang et al. showed sensitivity and specificity that were 86% and 80%, respectively with higher sensitivity and specificity (93% and 90% respectively) for detecting recurrences [27]. Hawkins et al. [28] and Salem et al. [29] conducted retrospective studies and showed that FDG-PETCECT scan correlates with histologic response, is prognostic and predictive of survival outcomes. In our study, there was significant difference between pre-chemotherapy and post-chemotherapy cT size (median cT size 11.40 cm vs 7.30 cm respectively, p = 0.034). However, the SUV max did not statistically differ in pre-chemotherapy and post-chemotherapy setting (10.34 vs 4.60 respectively, p = 0.31). This could either be because of less number of patients, or biological differences in behaviors between abdominal ES/PNETs with its skeletal counterparts. We also did not find a strong positive correlation between pT size and cT size or post-chemotherapy SUV max. However, chemotherapy did bring about significant reduction in cT size (p = 0.034).

The NCDB retrospective cohort study analyzed 2660 ES/PNET patients (2004–2014) containing 1682 (63.2%) skeletal ES/PNET and 978 (36.8%) ESS patients. They concluded that EES patients have well-differentiated tumors, more likely to be in stage I disease with no significant difference in metastatic disease at diagnosis, perioperative outcome and no significant difference in OS of the entire cohort (median OS: 47.5 months for skeletal ES/PNET and 48.2 months for EES) [25]. In the SEER database of 2202 patients, the OS for localized ESS was superior to skeletal ES/PNET in both 5- and 10-year OS; 69.7% vs 62.6% and 65.2% vs 55.3%, respectively. However, when localized and metastatic ES/PNETs were analyzed together or in metastatic disease alone, the OS is similar in ESS and skeletal ES/PNET. They also suggested that patients with ESS had worse OS in the first 24 months but better OS after 24 months as compared to skeletal ES/PNET patients, after controlling for other known prognostic factors suggesting that extra-skeletal site has greater impact on the long-term outcomes [26]. Studies with small number of patients also showed no significant differences in survival outcomes [30, 31].

In our study, the 1- and 2-year RFS were 88.90% and 76.20%, respectively. This survival figures are impressive since they represent patients who were operated upon and did not include metastatic/inoperable patients. With multimodality uniform treatment comprising of induction chemotherapy, surgery, maintenance chemotherapy, and adjuvant radiation (when indicated) superior outcomes can be expected.

ES/PNET is considered a systemic disease with majority patients have subclinical metastatic disease at the time of diagnosis even in the absence of overt metastasis. Modern treatment protocols utilize initial chemotherapy (induction/neoadjuvant) followed by local treatment and further chemotherapy (maintenance/adjuvant). Induction chemotherapy is important to downstage the tumor to achieve R0 resection, improve local control and control metastatic disease [32,33,34]. EES patients respond to the same chemotherapy regimens as skeletal ES and should be treated similarly [35,36,37,38]. The chemotherapy protocols have evolved over years with the IESS-I-III studies leading to the current regimen which constitutes alternating cycles of ifosfamide and etoposide (I/E) to vincristine, adriamycin, and cyclophosphamide (VAC) backbone (VAC/IE) [39,40,41].

All of our patients received chemotherapy as per EFT 2001 protocol. Three patients did not receive induction chemotherapy (patient B underwent emergency surgery in view of bleeding while the reason for patients A and D were not available).

ES/PNET is a radiosensitive tumor. Radiation is more commonly used in skeletal ES/PNET as compared to EES [26]. There are no randomized control trials comparing surgery and radiation. Surgery is the preferred modality of choice for local control if the tumor can be resected with negative margins without excess morbidity and with a reasonable functional result. Surgery also avoids the risk of secondary radiation-induced sarcomas and provides for analysis of degree of necrosis. However, if a tumor is unresectable following induction chemotherapy, the patient should be referred for definitive radiation. Data from skeletal ES/PNET tumors provide for the indications of adjuvant radiation in the following settings:

  1. i.

    Bulky tumors in difficult sites

  2. ii.

    Macro- or microscopic residual viable tumor after surgery

  3. iii.

    Positive or inadequate surgical margins

  4. iv.

    High risk chest wall tumors

In our series, 4/10 patients (40%) received adjuvant radiation, two patients had retroperitoneal origin, and two had abdominal wall origin.

In our study, presence of peritoneal metastasis was found to be statistically significant factor (p = 0.0026) and R0 resection (p = 0.11) was found to be non-significant factor but with trend towards significance (Table 3).

The limitations of our study are its small number of cases, the lack of events observed, and lack of long-term data. However, as we have limited our study to exclusively include abdominal ES/PNETs who underwent surgical resection, this number is still significant. The lack of events observed in our study shows that this subset of patients have good prognosis with multimodality approach, aggressive surgical resection, and salvage surgery whenever feasible.

Conclusion

Abdominal ES/PNET being a rare entity requires special attention with accurate histopathological diagnosis, IHCs and molecular studies (in special cases) and baseline imaging with FDG-PETCECT. Tumor size on FDG-PETCECT is predictive of tumor response to chemotherapy. All patients must be offered multimodality treatment including induction chemotherapy, surgery, and maintenance chemotherapy and adjuvant radiation when indicated. Every attempt to obtain R0 resection must be made and induction chemotherapy significantly reduces the tumor size. Presence of pathological peritoneal metastasis and inability to achieve R0 resection affect survival significantly.