Long- term survivors with desmoplastic small round cell tumor (DSRCT): Results from a retrospective single- institution case series analysis

Objective: To report on a retrospective study of primary DSRCT aiming at char-acterizing long- term survivors (LTS). Methods: All consecutive patients treated at our institution for a primary DSRCT between 2000 and 2021 were retrospectively identified. Patients received multiagent chemotherapy ± surgery ± hyperthermic intraperitoneal chemotherapy (HIPEC) ± whole abdomino- pelvic radiotherapy (WAP- RT) ± high- dose chemotherapy ± maintenance chemotherapy (MC). Event- free survival (EFS) and overall survival (OS) were estimated by Kaplan– Meier method. Patients alive, without evidence of disease at ≥ 36 months from diagnosis, were defined as LTS. Results: Thirty- eight patients were identified. All received multiagent chemotherapy; 27/38 (71%) surgery (7/27 [26%] plus HIPEC), 9/38 (24%) WAP- RT, 12/38 (32%) MC. At a median- follow- up of 37 months (IQR 18– 63), overall median- EFS


| INTRODUCTION
Desmoplastic small round cell tumor (DSRCT) is a high-grade ultra-rare soft tissue sarcoma (incidence 0.2/1,000,000/year), 1,2 characterized by the presence of the EWSR1-WT1 fusion gene. [3][4][5] DSRCT mostly affects male adolescents and young adults and usually presents with intra-abdominal multiple nodules originating from the peritoneal surface; it is frequently metastatic at diagnosis and the most common sites of metastasis are lymph nodes (LN), liver, and lungs. 3,4,6,7 Currently, there is no standard treatment for patient with DSRCT, even if an aggressive multimodality approach is commonly used strategy. This is based on results derived only from retrospective and often conflicting series. Primary tumor treatment includes multiagent chemotherapy, complete cytoreductive surgery and, in some cases, post-operative whole abdominopelvic radiotherapy (WAP-RT) and maintenance chemotherapy (MC). [8][9][10][11][12] Hyperthermic peritoneal perfusion with chemotherapy (HIPEC) can be combined to a macroscopically complete resection. 13 Other attempted strategies are high-dose chemotherapy with stem cell rescue. 14,15 Unfortunately, the vast majority of patients ultimately relapse and die of disease, with a reported median overall survival (OS) in the range of 17-60 months. 3,[8][9][10][11][16][17][18] A minority of patients are long-term survivors (LTS), although their features have not been characterized yet (Table SS1).
Herein, we report the results of a retrospective analysis of patients affected by primary DSRCT treated at Fondazione IRCCS Istituto Nazionale Tumori (Milan, Italy) (INT), starting from 2000, aiming to define the proportion of LTS, to describe their characteristics and identify predictors of a better outcome.

| METHODS
All consecutive patients of any age, treated at INT between January 2000 and December 2021 for primary DSRCT, were retrospectively identified. Diagnosis was reviewed according to the WHO classification, 5 and confirmed in all cases by the presence of EWSR1-WT1 fusion investigated by molecular and/or immuno-histochemical assessment.
Data were extracted from a prospectively maintained institutional database and confirmed through a review of patient files. All patients provided a written informed consent to their treatment. Approval by the institutional review board was obtained.

| Treatment
All patients were evaluated within the institutional multidisciplinary tumor board. Their treatment consisted of a multimodal approach including multiagent chemotherapy, cytoreductive surgery ± WAP-RT. Patients were treated according to different multi-agent chemotherapy regimens, all containing anthracyclines and ifosfamide. Three weeks of treatment was counted as one cycle. Data on MC (Table S2) or high-dose chemotherapy were also collected. MC was administered on a case-by-case basis following the first-line treatments until progression/unacceptable toxicity for a maximum of 12 months. Response to first-line chemotherapy was retrospectively assessed by RECIST 1.1. 19 Cytoreductive surgery was performed when technically feasible and combined with HIPEC using cisplatin upon decision of the multidisciplinary tumor board. HIPEC was performed mostly in case of disseminated peritoneal involvement and surgery was not supposed and median-OS were 15 and 37 months, respectively. All events occurred within 35 months. In patients who underwent surgery, median-EFS and median-OS were 19 and 37 months (23 and 43 months after R0/R1, and 10 and 19 months after R2 resection), respectively. LTS were 5/38 (13%), alive at 37, 39, 53, 64, 209 months.
Conclusions: In our series cure was likely achieved in 13% of DSRCT. LTS had no liver/extra-abdominal disease, were treated with complete surgery, and possibly WAP-RT/MC.

K E Y W O R D S
chemotherapy, prognosis, prognostic factors, radiotherapy, sarcoma, surgery, survival to be macroscopically complete. Peritoneal cancer index (PCI) was not used as selection criteria for surgery and/or HIPEC in our series. Surgery was classified as macroscopically complete (R0/R1) or incomplete (R2). Postoperative WAP-RT was mostly prescribed in pediatric patients, when feasible (Table S3). Total radiation dose (range between 15 and 32 Gy, with a median dose of 25.5 Gy) was based on the radiation techniques available during the study period (three-dimensional conformal vs. volumetric modulated arc therapy-VMAT) WAP-RT delivered a high radiation dose to the whole abdomen respecting dose constraints for the kidney and the liver. The kidneys were shielded after 12-15 Gy and the liver after 25-26 Gy. Treatment was delivered with fractions of 1.5 Gy.

| Statistical analysis
Patients were stratified based on disease extent at diagnosis: single mass, multiple peritoneal nodules, peritoneal nodule(s) plus LN involvement, liver metastases, supradiaphragmatic disease.
Event-free survival (EFS) was defined as the interval from the date of diagnosis to that of disease relapse, first progression or death, whichever occurred first, or last follow-up for patients without events. OS time was defined as the interval from the date of diagnosis to that of death or last follow-up. EFS and OS curves were estimated by the Kaplan-Meier method. The outcome comparison between patients treated versus not treated with WAP-RT was performed without conducting any statistical testing due to the limited number of the first subgroup. However, we conducted a matched-paired analysis by pairing each WAP-RT-treated patient with untreated patient matched by disease extent at diagnosis and the surgical resection. Patients alive and with no evidence of disease at ≥36 months from diagnosis were defined as LTS. Statistical analyses were performed using IBM SPSS 27.

| RESULTS
Between 2000 and 2021, 38 consecutive patients with primary DSRCT were retrospectively identified. At the time of diagnosis, the median age was 25 years (range 7-64). Table 1 summarizes patient characteristics and treatments. Only 12/38 (32%) patients presented with peritoneal disease alone.

| Matched-pair analysis
We successfully paired eight of the nine patients who received WAP-RT with a patient who did not present with the same disease extent at diagnosis and the same type of surgical resection. All the pairs but two had also comparable age at diagnosis. In this analysis the m-EFS and m-OS (95% CI) were 13 (8-18) and 32 (12-52) months in the no-WAP-RT group, respectively; 23 (12-34) months and not reached in the WAP-RT group ( Figure SS1-D). Without considering the two couples not matched for age at diagnosis, leading to select the above mentioned 21 R0/ R1 resected patients, m-EFS and m-OS (95% CI) in the no-WAP-RT group were 13 (range 9-19) and 24 (range 12-43) months, respectively. This exclusion did not affect the WAP-RT group (estimates already mentioned above) ( Figure 2).
The small number of patients coupled with variability of their characteristics and treatment combinations prevented additional matching for MC.

| DISCUSSION
In this series of 38 patients treated over two decades at a single institution, five patients (15%) were LTS with an m-follow-up of 53 months. LTS were managed with a multimodal approach, including multiagent chemotherapy, which in a case included high-dose chemotherapy, and macroscopic complete surgery. WAP-RT was used in 3/5 LTS, followed by MC in 2/3, while none had HIPEC. LTS did not have metastases to the liver or above the diaphragm.
This is a small, retrospective, single-institution study with all the limitations thereof. In particular, due to all the caveats of assessing radiologic and pathologic response retrospectively in soft tissue sarcoma, no correlation of response with the outcome could be assessed. In addition, the matched-pair analysis could not expand the set of matching characteristics beyond a few that were judged to have a major prognostic impact. However, DSRCT is an ultra-rare sarcoma in which prospective studies were never conducted and retrospective studies can generate hypothesis that can be then validated prospectively. 1,2 Our study confirms that DSRCT has a poor prognosis, with m-OS of 37 months and <25% of patients alive at 5 years, as reported in previous series, in which m-OS ranged from 17 to 60 months and 5-year OS from 15% to 25%. 3,[8][9][10][11]16,17,20,21 However, LTS do exist, though their proportion is low, ranging from 3% to 33% (Table SS1). Comparison of results from available studies is challenged by their retrospective design, variation in length of follow-up and, in most cases, lack of molecular confirmation of DSRCT diagnosis. The prognostic role of disease extent at diagnosis and complete surgical resection combined with multiagent chemotherapy is widely acknowledged. 3,6,11-15,21 It is also well known that recurrence or progression is common in DSRCT even after complete cytoreductive surgery, mostly within the peritoneum, reflecting issues with achieving a complete removal of the microscopic residual tumor. 3,10,22,23 In essence, our study confirms these findings. In fact, our LTS had only localized disease and were all treated by multimodal therapy, including multiagent chemotherapy and complete surgical resection (R0/R1 resection). In our series, R0/R1 surgery correlated with a better OS and EFS compared to R2 or no surgery, adding to the notion that a macroscopic complete resection may be crucial. Despite that, 15/21 patients in the R0/R1 group relapsed. Of course, our data do not exclude that R2 resection can be considered in selected conditions with palliative intent. [24][25][26] We also found that additional treatments (WAP-RT and/or MC) after R0/ R1 resection may improve survival, compared to R0/R1 surgery alone (data not shown). Given the retrospective nature of the study, we could not evaluate the impact of the prolonged multimodality approach received by the DSRCT patients treated within this study on their quality of life in the short and long term. However, this is certainly a sensitive that deserves to be investigated in future prospective studies.
Interestingly, in our study 3/5 LTS received WAP-RT, followed by MC in 2/3. During the period covered by this retrospective study, the institutional policy differed between pediatric and adult patients. Due to limited data about the role of WAP-RT and MC, in fact, these approaches in adult patients were usually not proposed. Therefore, our series cannot provide any result on the tolerability of WAP-RT in patients above 18 years, that is indeed left to be assessed prospectively. The clinical relevance of postoperative WAP-RT, which currently does not represent a standard treatment in DSRCT following a complete surgical resection, is described only in some series, with conflicting results ( Table 2). Some patients had prolonged OS and progression-free survival with WAP-RT. Conversely, two studies 10,23 comparing outcome of patients treated with or without RT did not show a survival improvement and the authors suggested WAP-RT only for selected patients within prospective clinical trials. Although small, our study would suggest a better prognosis when WAP-RT, possibly combined with MC, is added to the treatment strategy following multiagent chemotherapy and complete surgical resection in patients presenting with peritoneal/LN disease. In fact, 3/7 patients who received multiagent chemotherapy plus R0/R1 surgical resection and adjuvant WAP-RT (± MC) did not relapse, while 3/14 (one with a short follow-up) did not recur among those treated with multiagent chemotherapy plus R0/R1 resection without adjuvant WAP-RT or MC. In the subgroup of 7/21 R0/R1 patients who also received WAP-RT (followed by MC in 5/7) m-EFS was 27 months, with 3-year EFS of 43%, while in 14/21 R0/R1 patients not treated with WAP-RT, m-EFS was 16 months, with 3-year EFS of 16%. A matched-pair analysis to compare patients treated and untreated with WAP-RT that considered few characteristics (i.e., age, disease extent at diagnosis, and type of surgery) that may have a deeper impact on prognosis to make subgroups homogeneous, showed a potential benefit for WAP-RT ± MC (m-EFS 27 vs. 13 months in treated and untreated patients, respectively, with 3-year EFS of 50% vs. 0%), suggesting that maximizing systemic and local treatments may pay off at least in selected patients with DSRCT. The potential activity of MC after R0/ R1 resection has been described only in few case reports or small series. 3,6,27 Scheer et al. showed that MC with metronomic chemotherapy consisting of cyclophosphamide/vinblastine after complete surgery correlated with prolonged time to relapse. 3 Our study also suggests that MC could provide a survival benefit in R0/R1 patients. However, we could not discriminate between the prognostic effect of WAP-RT and MC as a matched-pair analysis was unfeasible due to number constraints. In terms of MC regimen, the combination of vinorelbine and low-dose oral cyclophosphamide was the most used approach (8/12 [67%]) in our patients. In the lack of prospective data, this was essentially borrowed from the experience in another sarcoma type that affects children/young adults, that is, pediatric rhabdomyosarcoma. 28 A prospective, larger study needs to be planned to shed light on this matter. Consistent with most available literature, 22,23,29 HIPEC did not result in any survival benefit in our cohort. Conversely, in the study by Hayes-Jordan, 13 R0/ R1 surgery and HIPEC significantly improved m-OS. However, the outcome differences may well be attributed to case selection and completeness of surgical resection. Finally, in a recent phase II trial 17 of 14 DSRCT treated with chemotherapy, R0/R1 surgery plus HIPEC, the combination of aggressive local therapies appeared to be beneficial in local control, but WAP-RT was incorporated as well.
The most effective chemotherapeutic regimen to consider in primary DSRCT is left to be defined. The most commonly used combinations contain anthracyclines, ifosfamide/cyclophosphamide, cisplatin, and etoposide. 3,11 Patients in our study received regimens based on anthracyclines and ifosfamide. Irinotecan was instead given only in two cases (already reported by Ferrari et al. 6 ) in combination with ifosfamide, vincristine and actinomycin-D (IrIVA), with short-interval regimens. These patients relapsed at 21 and 27 months from diagnosis. Also Liu et al. 30 reported that 3/6 patients treated with irinotecan-containing regimens, R1 surgery, and WAP-RT were alive and disease-free at 21, 47, and 60 months (m-follow-up 47 months). Concurrently, we assessed the activity of irinotecan in combination with trabectedin in a new patient-derived xenograft (PDX) model of DSRCT observing a complete response to this regimen. Finally, we saw irinotecan plus trabectedin stabilizing the disease in two patients with advanced DSRCT who failed two previous treatment lines. 31 Although very preliminary, these results suggest incorporating irinotecan in the first-line regimen for DSRCT. Indeed, a novel dose-density combination strategy using irinotecan has been recently described in high-risk sarcomas such as DSRCT. 30,32 Of note, vinorelbine also showed major antitumor effect in the DSRCT PDX model (submitted data).
In conclusion, with all the aforementioned limitations, our findings show that some patients with DSRCT become LTS, essentially if having a disease limited to the abdominal-pelvic region (peritoneum/LN). Complete surgery seems to play a crucial role while the addition of WAP-RT +/− MC to R0/R1 resection deserves further investigation. Incorporating irinotecan in the primary disease treatment also deserves further evaluation.
The retrospective nature of this analysis and the small number of patients do not allow to generate strong recommendations, but rather suggest hypotheses in order to guide new further studies. Indeed, prospective studies in homogeneous patient populations on integrated approaches maximizing local control are needed in DSRCT patients with abdominal-pelvic extent. Besides, new systemic agents are urgently needed to improve the prognosis of this ultra-rare aggressive disease.

FUNDING INFORMATION
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.