Reduced Intensity Allogeneic Stem Cell Transplantation in Children and Young Adults with Ultra-High Risk Pediatric Sarcomas

Abstract

Some subsets of pediatric sarcoma patients have very poor survival rates. We sought to determine the feasibility and efficacy of allogeneic HSCT (alloHSCT) in pediatric sarcoma populations with <25% predicted overall survival. Patients with ultra-high risk Ewing’s sarcoma family of tumors (ESFT), alveolar rhabdomyosarcoma or desmoplastic small round cell tumor received EPOCH-fludarabine induction, a cyclophosphamide/fludarabine/melphalan preparative regimen and HLA matched related peripheral blood stem cells. Thirty patients enrolled; 7 did not undergo alloHSCT due to progressive disease with diminishing performance status during induction. All 23 alloHSCT recipients experienced rapid full donor engraftment, with no peri-transplant mortality. Five of 23 alloHSCT recipients (22%) remain alive (overall survival of 30% by Kaplan-Meier analysis at 3 years), including 3 of 7 (42%) transplanted without overt disease (median survival 14.5 vs. 29.0 months from alloHSCT for patients transplanted with vs. without overt disease, respectively). Among the 28 patients who progressed on the study, the median survival from date of progression was 1.9 months for the 7 who did not receive a transplant compared to 11.4 months for the 21 transplanted (p=0.0003). We found prolonged survival after post-transplant progression with several patients exhibiting indolent tumor growth. We also saw several patients with enhanced anti-tumor effects from post-transplant chemotherapy (objective response to pre-transplant EPOCH-F was 24% vs. 67% to post-transplant EOCH), however this was associated with increased toxicity. This largest reported series of alloHSCT in sarcomas demonstrates that alloHSCT is safe in this population, and that patients undergoing alloHSCT without overt disease show higher survival rates than reported using standard therapies. Enhanced chemo- and radio-sensitivity of tumors and normal tissues was observed post-transplant.

INTRODUCTION

Approximately 25% of children and young adults with newly diagnosed sarcomas have metastatic disease at presentation and long-term survival for these patients using standard therapies ranges from 5–35%^1,2. We defined a population of “ultra-high risk patients” as those with expected survival of less than 25%. Among pediatric sarcoma patients with metastases, patients with bone and bone marrow metastases typically experience survival rates <25%^3–5. Within this ultra-high risk group, older patients and/or patients with multifocal bony ESFT, and other risk factors can be identified that demonstrate survival rates of <10%³. In metastatic rhabdomyosarcoma (RMS), risk factors for very poor survival include older age, alveolar histology, bone or bone marrow involvement, unfavorable primary sites and three or more metastatic sites. Patients with greater than two of these risk factors have <15% overall survival^4,6. Very few pediatric sarcoma patients who experience poor responses to frontline therapy, early recurrence (within 12 months of completing frontline therapy)^7,8, or multiple recurrences survive long-term^9–11. Patients with desmoplastic small round cell tumor (DSRCT) have similarly poor outcomes, especially those with metastatic or primary progressive disease (PPD)^12–14. Thus, patients with ESFT or alveolar rhabdomyosarcoma (aRMS) who have bone or bone marrow metastases (B/BM), early recurrence (ER) or PPD, and patients with metastatic or progressive DSRCT, can be classified as ultra-high risk, based upon long term survival rates of <25%.

High-dose therapy with autologous stem cell rescue (autoHSCT) in patients with high-risk sarcomas has been widely tested in single arm studies, without conclusive evidence for clinical benefit^{2,3,6,15–21}. Immune mediated killing post alloHSCT provides antileukemic effects²² and allogeneic GVT effects have been observed in a variety of solid tumors including breast^23,24, colon²⁵, renal^25,26, ovarian, and pancreatic cancers²⁷. Some case reports of reduced-intensity alloHSCT in neuroblastoma, ESFT, and RMS have suggested a possible graft versus tumor (GVT) effect in these diseases^28–33. Historically, myeloablative allogeneic hematopoietic stem cell transplant (alloHSCT) for pediatric sarcomas was associated with significant transplant related mortality (TRM) with rates of TRM up to 40%³⁴. However improvements in supportive care and the introduction of reduced-intensity conditioning regimens has made alloHSCT safer, which may diminish toxicity and improve outcomes in this heavily pre-treated population³⁴. We sought to test the feasibility and efficacy of a reduced intensity alloHSCT-based regimen in a series of patients with ultra high-risk pediatric sarcomas.

METHODS

Enrollment and Eligibility

Eligible patients had ESFT or alveolar RMS in any of the following categories: initial diagnosis with bone or bone marrow metastases and enrolled after completion of standard frontline therapy (see supplemental methods for details of acceptable standard frontline therapy); progressive disease during standard frontline therapy, tumor recurrence within 1 year after completing standard frontline therapy and enrolled at the time of recurrence; second or subsequent recurrence. Patients with DSRCT were eligible if they had unresectable or metastatic disease (extra-abdominal and abdominal), progressive or persistent disease with standard frontline therapy, or recurrence within 1 year of completing standard frontline therapy. Eligibility was not affected by the presence or absence of overt disease at the time of enrollment. The Institutional Review Board of the National Cancer Institute (NCI) approved this study and all patients, or their guardians, provided informed consent.

Treatment Regimen

All patients received 1 to 3 cycles of pre-transplant immune depleting “induction” chemotherapy (EPOCH-F) every 21 days (previously described by Bishop et al.²³). The goal of the induction chemotherapy was to reduce host T-cells to facilitate rapid donor engraftment and provide tumor control prior to proceeding to transplant. EPOCH-F consisted of: etoposide, 50 mg/m² per day by continuous IV infusion days 1–4; prednisone, 60 mg/m² per day in 2–4 divided doses PO days 1–5; vincristine, 0.4 mg/m² per day by continuous IV infusion days 1–4; cyclophosphamide, 750 mg/m² IV over 30 minutes on day 5; doxorubicin, 10 mg/m² per day by continuous IV infusion days 1–4; fludarabine, 25 mg/m² per day IV over 30 minutes days 1–3; and filgrastim (G-CSF), 5 μg/kg per day SQ from day 6 until ANC >1000/μl × 2 days. Initially, CD4+ counts were measured following each cycle and a target of < 50 CD4+ cells/mL was the determining factor of cycle number up to a maximum of 3 cycles. Patients with progressive disease, however were taken after recovery from chemotherapy cycle regardless of cycle number or CD4+ count. However, because of brisk and sustainable engraftment, the protocol was later amended, such that disease response was the primary determinant of cycle number. Patients received three cycles of induction EPOCH-F unless they developed progressive disease, at which time determination of eligibility for proceeding to transplant was based upon performance status (patients with ECOG 3–4 were removed from study whereas patients with ECOG 0-2 were eligible to proceed to transplant). Eligibility for transplant was not impacted by the presence of overt disease.

Transplant commenced as soon as possible following hematologic recovery from the final EPOCH-F cycle. Transplant conditioning consisted of: cyclophosphamide 1,200 mg/m²/d on days −6 through −3; fludarabine 30 mg/m²/d and days −6 through −3; and melphalan 100 mg/m² on day −2. Grafts were G-CSF mobilized, unmodified peripheral blood stem cells from 6/6 HLA-matched siblings (and 1 6/6 HLA-matched father) with successful collection of all donors during a single apheresis procedure. Recipients received G-CSF (5 μg/kg/day) until recovery of the absolute neutrophil count (ANC) to 5,000/μL × 3 days.

Cohort 1 (n=13) received single agent cyclosporine-A for GVHD prophylaxis, and a second cohort, Cohort 2, (n=10) received tacrolimus plus sirolimus (see supplemental methods for dosing). Where possible, donor lymphocyte infusions (DLI) (CD3 dose range 5 × 10⁵ – 1 × 10⁸/kg) were administered to patients with disease recurrence and no evidence of GVHD after discontinuation of GVHD prophylaxis. Grading of acute GVHD (aGVHD) was based on a modified Glucksburg Scale³⁵ and the classic Seattle Criteria were used for chronic GVHD (cGVHD) grading and scoring³⁶.

Statistical Analysis

Sample size was based on a Simon’s optimal phase II trial design, with alpha=0.10 and 90% power, to target an acceptable engraftment rate (>95% donor engraftment at Day 100 in >75% of patients) and to rule out an unacceptably low engraftment rate of 50% (p1=0.75, p0=0.50)³⁷. The study also sought to measure clinical outcomes such as relapse free and overall survival. Survival was determined from the on-study date until date of death or last follow-up. When analyses are restricted to patients who underwent alloHSCT, survival from date of transplant was reported. The probability of survival as a function of time was determined by the Kaplan-Meier method³⁸. Log-rank tests were used to assess the impact of various factors on the probability of survival for patients who underwent a alloHSCT³⁹. Other analyses, such as progression free survival, based on the time from on-study date until progression or last follow-up, and survival following progression, were performed. The association between patient cohort and grade or extent of GVHD was determined by an exact Cochran-Armitage test⁴⁰. All p-values are two-tailed and reported without adjustment for multiple comparisons.

RESULTS

Patient characteristics, pre-transplant EPOCH-F

Between September 2002 and November 2008, 30 pediatric and young adult patients (median age 19, range 8–32 years) were enrolled. Clinical characteristics of all patients are summarized in Table 1. Seventeen patients had ESFT, seven had aRMS, and six had DSRCT. Patients received a mean of 2.2 EPOCH-F cycles/pt (median 3 cycle/pt), Table 2. Immune depletion was achieved as evidenced by a median CD4 count of 0 cells/mcl at the time of transplant (Figure 1). Pre-transplant EPOCH-F was well tolerated. Of 66 EPOCH-F cycles administered pre-transplant, there were 5 episodes of fever and neutropenia in 4 patients and only 1 of 66 cycles was associated with mucositis (Table 2).

Table 1.

Demographic Information for All Patients Enrolled

Patient #	Age (yrs)	Sex	Diagnosis	High-risk Features	# Prior Regimens	Disease Status at Enrollment	Disease status at HSCT	Cohort	Time to progression (days)	Survival post-HSCT (mos)
Received SCT
1	20	M	aRMS	MR	5	PD	PR	I	100	34
2	19	M	aRMS	MR	3	PD	PR	I	42	4
3	10	F	ESFT	ER	2	PD	PR	I	28	4
4	14	F	ESFT	ER, B/BM, MR	3	PD	SD	I	100	6
6	18	F	aRMS	B/BM	2	PD	SD	I	100	15
8	20	M	ESFT	PP, B/BM	1	PD	PD	I	160	21
9	19	F	aRMS	MR	2	PD	SD	I	60	7
10	30	M	ESFT	B/BM	2	PD	SD	I	100	25
12	16	M	aRMS	B/BM	2	PD	SD	I	70	13
13	12	M	aRMS	MR	3	PD	PD	I	100	77+
14	19	F	ESFT	PP	1	PD	SD	I	42	75+
15	25	M	ESFT	MR	2	PD	SD	I	100	15
17	8	F	ESFT	ER	2	NED	NED	I	NED	b9
18	25	M	DSRCT	Metastatic	2	SD	CR(NED)	II	150	45
20	18	M	ESFT	B/BM	3	PD	PD	II	38	4
a21	15	M	ESFT	B/BM	1	NED	NED	II	NED	61+
22	21	M	ESFT	B/BM	1	PD	PR	II	100	19
24	14	M	DSRCT	Metastatic	2	NED	NED	II	300	48+
25	19	M	aRMS	MR	2	PD	CR(NED)	II	170	24
27	26	M	DSRCT	PP/Metastatic	1	PD	CR(NED)	II	100	39+
a28	14	F	ESFT	B/BM	1c	NED	NED	II	271	17
29	32	M	DSRCT	Metastatic	2	PD	SD	II	237	17
30	23	M	DSRCT	Metastatic	1	PD	SD	II	180	10
Did Not Receive SCT
5	26	M	ESFT	ER, B/BM	2	PD	PD
7	17	M	ESFT	PP, B/BM	1	PD	PD
11	13	M	ESFT	ER	3	PD	PD
16	28	F	ESFT	PP	2	PD	PD
19	17	F	ESFT	ER	2	PD	PD
23	15	F	ESFT	ER	4	PD	PD
26	23	M	DSRCT	PP	3	PD	PD

^aAmong the 2 patients with ESFT enrolled in first remission, prognostic scores according to Ladenstein et al were: 3.5 points due to bone marrow, lung, multiple bone mets (Pt #21) and 5.0 points due to lung, bone marrow, multiple bone and large primary (Pt. #28)

^b– death in remission due to radiation induced bronchiolitis obliterans,

^c- up front regimen included autologous stem cell transplant.

Table 2.

Non-infectious Toxicity of Pre-HSCT EPOCH-F, Post-HSCT EOCH and Post-HSCT Radiotherapy in Patients Undergoing HSCT.

Patient #	Diagnosis	Recurrence Post-HSCT	Pre-HSCT EPOCH-F			Post-HSCT EOCH			Post-HSCT Radiotherapy
			Cycles	Best Response	Toxicity	Cycles	Best Response	Toxicity	Site	Toxicity
1	aRMS	Y	3	PR		6	CR	G3 mucositis	Chest wall	G2 skin
2	aRMS	Y	3	PR		1	PD		Abdomen	G4 GI
3	ESFT	Y	3	PR					Pancreas	G4 LFTs, G4 pancreatitis
4	ESFT	Y	3	SD		1	PD	G3 mucositis
6	aRMS	Y	3	SD					Pleura, mediastinum	G4 LFTs G2 mucositis
8	ESFT	Y	1	PD	G2 mucositis	5	VGPR	G3 mucositis, G3 esophagitis, G3 liver toxicity
9	aRMS	Y	1	SD					Chest wall	G4 skin, G3 mucositis
10	ESFT	Y	2	SD		9	PR	G2 myalgias	Spine, skull	G2 nausea + vomiting, G2 fatigue
12	aRMS	Y	3	SD					Pelvis	G4 enteritis
13	aRMS	Y	2	PD
14	ESFT	Y	3	SD		6	PR		Pulmonary (cyberknife)
15	ESFT	Y	3	SD		1	PD		Brain	G3 mucositis
17	ESFT	N	3	NED					Whole lung	G3 mucositis, G3 skin, G5 lung
18	DSRCT	Y	3	CR		10	PR	G1 mucositis
20	ESFT	Y	1	PD					L arm, R shoulder, B/L femur
21	ESFT	N	3	NED
22	ESFT	Y	3	PR		2	PR	G3 mucositis
24	DSRCT	Y	1	NED		2	PD
25	aRMS	Y	3	CR		6	CR
27	DSRCT	Y	3	CR (with XRT)		8	VGPR
28	ESFT	Y	1	NED
29	DSRCT	Y	3	SD
30	DSRCT	Y	3	SD

aRMS – alveolar rhabdomyosarcoma, ESFT – Ewing’s Sarcoma Family of Tumors, DSRCT - Desmoplastic Small Round Cell Tumor, HSCT – hematopoietic stem cell transplantation, PR-partial response, SD – stable disease, VGPR – very good partial response, CR – complete response, NED – no evidence of disease, XRT – radiotherapy, G1 – grade 1, G2 – grade 2, G3 – grade 3,, G4 – grade 4, G5 – grade 5, R – right, L – left, B/L – bilateral. Responses based on RECIST criteria, a VGPR is a >75% reduction in disease.

Figure 1.

Immune Recovery. Median cell count recovery of lymphocyte subsets over time. The “Post” time point reflects cell counts following EPOCH-F cycles. Day 0 reflects cell counts on the day of transplant following the preparative regimen, where effective immune depletion was achieved and reflected by mean CD 4 counts of 0. Patients achieved normal levels by 6 months post transplant and sustained these levels over time (norms are represented by dashed lines).

All tumor response was based on RECIST criteria. Four patients enrolled without evidence of disease and all remained NED during EPOCH-F cycles. Of 26 patients enrolled with measurable disease, 1 had a complete response (CR) following EPOCH-F with radiation therapy to sites of disease prior to transplant, and is thus inevaluable for response to EPOCH-F. Among 25 patients evaluable for response to pre-transplant EPOCH-F, there were 2 CRs, 4 partial response (PRs), 9 stable disease (SD), and 10 progressive disease (PD), for an overall objective response rate of 24%. Among the 10 patients who experienced progressive disease during EPOCH-F, seven had a rapidly diminishing performance status due to disease progression and were removed from study prior to alloHSCT, whereas three proceeded to transplant despite PD.

Clinical Outcomes

Twenty-three patients underwent alloHSCT. The transplant regimen was well tolerated with no peri-transplant mortality (within 100 days). Rapid conversion to complete donor chimerism was observed in all patients (> 95% chimerism in 22/23 by day +14, 22/22 by day +28). Median time to reach an absolute neutrophil count of 500/mm³ was 9 days (range 8–12) and a platelet count of 50,000/mm³ was 15 days (range 10–43). Immune recovery was brisk and sustained. CD4, CD8, and NK cell numbers levels approached or exceeded normal values by day +28 (median CD4+ count 284/mm³ on day +28–42) and B-cell recovery reached normal values by 6 months (Figure 1, supplemental table).

Using an intent-to-treat analysis, median survival from enrollment for all patients was 15.9 months (range 2.2 – 77.0+ months, n=30), with a median progression free survival of 5.5 months from enrollment (Figure 2A). The seven patients who did not undergo transplant due to rapidly progressive disease had a median survival of 1.9 months from progression date and 3.3 months (range 2.2–11.2) from enrollment. The 23 patients who received alloHSCT had a median survival from enrollment of 19.1 months (range 5.6 – 77.0+), and had a median survival of 16.9 months from their date of transplant. Two-year survival rates for patients undergoing alloHSCT were 39.1% from date of enrollment (Figure 2A), and 34.8% from date of transplant. Figure 2A is an illustration of survival from study enrollment for those transplanted while on study and for those unable to proceed to transplant. As transplantation occurs months following enrollment date and ability to proceed is not known at the time of enrollment, the survival time is for both those transplanted and those not transplanted and includes time patients were receiving lymphodepleting chemotherapy. At the time of last follow-up, 22% (5/23) of transplanted patients are alive with a median potential follow up of 6.0 years. Patients transplanted without overt evidence of disease had better survival outcomes than those transplanted with overt disease, (median survival 29.0 versus 14.5 months from date of transplant, Figure 2B). We detected no difference in survival based upon high-risk features, as patients with multiply recurrent tumors had similar survival rates compared to patients with progressive or persistent disease and those with high risk metastatic disease (Figure 2C). Though not statistically significant, patients transplanted with DSRCT had a longer median OS from date of transplant when compared to patients with ESFT or RMS, (23.8 mos for DSRCT vs. 15.8 mos for ESFT vs. 13.9 mos for aRMS respectively, Figure 2D). We detected no association between age, gender, or GVHD prophylaxis and survival.

Figure 2.

Clinical Outcomes. Overall survival (OS) Kaplan-Meier survival curves are shown for: (A) all patients enrolled on study; showing patients who were transplant and non-transplanted. (B) Comparison of survival from transplant for NED patients vs. those with overt disease at transplant (p=0.074). (C) Survival post-alloHSCT based upon high-risk clinical characteristics (p=NS). (D) Survival post-alloHSCT based upon histology (p=NS).

Definitive GVT effects (defined as tumor response after day 42 post-transplant without cytotoxic therapy) were not observed, and DLI were administered to 8 patients (1–7 doses), without objective evidence for anti-tumor effects. However, despite early post-transplant relapse (median 100 days, range 28 – 300 days), prolonged survival was experienced by a subset of these patients (Table 1, Figure 3). Among the 28 patients who progressed on the study, the median survival from date of progression was 1.9 months for the 7 who progressed during induction and did not undergo alloHSCT compared to 11.4 months for the 21 alloHSCT recipients who progressed (p=0.0003 for actuarial analysis; figure 3 presents a more detailed view by patient). Thus, direct evidence for graft-versus-tumor effects were not observed, but pre-transplant progression was associated with rapid death in all patients, whereas prolonged survival occurred despite post-transplant disease progression in some patients.

Figure 3.

Prolonged Survival Despite Early Post-transplant Relapse. Survival (mos) for individual patients are shown by horizontal bars. Patients not receiving alloHSCT (red) have short survivals following relapse, whereas transplanted patients (blue, green) experienced prolonged survival following relapse. Patients in green represent surviving patients (only #21 has not relapsed). Patient #17* died of bronchiolitis obliterans following post-transplant lung irradiation.

Post Transplant Response to Treatment and Toxicity

We sought to evaluate potential causes for prolonged survival despite early post-transplant disease progression. First, post-transplant progression was accompanied by indolent tumor growth in some cases as illustrated in Figure 4. Second, post-transplant progression was often responsive to chemotherapy. All patients who relapsed received further therapy, which was variable and included, but not limited to: ECOH, irinotecan, radiation therapy, and IGFR-1 antibody. Determination of chemotherapy regimen was decided by multiple factors including patient organ function, GVHD, and history of prior therapy. The mostly commonly employed regimen and one that allowed pre and post alloHSCT comparison was EOCH (fludarabine and prednisone were eliminated to minimize potential impact on GVT effects). EPOCH-F was modified to EOCH and administered to 12 patients for post-transplant disease progression (Table 2). The anti-tumor response to post-transplant EOCH was substantial: 2 CR, 6 PR (including 2 very good PRs with >75% disease reduction) for an objective response rate of 67% to post-transplant EOCH compared to a 24% objective response rate to pre-transplant EPOCH-F. As shown in table 2, of ten patients with disease evaluable for response following pre-transplant EPOCH-F vs. post-transplant EOCH, four (patients #1,8,10,14) demonstrated enhanced chemoresponsiveness post-transplant, despite equal or decreased dose of the same cytotoxic agents.

Figure 4.

Indolent Growth of Recurrent Ewing’s Sarcoma following alloHSCT.

Patient #15 showed persistent disease at the first post-transplant restaging (Day 42). No systemic therapy was administered due to patient choice and he showed minimal progression of disease over an 8-month period.

The majority of patients also experienced greater toxicity following post-transplant EOCH than was experienced with the EPOCH-F regimen, despite the same or lower doses of cytotoxic drugs administered post-transplant. Mucositis was a common and severe side effect, occurring in 5 of 12 patients (3 grade 3, 1 grade 2, 1 grade 1) receiving post-transplant EOCH, whereas mucositis (grade 2) occurred in only 1 of 30 patients receiving EPOCH-F pre-transplant. Other toxicity following post-transplant EOCH included: presumed GVHD flares of the skin, mouth, musculoskeleton and liver; 20 documented infectious episodes in 6 patients, including 7 pneumonias (4 viral, 2 fungal and 1 bacterial), 3 bacteremias and 2 episodes of cholecystitis, one of which was associated with septic shock. Enhanced toxicity was not limited to EOCH chemotherapy, as one patient developed grade 3 mucosal toxicity and grade 4 skin toxicity in a site of previous radiation following a single dose of gemcitabine chemotherapy (675 mg/m²). Notably, these toxicities were not associated with or attributed to increased hematologic toxicity, as reductions in platelet count and absolute neutrophil counts were similar following pre-transplant EPOCH-F and post-transplant EOCH (data not shown).

Graft versus Host Disease

High-grade aGVHD was not observed in cohort 1 (cyclosporine alone as GVHD prophylaxis), but 92% (12/13) of patients developed grade 1 – 2 aGVHD, and cGVHD occurred in all (12/12) evaluable patients (supplemental figure). Chronic GVHD manifestations were generally mild and easily treated, but the near universal requirement for systemic immunosuppression potentially limited GVT effects and prevented administration of DLI. We therefore sought to diminish the rate of GVHD in Cohort 2 by modifying the GVHD prophylaxis to tacrolimus and sirolimus. The incidence of cGVHD in cohort 2 diminished to 50% (5/10), with 3/10 graded as extensive cGVHD and 2/10 graded as limited cGVHD. (supplemental figure). Thus, we observed a significant reduction in the incidence of extensive cGVHD with tacrolimus/sirolimus compared to cyclosporine A alone (p = .0007). No deaths were attributable to GVHD in this series.

Discussion

Here we present the largest series of alloHSCT undertaken for high-risk pediatric sarcomas. Several conclusions can be drawn. First, while sarcoma populations were reported to suffer high rates of peri-transplant mortality with traditional myeloablative conditioning⁴¹, our current regimen accomplished rapid full donor engraftment following alloHSCT without peri-transplant mortality. Second, median survival following alloHSCT appears comparable to ultra-high-risk pediatric sarcoma patients treated with standard salvage therapy, with 22% of patients (5/23) experiencing prolonged survival. Third, 42% of patients who underwent alloHSCT with no evidence of disease remain alive, which is higher than is typically reported for this population. The favorable toxicity profile and promising results for patients rendered NED prior to alloHSCT, provides a rationale for testing this approach in a larger series of ultra-high risk sarcoma patients rendered NED, since this population has not seen advances in clinical outcomes in the last 40 years.

Our clinical observations suggest that the alloHSCT may have altered the natural history of these very aggressive tumors. Tumor progression pre-transplant was associated with rapid demise, whereas tumor progression post-transplant was associated with more prolonged survival, which may reflect heretofore unappreciated interactions between the immune milieu, tumor growth rates and chemoresponsiveness. We saw clear evidence for increased chemosensitivity of both normal and tumor tissues post-transplant compared to that observed pre-transplant. Given that fludarabine and corticosteroids have no known antitumor effects against pediatric sarcomas, the EPOCH-F vs. EOCH regimens would be expected to mediate equivalent antitumor effects. However, in this series, we observed an impressive chemotherapy response rate post-transplant of 67%, compared to a 24% response rate to essentially the same regimen prior to transplant. This is an unexpected finding, since successive tumor recurrences are typically associated with increased chemoresistance, especially to regimens that were previous administered. Selection bias could contribute to these observations, as those with the most aggressive disease did not proceed to alloHSCT, but we also observed enhanced antitumor responses post-transplant in four individual patients, suggesting that the enhanced chemotherapy response rate may be due to alloHSCT induced alterations in chemoresponsiveness.

Furthermore, chemotherapy and radiation were also associated with enhanced toxicity, demonstrating that cytotoxic effects on normal tissues were also modulated post-transplant. The enhanced post-transplant chemotoxicity was manifested primarily as mucositis and skin toxicity, without evidence for increased myelosuppression. One patient died of bronchiolitis obliterans after receiving a total dose of 900 rads prophylactic whole lung irradiation, a dose that is usually well tolerated for prevention of lung metastases in ESFT⁴². Zitvogel and coworkers have demonstrated that immune mediated events contribute to the anti-tumor effects of radiotherapy and several cytotoxic agents, including doxorubicin as used here, are associated with chemotherapy induced calreticulin exposure on tumor cells that augments innate immunity^43,44. More recently, work has also implicated IL-1 signaling in the anti-tumor effects of chemotherapy⁴⁵. We postulate that immune priming due to alloreactivity effects result in both enhanced effects of chemotherapy and/or radiation therapy targeting both normal and tumor tissues. Regardless of the mechanism, these clinical observations should serve as a note of caution for clinicians treating patients with cytotoxic agents after alloHSCT. Even very small doses of radiation or chemotherapy can mediate unexpected toxicity resulting in substantial clinical consequences. We avoid post-transplant radiation to visceral organs in our current trials and substantially reduce the doses and dose intensity of cytotoxic regimens. Our clinical observations suggest that if cytotoxic therapy is initiated with lesser doses and the doses are increased gradually on subsequent cycles, toxicity will be mitigated but clinical response may be reduced as well (data not shown).

While pre-transplant vs. post-transplant tumor growth rates were not formally evaluated in this study, our clinical observations demonstrated that in some cases tumor recurrence was associated with slow progression rates, even in the absence of cytotoxic therapy (Figure 4). It remains possible that alloreactivity mediated alterations in the immune milieu modulated tumor growth rates in this study. Indeed, recent studies of immune based therapies for prostate cancer, have demonstrated slower growth rates without tumor shrinkage, which translate into prolongations in overall survival, similar to the findings observed here^48–51. It is known that HLA class I antigens are expressed on Ewing’s sarcoma with reports that expression level can vary with progression and depends on site^46,47. While the targets of allogeneic recognition are not entirely clear, the favored hypothesis is that minor histocompatibility mismatched antigens expressed on tumor tissues serve as targets of an allogeneic response in HLA-matched HSCT and may produce some anti-tumor effects. Additionally, there is in vivo evidence of natural killer (NK) cell tumor lysis in pediatric sarcomas and evidence of allogeneic GVT effect in a single patient following KIR-mismatched transplant^52,53. This data suggests utilizing an approach potentiating NK cell lysis via KIR-mismatch, or stem cell sources enriched for NK populations may maximize GVT impact.

Using the transplant regimen presented here, GVHD rates were substantial although the severity of GVHD was modest. In this non-randomized sequential series, we saw a significant decrease in the rate of cGVHD with the addition of sirolimus to a calcineurin inhibitor. These results are consistent with anti-GVHD activity of sirolimus reported by Cutler, Antin and colleagues⁵⁴. In some series, the incidence of transplant-associated microangiopathy (TAM) was increased with the combined regimen⁵⁵. However, we observed TAM in only one patient, who had previously received total body irradiation for autoHSCT. Although sirolimus has been reported to have preclinical and clinical activity in sarcomas⁵⁶ and one could postulate that potential anti-tumor effects could impact outcomes, we could discern no evidence of anti-tumor activity with sirolimus in this study and the cohort of patients receiving sirolimus did not experience diminished tumor recurrence compared to those treated without. Thus, the addition of sirolimus to this regimen did not result in any apparent increase in tumor control, but sirolimus did significantly reduce the incidence of GVHD.

In summary, we provide evidence that reduced intensity alloHSCT can be safely administered to patients with ultra-high risk pediatric sarcomas. Patients with ultra-high risk sarcomas rendered into a state of no evident disease prior to alloHSCT experienced higher survival rates than previously reported in this population, raising the prospect that alloHSCT should be tested in a larger series of these patients. The results for patients who received alloHSCT in the presence of overt disease are also of interest, as they suggest that alloHSCT may have altered the natural history and/or chemoresponsiveness of this group of very aggressive tumors. Clinicians caring for patients treated in this manner need to be aware of the potential for enhanced toxicity when cytotoxic agents are used following alloHSCT and future studies are needed to more fully investigate whether chemoresponsiveness can be altered by changing the immunologic milieu.