Lack of complete response pre-transplant is not associated with inferior overall survival for stage 4a metastatic retinoblastoma

Abstract

Background:

Stage 4a metastatic retinoblastoma (RB) is curable with intensive multimodality therapy including myeloablative chemotherapy with autologous stem cell transplant (HDC-ASCT) and involved field radiation therapy (IFRT). To our knowledge, no data exists on the impact of (1) pre-ASCT disease status and (2) IFRT to sites of metastatic disease post-ASCT on survival.

Procedure:

We retrospectively reviewed patients with stage 4a metastatic RB who underwent induction chemotherapy followed by HDC-ASCT with or without IFRT to residual tumor sites at MSKCC (n=24).

Results:

The degree of post-induction response prior to ASCT did not affect outcome with 5-year OS of 68% and 86% in patients who achieved CR and VGPR/PR prior to ASCT, respectively. IFRT administered post-ASCT in patients with possible residual bony metastatic disease increases the likelihood of developing osteosarcoma in the radiation field.

Conclusion:

Overall survival for patients with stage 4a metastatic retinoblastoma treated with ASCT with very good partial response or partial response to pre-transplant chemotherapy was not significantly different from patients with complete response. In addition, involved field radiation therapy does not seem to be required for bony disease control and increased the likelihood of developing osteosarcoma.

Introduction

Metastatic retinoblastoma (RB) is a highly lethal disease and the prognosis with conventional therapy has been poor.^{1, 2} Recent reports^3–9 including a prospective, multi-center, international study (ARET0321)¹⁰ demonstrated excellent survival results for patients with stage 4a RB (distant metastatic disease not involving the central nervous system) when treatment was intensified to include high-dose chemotherapy (HDC) with autologous stem cell transplant (ASCT) combined with involved field radiation therapy (IFRT) for patients with residual disease post chemotherapy. To our knowledge, no data exists on the impact of (a) pre-ASCT disease status and (b) IFRT to sites of metastatic disease post-ASCT on survival. We previously reported survival data for a small series of patients (n=15) with stage 4a RB who received intensive multimodality treatment with intent to include HDC-ASCT, even if patients did not proceed to ASCT⁸. In this manuscript, we describe an expanded cohort of patients (n=24) who underwent HDC-ASCT, including some that were previously reported⁸, with longer follow-up coupled with an evaluation of the impact of pre-ASCT disease status and IFRT on patient outcomes.

Methods

We performed an IRB-approved retrospective chart review of all patients with stage 4a metastatic RB who underwent induction chemotherapy followed by myeloablative consolidation chemotherapy with ASCT with or without IFRT to residual tumor at MSKCC between 1993–2020. Patients with central nervous system disease (trilateral and stage 4b retinoblastoma) were excluded.

At the time of diagnosis of metastatic disease, extent of disease evaluations consisting of magnetic resonance imaging of the brain and orbit, lumbar puncture for cerebrospinal fluid cytology, bone marrow aspirates and biopsies, bone scans and abdominal MRI or CT scans were performed. All patients received induction chemotherapy and HDC-ASCT as per or enrolled upon clinical research protocols that were approved by the MSKCC IRB.

Response criteria

A complete response (CR) was defined as resolution of the metastatic disease on bone marrow examinations and relevant imaging studies (those that revealed an abnormality prior to enrollment on this study). A very good partial response (VGPR) was defined as a > 90% decrease in the sum of the 2-dimensional measurements of the tumor(s) on relevant (CT and/or MRI) imaging studies, no new lesions, bone marrow free of tumor and improved bone scan. A partial response (PR) was defined as a 50–90% decrease in the sum of the 2-dimensional measurements of the tumor(s) on relevant (CT and/or MRI) imaging studies, no new lesions, and bone marrow free of tumor.

Induction Chemotherapy

The induction chemotherapy was intensive and alkylator-based in all cases. All patients received vincristine, cisplatin and/or carboplatin, cyclophosphamide, and etoposide. Two patients received doxorubicin (subjects#13 and #21) and one patient received additional topotecan (subject #17). All patients who proceeded to ASCT achieved an objective response to chemotherapy (CR=8, VGPR=13, PR=3). Chemotherapy doses and schedules are provided in Supplementary table 1.

High-Dose Chemotherapy with Autologous Hematopoietic Stem Cell Transplant

Twenty-four patients received HDC-ASCT at a median of 5.3 months post-diagnosis of metastatic relapse (range 3.8–8.3 months) after achieving a partial or complete response to induction chemotherapy and all patients experienced hematopoietic recovery post ASCT. All 24 received carboplatin and thiotepa; 23 also received a third agent, either etoposide (n=16) or topotecan (n=7). Details of the high-dose carboplatin, thiotepa and etoposide¹¹ and high-dose thiotepa, topotecan and carboplatin regimens¹² have been previously reported.

Radiation Therapy

Following recovery from HDC-ASCT, twelve patients received IFRT to prior sites of bony disease (3 CR, 8 VGPR, 1 PR pre-ASCT) with doses ranging from 2100 – 4500 cGy. Details of pre-ASCT disease status, receipt of IFRT and survival outcomes are provided in table 1. Two patients received IFRT to liver (1,950 cGy).

Table 1.

Treatment details and outcomes for patients with stage 4a retinoblastoma who underwent cytoreduction with ASCT

No.	Induction chemo	Cytoreduction	Status of disease prior to ASCT	Status of disease post ASCT	RT post ASCT	Osteo-sarcoma	Relapse	Time to relapse in months	OS from HSCT in months	Status
1	92–016	CTE (1)	VGPR	CR	L orbit (4000), liver (1950), femur (3000)	N	N		314.0	RBFS
2	ARET0321	CTE (0)	CR	CR	N	N	N		16.0	RBFS
3	ARET0321	CTE (1)	PR	PR	Tibia (3600)	N	N		30.6	RBFS
4	92–148	CTTo (0)	VGPR	VGPR	n/a	N	N		180.6	RBFS
5	ARET0321	CTE (0)	VGPR	CR	N	N	N		144.4	RBFS
6	92–148	CTTo (0)	CR	CR	N	N	Y(cerebellar mass)	9.3	23.3	DOD
7	ARET0321	CTE (0)	VGPR	CR	N	N	N		25.3	RBFS
8	ARET0321	CTE (1)	VGPR	CR	R orbit (3600)	N	N		97.2	RBFS
9	ARET0321	CTE (1)	VGPR	VGPR	L orbit (4410)	N	N		73.6	RBFS
10	ARET0321	CTE (0)	PR	CR	N	N	N		6.4	RBFS
11	ARET0321	CTE (0)	CR	CR	N	N	N		44.8	RBFS
12	ARET0321	CTE (1)	VGPR	PR	36Gy right lateral orbital wall/sphenoid bone and right base of skull	N	N		101.0	RBFS (GBM)
13	N6 (VDoCyPE)	CT (0)	VGPR	CR	skull base, cavernous sinus (2100)	Y	N		198.7	RBFS (OS - dead)
14	92–148	CTTo (1)	VGPR	VGPR	N	N	N		172.9	RBFS
15	ARET0321	CTE (0)	CR	CR	N	N	Y (lepto)	7.2	12.8	DOD
16	92–148	CTTo (1)	VGPR	VGPR	L4-Sacrum (3600)	N	Y (distal Mandible)	12.6	22.8	DOD
17	ARET0321 + topotecan	CTE (1)	PR	CR	180 cGy left orbit *prior to transplant	N	N		121.7	RBFS
18	92–148	CTTo (1)	CR	CR	R femur (4500)	Y	N		119.2	RBFS (OS - dead)
19	92–148	CTTo (1)	VGPR	CR	R Tibia (3960)	Y	N		77.1	RBFS (OS- dead)
20	V,P,Cy,E	CTE (1)	CR	CR	L orbit (4000) Liver (1950)	N	N		189.8	RBFS
21	resections x2, N8, ARET0321	CTE (0)	VGPR	PD	N	N	Y (multiple mets)	1.1	5	DOD
22	ARET0321	CTE (0)	VGPR	VGPR	L mandible (3600)	N	N		17.0	RBFS
23	ARET0321	CTTo	CR	CR	N	N	N		3.6	RBFS
24	V,C,Cy,E	CTE (1)	CR	CR	L Orbit (4000) R humerus (4000)	Y	N		156.3	RBFS (OS - alive)

CR- Complete response, VGPR – Very good partial response, PR- Partial response, RBFS- Retinoblastoma free survivor, DOD – Dead of disease

Cytoreduction regimens: CTE- Carboplatin, Thiotepa, Etoposide, CT – Carboplatin, Thiotepa, CTTo – Carboplatin, Thiotepa, Topotecan.

(0) – as per protocol

(1) – Enrolled on protocol

Statistics

Overall survival (OS) was calculated from the date of ASCT until the date of death, or last follow up. Retinoblastoma-free survival was calculated from the date of ASCT until the date of disease progression, death, or last follow-up. Event-free survival (EFS) was calculated from the date of ASCT until the date of disease progression, second malignancy, death, or last follow-up. RBFS, and EFS probabilities were estimated using the Kaplan–Meier method. Descriptive statistics were used for analysis and Kaplan-Meier survival curves were generated using Prism software.

Results

Twenty four patients (11 females, 13 male) were identified. Two patients with stage 4a RB were excluded from this analysis because they did not receive HDC-ASCT due to significant intracranial hemorrhage (n=1) and CNS leptomeningeal disease recurrence/progression at the completion of induction chemotherapy (n=1). Nineteen patients had unilateral RB and five patients had bilateral RB. Patients were tested for germline RB1 alterations and results were positive (n=5), negative (n=13). Of the remaining patients with unknown or uncertain RB1 germline testing, one patient had bilateral disease and five patients had unilateral disease with a negative family history. Treatment of the intra-ocular disease included enucleation (n=16), intra-arterial chemotherapy (n=9), systemic chemotherapy (n=4) and IFRT (n=4) (see Supplementary table 2). Metastatic disease was diagnosed at a median of 0.8 (0.3–7.7) years following the initial diagnosis of RB. The median age at diagnosis of stage 4a metastatic RB and ASCT was 2 (0.1 – 7.4) and 3.7 (2.3 – 9.8) years, respectively. The median time between diagnosis of stage 4a metastatic RB and ASCT was 0.4 (range 0.3–0.7) years. The median follow-up time from ASCT (n=24) was 6.3 (0.4–27.7) years. Sites of metastatic disease included bone (n=20), bone marrow (n=20) and liver (n=4). Table 1 and Supplementary table 1 summarize the clinical characteristics, treatment modalities and patient outcomes.

Recurrences

Four patients suffered recurrences at 1.1, 7.2, 9.3 and 12.6 months post ASCT and eventually died of disease. Sites of recurrence included the central nervous system (n=2), mandible (n=1) and multifocal relapse (n=1).

Retinoblastoma-Free Survival

Sixteen patients remain retinoblastoma-free in remission with a median follow-up of 7.1 years (range 1.2 – 27.1 years) from ASCT. Figure 1A demonstrates that the Kaplan–Meier estimate of 5-year RBFS is 76.4%. Figure 1B demonstrates that the Kaplan–Meier estimate of 5-year OS was 81% ± 8.6 and the 10-year OS 59.3% ± 12.4 with early deaths due to recurrent RB (n=4; 2 CR, 2 VGPR) and late deaths due to subsequent malignant neoplasms (SMNs) (three osteosarcoma, one glioblastoma).

Figure 1.

(A) Kaplan Meier estimate of retinoblastoma free survival (n=20). This curve does not include 4 patients that died from alternative causes (including osteosarcoma and glioblastoma). (B) Kaplan–Meier estimate of 5-year OS was 81% ± 8.6 and the 10-year OS 59.3% ± 12.4 with early deaths due to recurrent RB (n=4; 2 CR, 2 VGPR) and late deaths due to subsequent malignant neoplasms (SMNs) (3 osteosarcoma, 1 glioblastoma). (C) Kaplan Meier estimate of 5-year OS was 68.6% ± 18.6 and 86.5%± 8.9 in patients who achieved CR (n=8) and VGPR/PR (n=16) prior to ASCT, respectively (p = 0.45).

Impact of pre-ASCT disease status on survival

Table 2 summarizes the demographic/clinical characteristics including RB germline status and patient outcomes among the two patient cohorts, i.e., those patients achieving CR (n=8) or VGPR/PR (n=16), respectively. Figure 1C demonstrates that the 5-year OS was 68.6% ± 18.6 and 86.5%± 8.9 in patients who achieved CR and VGPR/P prior to ASCT, respectively (p = 0.45).

Table 2.

Demographic and clinical characteristics of patients by subgroup achieving CR or VGPR/PR prior to ASCT.

CR vs PR/VGPR pts	CR	PR/VGPR
Subtotal	n=8	n=16
Gender (F:M)	3:5	8:8
Age (in years)
at diagnosis	1.4 (0.1–2.6)	2.1 (0.3–7.3)
at ASCT	3.2 (2.3–9.5)	3.8 (2.8–9.7)
RB germline status Positive Negative Unknown	2 2 4	3 11 2
Time btw dx and relapse (in months)	7.4 (4.0–92.4)	11.8 (3.1–83.9)
RT post ASCT (outcome) Yes No	3 (all RBFS) 5 (3 RBFS)	9 (8 RBFS) 7 (6 RBFS)
Overall survival at 5 years (P = 0.4531)	68.5%	86.5%

Impact of Radiation Therapy on Survival

Twenty patients had bony metastases detected at the time of diagnosis of Stage 4a RB. Imaging responses pre-ASCT were as follows: 6 CR, 11 VGPR and 3 PR; post ASCT: 14 CR, 4 VGPR, 2 PR. All patients classified as VGPR or PR had bone scan and/or MRI evidence of residual abnormalities (persistent imaging abnormalities of unknown significance). Twelve patients received IFRT; none of these patients experienced a bony relapse at site of IFRT (median follow up of 8.2 and range of 1.4 – 26.1 years post ASCT). However, four of twelve patients developed osteosarcoma at 3.9, 7.2, 8.8, and 14 years post IFRT within the radiation field (germline RB1 positive in one patient, negative in one patient and unknown in two patients). Eight patients did not receive IFRT; none of these patients developed bony relapse or osteosarcoma (median follow up of 2.0 and range of 0.5–14.4 years post ASCT).

Discussion

To our knowledge, the current single institution series represents the largest cohort of patients with stage 4a RB published to date and includes some patients who enrolled on COG ARET0321. Our findings suggest that intensive multi-modality therapy including HDC-ASCT is curative in the majority of stage 4a metastatic RB patients whose treatment includes HDC-ASCT, in concordance with the preliminary results reported from the prospective COG clinical trial, ARET0321¹⁰. However, it remains unclear if patient outcomes are affected by the presence or absence of residual disease pre-ASCT, which is critical in determining which patients should proceed to ASCT. Our data suggest that the degree of post-induction response prior to ASCT (as determined with standard bone marrow, MRI and bone scan for disease evaluations) did not affect outcome. The role of IFRT to sites of bony disease is not clear and we initially chose to administer IFRT to all patients regardless of response, although our practice evolved to be more restrictive in recommending RT only to the sites with residual disease pre-ASCT in accordance with treatment guidelines on COG ARET0321. However, assessment of residual bony disease with standard imaging (MRIs and nuclear medicine bone scans) represents a challenge and it is likely that the residual abnormalities in our patients may not have represented active disease. Survivors of RB have a high risk of developing SMNs with the cumulative incidence of SMNs reported to be 26% in nonirradiated patients and 58% in irradiated patients by 50 years after diagnosis of RB¹³. More contemporary cohorts of patients treated with advanced radiation planning and delivery technology¹⁴, including proton therapy¹⁵ may carry lower risk, but longer follow-up is required. Certainly, avoiding IFRT is preferable, particularly in patients harboring germline RB1 mutations¹⁶, if that would not compromise survival from the metastatic disease. All four patients who have developed secondary osteosarcoma had tumors arise in irradiated fields, while none of the non-irradiated patients with residual bony disease suffered bony relapses or developed osteosarcoma, albeit with shorter follow up. None of the 20 patients with bony disease at diagnosis of Stage 4a RB relapsed at the original bony site of involvement, and our non-randomized data suggest that IFRT may not be necessary in patients with bony metastases noted to be in CR or VGPR post ASCT. More sensitive measures for assessment of residual disease are desirable and emerging plasma cell free DNA technologies merit evaluation to risk stratify patients at high risk of relapse^{17, 18}.

Our study has limitations given the retrospective design and the rarity of disease precludes a randomized trial. Further, the small sample size with few events (relapses) precludes multivariate analysis, nevertheless, the two important observations contribute to the management of stage 4a RB patients.

In conclusion, intensive multimodality therapy is curative in most patients with stage 4a metastatic RB and outcome for patients with VGPR/PR prior to ASCT was not significantly different from patients with CR. IFRT does not seem to be required for bony disease control and increases the likelihood of developing osteosarcoma.

Abbreviations:

ASCT

autologous stem cell transplant

CR

complete response

HDC

high-dose chemotherapy

IFRT

involved field radiation therapy

OS

overall survival

PR

partial response

RB

retinoblastoma

SMN

subsequent malignant neoplasms

VGPR

very good partial response

Data Availability Statement:

Data can be made available upon reasonable request to the corresponding author.