Survival of pediatric Hodgkin lymphoma patients treated with doxorubicin, bleomycin, vincristine, etoposide, prednisone, and cyclophosphamide (ABVE-PC) versus Adriamycin, bleomycin, vinblastine, and dacarbazine (ABVD) at a single institution

Abstract

Background:

Adriamycin, bleomycin, vinblastine, dacarbazine (ABVD), the de-facto standard-of-care in adult-onset Hodgkin lymphoma (HL), has not been directly compared to doxorubicin, bleomycin, vincristine, etoposide, prednisone and cyclophosphamide (ABVE-PC), a pediatric-aimed regimen designed to reduce late-effects. We aimed to describe the single institution experience of using both regimens in patients with pediatric HL.

Methods:

This retrospective cohort study evaluated a total of 224 patients diagnosed with HL between 1999 and 2018 at Children’s Hospital Los Angeles (CHLA), of which 93 patients were eligible having received ABVD (n=46) or ABVE-PC (n=47) chemotherapy as their initial treatment. Descriptive analyses were performed using the Student’s t-test or Fisher’s exact test. Survival analysis used the Kaplan-Meier method. Events included death, relapse and secondary malignancy. We also describe the use of radiation therapy, pulmonary toxicity and cardiomyopathy determined by shortening fraction <29%. Analyses followed an intention-to-treat principle.

Results:

There was no difference in baseline characteristics between the patients receiving ABVE-PC or ABVD in regard for stage, risk group or prognostic variables, such as the presence or absence of “B” symptoms, bulky disease, and extra-nodal involvement. A greater proportion of patients treated with ABVE-PC received consolidating external beam radiation treatment (XRT) either by randomization or by response compared to ABVD (59.6% vs 32.6% respectively, p=0.01). While not statistically significant, response to therapy, assessed by PET/CT where available, mirrored the use for radiation (rapid response 58.3% vs 90.0%, n=34, p=0.11). The median dose of anthracycline (doxorubicin) was the same in patients receiving ABVE-PC vs ABVD (200 vs. 200 mg/m², interquartile range 200–250 vs 200–300 mg/m², p=0.002). There was no difference in event-free survival (p=0.63) or overall survival (p=0.37) with a median follow up length of 3.9 years.

Conclusions:

ABVD and ABVE-PC achieved similar survival outcomes in our single-institution cohort

Introduction

Hodgkin Lymphoma (HL) in children is treated via a number of separate protocols in the low, intermediate and high risk strata.^1,2 In adults with HL however, ABVD has been considered the de-facto standard of care for several decades due to its effectiveness and excellent toxicity profile.^3,4 Late effects including cardiotoxicity and reduced fertility remain a concern although overall the late effects profile is favorable.^5–7 Of note, there is variability in the number of cycles utilized depending on the risk group which would impact the incidence of late effects. There is also evidence that radiotherapy can be eliminated in the low risk stratum, which is an attractive approach to reduce risk of organ toxicities and secondary malignancies.⁸

In the pediatric patient population, ABVE-PC, a dose-dense regimen has been more commonly used, especially in trials conducted by the Children’s Oncology Group (COG).^1,2,9 ABVE-PC therapy is risk-adapted and based on interim disease response.^9,10 Given the ability of children to tolerate more short-term toxicities, ABVE-PC is able to reduce exposure to anthracyclines and alkylators.⁹ The COG study AHOD0031 has also shown that external beam radiation treatment (XRT) can be safely eliminated in patients with interim rapid early response and in patients who were in complete remission after chemotherapy, which was per study criteria.⁹

Henderson et al. presented findings comparing AYA patients in two different clinical trials, and found that AYA patients do worse on adult protocols (ABVD vs Stanford V) when compared to pediatric protocol (ABVE-PC), with worse failure-free survival (FFS) and overall survival (OS).¹¹ We conducted a study to describe the efficacy and toxicity of ABVD to ABVE-PC for patients with HL within different disease risk strata at our institution. We hypothesized that ABVE-PC has a similar outcome as measured by EFS and OS when compared to ABVD in pediatric patients with newly diagnosed HL at our institution.

Methods

This retrospective cohort study evaluated pediatrics patients diagnosed with HL (classic or nodular lymphocyte predominant), ages 4–20 years old at diagnosis, between 1999 and 2018 at Children’s Hospital Los Angeles (CHLA). Patients were identified by reviewing an already-existing database of patients maintained for the purposes of providing direct patient care. All patients who received ABVD or ABVE-PC as initial treatment for newly diagnosed HL were included. Patients were excluded due to either receiving a different therapy (e.g. bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine and prednisone {BEACOPP}), incomplete therapy, receiving therapy at an outside institution of CHLA with incomplete data including pathology reports, treatment documentation and disease response evaluations. All decisions for therapeutic regimens, number of courses, radiation therapy and/or to offer clinical trial were made by the treating oncologists in conjunction with patient and family wishes.

Demographic information collected included: age at diagnosis as well as self-reported gender, race, and ethnicity. Prognostic variables abstracted from the medical record included Ann Arbor staging, presence of B-symptoms (unexplained weight loss > 10% in the preceding 6 months, unexplained recurrent fever ≥38°C in the preceding month or recurrent drenching sweats in the preceding month), presence of splenic involvement, presence of extra-nodal disease. Erythrocyte sedimentation rate at diagnosis, peripheral blood leukocyte count at diagnosis, and hemoglobin concentration at diagnosis are historical prognostic factors but reported to be complete. As risk classification has changed multiple times during the time period examined, patients were stratified retrospectively and uniformly into comparable risk groups utilizing the criteria employed by the COG AHOD0031 study.⁹ Intermediate group was defined on the study as: all Stage IB regardless of bulk disease, all Stage IIB regardless of bulk disease, Stage IA only with bulk disease, Stage IIA only with bulk disease, all Stage IA with extra-nodal involvement (E), IIAE regardless of bulk disease, all Stage IIIA, IIIAE, IIIAS, IIIAE + splenic involvement regardless of bulk disease, all Stage IVA, IVAE regardless of bulk disease. Patients with less disease were assigned to the low-risk group. Patient with more disease involvement were assigned to the high-risk group. All chemotherapy regimens, including changes and deviations, were recorded. The total number of courses of any type of chemotherapy employed in first remission were recorded, including non-ABVD / non-ABVE-PC courses. Results of positron emission tomography/computed tomography (PET/CT) were recorded where available. Total anthracycline doses were recorded in doxorubicin-equivalent units and reported as mg/m². Echocardiogram and pulmonary function testing results were recorded where available. OS was defined as the time from diagnosis to death of any cause. EFS was defined as the time from diagnosis to death, relapse/progression of disease, or diagnosis of a secondary malignancy.

Descriptive analyses were performed using the Student’s t-test or Fisher’s exact test for continuous or dichotomous variables, respectively, except for anthracycline dose comparison for which the Wilcoxon rank sum test was employed due to data skew. Survival analysis was performed using the Kaplan-Meier life table method. All analyses followed an intention-to-treat principle, guided by the initial therapy chosen. Late effects were analyzed and including secondary malignancy, pulmonary toxicity and cardiomyopathy determined by shortening fraction <29% measured on the most recent follow up echocardiogram. We assessed treatment response from the available radiology reports, which included CT and PET/CT. PET positivity was defined as greater 18-fluorodeoxygluose uptake than the liver (Deauville score 4–5).¹² All analyses were carried out on the R Project for Statistical Computing version 3.

All study protocols were carried out under supervision of the IRB and in compliance with the Declaration of Helsinki.

Results

224 patients with HL were screened for inclusion. Of the total charts examined 93 patients were found to be eligible; 46 patients had received ABVD and 47 patients had received ABVE-PC chemotherapy as initial chemotherapy. 129 patients were not included due to either receiving either a different therapy or a variation of the two therapies in question, and 2 patients were excluded due to having received initial therapy at a different institution. There were no significant differences in the two groups regarding gender, ethnicity, and race (Table 1). Of note, the ABVE-PC group contained statistically significantly older patients on average compared to the ABVD group (14.5 vs 12.7 years, p=0.03).

Table 1 – Demographics of study subjects.

p-values were determined by Student’s t-test (age) and Fisher’s exact test (all others). ABVD: Adriamycin, Bleomycin, Vinblastine, Dacarbazine. ABVE-PC: Adriamycin, Bleomycin, Vincristine, Etoposide, Prednisone, Cyclophosphamide. SD: Standard deviation.

	ABVD (%), n=46	ABVE-PC (%), n=47	p-value
Average age, years (SD)	12.66 (4.13)	14.45 (3.46)	0.03
Gender			1
Female	19 (41.3%)	19 (40.43%)
Male	27 (58.7%)	28 (59.57%)
Race			0.54
White	17 (36.96%)	14 (29.79%)
Black	4 (8.7%)	3 (6.38%)
Asian	0 (0%)	2 (4.26%)
Other	25 (54.35%)	28 (59.57%)
Ethnicity			0.41
Hispanic	23 (50%)	28 (59.57%)
Not Hispanic	23 (50%)	19 (40.43%)

Known prognostic factors are summarized in Table 2. There was even distribution of each risk group for the 2 regimens and no statistically significant differences found in risk stratification (Table 2, p=0.56). There were also no differences in splenic involvement, extra-nodal extension, erythrocyte sedimentation rate, leukocyte count, or hemoglobin concentration.

Table 2 – Prognostic factors of study subjects.

p-values were determined by Student’s t-test (Erythrocyte sedimentation rate, White blood cell count, Hemoglobin) and Fisher’s exact test (all others). ABVD: Adriamycin, Bleomycin, Vinblastine, Dacarbazine. ABVE-PC: Adriamycin, Bleomycin, Vincristine, Etoposide, Prednisone, Cyclophosphamide. COG: Children’s Oncology Group. SD: Standard deviation.

	ABVD (%), n=46	ABVE-PC (%), n=47	pvalue
Stage			0.3
1A	4 (8.7%)	3 (6.38%)
1B	1 (2.17%)	2 (4.26%)
2A	25 (54.35%)	17 (36.17%)
2B	6 (13.04%)	4 (8.51%)
3A	4 (8.7%)	9 (19.15%)
3B	3 (6.52%)	2 (4.26%)
4A	2 (4.35%)	5 (10.64%)
4B	1 (2.17%)	5 (10.64%)
Risk Group (COG AHOD0031)			0.56
Low Risk	12 (26.67%)	11 (23.4%)
Intermediate Risk	29 (64.44%)	28 (59.57%)
High Risk	4 (8.89%)	8 (17.02%)
B symptoms	11 (23.91%)	13 (27.66%)	0.81
Bulky Disease	27 (58.7%)	23 (48.94%)	0.41
Splenic Involvement	7 (15.22%)	10 (21.28%)	0.59
Extra-nodal Involvement (not including spleen)	8 (17.78%)	13 (27.66%)	0.32
Histology			0.62
Classic Hodgkin Lymphoma	45 (97.83%)	44 (93.62%)
Nodular Lymphocyte Predominant Hodgkin Lymphoma	1 (2.17%)	3 (6.38%)
Erythrocyte sedimentation rate (mm/hr) at diagnosis	44.48 (35.43)	50.49 (35.99)	0.43
White blood cell count (x109/L) at diagnosis	10 (4.14)	10.68 (4.38)	0.44
Hemoglobin (g/dL) at diagnosis	11.5 (1.67)	11.54 (1.95)	0.91

Patients were followed for a median of 3.9 years from diagnosis (range 0.3–18.1 years, interquartile range 1.5–5.8 years). The 4 year EFS was 88.0% (95% CI 78.7–98.5%) for patients receiving ABVD vs 83.4% (95% CI 72.0–96.7%) for patients receiving ABVE-PC. The 4 year OS was 100% (95% CI 100.0–100.0%) for patients receiving ABVD vs 100% (95% CI 100.0–100.0%) for patients receiving ABVE-PC. There was no difference found in EFS and OS between patients who received ABVE-PC and those who received ABVD (Figures 1A and 1B, p=0.46 and p=0.37, respectively). We saw no difference in the proportion of patients requiring more than four cycles of chemotherapy or requiring a chemotherapy regimen change, although any differences or similarities may be explained by provider decision variation (Table 3). Notably, more patients in the ABVE-PC group received external beam radiation when compared to the ABVD group in intention-to-treat analysis (59.6% vs 32.6%, Table 3, p=0.01), although 6 patients achieved complete response but were randomized to receive radiation on the AHOD0031 study. When the 6 patients were analyzed as not having radiation, the difference was not statistically significant (46.8% vs 32.6%, Table 3, p=0.21). Response by PET-CT was only available for a subset of the cohort (n=34). 90% of patients receiving ABVD (n=9 of 10 total) and 58.3% of patients receiving ABVE-PC (n=14 of 24 total) had rapidly-responsive lesions on PET-CT following 2 cycles of therapy (i.e., Deauville score of 1, 2, or 3) (p=0.11).

Figure 1 – Event-free survival and overall survival.

A: Event-free survival as defined by time to first relapse or last-known follow-up if still in first remission. B: Overall survival, as defined by time to death or to last-known follow-up if still alive. ABVD: Adriamycin, Bleomycin, Vinblastine, Dacarbazine. ABVE-PC: Adriamycin, Bleomycin, Vincristine, Etoposide, Prednisone, Cyclophosphamide

Table 3 – Treatment details of study subjects.

p-values were determined by Student’s t-test(number of cycles received) and Fisher’s exact test (all others). ABVD: Adriamycin, Bleomycin, Vinblastine, Dacarbazine. ABVE-PC: Adriamycin, Bleomycin, Vincristine, Etoposide, Prednisone, Cyclophosphamide. PET: Positron Emission Tomography. SD: Standard deviation.

	ABVD (%), n=46	ABVE-PC (%), n=47	pvalue
Number of chemotherapy courses received			0.28
4 or less	28 (60.87%)	34 (72.34%)
5 or more	18 (39.13%)	13 (27.66%)
Switched regimen (escalated or crossed over)*	5 (10.87%)	4 (8.51%)	0.74
External Beam Radiation	15 (32.61%)	28 (59.57%)	0.01
–received radiation for non-randomization reason	15 (32.61%)	22 (46.81%)	0.21
Response by PET (n=34)			0.11
Rapidly Responding Lesion (Deauville 1,2,3 at PET-2)	9 (90%)	14 (58.33%)
Slowly Responding Lesion (Deauville 4,5 at PET-2)	1 (10%)	10 (41.67%)

^*ABVD: also received BEACOPP (n=5). ABVE-PC: also received DECA (n=2), BEACOPP (n=1), ABVD (n=1)

The median dose of anthracycline (doxorubicin) was lower in patients receiving ABVE-PC vs ABVD (200 vs. 200 mg/m², interquartile range 200–250 vs 200–300 mg/m², p=0.002, Table 4). Late effects are also summarized in Table 4; statistics are not available due to extremely low event rates. There were no apparent differences seen in reduced cardiac shortening fractions (SF, reduced SF defined as <29%), reduced diffusing capacity of the lung for carbon monoxide (DLCO), second malignancies or evidence of early gonadal dysfunction (as measured by follicle stimulating hormone above gender-defined norms).

Table 4 – Late effects in study subjects.

Statistical analysis is not reported as our study is not sufficiently powered to detect differences in rare events. FSH elevation was defined as >10 units/mL in men and >21.5 units/mL in women. ABVD: Adriamycin, Bleomycin, Vinblastine, Dacarbazine. ABVE-PC: Adriamycin, Bleomycin, Vincristine, Etoposide, Prednisone, Cyclophosphamide. DLCO: Diffusing capacity for carbon monoxide reported as % of reference. Follicle-stimulating hormone. SD: Standard deviation.

	ABVD (%), n=46	ABVE-PC (%), n=47
Median total anthracycline exposure in doxorubicinequiv/m2 (inter-quartile range)	200 (200−300)	200 (200−250)
Last known shortening fraction <29% (n=88)	0 (0%)	0 (0%)
Last known DLCO (SD) (n=51)	85.79 (13.85)	98.33 (29.1)
Second Malignancy*	1 (2.17%)	1 (2.13%)
FSH elevation (n=48)
Yes	1 (5%)	2 (7.14%)
No	19 (95%)	26 (92.86%)

Discussion

In this study, we aimed to describe the use of two chemotherapy regimens commonly utilized at our institution for pediatric HL regarding efficacy and toxicity. We found that there was no significant difference in EFS (p=0.46) or OS (p=0.37), with a median follow up length of 3.9 years. Our four year survival numbers are similar to other published outcomes in pediatrics groups such as COG AHOD0031 (EFS of 85.0% (95% CI, 83.2% to 86.7%). OS of 97.8% (95% CI, 97.0% to 98.5%).⁹ Our results corroborate prior published literature concerning the high overall success rate of both medication regimens and allow for direct comparison between the two regimens.^8,9

Since most of the treatment for HL is comparable in efficacy, the main challenge in developing pediatric clinical trials is to develop strategies to maintain OS while avoiding or reducing long-term morbidity. ABVE-PC, a pediatric regimen specifically designed to reduce late-effects in comparison to regimens developed in adult populations, was examined in the Children’s Oncology Group AHOD0031 trial in an effort to eliminate radiation for a subset of intermediate-risk patients with rapid early response. The 10-year follow up for the AHOD0031 trial found excellent outcomes for rapid early responders, and radiation therapy did not provide significant benefit in EFS over 10 years (RT 83.8% vs no RT 82.5%, p=0.26) in patients with interim rapid early response and in patients who were in complete remission after chemotherapy.¹³ Since completion of the AHOD0031 trial however, several case series have shown that ABVD without radiation is feasible and associated with excellent outcomes in children with complete metabolic response on PET/CT imaging after chemotherapy.^14,15 When looking at the two regimens in similar populations, we found in our cohort that a greater proportion of patients treated with ABVE-PC received consolidating external beam radiation treatment (XRT) compared to ABVD in intention-to-treat analysis, but the difference was no longer statistically significant when patients randomized to radiation were instead analyzed as not having received radiation. Response by PET, where available, mirrored the need for radiation although the numbers available were limited.

Ultimately the goal of risk-adapted therapy, selective elimination of radiotherapy and overall reduction in therapeutic intensity is to reduce the risk of late effects and subsequent malignant neoplasms.¹⁶ Numerous late effects such as hyper/hypothyroidism, thyroid nodules, heart disease, pulmonary fibrosis/pneumonitis, skin cancer in radiation field, chronic fatigue, perceived cognitive change, peripheral neuropathy, reproductive late adverse effects such as early menopause and infertility, and osteoporosis have been described following therapy for pediatric HL.¹⁷ Subsequent malignant neoplasms are also seen, as noted in for example in the long-term results of AHOD0031 where study authors found an excess absolute risk of 1.2 malignant neoplasms / 1000 person-years.¹³ Reductions in therapy intensity have been shown to reduce the rate of late complications as well.¹⁸ Although the follow up time was limited, we did not find significant differences commonly screened late effects or subsequent malignant neoplasm.

One of the stated goals in designing ABVE-PC was to reduce the exposure of children with HL to high cumulative doses of anthracyclines.¹⁰ We did find that overall, the exposure to anthracyclines was successfully reduced, although this did not result in any notable change in the number of patients with detectable abnormalities on follow-up echocardiogram; this may have been due to small sample size and/or short follow-up, however (late-onset cardiomyopathy may evolve over more than a decade).¹⁹ Of note, other regimens such as OEPA have successfully reduced anthracycline exposure even further in exchange for more exposure to glucocorticoids and alkylators, with overall good outcomes in pediatric patients.²⁰

Of note, patients with nodular lymphocyte predominant Hodgkin lymphoma are often treated differently.²¹ These patients are often younger males with localized disease, without “B” symptoms, non-bulky. If the disease if localized, it can be treated with surgical resection alone. Chemotherapy, which include regimens such as ABVD per the National Cancer Center Network guidelines,²² and radiotherapy are reserved for more advanced disease, which was applicable to the patients that were included in this study.

There are several limitations to our study. Primarily, as this was not a clinical trial and chemotherapy regimens were chosen individually, subtle selection biases may have been introduced along with variations in number of chemotherapy courses and decisions regarding the use of radiotherapy. Second, despite collecting all the HL patients at one of the largest children’s hospitals in the nation over nearly two decades, the numbers remain small due to the rarity of the disease; such numbers continue to illustrate the need for multicenter cooperative group trials. Also, following the intention-to-treat principle meant all patients randomized to receive radiation were counted as having received radiation, although the published findings of AHOD0031 likely will similar patients in the future (RER) will not receive radiation. Additionally, the small number of late effects precluded meaningful comparisons, although overall the rate was low with both regimens. We were limited to length of follow-up and surveillance study results available given the nature of this retrospective chart review, as we do not have strict protocols for late effect monitoring in our institution. Finally, due to the nature of patient heath documentation maintenance, and the lack of electronic medical records for the patients who were treated in the earlier years, the data collection, such as radiology and outside records, was limited for some.

In conclusion, our single institution experience further supports other published data that the ABVD and ABVE-PC regimens has similar survival outcomes without excess late effects. In our particular cohort, more patients treated with the ABVE-PC regimen received radiation, although several received radiotherapy due to randomization on clinical trial. Nonetheless, the rate of second malignant neoplasms and other late effects appeared to be low.

Abbreviations:

ABVE-PC

Doxorubicin, bleomycin, vincristine, etoposide, prednisone, cyclophosphamide

ABVD

Adriamycin, bleomycin, vinblastine, dacarbazine

AYA

Adolescent and young adult

BEACOPP

bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, prednisone

CHLA

Children’s Hospital Los Angeles

COG

Children’s Oncology Group

DLCO

Diffusing capacity of lung for carbon monoxide

EFS

Event free survival

FFS

Failure free survival

FSH

Follicle-stimulating hormone

HL

Hodgkin Lymphoma

IRB

Institutional Review Board

OS

Overall survival

PET/CT

Positron Emission Tomography/Computerized Tomography

RER

Rapid early responders

SF

Shortening fraction

XRT

External beam radiation treatment