Efficacy and safety of front-line treatments for advanced Hodgkin lymphoma: a systematic literature review

ABSTRACT

Objective

To assess evidence on the safety and efficacy of ABVD (doxorubicin [Adriamycin®], bleomycin, vinblastine, and dacarbazine), BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone), and A+AVD (brentuximab vedotin, with doxorubicin, vinblastine, and dacarbazine) for advanced-stage Hodgkin lymphoma (HL).

Methods

A systematic literature review (SLR) was conducted on 29 July 2016 (updated 26 July 2018) to identify randomized controlled trials (RCTs) and non-RCTs assessing the treatment of newly-diagnosed advanced-stage HL with ABVD and BEACOPP (and their variants), and A+AVD.

Results

The SLR identified 62 RCTs and 42 non-RCTs. Five-year overall survival rates for ABVD and BEACOPP were 60–97% and 84–99%, and 5-year progression-free survival rates were 58–81% and 83–96%, respectively. Both regimens were associated with tolerability issues and side effects. Discontinuation or dose reduction of bleomycin resulted in fewer adverse events, without significantly affecting efficacy. A head-to-head trial demonstrated improved efficacy for A+AVD vs ABVD, with an acceptable tolerability profile. No data from head-to-head trials comparing A+AVD with BEACOPP were available, and an indirect treatment comparison was not feasible.

Conclusion

New therapies, such as A+AVD, maintain the efficacy observed with current treatments, and may provide a more tolerable treatment option for patients with advanced-stage HL.

KEYWORDS:

• Advanced stage Hodgkin lymphoma
• ABVD
• BEACOPP
• brentuximab vedotin
• bleomycin
• systematic review

1. Introduction

Hodgkin lymphoma (HL) is categorized as either classical HL or nodular lymphocyte-predominant HL (NLPHL) (based on lymphoma cell morphology and phenotype, and the composition of the cellular infiltrate), which account for approximately 95% and 5% of cases, respectively [1]. The Ann Arbor system categorizes HL into Stages I–IV; although not uniformly used, Stages I/II with no risk factors are considered limited stages, Stages I/II with at least one risk factor (such as large mediastinal mass, elevated erythrocyte sedimentation rate [ESR] or ≥4 nodal areas) are considered intermediate stages, and Stages III and IV are advanced [2]. Stage IIB is considered advanced if associated with the risk factors of large mediastinal mass and/or extra-nodal disease [2]. Extra-nodal involvement is less common in HL compared with non-Hodgkin lymphoma (NHL), with approximately 5% of HL occurring at extra-nodal sites [3].

Current treatment of HL is considered a clinical success, with cure rates greater than 90% and 70% for early and advanced HL, respectively [4]. The first successful combination therapy for the treatment of advanced HL was introduced in the 1960s and contained mustard, vincristine (Oncovin®), procarbazine, and prednisone (MOPP) [5]. Today, the current standard of care, front-line treatments for advanced classical HL are the chemotherapy (CT) regimens ABVD (doxorubicin [Adriamycin®], bleomycin, vinblastine, and dacarbazine) and BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone), often in combination with radiotherapy (RT) [4]. The most commonly prescribed treatment in Asia, Latin America, the US and UK is ABVD, with BEACOPP more common in Germany and other parts of Europe [6]. Use of treatment regimens may also depend on disease stage; ABVD is often regarded as standard of care for early-stage, while the BEACOPP escalated dose regimen (BEACOPP_ESC), has greater efficacy in advanced-staged HL, compared with other regimens, although this comes at the cost of greater treatment-related toxicity [7].

While current front-line treatments have high survival rates, many patients experience short- and long-term health problems due to therapy-related side effects. Myelosuppression and associated cytopenias are the most common acute side effects, resulting in an increased risk of infection [8]. Common serious long-term toxicities associated with HL treatment include secondary malignancies, sterility, RT-associated hypothyroidism, and cardiovascular disease [8], and bleomycin in particular is associated with pulmonary toxicity [9,10].

Novel targeted agents have been developed for HL, including brentuximab vedotin (Adcetris®), an antibody–drug conjugate that targets CD30 [11,12]. In the US, approval for brentuximab vedotin, in combination with doxorubicin, vinblastine, and dacarbazine (A+AVD), has recently been expanded to include first-line treatment of Stage III or IV classical HL [13]. In Europe in 2019, A+AVD was approved as a treatment for patients with previously untreated CD30+ Stage IV HL [14].

The objective of this systematic literature review (SLR) was to determine the efficacy and safety of current treatments for newly-diagnosed advanced-stage HL. Particular focus was given to the front-line treatments ABVD, BEACOPP, and A+AVD. The objective was also to conduct a feasibility assessment for an indirect treatment comparison (ITC) between BEACOPP and any new targeted therapy-chemo combination approved.

2. Methods

The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [15].

2.1. Search methods

The original search was conducted on the 29 July 2016, and was updated on 26 July 2018, both using the following electronic databases via the OVID platform: MEDLINE®, Embase, and the Cochrane Library. The database searches were supplemented with hand-searches of relevant conference proceedings, previous health technology assessment (HTA) submissions, clinical trial registries, and reference lists of included studies. Search strategies and additional details are provided in the supplementary material.

Citations identified through the searches were assessed by a reviewer based on title and abstract using pre-defined eligibility criteria. Any uncertainties were resolved by discussion with a second reviewer. Full publications of potentially relevant citations were obtained and examined in full by two reviewers to identify publications eligible for inclusion in the systematic reviews. Disputes were resolved via discussion with a third party.

2.2. Eligibility criteria

Key eligibility criteria were recorded a priori in a review protocol (not registered) and are provided in the supplementary material (Supplementary Table 1). Eligibility criteria included newly diagnosed patients with advanced-stage HL (defined as Stage IIB, III, and IV), treated with CT regimens with/without RT. This manuscript focuses on data for ABVD and BEACOPP (and their variants), and A+AVD.

2.3. Quality assessment

Quality (risk of bias) assessment for the publications identified in the review was performed using the seven criteria checklist provided in Section 4.6 of the National Institute for Health and Care Excellence (NICE) single technology appraisal (STA) user guide [16]. This approach is based on guidance provided by the Centre for Reviews and Disseminations for assessing the quality of studies included in systematic reviews [17].

2.4. Feasibility assessment for treatment comparisons

The Phase III clinical trial of A+AVD (ECHELON-1) included an ABVD treatment arm, and therefore a published direct comparison of A+AVD and ABVD is available. BEACOPP_ESC is the standard of care BEACOPP regimen for patients <60 years of age with advanced stage HL in some European countries (BEACOPP_ESC has greater efficacy in advanced-staged HL compared with other therapies) [2,7]. In the absence of a head-to-head trial, an ITC is necessary in order to estimate the comparative efficacy and safety of A+AVD with BEACOPP_ESC. An assessment was conducted to determine the feasibility of a conventional ITC, which preserves randomization, using studies identified in the SLR. The population of interest included newly-diagnosed adult patients with advanced stage HL (Stage III or IV). All randomized controlled trials (RCTs) and non-randomized studies assessing BEACOPP_ESC (non-positron emission tomography [PET] guided therapies or PET-guided therapies) were considered for inclusion in the ITC. Study heterogeneity was assessed to confirm suitability of studies for pooling in an ITC.

3. Results

3.1. Systematic review – RCTs

Across the original and update searches, 62 publications reporting on 43 RCTs were included [18–79]. A PRISMA diagram for the RCT search is presented in Figure 1, and identified treatment regimens are presented in Table 1. Study summaries are presented in the supplementary data (Supplementary Table 14).

Figure 1. A PRISMA diagram for the RCT search.

Table 1. Treatment regimens in identified RCT and non-RCT/observational studies.

Treatment regimen	Number of studies	References
RCT
ABVD	40	[18–22,24–31,33,35–43,45–48,59,64–73,78,121]
BEACOPP	21	[21,23,26,27,31,32,34–36,44,46,59–63,67,68,72,73,78]
BEACOPPBASE	3	[21,31,32]
BEACOPPESC	5	[23,26,31,32,34]
BEACOPP4+4	4	[23,26,46,59]
BEACOPP4+2	1	[35]
BEACOPP14	1	[34]
Non-RCT/observational
ABVD	23	[81–83,85,86,88–90,92–95,97–99,101,110,112,114,115,118–120]
BEACOPP	12	[80,84,87–89,91,94,96,100,114,115,120]
BEACOPPBASE	1	[88]
BEACOPPESC	5	[88,89,91,94,100]
BEACOPP4+4	3	[80,84,91]
BEACOPP2+6	3	[80,84,87]
BEACOPP14	1	[91]

Abbreviations: ABVD, doxorubicin (Adriamycin), bleomycin, vinblastine, dacarbazine and methylprednisolone; BEACOPP, bleomycin, etoposide, doxorubicin (Adriamycin), cyclophosphamide, vincristine, procarbazine and prednisone; RCT, randomized controlled trial. Note: BEACOPP_BASE, standard baseline treatment; BEACOPP_ESC, escalated dose; BEACOPP₄₊₄, four cycles of BEACOPP_ESC followed by four cycles of BEACOPP_BASE; BEACOPP₄₊₂, four cycles of BEACOPP_ESC followed by four cycles of BEACOPP_BASE; BEACOPP₁₄, a time-intensified regimen of 14 intervals; BEACOPP₂₊₆, two cycles of BEACOPP_ESC followed by six cycles of BEACOPP_BASE.

3.2. Systematic review – non-RCTs/observational studies

Across the original and update searches, 42 non-RCT/observational studies were identified [80–121]. A PRISMA diagram for the non-RCT search is presented in Figure 2, and identified treatment regimens are presented in Table 1. Study summaries are presented in the supplementary data (Supplementary Table 15).

Figure 2. A PRISMA diagram for the non-RCT search.

3.3. Efficacy outcomes

The definition and results of efficacy outcomes (overall survival [OS], progression-free survival [PFS], relapse-free survival [RFS], and failure-free survival [FFS]) for the RCTs and non-RCTs identified in the SLR that included ABVD and/or BEACOPP treatment regimens are listed in the supplementary data (Supplementary Table 16 and Supplementary Table 17, respectively). Of the eight RCTs that compared ABVD with BEACOPP [21,26,32,35,46,59,68,78], none found significant differences between ABVD and BEACOPP with regard to OS rates (4-year, 5-year, and 7-year rates) [21,26,46,59,68,78,122], although there was a significant difference in 10-year OS across three treatment arms (ABVD, BEACOPP_BASE, and BEACOPP_ESC) in one study (p < 0.001) [32]. Of the three studies comparing ABVD and BEACOPP that reported PFS [26,46,78,122], all had significantly better 5-year PFS rates after treatment with the BEACOPP regimens (BEACOPP₄₊₄ and BEACOPP₄₊₂), vs ABVD [26,46,122] (p < 0.05). One non-RCT compared ABVD and BEACOPP, and reported that 2-year PFS was significantly higher for ABVD with BEACOPP than ABVD alone (p = 0.00015) [114]. Among the included non-RCTs, patients treated with ABVD alone, with or without RT, had a 5-year OS rate ranging from 60% to 97% [92,119]. This was comparable with the RCT data, which had 5-year OS rates for ABVD ranging from 64% to 92% [41,46]. A follow-up study reported the 10-year OS for ABVD as 87% [29].

Reported 5-year PFS in non-RCTs for standard ABVD ranged from 58% to 81% [92,119]. Among the non-RCTs evaluating different combinations of BEACOPP_BASE/BEACOPP_ESC, the 5-year OS ranged from 84% to 97% [80,120] and the 10-year OS was 90% [84]. This compares with data from RCTs in which 5-year OS ranged from 90% to 99% [23,46]. Follow-up studies found the 10-year OS rates were 80% and 86% with BEACOPP_BASE and BEACOPP_ESC, respectively [32]. Five-year PFS for BEACOPP for non-RCTs was 84% for BEACOPP₂₊₆ [87], and 87–96% with a variety of BEACOPP doses [91,120], compared with the 5-year PFS rates for RCTs, which ranged from 83% to 93% [26,46].

Several studies investigated Stanford V vs ABVD. Two studies found that ABVD was superior to Stanford V in terms of 5-year FFS, and 10-year FFS (78% vs 49%, and 75% vs 49%, respectively) [29,39]. However, one study found that 5-year FFS was similar between treatment groups (74% and 71% for ABVD and Stanford V, respectively) [40]. Five-year PFS was also comparable between treatment groups in another study (76% and 74%, respectively) [42].

3.3.1. Stage IV disease and extra-nodal involvement

To examine whether further advanced-stage HL had an impact on treatment efficacy, the review focused on studies reporting outcomes for the patient groups with Stage IV disease and extra-nodal involvement. One RCT involved patients exclusively with Stage IV disease, in which the complete remission rate with a treatment regimen of MOPP/ABVD was 88.9% [24]. A further 10 RCTs [22,25,33,37,40,42,48,62,64,68] and three non-RCTs reported specific efficacy outcomes for patients with Stage IV and/or extra-nodal involvement [81–83].

Disease stage at the beginning of treatment influenced efficacy outcomes; several studies found lower efficacy in patients with Stage IV HL. In patients treated with either MOPP or an alternating MOPP/ABVD regimen, OS was significantly lower in Stage IV vs Stage III patients; 52% and 77%, respectively (p = 0.05) [37]. With either MOPP, ABVD, or alternating MOPP/ABVD, FFS was significantly worse in patients with Stage IV vs Stage III disease [25], and a significantly lower proportion of patients with Stage IV disease achieved complete response (54.0% vs 85.5%, respectively), although these results were not statistically significant (p = 0.1). Ten-year FFP and survival was lower in Stage IV vs Stage III disease in a treatment regimen involving ABVD and RT [83]. A further study examining ABVD and BEACOPP treatment regimens found 3-year PFS was significantly lower in Stage IV vs Stage II and Stage III (79.6%, 90.0%, and 83.1%, respectively, p < 0.001) [68]. However, in the HD18 trial, 3-year estimates for PFS and OS for patients treated with BEACOPP_ESC or R-BEACOPP_ESC remained largely unchanged when analysis was restricted to patients with Stage IV disease and compared with patients with Stage II–IV disease (OS: BEACOPP_ESC 95.0% vs 96.5%, R-BEACOPP_ESC 89.0% vs 94.4%; PFS BEACOPP_ESC 93.1% vs 91.4%, R-BEACOPP_ESC 89.9% vs 93.0%) [62]. This suggests that intensified polychemotherapy may reduce or eliminate the prognostic impact of staging in HL.

Efficacy and response rates were also lower in patients with extra-nodal involvement. In patients treated with combined ABVD and MOPP, there were lower efficacy rates in patients with extra-nodal disease compared with nodal only for complete response (86% vs 90%, respectively), FFP (62% vs 70%, respectively), and RFS rates (73% vs 76%, respectively) [48]. Similarly, in a study of patients receiving ABVD, MOPP, or MOPP alternating with ABVD, there were lower rates of FFS and OS with the presence of ≥2 extra-nodal sites, vs <2 (p < 0.001 and p = 0.03, respectively) with a treatment regimen involving MOPP and/or ABVD [25]. The results from non-RCTs showed that, when using a treatment regimen involving ABVD, complete response was significantly lower in the patient group that had >1 extra-nodal site compared with ≤1 (60% vs 85.4%, respectively, p < 0.01) [82].

Novel targeted therapies may afford benefits in patients with more advanced disease. The ECHELON-1 study suggests that patients with Stage IV disease and patients with involvement of >1 extra-nodal sites have improved outcomes with A+AVD vs a standard ABVD regimen [64]. A subgroup analysis of this study demonstrated a greater treatment benefit of A+AVD vs ABVD in patients with Stage IV compared with Stage III HL; the hazard ratio for PFS was 0.92 (95% CI: 0.60, 1.42) and 0.71 (95% CI: 0.53, 0.96) for patients with Stage III and Stage IV HL, respectively [64,69,71].

3.3.2. PET-guided treatment

Recently, PET response-adapted therapies have been assessed to improve efficacy and avoid toxicity. Four studies reported patients’ response to PET response-adapted therapy [60,61,68,73,78].

Patients in the Phase III HD18 trial with a negative PET-2 scan received a reduction in BEACOPP_ESC from eight to four cycles without a loss of efficacy (as determined by PFS) [60]. For patients with postive PET-2 findings following two cycles of BEACOPP_ESC, the addition of rituximab to standard BEACOPP_ESC treatment did not improve PFS [61].

Omission of bleomycin from the ABVD regimen after negative PET-2 findings following two cycles of ABVD resulted in a lower incidence of pulmonary toxicity than with continued ABVD in Phase III RATHL trial, and did not significantly lower efficacy (PFS) [68].

In the Phase III AHL 2011 LYSA trial, de-escalation of treatment from BEACOPP_ESC to ABVD in patients with a negative PET-2 scan following two cycles of BEACOPP_ESC did not impair efficacy (PFS). Treatment toxicity was significantly higher in the patients who received four additional cycles of BEACOPP_ESC following PET-2 scan compared with those de-escalated to ABVD (p < 0.002) [73]. For patients with a positive PET-2 scan following two cycles of ABVD, the GITIL/FIL HD 0607 Phase II study demonstrated that early treatment intensification with BEACOPP_ESC is a feasible, safe, and effective therapeutic strategy in advanced HL [78].

3.3.3. Patients aged ≥60 years

For patients with HL aged ≥60 years (median age 69 [range 62–88]) who were ineligible for frontline chemotherapy, a Phase II open-label study investigated brentuximab vedotin in combination with dacarbazine or bendamustine [117]. The objective response rate (ORR) for brentuximab vedotin plus dacarbazine was 100% and the complete remission (CR) rate was 62%. Median PFS was 17.9 months. For brentuximab vedotin plus bendamustine, the ORR was 100.0% and the CR rate was 88%. Neither the median PFS or OS was reached. Despite disease activity, brentuximab vedotin plus bendamustine treatment arm was stopped prematurely due to an unacceptably high rate of serious adverse events (SAEs) (65%) and two deaths (10%). A separate study investigating treatment with brentuximab vedotin (before and after AVD) in patients aged ≥60 years of age. The 2-year survival rates were 80%, 84%, and 93%, for EFS, PFS, and OS, respectively [123].

The HD9_elderly RCT compared BEACOPP_BASE and COPP-ABVD in elderly patients with HL (75 patients aged 66–75 years) and found that at 5 years there were no significant difference between treatment groups for CR, OS, and freedom from treatment failure [21].

The E2496 RCT compared ABVD and Stanford V treatment regimens. In the elderly subgroup population (aged ≥60 years, n = 44), 5-year failure-free survival and 5-year OS, were lower in the older group compared with the younger patients (48% vs 74% [p = 0.002], and 58% vs 90% [p < 0.0001], respectively). In contrast, time to progression did not differ between age groups [41].

3.4. Safety outcomes

3.4.1. Mortality

The mortality outcomes for ABVD and/or BEACOPP regimens are displayed in the supplementary data (Supplementary Table 18). Reported all-cause mortality ranged from 1–54% and 1.5–55% for ABVD- and BEACOPP-containing regimens, respectively [21,39,46,64,68].

3.4.2. Pulmonary toxicity

In total, 19 RCTs and nine non-RCTs reported data associated with pulmonary toxicity (Table 2). Any-grade pulmonary toxicity ranged from 0% to 36% across both the RCT and non-RCT studies [39,76,101]. Seven RCTs and four non-RCTs reported a causal relationship between bleomycin and pulmonary toxicities [26,30,33,38,41,42,48,68,91,93,101]. Few studies reported the incidence of specific pulmonary toxicities, including interstitial lung disease, lung fibrosis, dyspnea, and pneumonitis. One RCT and one non-RCT reported data pertaining to interstitial lung disease [64,101], with any-grade interstitial lung disease reported by 4% of patients who received treatment with brentuximab vedotin + ABVD (A+ABVD) compared with 0% of patients treated with A+AVD [101]. The incidence of lung fibrosis was reported in one RCT [18]. In the RCT, a similar proportion of patients who received ABVD/MOPP and ABVD/OPP reported any-grade lung fibrosis (4% vs 5%, respectively) [18].

Table 2. Proportion of patients with pulmonary toxicity with ABVD- and BEACOPP-containing treatment regimens.

Study	Treatment	Any pulmonary toxicity, %		Specific type (%)
		Any grade	Grade 3/4
RCTs
Anselmo et al, 1998 [18]	ABVD/MOPP	NR	NR	Lung fibrosis (4.0)
	ABVD/OPP	NR	NR	Lung fibrosis (5.0)
HD9elderly trial Ballova et al, 2005 [21]	COPP/ABVD	NR	0.0	NR
	8x BEACOPPBASE	NR	8.0	NR
Canellos et al, 1992 [25]	ABVD	NR	7.0	NR
	MOPP/ABVD	NR	4.0	NR
EORTC 20,012 trial Carde et al, 2016 [26]	ABVD	NR	4.0	NR
	BEACOPP4+4	NR	7.1	NR
Connors et al, 1997 [30]	MOPP/ABVD	NR	0.0	NR
Duggan et al, 2003 [33]	ABVD	NR	24.5	NR
HD15 trial Engert et al, 2012 [34]	8x BEACOPPESC	NR	6.4	NR
	6x BEACOPPESC	NR	3.7	NR
	8x BEACOPP14	NR	9.2	NR
E2496 trial Evens et al, 2013 [41]	ABVD	NR	8.0^†	Grade 3/4 dyspnea (17.0), Grade 3/4 pneumonitis (8.0)
Intergroup trial Glick et al, 1998 [38]	MOPP-ABVD sequential	NR	1.0	NR
Gobbi et al, 2005 [39]	ABVD	0.0	NR	NR
E2496 trial Gordon et al, 2013 [40]	ABVD	No difference compared with Stanford V	NR	NR
ISRCTN 64,141,244 trial Hoskin et al, 2009 [42]	ABVD	10.0	NR	NR
LYSA H34 trial Mounier et al, 2014 [46]	ABVD	NR	1.0	NR
	BEACOPP4+4	NR	1.0	NR
HD6 trial Sieber et al, 2004 [47]	COPP/ABVD	NR	0.0	NR
NCT01251107 trial Viviani et al, 2011 [59]	ABVD	NR	<1.0	NR
	BEACOPP4+4	NR	4.0	NR
ECHELON-1 Connors et al, 2018 [64]	ABVD A+AVD	7.0 2.0	3.0 <1.0	NR NR
NCT01569204 Eichenauer, 2017 [76]	BrECAPP BrECADD	0.0 0.0	NR	NR
GITIL/FIL HD 0607 Gallamini et al, 2018, [78]	PET +ve: ABVD/BEACOPPESC + BEACOPPBASE, ± rituximab	7.0	1.0	NR
	PET -ve: ABVD/ABVD	7.0	2.0	NR
HD18 Borchmann et al, 2017 [62]	PET -ve: 2x BEACOPPESC/6x BEACOPPESC 2x BEACOPPESC/4x BEACOPPESC 2x BEACOPPESC/2x BEACOPPESC	NR		NR
			6.0
			2.0
			2.0
RATHL Johnson et al, 2016 [68]	ABVD Cycles 1 and 2	NR	NR	Grade 3/4 dyspnea (<0.5), pneumonitis (0)
	ABVD Cycles 3–6 (PET-ve)	NR	NR	Grade 3/4 dyspnea (2.0), pneumonitis (1.0)
	BEACOPP14 (PET +ve)	NR	NR	Grade 3/4 dyspnea (2.0), pneumonitis (0)
	BEACOPPESC (PET +ve)	NR	NR	Grade 3/4 dyspnea (3.0), pneumonitis (3.0)
Non-RCTs
Belada et al, 2015 [84]	BEACOPP4+4	NR	NR	Any grade pneumonitis (2.5)
Fossa et al, 2012 [87]	BEACOPP2+6	NR	9.0	NR
Haverkamp et al, 2015 [91] Analysis of HD12 and HD15 trials	BEACOPP (varied doses); cycles of Bleomycin ≤4	NR	24.8	NR
	BEACOPP (varied doses); cycles of Bleomycin >4	NR	4.7	NR
	BEACOPP (varied doses); cycles of Vincristine ≤3	NR	6.0	NR
	BEACOPP (varied doses); cycles of Bleomycin >3	NR	5.7	NR
Jalali et al, 2016 [93]	ABVD	15.0	NR	NR
Russo et al, 2014 [98]	ABVD	NR	NR	Any grade pneumonitis (5)
A retrospective analysis of HD9, HD12 and HD15 trials Wongso et al, 2013 [100]	BEACOPPESC	0.1	NR	NR
NCT01060904 Younes et al, 2013 [101]	A+ABVD	36.0	20.0	Any grade interstitial lung disease (4), any grade dyspnea (32), Grade 3/4 dyspnea (12), any grade pneumonitis (4)
NCT00504504 Younes et al, 2012 [110]	Rituximab + ABVD	NR	NR	Any grade dyspnea (21), Grade 3/4 dyspnea (1)

Abbreviations: A+ABVD, brentuximab vedotin + Adriamycin, bleomycin, vinblastine and dacarbazine; ABVD, Adriamycin, bleomycin, vinblastine and dacarbazine; BEACOPP, cyclophosphamide, Adriamycin, etoposide, procarbazine, prednisolone, vincristine and bleomycin; BEACOPP₄₊₄, 4x BEACOPP_ESC + 4x BEACOPP_BASE; BEACOPP_BASE, BEACOPP baseline cycles; BEACOPP_ESC, BEACOPP escalated cycles; BReCADD, etoposide, cyclophosphamide, doxorubicin, dexamethasone, dacarbazine, and brentuximab vedotin; BReCAPP, etoposide, cyclophosphamide, doxorubicin (Adriamycin), prednisone, procarbazine, and brentuximab vedotin; COPP, cyclophosphamide, vincristine, procarbazine and prednisone; MOPP, mechlorethamine, vincristine, procarbazine and prednisone; OPP, vincristine, procarbazine and prednisone; NR, not reported; RT, radiotherapy. †Grade 5

Dyspnea was reported in two RCTs [41,68] and two non-RCTs [101,124]. In one non-RCT, patients receiving treatment with A+ABVD reported a higher incidence of any-grade dyspnea than those receiving A+AVD (33% vs 22%, respectively) [101]. Patients receiving bleomycin also experienced a higher incidence of Grade 3/4 dyspnea in one RCT (ABVD vs Stanford V, 17% vs 0%, respectively) [41] and one non-RCT (A+ABVD vs A+AVD, 12% vs 4%, respectively) [101]. Pneumonitis was reported in two RCTs [41,68] and three non-RCTs [84,98,101]. Any-grade pneumonitis ranged from 0% to 5% [68,84,98,101,102,124]. Grade 3/4 pneumonitis was reported by 8% of patients who received ABVD vs 0% who received Stanford V [41].

3.4.3. Bleomycin discontinuation – impact on efficacy/safety

Four studies reported the impact of bleomycin on efficacy and safety outcomes; a reduction in the dose of bleomycin or discontinuation of bleomycin was associated with a decrease in the incidence of adverse events (AEs) such as pulmonary toxicity [48,64,68,91,93].

Results from the RATHL study demonstrated that excluding bleomycin from the ABVD regimen reduced the incidence of pulmonary toxicity in patients with advanced HL without significant impact on the efficacy outcomes (absolute difference in the 3-year PFS rate [ABVD minus AVD]: 1.6%) [68]. At a median follow-up of 41 months, the 3-year PFS and OS rates in the ABVD group were 85.7% and 97.2%, respectively, and the corresponding rates in the AVD group were 84.4% and 97.6%, respectively [68].

Jalali et al., 2016 reported that in the ABVD regimen, bleomycin was the most commonly reduced or delayed intervention due to pulmonary toxicity. In the univariate analysis, significantly inferior 5-year OS rates were observed in patients with advanced disease with any dose change during treatment (either delay, reduction, or both) compared with those without any dose change during treatment (75% vs 93%, respectively, p < 0.01). However, no significant difference could be detected between the groups (with or without dose change during treatment) in the multivariate analysis [93]. Haverkamp et al., 2015 evaluated the impact of dose reduction or delay due to drug-specific toxicity of bleomycin in patients with advanced HL receiving treatment with BEACOPP in the German Hodgkin Study Group HD12 and HD15 trials. Bleomycin discontinuation did not affect the overall efficacy of treatment. The authors reported that there were no significant differences in PFS or OS in patients receiving ≤4 or >4 cycles of bleomycin (5-year PFS difference: 1.7%; 5-year OS difference: 1.5%) [91].

The ECHELON-1 trial compared A+AVD and ABVD regimens. A+AVD had improved efficacy compared with ABVD in patients with Stage III–IV HL; the 2-year modified PFS rate was significantly higher (82.1% vs 77.2%, respectively, p = 0.04), resulting in a 23% risk reduction in the A+AVD group [64]. Furthermore, patients treated with A+AVD vs ABVD had a lower incidence of pulmonary toxicity; 2% vs 7%, respectively, with less than 1% of patients in the A+AVD group experiencing Grade 3 pulmonary events compared with 3% of patients in the ABVD group [64].

3.4.4. Secondary malignancies

The incidence of secondary malignancies ranged from 0–15% and 0.61–14% for ABVD- and BEACOPP-containing treatments, respectively (Supplementary Table 19) [21,39,59,62]. Commonly reported secondary malignancies in the RCTs were the hematological malignancies NHL, acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS). A study comparing BEACOPP_BASE and COPP-ABVD in patients aged 66–75 years found that at a median follow-up of 80 months, 14% and 12% had secondary malignancies in each treatment group, respectively [21].

Limited evidence on secondary malignancies was available from non-RCTs in advanced HL, with one study reporting three cases of large B-cell lymphoma with a BEACOPP treatment regimen [84]. In the ABVD-containing treatment regimens, secondary malignancies included NHL, melanoma, colorectal cancer, lung cancer, breast cancer, second lymphoma, lymphoblastic leukemia, and carcinoma of the tongue [83,95]. Dann et al., 2017 reported that of 355 patients with early and advanced HL whose therapy was individualized based on initial prognostic factors and PET, 1.9% developed second malignancies during the study period (2006–2013), including secondary leukemia, lymphoblastic lymphoma, and breast cancer [115].

3.4.5. Neutropenia

Table 3 lists the incidence of all neutropenia, and specifically Grade 3/4 events. Incidence of total neutropenia ranged from 21% to 88% for ABVD [25,41,46,64,68] and was reported in one study for BEACOPP (ABVD/BEACOPP regimen; 63.0%) [68]. Grade 3/4 neutropenia incidence ranged from 8% for ABVD [81] to 80% for A+ABVD [101]. The proportion of BEACOPP-treated patients with Grade 3/4 neutropenia was 90%, 63%, and 67% for BEACOPP₄₊₄, BEACOPP₁₄, and BEACOPP_ESC, respectively [68,84]. In the ECHELON-1 trial comparing A+AVD and ABVD, the incidence of neutropenia was 58% and 45%, respectively, and febrile neutropenia was 19% and 8%, respectively [64].

Table 3. Incidence of neutropenia in ABVD and BEACOPP treatment regimens in RCTs and non-RCTs.

Study	Treatment	Neutropenia, all grades, %	Grade 3/4 neutropenia, %
RCTs
Canellos et al, 1992 [25]	ABVD	21.0	NR
E2496 trial Evens et al, 2013 [41]	ABVD	88.0	NR
LYSA H34 trial Mournier et al, 2014 [46]	ABVD	42.0	NR
NCT01251107 trial Viviani et al, 2011 [59]	ABVD	NR	42.0
ECHELON-1 Connors et al, 2018 [64]	ABVD	45.0	39.0
	A+AVD	58.0	54.0
RATHL Johnson et al, 2016 [68]	PET -ve:	PET -ve:
	ABVD	ABVD, Cycles 1 and 2, 58.0	NR
	ABVD/AVD	ABVD/AVD, 59.0	NR
	PET +ve:	PET +ve:
	ABVD/BEACOPPESC/BEACOPP14	ABVD/BEACOPPESC/BEACOPP14, Cycles 1 and 2, 63.0	NR
Non-RCTs
Andjelic et al, 2012 [81]	ABVD	NR	8.0
Belada et al, 2015 [84]	BEACOPP4+4	NR	90.0
Jalali et al, 2016 [93]	ABVD	NR	54.0
Tao et al, 2014 [99]	ABVD	NR	39.0
	Dose dense ABVD	NR	58.0
Russo et al, 2014 [98]	ABVD	NR	17.0
NCT01060904 Younes et al, 2013 [101]	A+ABVD	80.0	80.0
NCT01060904 Younes et al 2012 [110]	Rituximab + ABVD	33.0	23.0
Friedburg et al, 2017 [117]	BV + dacarbazine BV + bendamustine	14.0 25.0	5.0 5.0

Abbreviations: A+ABVD, brentuximab vedotin + doxorubicin (Adriamycin), bleomycin, vinblastine, dacarbazine; A+AVD, brentuximab vedotin + Adriamycin, vinblastine and dacarbazine; ABVD, doxorubicin (Adriamycin), bleomycin, vinblastine, dacarbazine; BEACOPP, bleomycin, etoposide, doxorubicin (Adriamycin), cyclophosphamide, vincristine (Oncovin), procarbazine, prednisolone; BV, brentuximab vedotin; NR, not reported; PET, positron emission tomography; RCT, randomized controlled trial.

Granulocyte-colony stimulating factor (G-CSF) can be administered to manage the risk of neutropenia [125]. G-CSF use in the included studies varied from use as primary prophylaxis from Day 8 of treatment [19], obligatory use for certain treatments such as BEACOPP_ESC [84], and secondary prophylaxis reserved for patients with severe neutropenia [39]. In the ECHELON-1 trial that compared A+AVD with ABVD, a protocol amendment was done after 75% of patients were enrolled and the patients enrolled thereafter in the A+AVD group received prophylactic G-CSF. It was observed that patients in the A+AVD group who received prophylactic G-CSF had a lower incidence of febrile neutropenia compared with those who did not (11% vs 21%, respectively) [64]. Adverse events associated with G-CSF use were not identified in this review.

3.4.6. Peripheral neuropathy

The incidence of reported peripheral neuropathy in ABVD and BEACOPP treatment regimens is reported in Table 4. Rates for all-grade peripheral neuropathy ranged from 1% to 76% after treatment with standard ABVD and CHOPE/ABVD, respectively [25,95]. Two studies reported peripheral neuropathy associated with BEACOPP (6% and 23% for Grades 1–2 and Grade 3, respectively) [84,87]. One study reported all-grade peripheral neuropathy associated with BrECAPP and BrECADD; 32% and 35%, respectively [76]. A trial comparing A+AVD with ABVD found that the incidence of all-grade peripheral neuropathy was 26.0% and 13.0%, respectively.

Table 4. Incidence of peripheral neuropathy in ABVD and BEACOPP treatment regimens.

Study	Treatment	Peripheral neuropathy, all-grade, %	Grade 3/4 peripheral neuropathy, %
RCTs
Canellos et al, 1992 [25]	ABVD	1.0	NR
Intergroup trial Glick et al, 1998 [38]	MOPP-ABVD sequential	8.0	NR
LY09 trial Johnson et al, 2005 [43]	ABVD	NR	3.0
ECHELON-1 Connors et al, 2018 [64]	ABVD	13.0	<1.0
	A+AVD	26.0	4.0
NCT01569204 Eichenauer, 2017 [76]	BrECAPP BrECADD	32.0 35.0	2.0 0.0
Non-RCTs
Belada et al, 2015 [84]	BEACOPP4+4	6.0	NR
Fossa et al, 2012 [87]	BEACOPP2+6	NR	23.0
CALGB 8856 trial Lester et al, 2001 [95]	CHOPE/ABVD	76.0	NR
NCT01060904 Younes et al, 2013 [101]	A+ABVD	72	0.0
NCT00504504 Younes et al, 2012 [110]	Rituximab + ABVD	NR	NR

Abbreviations: A+ABVD, brentuximab vedotin + doxorubicin (Adriamycin), bleomycin, vinblastine, dacarbazine; A+AVD, brentuximab vedotin + Adriamycin, vinblastine and dacarbazine; ABVD, doxorubicin (Adriamycin), bleomycin, vinblastine, dacarbazine; BEACOPP, bleomycin, etoposide, doxorubicin (Adriamycin), cyclophosphamide, vincristine (Oncovin), procarbazine, prednisolone; BReCADD, etoposide, cyclophosphamide, doxorubicin, dexamethasone, dacarbazine, and brentuximab vedotin; BReCAPP, etoposide, cyclophosphamide, doxorubicin (Adriamycin), prednisone, procarbazine, and brentuximab vedotin CHOPE, cyclophosphamide, doxorubicin hydrochloride, vincristine (Oncovin), prednisone, and etoposide; MOPP, mechlorethamine, vincristine, procarbazine and prednisone; RCT, randomized controlled trial.

3.4.7. Adverse events commonly associated with RT

Radiotherapy is a common component of frontline treatment of patients with HL (either as adjuvant or consolidation therapy) and was used in majority of the studies included in the review (details provided in Supplementary Table 14 and 15). While studies have reported commonly observed AEs, no direct association with RT was reported in any study. The AEs from the included studies that have been commonly observed in the literature as being linked to RT are therefore presented in Supplementary Table 20. Long-term AEs such as cardiac and lung toxicity were frequently reported in the included studies. Cardiovascular toxicity was reported in 16 studies and ranged from 0% to 10% with ABVD [18,98] and 0–15% with BEACOPP [21]. Grade 3/4 cardiovascular toxicity ranged from 1% to 15% [21,47]. Similarly, pulmonary toxicities were reported in 21 studies and ranged from 0% to 36% [39,101], with 0–24.5% of patients reporting Grade 3/4 toxicity [21,33,62,64,78]. The most commonly reported short-term adverse effects were nausea/vomiting and infections. In one study, Grade 3 and 4 nausea/vomiting was reported in 18% of patients [21]. Infection was reported in 16 studies in 0–42% of patients [59,68].

3.5. Feasibility assessment for indirect treatment comparisons

The SLR identified 10 RCTs (eight non-PET and two PET-guided studies) and three retrospective studies that assessed a BEACOPP_ESC-containing regimen Table 5. Based on data for A+AVD from ECHELON-1, no BEACOPP_ESC studies were suitable for inclusion in a conventional ITC of BEACOPP_ESC with A+AVD as front-line therapy in patients with Stage III–IV HL. The reasons for unsuitability for inclusion for each study in a conventional ITC are listed in Table 5. Heterogeneity was also observed across trials in terms of the key baseline patient characteristics (e.g. age, gender, Eastern Cooperative Oncology Group [ECOG] Performance Status score), as well as definitions of reported outcomes – any study not reporting Stage III–IV patient outcome data were too dissimilar to the ECHELON-1 trial to ensure appropriate patient outcome data would be compared (similarity assumption). Only LYSA H34 and EORTC 20,012 reported this subgroup data Table 5. ECHELON-1 compared A+AVD with 6xABVD. However, the only ABVD treatment arms in LYSA H34 and EORTC 20,012 were 8xABVD. Even if an assumption was made that these treatment arms were similar enough to be pooled into a single treatment node, the BEACOPP arm in each study was, not the comparator of interest, BEACOPP_ESC. BEACOPP_ESC has been shown to have greater efficacy in advanced-staged HL compared with BEACOPP_BASE [7]. Consequently, two assumptions about the equivalence of treatment regimens were required to conduct an ITC:

• Considering 8xABVD (LYSA H34, EORTC 20,012) clinically equivalent to 6xABVD (ECHELON-1)

• Considering the 4xBEACOPP_ESC + 4xBEACOPP_BASE treatment arm clinically equivalent to BEACOPP_ESC

Table 5. Feasibility assessment of identified trials for indirect treatment comparison of BEACOPP_ESC with A+AVD.

Study name, ref	Recommendation for inclusion	Reason
BEACOPPESC RCTs (non-PET therapies)
HD12, Borchmann et al, 2011 [23]	Not suitable	Subgroup data for Stage III–IV patients not available
HD15, Engert et al, 2012 [34]	Not suitable	Subgroup data for Stage III–IV patients not available, no common comparator
Kriachok et al, 2013 [44]	Not suitable	No common comparator, conference abstract with limited information
HD9, Engert et al, 2009 [32]	Not suitable	Subgroup data for Stage III–IV patients not available, no common comparator
HD2000, Federico et al, 2009; Merli et al, 2016 [35,36]	Not suitable	Subgroup data for Stage III–IV patients not available
Viviani et al, 2011 [59]	Not suitable	Subgroup data for Stage III–IV patients not available
LYSA H34, Mounier et al, 2014 [46]	Not suitable	Investigated 8x ABVD, which is not comparable with ECHELON-1 (6x ABVD), investigated 4x BEACOPPESC + 4x BEACOPPBASE (not pure BEACOPPESC), follow-up data are reported at Year 5 (vs Year 3 for ECHELON-1)
EORTC 20,012, Carde et al, 2016 [26]	Not suitable	Investigated 8x ABVD, which is not comparable with ECHELON-1 (6x ABVD), investigated 4x BEACOPPESC + 4x BEACOPPBASE (not pure BEACOPPESC), follow-up data are reported at Year 4 (vs Year 3 for ECHELON-1)
BEACOPPESC RCTs (PET-guided therapies)
AHL2011 LYSA, Casanovas et al, 2018 [72]	Not suitable	Subgroup data for Stage III–IV patients not available, no common comparator, data reported for overall standard vs experimental (PET-directed) treatment arms but not according to treatment assigned for PET positive and PET negative populations
HD18, Borchmann et al, 2017 [62]	Not suitable	No common comparator
BEACOPPESC non-randomized studies
Belada et al, 2015 [84]	Not suitable	Subgroup data for Stage III–IV patients not available; single-arm retrospective study
Fossa et al, 2012 [87]	Not suitable	Subgroup data for Stage III–IV patients not available; single-arm retrospective study
Russell et al, 2017 [120]	Not suitable	Subgroup data for Stage III–IV patients not available; retrospective study

Abbreviations: ABVD, doxorubicin (Adriamycin), bleomycin, vinblastine and dacarbazine; BEACOPP, bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine and prednisone escalated cycles; PET, positron emission tomography; RCT, randomized controlled trial.

These assumptions were considered not clinically justifiable, due to potential differences in the efficacy and safety profiles of the regimens, which could lead to considerable bias in the analysis results; the longer 8xABVD regimen may result in a greater number of AEs and discontinuations and thus could bias the safety results in favor of BEACOPP_ESC. Therefore, a conventional ITC was not considered feasible. In the absence of a suitable common comparator in the trials, an alternative approach would be to consider an unanchored matching-adjusted indirect comparison (MAIC) or simulated treatment comparison (STC) analysis. However, the strong assumption on which this method is based is that all treatment effect modifiers and prognostic factors are accounted for in the model. This is considered very difficult, the only factors that are available for matching are those reported in the baseline characteristics table in the comparator trial. This assumption also cannot hold if there are unobserved differences between treatment arms that impact patient outcome; an unknown amount of residual bias may be introduced to the results, making it difficult to justify that this approach will produce a less biased estimate of the relative treatment effect than a conventional ITC [126].

In the context of the current data set, the only study identified as potentially suitable to conduct an unanchored indirect comparison of A+AVD with BEACOPP_ESC was HD18 which reported relevant outcome data in the Stage III–IV patient subgroup for BEACOPP_ESC treatment. However, in the absence of baseline characteristic data for the Stage III–IV subgroup from the HD18 trial, it would not be possible to match the Stage III–IV group exactly. Matching patient characteristics are one of the key steps of an unanchored MAIC or STC, and therefore these analyses were also not considered feasible.

4. Discussion

The aim of this SLR was to assess the published evidence for the efficacy and safety of treatments for newly-diagnosed advanced-stage HL. Specifically, the front-line treatments for HL, ABVD and BEACOPP, with and without RT, in addition to the newly approved therapy A+AVD. This SLR was conducted according to PRISMA guidelines, however, there were some limitations. The same efficacy outcomes were not reported for all included studies and treatment regimens were not always consistent between trials, with some involving different cycles of treatment. This review focused primarily on 5-year OS and PFS rates as these were the most commonly reported. However, not all studies assessed these specific outcomes, and some reported values at different time-points, limiting comparisons between studies. A particular focus was Stage IV disease or extra-nodal involvement, however, subgroup analyses for these populations were not reported for most studies. Inclusion criteria were largely consistent across studies included in the SLR. Most studies included patients aged ≥18–60 years, with the exception of a small number of studies that also included patients aged >60 years. The focus of this SLR was to assess the efficacy and safety of current standard of care frontline therapies used in the treatment of patients with advanced HL and therefore other therapies such as Stanford V and COPP are discussed briefly. Additionally, results of a Phase II study investigating BEACOPP backbone plus brentuximab vedotin have been presented; however, this therapy has not been described in detail as results from the Phase III trial (HD21) are pending.

Results from included studies showed high efficacy for ABVD and BEACOPP. Several of the studies observed lower complete response, OS, FFS, and FFP rates in patients with Stage IV treated with MOPP/ABVD compared with other disease stages [25,37,83]. Progression-free survival rates were also lower in patients with Stage IV HL treated with ABVD/BEACOPP [68] and A+AVD/AVBD [64,69,71], compared with less-advanced disease. Following treatment with MOPP and/or ABVD, patients with extra-nodal involvement had lower complete response, FFP, FFS, and OS rates than those with no or less extra-nodal involvement [25,48,82]. However, in a study of patients treated with BEACOPP_ESC or R-BEACOPP_ESC, 3-year estimates for PFS and OS remained largely unchanged in patients with Stage IV vs Stage II–IV disease.

The current review has similarities with a recent Cochrane review by Skoetz et al., 2017 comparing the efficacy and safety of BEACOPP vs ABVD for treatment of HL [6]; however, there are several key differences which are of relevance when considering the results of each review (Supplementary Table 21). Five RCTs [26,32,35,59,127] were identified by the Skoetz review, all of which met the criteria for inclusion in the current review, with the exception of HD14, which enrolled patients with early disease (Stage IA, IB, IIA, or IIB), and was therefore outside the scope of our review [127]. Results from the Skoetz meta-analysis indicate that BEACOPP_ESC was associated with improvements in both OS and PFS vs ABVD [6]. A network meta-analysis was not performed in the current review. Results from individual trials that compared directly ABVD and BEACOPP did not find significant differences in 4-, 5-, or 7-year survival [21,26,46,59,122]. However, a significant difference in 10-year OS across treatment arms (ABVD, BEACOPP_BASE, and BEACOPP_ESC) was observed (p < 0.001) [32], and there were significantly better 5-year PFS rates after treatment with BEACOPP regimens (BEACOPP₄₊₄ and BEACOPP₄₊₂), compared with ABVD (p < 0.05) [26,46,122]. The current review found one additional head-to-head RCT that was not included in the Skoetz review [21], which was part of the HD9 trial, but included only patients aged 66–75 [21] and was excluded as it did not include BEACOPP_ESC, a regimen that is not well tolerated in older patients [6,21,32]. In contrast to the original study, older patients treated with BEACOPP_BASE did not have significantly higher OS and complete response rates compared with the alternating ABVD/COPP regimen [21,32].

Although ABVD and BEACOPP are efficacious treatment options for advanced-stage HL, both are associated with several tolerability issues and short- and long-term side effects. The Skoetz review found that use of BEACOPP resulted in a statistically significant increase in incidence of AML and MDS (p = 0.02), and BEACOPP_ESC in an increase in Grade 3/4 hematological events compared with ABVD [6]. This was also true in the additional studies found in our review. Elderly patients treated with BEACOPP_BASE had higher rates of Grade 4 acute toxicity vs ABVD/COPP [21]. Mounier et al., 2014 showed higher rates of anemia and leukopenia in the BEACOPP₄₊₄ treatment group vs ABVD [46]. Chemotherapy-induced neutropenia is another common side-effect of HL treatment, and can result in dose-modifications or treatment delays [35,128]. One study reported neutropenia in 90% of patients treated with BEACOPP₄₊₄ [84]. This compares with results from the Skoetz meta-analysis, in which patients treated with BEACOPP_ESC vs ABVD had a statistically significant increase in neutropenia (p < 0.0001) [6]. Risk of neutropenia can be managed using G-CSF [125], which in our review ranged from use as primary prophylaxis [19] and use with certain treatments such as BEACOPP_ESC [84], to secondary prophylaxis for patients with severe neutropenia [39]. ABVD and BEACOPP may also be associated with side effects not identified in this review. For example, alkylating agents in the BEACOPP regimen can affect fertility [129].

Radiotherapy is also associated with a range of early (skin changes, fatigue, dry mouth, nausea, and diarrhea) and late (hypothyroidism) adverse effects [8]. Secondary malignancies are higher in patients treated with ABVD that receive RT, compared with those that do not; as patients receiving BEACOPP_ESC compared with ABVD require less additional RT, they may be less likely to develop solid tumors [130]. The studies identified did not report direct associations of toxicity with RT; however, many treatment-related AEs were reported that have commonly been linked to RT, including nausea, vomiting, infections, skin-related complications, cardiovascular, and pulmonary toxicity. Chemotherapy agents may exacerbate these RT-associated conditions; use of anthracyclines, such as doxorubicin, are associated with an increased risk of heart failure [129,131].

The most serious long-term effect of RT is increased risk of secondary malignancy, which was reported in 8.3% and 1.7–4% of patients treated with ABVD [83] and BEACOPP [62,68,84], respectively. Consistent with published data, AML and NHL were the most commonly observed secondary malignancies [8]. The incidence of secondary malignancies is higher in patients who had combined RT and chemotherapy, vs RT alone [132]; however, this was not specifically reported in the included studies.

Previous studies have associated bleomycin with pulmonary toxicity [9,10]. Eleven studies in our review identified a causal relationship between pulmonary toxicity and bleomycin [26,30,33,38,41,42,48,68,91,93,101,124]. G-CSF may increase the incidence of bleomycin-associated lung toxicity [9,133]. Bleomycin discontinuation, or dose-reduction, resulted in fewer adverse events such as pulmonary toxicity, without a significant effect on efficacy (OS and PFS) [48,68,91,93].

Recently, PET response-adapted therapies have been assessed to improve efficacy and avoid toxicity. In addition to four studies reporting patients’ response to PET response-adapted therapy [60,61,68,73,78], two recent SLRs have investigated PET and HL; one on interim PET results and prognosis concluded that negative interim PET results have a large advantage in terms of OS (moderate certainty evidence) [134]. A review of PET-adapted therapy for HL recommended a PET treatment approach as standard of care for advanced HL [135].

Superior efficacy results were reported for A+AVD vs ABVD in the ECHELON-1 trial based on 2-year modified PFS. Modified PFS, in addition to disease progression or death, included modified progression (defined as evidence of noncomplete response after the completion of frontline chemotherapy based on independently assessed PET results, followed by subsequent anticancer therapy) as an event, which more accurately assesses the curative potential of frontline chemotherapy for HL than the standard endpoint PFS. The conventional end point of PFS does not accurately assess a cure by frontline chemotherapy; PFS would not recognize the persistence of disease when it does not cause progression or death, a state which requires further intervention [64].

Patients aged ≥60 years with HL have limited treatment options and inferior survival due to treatment-related toxicities and comorbidities. Brentuximab vedotin plus dacarbazine may be a frontline option for patients aged ≥60 years based on tolerability and response duration observed in a Phase II trial [117]. Brentuximab vedotin plus bendamustine is not considered a tolerable treatment regimen for older patients, as treatment was stopped prematurely due to an unacceptably high rate of SAEs and deaths, highlighting the importance of dedicated trials evaluating novel regimens in elderly patients. However, A+AVD is a treatment option for older patients with advanced HL, as it provides a tolerable treatment option with equivalent efficacy to ABVD [64]. In the ECHELON-1 study, lower rates of pulmonary toxicity were reported for A+AVD vs ABVD. Although higher rates of myelotoxicity and neuropathy were observed in the A+AVD vs ABVD arm, it was noted that myelotoxicity can be ameliorated with prophylactic G-CSF and neurotoxicity was largely reversible.

Although good efficacy outcomes can be achieved in most patients with advanced-stage HL with ABVD and BEACOPP, patients with Stage IV disease, extra-nodal involvement, or elderly patients are likely to benefit from treatments with enhanced efficacy. There is significant scope for improvements in safety and tolerability for all patients, particularly with regards to nausea/vomiting, neutropenia, and RT-associated AEs. Brentuximab vedotin has been recently approved for the treatment of Stage III (US)/Stage IV (US and Europe) HL, in combination with AVD (A+AVD), which does not contain bleomycin and provides an additional option for clinicians. No head-to-head trial data comparing A+AVD with BEACOPP_ESC are available, and based on available published evidence, it is not possible to compare these two therapies by ITC. However, the safety and efficacy of A+AVD and ABVD were directly compared in the Phase III RCT ECHELON-1 and demonstrated that PFS (based on 2-year modified PFS) and OS were higher and pulmonary toxicity was decreased in the A+AVD vs ABVD group [64]. Furthermore, the treatment benefit was higher for patients with Stage IV vs III HL and for patients with extra-nodal involvement [64,69,71].

5. Conclusion

ABVD and BEACOPP are efficacious treatment options for advanced-stage HL, with reported 5-year OS rates ranging from 60–97% and 84–99%, and 5-year PFS rates ranging from 58–81% and 83–96%, respectively. However, both treatments are associated with several tolerability issues and short- and long-term side effects, particularly with regards to nausea/vomiting, neutropenia, and RT-associated AEs. New therapies, such as A+AVD, may provide a more tolerable treatment option for patients with advanced stage HL.

Author contributions

Mehul Dalal: Conceptualization, Project administration, methodology, Writing – Review & Editing. Jatin Gupta: Methodology, investigation, validation, Writing – Review & Editing. Kim price: Supervision, Writing – Review & Editing. Athanasios Zomas: Writing – Review & Editing. Harry Miao: Writing – Review & Editing. Ajibade Ashaye: Project administration, Supervision, validation, Writing – Review & Editing.

Declaration of interest

M Dalal, A Ashaye, A Zomas and H Miao are currently employed by Takeda and owns stock in the aforementioned. J Gupta is currently employed by Decision Resources Group, who were commissioned to conduct the systematic review by Takeda. K Price is a former employee of Decision Resources Group, who were commissioned to conduct the systematic review by Takeda. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

The authors would like to thank Laura Eaton (Decision Resources Group) for writing assistance.

Data availability statement

All data generated or analyzed during this study are included in this published article as supplementary information.

Supplementary data

Supplemental data for this article can be accessed here.

Additional information

Funding

This study was funded by Millennium Pharmaceuticals, Inc., Cambridge, MA, USA, a wholly owned subsidiary of Takeda Pharmaceuticals.