PET-guided treatment for personalised therapy of Hodgkin lymphoma and aggressive non-Hodgkin lymphoma

Abstract

FDG-PET scanning has a central role in lymphoma staging and response assessment. There is a growing body of evidence that PET response assessment during and after initial systemic therapy can provide useful prognostic information, and PET response has an evolving role in guiding patient care. This review provides a perspective on the role of PET response assessment for individualised management of patients with the most common aggressive lymphomas, Hodgkin lymphoma and diffuse large B-cell lymphoma.

Introduction

Lymphoma management is influenced by individual patient and disease characteristics, including prognostic indices, disease stage and response to initial therapy. Fluorodeoxyglucose Positron Emission Tomography (PET) plays a vital role in both disease staging and response assessment. PET response to initial therapy is a major prognostic factor and has increasingly been used in clinical practice to guide treatment.

The establishment of standardised response criteria has facilitated the development of PET-directed treatment.^1,2 The 2007 International Harmonisation Project (IHP) consensus criteria have been replaced by the more recent 5-point Deauville scale (DS), which provides a pragmatic and reproducible method for response assessment.^1,3 The threshold for PET-positivity can be set according to performance requirements. A lower threshold (i.e., DS 3, i.e. level of mediastinal blood pool) may be suitable for de-escalation studies where high sensitivity is required, whereas a higher threshold (i.e., DS 4, i.e. liver uptake) provides greater specificity in settings where treatment escalation is being considered.^2–4

This review focuses on the practical application of PET response to modify frontline management of Hodgkin lymphoma (HL) and diffuse-large B-cell lymphoma (DLBCL).

Early-stage Hodgkin lymphoma

Key studies

Favourable-risk disease

For early-stage (ES) HL, combined modality therapy (CMT) with 2–4 cycles of ABVD chemotherapy (doxorubicin, bleomycin, vinblastine, dacarbazine) followed by radiotherapy is widely considered standard of care. Modern involved-node/site radiotherapy (INRT/ISRT) is expected to be associated with less late toxicity than older mantle and involved-field radiotherapy (IFRT), however, potential residual toxicity risks have led to ongoing interest in treatment with chemotherapy alone. Three randomised trials (RAPID, H10 and HD16) failed to demonstrate non-inferiority of chemotherapy alone for patients with complete metabolic response (CMR) to initial chemotherapy, with a PFS difference of 3.8–11.9% favouring consolidative radiotherapy (Table 1).^5,7,10 However, no difference in 5-year overall survival rates has been observed and current follow-up is too short to evaluate late radiotherapy-related cardiac toxicity and second malignancies. Thus, while a matter of ongoing debate, PET-adapted omission of radiotherapy is considered acceptable for selected favourable-risk patients with CMR.^11,12

Table 1.

Key Trials of PET-Adapted Therapy in Early-Stage Hodgkin Lymphoma

Trial	Phase	Inclusion	PET assessment		Treatment		N	Key endpoints	Comments
			Timing	Definition of CMR	PET status	Regimen
Favourable
UK NCRI RAPID5	3	Stage IA/IIA No mediastinal bulk	Post C3	DS 1–2	CMR	3x ABVD +30 Gy IFRT	209	3y PFS 94.6%	Did not meet non-inferiority criteria for omission of radiotherapy
						3x ABVD	211	3y PFS 90.8%
					PET+	4x ABVD +30 Gy IFRT	145	5y PFS 87.6%
LYSA/EORTC/FIL H106	3	EORTC favourable risk	Post C2	IHP criteria	CMR	3x ABVD +INRT	227	5y PFS 99.0%	Did not meet non-inferiority criteria
						4x ABVD	238	5y PFS 87.1%
HD167	3	GHSG favourable risk	Post C2	DS 1–2	CMR	2x ABVD +20 Gy IFRT	300	5y PFS 93.4%	Did not meet non-inferiority criteria
						2x ABVD	328	5y PFS 86.1%
					PET+	2x ABVD +20 Gy IFRT	340	5y PFS 88.4%
CALGB 506048	2	Stages I/II without bulk	Post C2	DS 1–3	CMR	4x ABVD	135	3y PFS 91%
					PET+	2x ABVD +2x eBEACOPP +IFRT	14	3y PFS 66%
Unfavourable
GHSG HD17  9	3	GHSG unfavourable risk Excludes Stage 2B with mediastinal bulk or extranodal disease	Post C4	DS 1–2	CMR	2x eBEACOPP +2x ABVD +IFRT 30 Gy	318	5y PFS 97.6%	Demonstrated non-inferiority for omission of radiotherapy
						2x eBEACOPP +2x ABVD	333	5y PFS 96.0%
					PET+	2x eBEACOPP +2x ABVD +IFRT 30 Gy	328	5y PFS 94.2%
LYSA/EORTC/FIL H106	3	EORTC unfavourable risk	Post C2	IHP criteria	CMR	4x ABVD +INRT	292	5y PFS 92.1%	Established PET-based treatment escalation for patients receiving frontline ABVD
						6x ABVD	302	5y PFS 89.6%
					PET+*	4x ABVD +INRT*	192	5y PFS 77.4%
						2x ABVD +2x eBEACOPP +INRT	169	5y PFS 90.6%

ABVD, Doxorubicin, bleomycin, vinblastine; C, Cycle; CMR, Complete metabolic response; DS, Deauville score; EORTC, European organisation for the research and treatment of cancer; FIL, Italian lymphoma foundation; GHSG, German hodgkin study group; IFRT, Involved field radiotherapy; IHP, International harmonisation project; INRT, Involved node radiotherapy; LYSA, Lymphoma study association; PFS, progression free survival; SWOG, South-Western oncology group; UK NCRI, United Kingdom National cancer research Institute; eBEACOPP, Escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine.

Results are given for intention-to-treat analyses. *27% of patients had favourable-risk disease and received 3x ABVD + INRT in the standard arm.

Patients with a positive PET scan (DS 3–5) generally have very good outcomes with CMT, with ~88% PFS following 2–4 cycles of ABVD plus radiotherapy.^5,7 In the RAPID trial, a worse prognosis was only associated with DS 5 (4% patients; 5 year PFS 61.9%).¹³ The H10 trial demonstrated that patients with a positive PET after two cycles (PET2) of ABVD benefited from treatment intensification to escalated (e) BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, prednisolone) prior to radiotherapy, although only a minority had favourable-risk disease.⁶

Unfavourable-risk disease

For ES-HL patients with risk factors, standard treatment has generally been four cycles of ABVD plus radiotherapy.¹⁴ However, the benefit of radiotherapy for patients with CMR is unclear. In the H10 trial, patients with unfavourable-risk ES-HL with CMR after two cycles of ABVD were randomised to receive either four cycles of ABVD plus INRT, or six cycles of ABVD.^6,10 Relapses occurred earlier in non-irradiated patients, with randomisation halted after an interim analysis suggested futility.¹⁰ The PFS curves have subsequently converged, with a PFS difference at 5 years of only 2.5% in favour of CMT.⁶ This small potential benefit of consolidation radiotherapy requires balancing against possible toxicity. Where chemotherapy alone is used, the UK NCRI RATHL trial (which both recruited unfavourable ES-HL and advanced-stage HL) demonstrated that bleomycin can safely be omitted for the final four cycles of ABVD for patients with CMR (DS 1–3) at PET2. Three-year PFS for patients with Stage 2 ES-HL plus risk factors (B symptoms, bulk or ≥3 involved sites) was 90%, similar to outcomes in H10 (Table 1).¹⁵

The only trial to demonstrate non-inferiority of chemotherapy alone is the German Hodgkin Study Group (GHSG) HD17 trial.⁹ Patients received two cycles of eBEACOPP plus two cycles of ABVD (“2 + 2 approach”). Patients with CMR (DS 1–2) following chemotherapy were randomised to receive 30 Gy IFRT or no further treatment, with only a 1.7% difference in 5-year PFS favouring CMT.⁹ Therefore, radiotherapy can safely be omitted for unfavourable-risk patients with CMR after frontline eBEACOPP. It is unclear whether the same applies with frontline ABVD.

Patients with a positive PET continued to receive radiotherapy in most ES-HL trials, apart from RATHL where patients received 4–6 cycles of BEACOPP-based therapy instead.¹⁵ For patients with a positive PET2 after frontline ABVD, the H10 trial demonstrated a clear PFS benefit from escalation to eBEACOPP (with radiotherapy), compared with ABVD alone, and a trend towards improved overall survival (OS).⁶

Practical considerations

PET thresholds

“PET-negativity” in most ES-HL trials was equivalent to DS 1–2. CALGB50604 was the only trial to omit radiotherapy for favourable-risk patients with DS 3. Although small numbers precluded firm conclusions, PFS for patients with DS three receiving ABVD alone was suboptimal, highlighting a need for caution when broadening criteria for de-escalation.⁸

Tumour size

Larger baseline tumours (>5–7 cm) are associated with an increased relapse risk, particularly for favourable-risk patients receiving chemotherapy alone.^16,17 In the RAPID trial, relatively few patients had larger masses, with a median maximum tumour diameter (MTD) of 3 cm for those with CMR.¹⁷ As all studies of unfavourable ES-HL include a mix of patients with bulky disease or other risk factors (e.g., B-symptoms or multiple low-volume sites), such trials may underestimate the benefit of radiotherapy in the subgroup with tumour bulk as the sole risk factor.

Baseline risk factors

Definitions of “unfavourable-risk” disease differ between trial groups.¹⁸ Some groups also consider mediastinal bulk, extranodal disease and/or B-symptoms as representing “advanced-stage” HL.¹⁸ Treatment strategies should be applied with consideration of the inclusion criteria of the relevant trial. Furthermore, stratification criteria are based on historical data, which require validation in the PET era. For patients with CMR in the RAPID trial, there was no difference in outcome for patients with “favourable” and “unfavourable” ES-HL by EORTC and GHSG criteria, albeit noting that patients with selected high-risk features were excluded (B-symptoms and mediastinal bulk).¹³

Personalised treatment

For patients with favourable-risk ES-HL who achieve CMR after ABVD, the reduced risk of relapse with radiotherapy should be balanced against potential radiotherapy toxicity, considering disease bulk/distribution, patient age, sex, co-morbidities, fertility and patient preferences. CMT is appropriate for many patients, but young patients with low-volume disease (MTD <5 cm) and DS 1–2 response can potentially be treated with ABVD alone.

For unfavourable-risk ES-HL, similar considerations influence the choice of frontline ABVD versus eBEACOPP (“2 + 2 approach”). The latter offers high progression-free survival (PFS) rates and a radiotherapy-free strategy for many patients,^9,14 but whether this justifies an increase in acute toxicity, infertility and resource use remains a matter of considerable debate.

Potential treatment pathways are illustrated in Figure 1.

Figure 1.

Treatment Pathways in Early-Stage Hodgkin Lymphoma.

Advanced-stage Hodgkin lymphoma

Key studies

Patients with Stage 3–4 HL have traditionally received 6–8 cycles of either ABVD or eBEACOPP.

Frontline ABVD

The RATHL trial demonstrated that for patients with CMR after two cycles of ABVD, de-escalation to AVD (doxorubicin, vinblastine, dacarbazine) was equally effective and less toxic than further ABVD, with less pulmonary toxicity, febrile neutropenia, and fatigue (Table 2). This strategy is widely considered standard of care.¹⁵

Table 2.

Key Trials of PET-Adapted Therapy in Advanced-Stage Hodgkin Lymphoma

Trial	Phase	Inclusion	PET assessment		Treatment		N	Key endpoints	Comments
			Timing	Definition of CMR	PET status	Regimen
FIL HD080119	3	Stage IIB to IV	Post C2	IHP criteria	CMR	6x ABVD ± RTa	354	2y PFS 81%	No definite benefit with RT for patients with baseline nodal masses ≥ 5 cm who achieve CMR
					PET+	2x ABVD +4x IGEV +SCT	158	2y PFS 76%
UK NCRI RATHL15	3	Stage IIB to IV Stage IIA with adverse features	Post C2	DS 1–3	CMR	6x ABVD	470	3y PFS 85.7%	Established PET-adapted escalation/de-escalation as standard care with frontline ABVD
						2x ABVD +4x AVD	465	3y PFS 84.4%
					PET+	2x ABVD plus 6x BEACOPP or 4x eBEACOPP	172	3y PFS 67.5%
FIL HD060720	2	Stage IIB to IV	Post C2	DS 1–3	CMR	6x ABVD ± RTa	630	3y PFS 87%	No definite benefit with RT for patients with baseline nodal masses ≥ 5 cm who achieve CMR
					PET+	2x ABVD +4x eBEACOPP +4x BEACOPP	76	3y PFS 57%
						2x ABVD +4x R-eBEACOPP +4x R-BEACOPP	72	3y PFS 63%
GHSG HD1821	3	Stage III to IV Stage IIB with mediastinal bulk or extranodal disease	Post C2	DS 1–2	CMR	4x eBEACOPP	501	5y PFS 92.2%	Established PET-based de-escalation as standard care with frontline eBEACOPP
						6–8x eBEACOPP	504	5y PFS 90.8%
					PET+	8x eBEACOPP	217	5y PFS 89.7%
						8x R-BEACOPP	217	5y PFS 88.1%
LYSA AHL201122	3	Stage III to IV Stage IIB with mediastinal bulk or extra nodal disease	Post C2	SUVmax <1.4x liver	CMR	6x eBEACOPP	349	5y PFS 88.4%	Established PET-based de-escalation as standard care with frontline eBEACOPP
						2x eBEACOPP +4x ABVD	346	5y PFS 89.4%
					PET+	6x eBEACOPP	100	5y PFS 70.7%
ECHELON-123	3	Stage III to IV	N/A	N/A	N/A	6x BV + AVD	664	3y PFS 83.1%	PET2 performed but not used to adapt treatment
						6x ABVD	670	3y PFS 76%

BV, Brentuximab vedotin; IGEV, Ifosfamide, gemcitabine, vinorelbine; N/A, Not applicable; SCT, Autologous stem cell transplant; SUV, Standardised uptake value.

^aPatients with large baseline nodal masses ≥ 5 cm were randomised to receive RT or no further treatment

For patients with a positive PET2 after frontline ABVD, three studies (HD0607, S0816 and RATHL) evaluated intensification with BEACOPP, reporting 3–5 year PFS of 60–67.5% compared to historical controls with PFS rates ≤28%.^{15,20,24–27} Although there is uncertainty about the true effect size, randomised data from the H10 trial in ES-HL provide further support of concept.⁶ This approach has widely been adopted in clinical practice for advanced-stage HL. Intensification with IGEV (ifosfamide, gemcitabine, vinorelbine) plus high-dose therapy and autologous stem cell transplant (HDT-ASCT) resulted in high PFS rates (75% at 2 years) in the HD0801 trial but has significant toxicity, with fewer salvage options for those that relapse.¹⁹

Frontline eBEACOPP

Although pre-PET era trials demonstrated better disease control with frontline eBEACOPP than ABVD, its use has been limited by toxicity concerns.²⁸ Response-adapted de-escalation of eBEACOPP has been evaluated in two randomised trials (Table 2).^21,22 The GHSG HD18 trial demonstrated that four cycles of eBEACOPP are non-inferior to 6–8 cycles for patients with CMR at PET2.²¹ Toxicity, resource utilisation, and treatment duration clearly favoured abbreviated treatment.²¹ However, due to the conservative definition of PET-positivity (DS 3–5), only 52% of patients were eligible for de-escalation.

The LYSA AHL2011 trial demonstrated that patients with CMR after 2 cycles of eBEACOPP can be de-escalated to four cycles of ABVD without loss of efficacy, and with less acute toxicity and better fertility preservation.^22,29 A higher threshold was used to define PET positivity (1.3x liver SUV_max) making the majority of patients (87%) eligible for treatment de-escalation. Furthermore, this approach is potentially more cost-effective than the other frontline advanced-stage HL treatment strategies described in this review.³⁰

In both HD18 and AHL2011, PET2-positive patients continued to receive six cycles of eBEACOPP.

Targeted agents

The ECHELON-1 trial demonstrated that six cycles of brentuximab vedotin (BV) plus AVD are superior to six cycles of ABVD, with a 7% improvement in 3-year PFS.²³ The absence of an OS benefit and cost has limited adoption of this strategy. ECHELON-1 has also been criticised for not allowing PET-directed intensification nor omission of bleomycin in the ABVD arm, where there were several pulmonary deaths. Response-adapted approaches may help to limit both the toxicity and financial burden of novel agents in future.

RT consolidation for advanced-stage HL

For patients with CMR after frontline ABVD, the HD0607 and HD0801 trials reported that irradiating initial masses ≥ 5 cm led to a small but non-significant increase in PFS (4–7.5% at 3–5 years).^20,31 With frontline eBEACOPP, the GHSG HD15 trial demonstrated that consolidative radiotherapy is not required for patients with CMR.³² GHSG trials limit use radiotherapy to PET-positive masses (DS 3–5) measuring ≥2.5 cm (~12% patients) with PFS rates only slightly lower than those with CMR.^21,32 In the RATHL and AHL2011 trials, patients with a positive post-chemotherapy PET generally received radiotherapy and/or salvage treatment according to clinician preference.^15,22 Randomised data are lacking but, in principle, radiotherapy is most likely to benefit patients with limited-extent, non-progressive residual FDG-avid lesions.

Interim PET has been used in paediatric trials to identify high-risk patients for consolidative radiotherapy.³³ However, PET2-positive patients in the EURONET-PHL-C1 trial received radiotherapy to all initially involved nodal sites, potentially creating large radiotherapy fields, so has not been widely adopted. The subsequent EURONET-PHL-C2 trial is comparing this approach with another strategy that limits radiotherapy to areas with persistent post-chemotherapy PET-positivity only (NCT02684708). The role of interim PET to direct consolidation radiotherapy in HL remains unclear.

Practical considerations

IPS score

The International Prognostic Score (IPS) is often used to risk-stratify patients and select frontline treatment.³⁴ Pre-PET studies demonstrated that eBEACOPP offers the greatest benefit over ABVD for high-risk patients (IPS score ≥3).²⁸ However, the predictive value of IPS is suboptimal and may be diminished with PET-adapted treatment.^15,25 It is unclear whether there is a survival benefit with frontline eBEACOPP for high-risk HL patients over a RATHL approach; these PET-adapted approaches have not been directly compared.

High-risk interim PET findings

Interim PET can identify selected patients who remain at high risk of failure, irrespective of treatment intensity. Patients with a positive PET after four cycles (PET4) of eBEACOPP had 5-year PFS of only 46·5% in AHL2011.²² Patients with DS 5 after two cycles of ABVD also have PFS rates < 50%, despite escalation to BEACOPP.^15,20 The optimum strategy for these patients remains uncertain and novel approaches are required.

Personalised treatment

Frontline options for advanced-stage HL include PET-adapted treatment with frontline ABVD or eBEACOPP, or six cycles of BV-AVD. There is therefore potential to personalise treatment selection, considering patient-specific factors (fertility, age and fitness), as well as disease-specific risk. The RATHL protocol offers lower-intensity treatment with ABVD alone for the majority of patients, but PET-adapted eBEACOPP may be preferred for high-risk HL. For patients with CMR after two cycles of eBEACOPP, there is a further choice between an HD18 or AHL2011 approach, considering toxicity, fertility, and patient preferences.^28,35 BV-AVD, where available, has utility for older, anthracycline-fit patients, who have a higher risk of bleomycin and eBEACOPP toxicity. Consolidative radiotherapy is largely reserved for patients with PET-positive masses after completion of chemotherapy.

Limited-stage DLBCL

Key studies

Radiotherapy following 3–4 cycles of R-CHOP chemotherapy (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisolone) has been a standard approach for early-stage DLBCL without bulk or IPI risk factors, and is curative in approximately 90% patients.^36–39 Similar results have been achieved with six cycles of R-CHOP, without radiotherapy.⁴⁰ Recent studies have explored the use of fewer chemotherapy cycles and the omission of radiotherapy, in an effort to minimise radiation- and chemotherapy-related toxicity (Table 3).

Table 3.

Key Studies of PET-Directed Radiotherapy in DLBCL

Trial	Phase	Inclusion	PET assessment		Treatment		N	Key endpoints	Comments
			Timing	Definition of CMR	PET status	Regimen
Limited-Stage Non-Bulky DLBCL
FLYER41	3	Stage I/II, non-bulky (7.5 cm), normal LDH	N/A	N/A	N/A	6x R-CHOP	295	3y PFS 94%	Established abbreviated immunochemotherapy as standard care
						4x R-CHOP +2x R	297	3y PFS 96%
LYSA/ GOELAMS38	3	Stage I/II Non-bulky (<7 cm)	Post C4	IHP criteria	CMR	4–6 x R-CHOP +IFRT	144	5y EFS 92%a	Suggested non-inferiority of chemotherapy alone
						4–6 x R-CHOP	137	5y EFS 89%a
Intergroup  S100142	2	Stage I/II Non-bulky (<10 cm)	Post C3	DS 1–3	CMR	4x R-CHOP	114	5y PFS 89%	Supports data from randomised LYSA trial
					PET+	3x R-CHOP +IFRT ibritumomab tiuxetan	14	5y PFS 86%
Advanced-Stage or Bulky DLBCL
UNFOLDER43	3	Bulk (≥7.5 cm) or extranodal disease	N/A	N/A	N/A	6 x R-CHOP	162	3y EFS 68% 3y PFS 89%	Suggests that RT is not required, but PET not performed. EFS not interpretable, due to an imbalance between arms.
						6 x R-CHOP +RT to bulky/extranodal sites	305	3y EFS 84% 3y PFS 81%
OPTIMAL  Substudy44	3	61-80y Bulk ≥ 7.5 cm	Post C6	Unknown	CMR	6x R-CHOP + 2 x R	187	2 year PFS 79%	Similar outcomes to preceding RICOVER-60 trial, despite modest reduction in RT use
					PET+	6x R-CHOP + 2 x R + 39.8 Gy ISRT
Vancouver45	Retros-pective	Stage III/IV Stage I/II with either B symptoms or bulk ≥10 cm	Post C6	IHP (pre-2014) then DS 1–3	CMR	6x R-CHOP	517	3y TTP 83%	For patients with CMR, outcomes were similar for patients with and without bulk >10 cm (3y EFS 82% and 84%, respectively), without RT
					PET+	6x R-CHOP +RT	109	3y TTP 76%
						6x R-CHOP without RT	97	3y TTP 34%

CMR, Complete metabolic response; DS, Deauville score; IFRT, involved field radiotherapy; IHP, International harmonisation project; LDH, Lactate dehydrogenase; PFS, progression free survival; R, rituximab; R-CHOP, Rituximab, cyclophosphamide, doxorubicin, vincristine; TTP, Time to progression; eBEACOPP, Escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine.

^aIncludes a small proportion of PET-positive patients who received RT in both arms.

The randomised FLYER trial for low-risk DLBCL (bulk <7.5 cm, age-adjusted International Prognostic score (IPI) 0) demonstrated non-inferiority of four cycles of R-CHOP (followed by 2x rituximab) to six cycles of RCHOP. These excellent outcomes were achieved without use of PET, and only 5% of patients received consolidation radiotherapy.⁴¹

In a LYSA/GOELAMS trial, patients with stage I-II (bulk <7 cm) DLBCL received 4–6 cycles of R-CHOP (depending on modified IPI). Patients with CMR at PET4 were randomised to receive IFRT or observation.³⁸ Radiotherapy reduced the number of local relapses but PFS and OS were not improved. Of note, 44% patients had a modified IPI score of 1 and received six cycles of R-CHOP, and 5% of those allocated to the radiotherapy arm did not receive it.

The US Intergroup S1001 trial treated patients with stage I-II non-bulky (<10 cm) DLBCL who achieved CMR at PET3 with four cycles of R-CHOP alone. Event-free survival (EFS) rates for those with CMR were similarly high, although interventions were not randomised.⁴²

Practical considerations

These trials suggest that four cycles of R-CHOP ( ± 2x rituximab) without radiotherapy is sufficient therapy for most patients with favourable-risk, limited-stage, non-bulky DLBCL. However, several features should be borne in mind when interpreting their results.

PET response

The LYSA/GOELAMS trial used IHP criteria to define PET response.³⁸ It remains unclear whether four cycles of R-CHOP are sufficient for patients with DS 3 response, without radiotherapy. In the S1001 trial, relapse rates were slightly higher in patients with DS 3 (3/30 patients) than those with DS 1–2 (1/80), but the difference was not statistically significant and patients were not randomised.⁴² However, the excellent results in FLYER challenge the necessity of PET-directed therapy in this setting.

Tumour size

Tumour size is an important determinant of outcomes for patients receiving chemotherapy alone,⁴⁶ with a higher risk of treatment failure for every centimetre increase in tumour diameter.^40,46 The median tumour diameter in S1001 was 3.5 cm, and was not reported for other trials.⁴² In the LYSA/GOELAMS study, 20% of patients had no evident disease at study entry. These observations should be considered when interpreting the study outcomes, and some caution is required when de-escalating treatment for patients with larger tumours within the “non-bulky” range (i.e., 5–7 cm). The presence of residual PET-negative masses (>2 cm) may also confer an increased risk of relapse with chemotherapy alone,⁴⁶ although there are conflicting data.^47,48

Extranodal disease

These trials included a slightly lower proportion of patients with extranodal disease (32–43%) than in clinical practice. Extranodal disease was not associated with adverse outcomes, and retrospective data also support omission of radiotherapy for patients CMR.⁴⁹ However, certain extranodal sites (e.g. testicular, breast, cutaneous) were either excluded or recruited in low numbers. For rare and/or high-risk extranodal sites, caution is required with omission of radiotherapy; site-specific literature should be used to inform treatment strategy.

Personalised treatment

For patients with favourable-risk, non-bulky (<7 cm) stage I-II DLBCL with CMR after R-CHOP, there is a choice between four cycles of R-CHOP (± 2 cycles of rituximab) and 3–4 cycles of R-CHOP plus ISRT. Disease-specific factors (distribution and bulk) and patient-related factors (age and comorbidities) are important in treatment planning. CMT may remain preferable for older and/or co-morbid patients, where toxicity from chemotherapy is higher and the incidence of late radiotherapy-related toxicity is low. Radiotherapy should also be considered for selected extranodal sites and those with larger tumour masses. For patients with a suboptimal PET response after 3–4 cycles of R-CHOP, CMT remains standard care.

Primary mediastinal B-cell lymphoma (PMBCL)

PMBCL presents a special case of largely localised large-cell lymphoma, traditionally treated with six cycles of R-CHOP plus radiotherapy. However, the role of radiotherapy has remained controversial, given its toxicity risks in a young patient population.⁵⁰ The recently completed, randomised IELSG37 trial is evaluating whether RT can be omitted for patients with CMR (DS 1–3) after R-CHOP (NCT01599559) and results are awaited. Dose-adjusted (DA)-EPOCH-R (etoposide, prednisolone, vincristine, cyclophosphamide, doxorubicin, rituximab) has been shown to be a highly effective regimen without consolidative radiotherapy, although randomised data are lacking.^51–53 Retrospective data also show high PFS rates for patients with CMR treated with R-CHOP alone.⁵⁴

Optimal management of PMBCL patients with a positive post-chemotherapy PET scan is less clear, as at least half of patients do not progress.⁵⁵ Patients with DS 5 have the highest risk of treatment failure, both with R-CHOP plus radiotherapy and DA-EPOCH-R, but patients with DS 4 have similar outcomes to those with CMR.^54,56,57 Notably, patients with DS 4 after DA-EPOCH-R appear to have excellent outcomes without further treatment, although reported numbers are small (N = 17).⁵⁷ More data are needed to inform management of these patients; options include consolidation radiotherapy or serial imaging ± repeat biopsy.

Advanced-stage DLBCL

Six cycles of R-CHOP are the standard of care for bulky/advanced DLBCL. Attempts to improve outcomes through treatment intensification and introduction of targeted agents have largely been unsuccessful to date.^58–60

Key trials

Interim PET-adapted treatment intensification

Multiple trials have evaluated treatment intensification for patients with inadequate early PET response after R-CHOP, or similar regimens. Interpretation of these trials is challenging due to variations in PET timing and thresholds, and, in most cases, the lack of a randomised comparator.

Interim PET response is prognostic in DLBCL with a NPV typically exceeding 80%. However, the reported PPV ranges between 20 and 74%, depending on the method of PET assessment.⁶¹ Interim PET response according to IHP and Deauville criteria (DS 4–5) is less discriminatory in DLBCL than in HL. DS 5 has greater ability to identify true non-responders, as do quantitative methods such as ΔSUVmax.⁶² The latter quantifies the relative reduction in maximum Standardised Uptake Value (SUV) between baseline and interim PET, with inadequate response generally defined as ΔSUVmax ≤ 66% at PET2 and ≤70% at PET4.^63,64 The timing of PET assessment is also important, with a recent meta-analysis reporting a higher PPV of a positive PET4 than PET2.⁶²

The most widely investigated approach for patents with inadequate PET response has been consolidation with HDT-ASCT. The GAINED trial escalated patients with inadequate PET response (PET2+/PET4- by ΔSUVmax) to high-dose methotrexate plus HDT-ASCT, with 4-year EFS (83.9%) similar to early responders who continued standard immunochemotherapy (83.0%).⁶⁵ However, PET4-positive patients were excluded, so it is not clear that those with “inadequate response” were truly high-risk; comparative data with standard immunochemotherapy in this group are lacking.⁶⁵ The ALLG NHL21 trial evaluated R-ICE (rituximab, ifosfamide, carboplatin, etoposide), followed by HDT-ASCT for PET-positive patients (IHP criteria) after four cycles of R-CHOP, with 2-year PFS (67%) comparable to those with CMR (74%) who continued R-CHOP.⁶⁶ The authors of both trials suggest that HDT-ASCT can overcome the adverse impact of early PET-positivity and rescue high-risk DLBCL. However, neither trial was randomised, and PET criteria for identifying high-risk patients were suboptimal. Other trials have reported discordant results, with PFS rates of 53–57%, despite escalation with HDT-ASCT.^67–69

The only published randomised trial to date is the PETAL trial.⁶⁴ Patients with aggressive non-Hodgkin lymphoma (mostly DLBCL) with inadequate response after two cycles of R-CHOP (ΔSUVmax ≤ 66%) were randomised to receive either a dose-intense Burkitt lymphoma regimen or six further cycles of R-CHOP. Accrual stopped early due to futility, with a 2-year EFS rate of 31.6% for Burkitt-like therapy compared with 42.0% for R-CHOP, and greater toxicity with the former.⁶⁴ Multiple single-arm trials have evaluated PET-based escalation to R-ICE, or related regimens, for those with inadequate interim response to R-CHOP. However, only 28–53% of patients achieved CMR after R-ICE.^66,67,70 Furthermore, up to 20% failed to complete planned chemotherapy due to toxicity.⁷⁰

These trials highlight the difficulties in delivering intensive therapy to this patient population and challenge the paradigm that intensification can overcome early treatment resistance. Therefore, PET-directed treatment intensification remains investigational and has not been widely adopted in DLBCL.

Image-guided radiotherapy

Multiple non-randomised studies have reported a benefit of radiotherapy following R-CHOP for advanced DLBCL, particularly for patients with initial disease bulk, which is an established prognostic factor.^{40,46,71–81} The randomised UNFOLDER trial, reported in abstract form, evaluated radiotherapy to sites of bulk and/or extranodal disease following six cycles of R-CHOP. There was a non-significant difference in PFS favouring radiotherapy (8%; p = 0.221) but no difference in OS.⁴³ The role of radiotherapy for advanced DLBCL remains controversial, particularly in the era of PET response assessment.⁸²

Several studies have evaluated post-chemotherapy PET, rather than disease bulk, to guide radiotherapy use.^44,83 In the OPTIMAL trial, patients with CMR after R-CHOP did not receive RT, and 2-year PFS (79%) was similar to that of irradiated patients in the historical RICOVER trial (75%).^44,71,84 The Vancouver group published a population-based study of PET-guided radiotherapy for advanced DLBCL. Radiotherapy was not given patients with CMR, and there was no difference in outcomes between bulky and non-bulky disease.⁴⁵ These studies have been interpreted to suggest that radiotherapy is not needed for patients with CMR, but are limited by their non-randomised design. Further work is needed to identify whether there are subgroups who might benefit from radiotherapy. Interim PET response has not been well studied as a guide for radiotherapy, however at least one early study suggested that progression may more often occur at interim PET-positive sites.^73,85–87

For patients with a positive post-chemotherapy PET scan, the role of radiotherapy is not well defined. However, several studies have demonstrated that radiotherapy alone is sufficient for a proportion of patients who remain PET-positive post-chemotherapy, particularly those with limited, non-progressive disease sites that can be encompassed within a radiotherapy field.^44,45,83 In the retrospective study by the Vancouver group, selected DLBCL patients who received radiotherapy to residual sites of PET-positive disease had similar outcomes to those achieving post-chemotherapy CMR, who did not receive radiotherapy.⁴⁵

Practical considerations

There is evidence to suggest an interaction between bulk, and stage that influences the benefit of radiotherapy.⁸⁸ On first principles, radiotherapy may be most beneficial when there is one dominant site of disease, with minimal tumour burden elsewhere. By contrast, if a bulky site occurs in the setting of widespread disease, consolidative radiotherapy may be less beneficial.

Residual masses following systemic therapy may also confer a higher risk of relapse.^{40,43,46,71,74–77,88–90} While not all residual masses represent disease, they may identify sites at elevated risk of harbouring disease and predict potential benefit from RT.^71,91

Personalised treatment

The reference treatment for patients with bulky and/or advanced-stage DLBCL remains six cycles of R-CHOP. Available data provide limited basis for personalisation of therapy but practice varies, particularly with respect to application of radiotherapy. It is reasonable to consider consolidative radiotherapy to sites of initial bulk, particularly when such sites represent a large proportion of the initial tumour burden and/or are associated with a residual mass of >2 cm. The anticipated toxicity of radiotherapy and the patient's suitability for salvage in the event of relapse should also be considered, as small gains from radiotherapy may be worthwhile where salvage options are limited.

End-of-treatment PET has an important role in identifying patients with suboptimal response who may benefit from radiotherapy. However, the role of interim PET remains to be defined.

Conclusions and future directions

PET is now an integral part of risk-adapted treatment for HL, where there are very clear benefits for PET-directed treatment de-escalation and intensification in terms of toxicity and efficacy, respectively. In DLBCL, post-chemotherapy PET plays a key part in selecting patients for CMT, while use of interim PET remains experimental. However, in both diseases, there is a need for additional biomarkers to aid risk-stratification and refine risk-adapted treatment strategies. Finally, in both HL and DLBCL, there are high-risk patients that have inadequate outcomes despite treatment intensification and/or radiotherapy. Novel treatment strategies are required to improve outcomes for those at highest risk of treatment failure.