review
Current status of prognostication in classical Hodgkin lymphoma Girish Venkataraman,1 M. Kamran Mirza,1 Dennis A. Eichenauer2,3 and Volker Diehl3 1
Department of Pathology, Section of Hematopathology, The University of Chicago Medicine, Chicago, IL, USA, 2First Department of Internal Medicine, University Hospital Cologne, and 3German Hodgkin Study Group (GHSG), Cologne, Germany
Summary Classical Hodgkin lymphoma (cHL) is characterized by a paucity of neoplastic Hodgkin/Reed Sternberg (HRS) cells within a complex cellular milieu that is rendered immunologically incapable of reacting against CD30+HRS cells due to a plethora of immune escape mechanisms initiated by the neoplastic cells. Accounting for 25% of all lymphomas and nearly 95% of all Hodgkin lymphomas, patients with cHL are typically young adults. Besides traditional prognostic factors, such as the International Prognostic Index (IPI), newer imaging and ancillary biomarkers (CD68, Galectin-1 and plasma microRNA) have shown promise. Furthermore, the evolution of gene expression profiling (GEP) in recent years has enabled the development of several practically feasible GEP-based predictors with prognostic relevance. This review discusses the current status of clinical prognostication in cHL, the critical role of histological evaluation in light of several mimicking entities, and the relevance of tissue as well as serum biomarkers pertaining to immune escape mechanisms and recent GEP studies. Keywords: prognosis, classical Hodgkin lymphoma, immunohistochemistry, genome expression profiling, serum biomarkers, clinical risk factors. Since the seminal description ‘on some morbid experiences of the absorbent glands and spleen’ by Thomas Hodgkin in 1832 (Hodgkin, 1832), we have come a long way in our understanding of the pathobiology of the histologically peculiar and unique lesion identified as Hodgkin lymphoma (HL). The 20th century witnessed refined morphological descriptions and several classification systems for HL. As of today, the World Health Organization (WHO) 2008 classification (Swerdlow et al, 2008) recognizes four histological subtypes of classical HL (cHL): nodular sclerosis (NSCHL), mixed
Correspondence: Girish Venkataraman, Section of Hematopathology, Department of Pathology, The University of Chicago Medicine, TW055B, MC0008, 5841 S. Maryland Avenue, Chicago, IL 60637, USA. E-mail: [email protected]
ª 2014 John Wiley & Sons Ltd British Journal of Haematology, 2014, 165, 287–299
cellularity (MCCHL), the lymphocyte rich (LRCHL) and lymphocyte-depleted (LDCHL). Overall, the fundamental description of HL has largely remained unchanged through the past century. Despite the separation of nodular lymphocytepredominant HL (NLPHL) as a distinct entity in the late 1990s, it is critical to note that both cHL and NLPHL share a common biological origin from B-cells. The two major differences between them are (i) the presence of a relatively intact germinal centre/B-cell programme and (ii) lack of a role for Epstein-Barr virus (EBV) in the pathogenesis of NLPHL. In addition, these entities differ in their clinical features and behavior. The epidemiology of cHL is notable for complex interplay of genetic, racial and geographic as well immune competence/EBV-driven factors, which are instrumental in its pathogenesis (Swerdlow et al, 2008). Clinically, HL patients have been traditionally divided into two or three prognostic groups, mainly according to the stage and presence of systemic symptoms at diagnosis. The definition of these clinical and pathological prognostic groups varies, as does the choice of treatment. Prognostic factors are rarely studied clinically but are usually discovered when analysing data from reliable, large clinical trials. Several recent reviews touch upon the Hodgkin/Reed Sternberg (HRS) cell biology, the role of the cHL microenvironment, the cytokine milieu and potential serum biomarkers in this disease (Kuppers et al, 2002; Skinnider & Mak, 2002; Kuppers, 2009; Steidl et al, 2011). In the following sections, this review focuses solely on clinical and biological markers that have been shown to have prognostic relevance with a focus on risk factors, histological/immunohistochemical findings and the advances made in gene expression profiling (GEP) over the last years. We start our discussion with an overview of established clinical risk factors for HL as well as more recently developed stratification tools, including interim fluorodeoxyglucose positron emission tomography (FDG-PET).
Clinical risk factors and clinical outcome prediction in HL Although several biomarkers have been shown to be predictive for a favourable or unfavourable clinical course in HL, to
First published online 4 February 2014 doi:10.1111/bjh.12759
Review our knowledge, none of them have been incorporated in any clinical risk stratification strategies for HL to date. In newly diagnosed HL, patients are allocated to different risk groups according to clinical features and laboratory parameters present at diagnosis. In relapsed disease, treatment stratification according to the presence or absence of risk factors has not become standard to date, so that patients with recurrent HL are usually treated with high-dose chemotherapy followed by autologous stem cell transplantation (ASCT) irrespective of the clinical stage or the presence of risk factors at time of relapse. However, a multitude of clinical risk scores for both newly diagnosed patients and patients with relapsed HL have been developed within the past 15 years. Furthermore, functional imaging with interim FDG-PET is increasingly recognized as an appropriate tool to distinguish good-risk from poor-risk patients. Thus, the clinical practice regarding the allocation of HL patients to different risk groups may change in the future.
Clinical risk factors in newly diagnosed HL Patients with newly diagnosed HL are usually allocated to different risk groups according to their clinical stage and the presence or absence of certain clinical and biological risk factors. Larger study groups, such as the European Organisation for Research and Treatment of Cancer (EORTC) and the German Hodgkin Study Group (GHSG), distinguish between early favourable, early unfavourable, and advanced HL. Early favourable HL is defined by stage I/II disease without risk factors such as large mediastinal mass, extranodal disease, elevated erythrocyte sedimentation rate (ESR) or involvement of three or more nodal areas. In contrast, patients with stage I/II disease presenting with one or more of these risk factors are usually allocated to the early unfavourable risk group while patients with stage III/IV disease have advanced HL (Table I) (Eichenauer et al, 2011; Hoppe et al, 2012). As the prognosis of patients with advanced HL is worse when compared with patients with early-stage disease, treatment strategies adapted to the patients′ risk profile have been established within the past decades. At present, patients diagnosed with early-stage disease mostly receive combined-modality approaches consisting of two to four cycles of chemotherapy followed by involved-field radiotherapy (IF-RT). ABVD (adriamycin, bleomycin, vinblastine, dacarbazine) represents the chemotherapy protocol most often used. Advanced HL is usually treated with six to eight cycles of chemotherapy optionally followed by localized RT. The question whether ABVD or escalated BEACOPP (bleomycin, etoposide, adriamycin, cyclophosphamide, vincristine, procarbazine, prednisone) should be routinely used in advanced HL is not ultimately answered although BEACOPP-escalated appears to be superior as recently revealed by a large network meta-analysis (Skoetz et al, 2013). Generally, the long-term survival rates in HL patients exceed 80% irrespective of the stage at diagnosis when adequate treatment is applied (Borchmann et al, 2012). 288
Table I. Allocation to treatment groups according the EORTC and GHSG criteria. Risk group
EORTC
GHSG
Early favourable
CS I-II without risk factors (supradiaphragmatic) CS I-II with ≥1 risk factors (supradiaphragmatic)
CS I-II without risk factors
Early unfavourable
Advanced
CS III-IV
Risk factors
(A) Large mediastinal mass (B) Age ≥50 years (C) Elevated ESR (D) ≥4 nodal areas
CS I, CS IIA with ≥1 risk factors; CS IIB with risk factors C/D, but not A/B CS IIB with risk factors A/B, CS III/IV (A) Large mediastinal mass (B) Extranodal disease (C) Elevated ESR (D) ≥3 nodal areas
GHSG, German Hodgkin Study Group, EORTC, European Organisation for Research and Treatment of Cancer, CS, Clinical stage, Elevated erythrocyte sedimentation rate (ESR), >50 mm/h without B symptoms, >30 mm/h with B symptoms, Large mediastinal mass, more than one-third of the maximum horizontal chest diameter, B symptoms, fever, night sweat, weight loss.
In 1998, the International Prognostic Score (IPS) for patients with advanced HL was introduced (Hasenclever & Diehl, 1998). This score is composed of seven adverse factors with similar independent prognostic effects, namely serum albumin of less than 40 g/l, a haemoglobin level of less than 105 g/l, male sex, age of 45 years or older, stage IV disease according to the Ann-Arbor classification, leucocytosis with a white-cell count of at least 15 9 109/l and lymphocytopenia (lymphocyte count of less than 06 9 109/l or less than 8% of the white-cell count or both). Data from 4695 patients aged between 15 and 65 years with advanced HL and mostly treated with anthracycline-containing chemotherapy protocols were considered for the analysis. While the 5-year freedom from treatment failure (FFTF) and overall survival (OS) rates for the whole patient group were 66% and 78%, respectively, FFTF rates according to the IPS ranged between 84% for patients presenting with no risk factor and 42% for patients presenting with all seven risk factors contained in the IPS. The corresponding OS rates ranged between 90% and 56%. Thus, the IPS is a valuable and reliable tool to predict the outcome of patients with newly diagnosed advanced HL. As the patients taken into account for the initial development of the IPS were treated before 1992, it has been unclear whether this tool is still applicable in patients currently undergoing HL treatment. Thus, the British Columbia Cancer Agency (BCCA) recently performed an analysis evaluating the validity of the IPS in 710 patients diagnosed with advanced HL between 1980 and 2010 and treated with ABVD or ABVD-like protocols. Within this study the IPS retained its ª 2014 John Wiley & Sons Ltd British Journal of Haematology, 2014, 165, 287–299
Review prognostic value and still discriminated high-risk from lowrisk patients. In general, the prognosis has apparently improved. Freedom from progression (FFP) estimates at 5 years ranged between 62% and 88%; the corresponding 5-year OS estimates ranged between 67% and 98% (Moccia et al, 2012). Given the relevant proportion of patients with advanced HL treated with escalated BEACOPP due to the improved tumour control associated with this protocol, it should also be mentioned that the IPS also works when this chemotherapy schema is given. In the final analysis of the pivotal HD9 trial by the GHSG comparing escalated BEACOPP with BEACOPP in baseline dosage and COPP (cyclophosphamide, vincristine, procarbazine, prednisone)/ABVD in newly diagnosed advanced HL, patients were divided into three groups according to their IPS (IPS 0-1 vs IPS 2-3 vs IPS 4-7) (Diehl et al, 2003). Among the patients treated with escalated BEACOPP, FFTF and OS rates at 5 years were 92% and 95%, respectively, for patients with an IPS of 0-1, 87% and 90% for patients with an IPS of 2-3 and 82% and 82% for patients with an IPS of 4-7 (Diehl et al, 2003).
Role of functional imaging in newly diagnosed HL Within the past decade, the use of FDG-PET has become increasingly common to distinguish good-risk from poorrisk HL patients. Hutchings et al (2006) reported 77 patients who were diagnosed with HL at different stages and mainly received ABVD chemotherapy. After two cycles, patients underwent interim FDG-PET; 16 patients were FDG-PET-positive while no residual metabolic activity was detected in 61 patients. At a median follow-up of 23 months, 11 of the 16 FDG-PET-positive patients had relapsed. In contrast, only three of the 61 patients who were FDG-PET-negative after two cycles of chemotherapy had disease recurrence. Thus, early interim FDG-PET was shown to be a good predictor for progression-free survival (PFS) in HL patients treated with ABVD-based chemotherapy (Hutchings et al, 2006). Another analysis including a larger cohort of 260 patients with newly diagnosed advanced HL receiving ABVD chemotherapy was performed by a joint group from Denmark and Italy (Gallamini et al, 2007). Early interim FDG-PET was performed after two cycles of chemotherapy. This study also revealed a strong predictive value for early interim FDG-PET in terms of PFS. At a median follow-up of 22 years, the 2year PFS rates were 128% for patients with a positive FDGPET after two cycles of chemotherapy compared to 95% for patients with a negative FDG-PET after the same amount of chemotherapy (Gallamini et al, 2007). Given these data, several prospective trials evaluating early risk-adapted treatment stratification strategies on the basis of interim FDG-PET were initiated worldwide. The majority of these trials are still ongoing and final analyses are pending. Thus, treatment ª 2014 John Wiley & Sons Ltd British Journal of Haematology, 2014, 165, 287–299
decisions based on early interim FDG-PET should not yet be made outside prospective clinical studies. The question of whether consolidating RT can be omitted in patients with a negative FDG-PET after chemotherapy is at least answered for patients receiving intensive chemotherapy with BEACOPP. Within the randomized GHSG HD15 trial for newly diagnosed advanced HL, all patients with residual lymphoma larger than 25 cm after BEACOPPescalated or BEACOPP-14 chemotherapy underwent FDG-PET evaluation (Engert et al, 2012). Patients with a positive FDGPET had localized RT while patients with a negative FDGPET had no additional RT. The negative predictive value for FDG-PET, defined as the proportion of PET-negative patients without progression, relapse or RT at 12 months, was 941% so this modality can safely be used as tool to decide whether patients with larger residual lymphoma after intensive first-line chemotherapy should be irradiated (Engert et al, 2012). The question of whether this is also true for advanced-stage patients treated with ABVD chemotherapy and patients with early-stage HL with a complete metabolic response after chemotherapy is unanswered to date. However, first results on the role of interim FDG-PET in early favourable HL have recently become available. Within the multicentre RAPID trial conducted in the UK, patients with a negative FDG-PET after three cycles of ABVD were randomized between IF-RT and no further treatment. At a median follow-up of 457 months, 3-year PFS rates were 938% in the IF-RT arm and 907% in the observation only arm. The corresponding OS rates were 970% among patients receiving IF-RT and 995% among patients receiving no RT (Radford et al, 2012). Despite of these data, final conclusions should not be drawn before mature data from further studies addressing this issue, such as the GHSG HD16 trial (NCT00736320) and the EORTC H10 (NCT00433433) study, are published.
Clinical risk factors and functional imaging in relapsed or refractory HL Similar to newly diagnosed disease, certain clinical factors as well as functional imaging were shown to be of value for outcome prediction in relapsed HL. The most comprehensive risk score established to date came from the GHSG. Data from 422 HL patients who relapsed after first-line treatment consisting of chemotherapy and/or RT were included. Time to relapse, clinical stage at relapse and anaemia (25 mg/l correlate with poor EFS compared to levels
Current status of prognostication in classical Hodgkin lymphoma Girish Venkataraman,1 M. Kamran Mirza,1 Dennis A. Eichenauer2,3 and Volker Diehl3 1
Department of Pathology, Section of Hematopathology, The University of Chicago Medicine, Chicago, IL, USA, 2First Department of Internal Medicine, University Hospital Cologne, and 3German Hodgkin Study Group (GHSG), Cologne, Germany
Summary Classical Hodgkin lymphoma (cHL) is characterized by a paucity of neoplastic Hodgkin/Reed Sternberg (HRS) cells within a complex cellular milieu that is rendered immunologically incapable of reacting against CD30+HRS cells due to a plethora of immune escape mechanisms initiated by the neoplastic cells. Accounting for 25% of all lymphomas and nearly 95% of all Hodgkin lymphomas, patients with cHL are typically young adults. Besides traditional prognostic factors, such as the International Prognostic Index (IPI), newer imaging and ancillary biomarkers (CD68, Galectin-1 and plasma microRNA) have shown promise. Furthermore, the evolution of gene expression profiling (GEP) in recent years has enabled the development of several practically feasible GEP-based predictors with prognostic relevance. This review discusses the current status of clinical prognostication in cHL, the critical role of histological evaluation in light of several mimicking entities, and the relevance of tissue as well as serum biomarkers pertaining to immune escape mechanisms and recent GEP studies. Keywords: prognosis, classical Hodgkin lymphoma, immunohistochemistry, genome expression profiling, serum biomarkers, clinical risk factors. Since the seminal description ‘on some morbid experiences of the absorbent glands and spleen’ by Thomas Hodgkin in 1832 (Hodgkin, 1832), we have come a long way in our understanding of the pathobiology of the histologically peculiar and unique lesion identified as Hodgkin lymphoma (HL). The 20th century witnessed refined morphological descriptions and several classification systems for HL. As of today, the World Health Organization (WHO) 2008 classification (Swerdlow et al, 2008) recognizes four histological subtypes of classical HL (cHL): nodular sclerosis (NSCHL), mixed
Correspondence: Girish Venkataraman, Section of Hematopathology, Department of Pathology, The University of Chicago Medicine, TW055B, MC0008, 5841 S. Maryland Avenue, Chicago, IL 60637, USA. E-mail: [email protected]
ª 2014 John Wiley & Sons Ltd British Journal of Haematology, 2014, 165, 287–299
cellularity (MCCHL), the lymphocyte rich (LRCHL) and lymphocyte-depleted (LDCHL). Overall, the fundamental description of HL has largely remained unchanged through the past century. Despite the separation of nodular lymphocytepredominant HL (NLPHL) as a distinct entity in the late 1990s, it is critical to note that both cHL and NLPHL share a common biological origin from B-cells. The two major differences between them are (i) the presence of a relatively intact germinal centre/B-cell programme and (ii) lack of a role for Epstein-Barr virus (EBV) in the pathogenesis of NLPHL. In addition, these entities differ in their clinical features and behavior. The epidemiology of cHL is notable for complex interplay of genetic, racial and geographic as well immune competence/EBV-driven factors, which are instrumental in its pathogenesis (Swerdlow et al, 2008). Clinically, HL patients have been traditionally divided into two or three prognostic groups, mainly according to the stage and presence of systemic symptoms at diagnosis. The definition of these clinical and pathological prognostic groups varies, as does the choice of treatment. Prognostic factors are rarely studied clinically but are usually discovered when analysing data from reliable, large clinical trials. Several recent reviews touch upon the Hodgkin/Reed Sternberg (HRS) cell biology, the role of the cHL microenvironment, the cytokine milieu and potential serum biomarkers in this disease (Kuppers et al, 2002; Skinnider & Mak, 2002; Kuppers, 2009; Steidl et al, 2011). In the following sections, this review focuses solely on clinical and biological markers that have been shown to have prognostic relevance with a focus on risk factors, histological/immunohistochemical findings and the advances made in gene expression profiling (GEP) over the last years. We start our discussion with an overview of established clinical risk factors for HL as well as more recently developed stratification tools, including interim fluorodeoxyglucose positron emission tomography (FDG-PET).
Clinical risk factors and clinical outcome prediction in HL Although several biomarkers have been shown to be predictive for a favourable or unfavourable clinical course in HL, to
First published online 4 February 2014 doi:10.1111/bjh.12759
Review our knowledge, none of them have been incorporated in any clinical risk stratification strategies for HL to date. In newly diagnosed HL, patients are allocated to different risk groups according to clinical features and laboratory parameters present at diagnosis. In relapsed disease, treatment stratification according to the presence or absence of risk factors has not become standard to date, so that patients with recurrent HL are usually treated with high-dose chemotherapy followed by autologous stem cell transplantation (ASCT) irrespective of the clinical stage or the presence of risk factors at time of relapse. However, a multitude of clinical risk scores for both newly diagnosed patients and patients with relapsed HL have been developed within the past 15 years. Furthermore, functional imaging with interim FDG-PET is increasingly recognized as an appropriate tool to distinguish good-risk from poor-risk patients. Thus, the clinical practice regarding the allocation of HL patients to different risk groups may change in the future.
Clinical risk factors in newly diagnosed HL Patients with newly diagnosed HL are usually allocated to different risk groups according to their clinical stage and the presence or absence of certain clinical and biological risk factors. Larger study groups, such as the European Organisation for Research and Treatment of Cancer (EORTC) and the German Hodgkin Study Group (GHSG), distinguish between early favourable, early unfavourable, and advanced HL. Early favourable HL is defined by stage I/II disease without risk factors such as large mediastinal mass, extranodal disease, elevated erythrocyte sedimentation rate (ESR) or involvement of three or more nodal areas. In contrast, patients with stage I/II disease presenting with one or more of these risk factors are usually allocated to the early unfavourable risk group while patients with stage III/IV disease have advanced HL (Table I) (Eichenauer et al, 2011; Hoppe et al, 2012). As the prognosis of patients with advanced HL is worse when compared with patients with early-stage disease, treatment strategies adapted to the patients′ risk profile have been established within the past decades. At present, patients diagnosed with early-stage disease mostly receive combined-modality approaches consisting of two to four cycles of chemotherapy followed by involved-field radiotherapy (IF-RT). ABVD (adriamycin, bleomycin, vinblastine, dacarbazine) represents the chemotherapy protocol most often used. Advanced HL is usually treated with six to eight cycles of chemotherapy optionally followed by localized RT. The question whether ABVD or escalated BEACOPP (bleomycin, etoposide, adriamycin, cyclophosphamide, vincristine, procarbazine, prednisone) should be routinely used in advanced HL is not ultimately answered although BEACOPP-escalated appears to be superior as recently revealed by a large network meta-analysis (Skoetz et al, 2013). Generally, the long-term survival rates in HL patients exceed 80% irrespective of the stage at diagnosis when adequate treatment is applied (Borchmann et al, 2012). 288
Table I. Allocation to treatment groups according the EORTC and GHSG criteria. Risk group
EORTC
GHSG
Early favourable
CS I-II without risk factors (supradiaphragmatic) CS I-II with ≥1 risk factors (supradiaphragmatic)
CS I-II without risk factors
Early unfavourable
Advanced
CS III-IV
Risk factors
(A) Large mediastinal mass (B) Age ≥50 years (C) Elevated ESR (D) ≥4 nodal areas
CS I, CS IIA with ≥1 risk factors; CS IIB with risk factors C/D, but not A/B CS IIB with risk factors A/B, CS III/IV (A) Large mediastinal mass (B) Extranodal disease (C) Elevated ESR (D) ≥3 nodal areas
GHSG, German Hodgkin Study Group, EORTC, European Organisation for Research and Treatment of Cancer, CS, Clinical stage, Elevated erythrocyte sedimentation rate (ESR), >50 mm/h without B symptoms, >30 mm/h with B symptoms, Large mediastinal mass, more than one-third of the maximum horizontal chest diameter, B symptoms, fever, night sweat, weight loss.
In 1998, the International Prognostic Score (IPS) for patients with advanced HL was introduced (Hasenclever & Diehl, 1998). This score is composed of seven adverse factors with similar independent prognostic effects, namely serum albumin of less than 40 g/l, a haemoglobin level of less than 105 g/l, male sex, age of 45 years or older, stage IV disease according to the Ann-Arbor classification, leucocytosis with a white-cell count of at least 15 9 109/l and lymphocytopenia (lymphocyte count of less than 06 9 109/l or less than 8% of the white-cell count or both). Data from 4695 patients aged between 15 and 65 years with advanced HL and mostly treated with anthracycline-containing chemotherapy protocols were considered for the analysis. While the 5-year freedom from treatment failure (FFTF) and overall survival (OS) rates for the whole patient group were 66% and 78%, respectively, FFTF rates according to the IPS ranged between 84% for patients presenting with no risk factor and 42% for patients presenting with all seven risk factors contained in the IPS. The corresponding OS rates ranged between 90% and 56%. Thus, the IPS is a valuable and reliable tool to predict the outcome of patients with newly diagnosed advanced HL. As the patients taken into account for the initial development of the IPS were treated before 1992, it has been unclear whether this tool is still applicable in patients currently undergoing HL treatment. Thus, the British Columbia Cancer Agency (BCCA) recently performed an analysis evaluating the validity of the IPS in 710 patients diagnosed with advanced HL between 1980 and 2010 and treated with ABVD or ABVD-like protocols. Within this study the IPS retained its ª 2014 John Wiley & Sons Ltd British Journal of Haematology, 2014, 165, 287–299
Review prognostic value and still discriminated high-risk from lowrisk patients. In general, the prognosis has apparently improved. Freedom from progression (FFP) estimates at 5 years ranged between 62% and 88%; the corresponding 5-year OS estimates ranged between 67% and 98% (Moccia et al, 2012). Given the relevant proportion of patients with advanced HL treated with escalated BEACOPP due to the improved tumour control associated with this protocol, it should also be mentioned that the IPS also works when this chemotherapy schema is given. In the final analysis of the pivotal HD9 trial by the GHSG comparing escalated BEACOPP with BEACOPP in baseline dosage and COPP (cyclophosphamide, vincristine, procarbazine, prednisone)/ABVD in newly diagnosed advanced HL, patients were divided into three groups according to their IPS (IPS 0-1 vs IPS 2-3 vs IPS 4-7) (Diehl et al, 2003). Among the patients treated with escalated BEACOPP, FFTF and OS rates at 5 years were 92% and 95%, respectively, for patients with an IPS of 0-1, 87% and 90% for patients with an IPS of 2-3 and 82% and 82% for patients with an IPS of 4-7 (Diehl et al, 2003).
Role of functional imaging in newly diagnosed HL Within the past decade, the use of FDG-PET has become increasingly common to distinguish good-risk from poorrisk HL patients. Hutchings et al (2006) reported 77 patients who were diagnosed with HL at different stages and mainly received ABVD chemotherapy. After two cycles, patients underwent interim FDG-PET; 16 patients were FDG-PET-positive while no residual metabolic activity was detected in 61 patients. At a median follow-up of 23 months, 11 of the 16 FDG-PET-positive patients had relapsed. In contrast, only three of the 61 patients who were FDG-PET-negative after two cycles of chemotherapy had disease recurrence. Thus, early interim FDG-PET was shown to be a good predictor for progression-free survival (PFS) in HL patients treated with ABVD-based chemotherapy (Hutchings et al, 2006). Another analysis including a larger cohort of 260 patients with newly diagnosed advanced HL receiving ABVD chemotherapy was performed by a joint group from Denmark and Italy (Gallamini et al, 2007). Early interim FDG-PET was performed after two cycles of chemotherapy. This study also revealed a strong predictive value for early interim FDG-PET in terms of PFS. At a median follow-up of 22 years, the 2year PFS rates were 128% for patients with a positive FDGPET after two cycles of chemotherapy compared to 95% for patients with a negative FDG-PET after the same amount of chemotherapy (Gallamini et al, 2007). Given these data, several prospective trials evaluating early risk-adapted treatment stratification strategies on the basis of interim FDG-PET were initiated worldwide. The majority of these trials are still ongoing and final analyses are pending. Thus, treatment ª 2014 John Wiley & Sons Ltd British Journal of Haematology, 2014, 165, 287–299
decisions based on early interim FDG-PET should not yet be made outside prospective clinical studies. The question of whether consolidating RT can be omitted in patients with a negative FDG-PET after chemotherapy is at least answered for patients receiving intensive chemotherapy with BEACOPP. Within the randomized GHSG HD15 trial for newly diagnosed advanced HL, all patients with residual lymphoma larger than 25 cm after BEACOPPescalated or BEACOPP-14 chemotherapy underwent FDG-PET evaluation (Engert et al, 2012). Patients with a positive FDGPET had localized RT while patients with a negative FDGPET had no additional RT. The negative predictive value for FDG-PET, defined as the proportion of PET-negative patients without progression, relapse or RT at 12 months, was 941% so this modality can safely be used as tool to decide whether patients with larger residual lymphoma after intensive first-line chemotherapy should be irradiated (Engert et al, 2012). The question of whether this is also true for advanced-stage patients treated with ABVD chemotherapy and patients with early-stage HL with a complete metabolic response after chemotherapy is unanswered to date. However, first results on the role of interim FDG-PET in early favourable HL have recently become available. Within the multicentre RAPID trial conducted in the UK, patients with a negative FDG-PET after three cycles of ABVD were randomized between IF-RT and no further treatment. At a median follow-up of 457 months, 3-year PFS rates were 938% in the IF-RT arm and 907% in the observation only arm. The corresponding OS rates were 970% among patients receiving IF-RT and 995% among patients receiving no RT (Radford et al, 2012). Despite of these data, final conclusions should not be drawn before mature data from further studies addressing this issue, such as the GHSG HD16 trial (NCT00736320) and the EORTC H10 (NCT00433433) study, are published.
Clinical risk factors and functional imaging in relapsed or refractory HL Similar to newly diagnosed disease, certain clinical factors as well as functional imaging were shown to be of value for outcome prediction in relapsed HL. The most comprehensive risk score established to date came from the GHSG. Data from 422 HL patients who relapsed after first-line treatment consisting of chemotherapy and/or RT were included. Time to relapse, clinical stage at relapse and anaemia (25 mg/l correlate with poor EFS compared to levels
