Because long-term survival of patients with nonbulky stage IA to IIA Hodgkin lymphoma is dependent on disease control and avoidance of late toxic effects associated with the treatment received, the initial choice of treatment can be associated with trade-offs that balance optimum disease control with avoidance of these late effect risks. Health professionals and patients face the dilemma of making treatment decisions without the benefit of completely understanding the risk-benefit balances associated with how current treatments affect all outcomes of interest. Optimum management of these patients requires careful multidisciplinary evaluation and communication strategies that account for patient preferences.
Key points
- •
More than 80% of patients with stage IA to IIA nonbulky Hodgkin lymphoma are cured with doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) chemotherapy alone.
- •
Death after a diagnosis of Hodgkin lymphoma is more commonly caused by factors unrelated to disease progression and includes treatment-related adverse late effects.
- •
Follow-up into the third decade after treatment is required to properly assess overall survival after radiation therapy.
- •
Refinement of early response assessment using positron emission tomography–computed tomography scans may assist individualizing therapy, including defining those who may benefit most from combined modality therapy.
Introduction
Current controversies in managing patients with Hodgkin lymphoma are a result of uncertainties in balancing the best measures to eradicate the disease while minimizing the risks of long-term adverse effects that are associated with available therapies. For patients with localized Hodgkin lymphoma and favorable-risk features, debates relate to the relative merits of treatment with doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) chemotherapy alone versus combining fewer cycles of this treatment with localized radiation therapy (RT); for those with advanced disease, debates center around use of ABVD versus bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, prednisone, and procarbazine (BEACOPP). Common to both debates is realization that most clinical trials that inform practices do not directly address the morbidities and mortality that become apparent in the second and third decades after treatment and that available data about these long-term outcomes relate to previous treatment strategies that are now outdated.
Reviews describing these debates benefit from placing current perspectives onto a background of previous accomplishments in managing patients with Hodgkin lymphoma. Early in the twentieth century, the diagnosis was fatal to all but a few with accessible, localized tumors, in whom radical surgical excision afforded freedom from disease. Pioneers in RT discovered means to control localized disease, over time observing fewer relapses with increasing radiation fields, leading to adoption of extended-field irradiation encompassing the mantle field (axillary, mediastinal, and axillary nodes), spleen, and para-aortic region. Simultaneously, chemotherapy evolved from single-agent nitrogen mustard to combinations resulting in the possibility of cure, even for those with advanced disease. Cure rates increased with coadministration of radiation and chemotherapy, termed combined modality therapy (CMT). By the turn of the millennium, long-term disease control was observed in 80% to 90% of those with favorable limited-stage Hodgkin lymphoma treated with CMT, and many reports described cure rates of greater than 90%. With long-term follow-up of young survivors, there was increasing recognition of the problem of late treatment-related toxicities, resulting in premature death, despite cure of Hodgkin lymphoma. Over recent decades, strategies attempting to reduce late effects and maintain or improve disease control in newly diagnosed patients have included adoption of ABVD as the standard chemotherapy regimen because it is more effective than historical regimens and is not associated with the risks of leukemogenesis or gonadal toxicity, reduction of radiation from extended to involved fields, abbreviation of the number of cycles of chemotherapy as part of CMT, and omission of RT by using ABVD chemotherapy alone.
The objective of this review is to describe the rationale and context of a treatment decision to use chemotherapy alone as treatment of patients with stage IA and IIA nonbulky Hodgkin lymphoma, building on evidence presented in a recent review. As described in that review, a focus of the current debate about treatment options has been the reporting of the final results of the NCIC Clinical Trials Group (NCIC CTG)–Eastern Cooperative Oncology Group (ECOG) HD.6 trial, in which patients with stage IA and IIA nonbulky Hodgkin lymphoma were randomized to receive 4 to 6 cycles of ABVD alone (the choice of 4 vs 6 cycles was based on disease-control assessment after 2 treatment cycles) or subtotal nodal radiation, given as a single modality to patients younger than 40 years who had no more than 3 nodal sites of disease and an erythrocyte sedimentation rate (ESR) of less than 50, or combined with 2 cycles of ABVD for patients not satisfying these 3 qualifying criteria. The main results of this trial were that chemotherapy alone was associated with superior 12-year overall survival (94% vs 87%; hazard ratio [HR] = 0.50, 95% confidence interval [CI], 0.25–0.99; P = .04) but inferior disease control as assessed by 12-year freedom from disease progression (87% vs 92%; HR = 1.91, 95% CI, 0.99–3.69; P = .05). Superior overall survival with ABVD alone was because fewer deaths were observed attributed to causes other than Hodgkin lymphoma.
Introduction
Current controversies in managing patients with Hodgkin lymphoma are a result of uncertainties in balancing the best measures to eradicate the disease while minimizing the risks of long-term adverse effects that are associated with available therapies. For patients with localized Hodgkin lymphoma and favorable-risk features, debates relate to the relative merits of treatment with doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) chemotherapy alone versus combining fewer cycles of this treatment with localized radiation therapy (RT); for those with advanced disease, debates center around use of ABVD versus bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, prednisone, and procarbazine (BEACOPP). Common to both debates is realization that most clinical trials that inform practices do not directly address the morbidities and mortality that become apparent in the second and third decades after treatment and that available data about these long-term outcomes relate to previous treatment strategies that are now outdated.
Reviews describing these debates benefit from placing current perspectives onto a background of previous accomplishments in managing patients with Hodgkin lymphoma. Early in the twentieth century, the diagnosis was fatal to all but a few with accessible, localized tumors, in whom radical surgical excision afforded freedom from disease. Pioneers in RT discovered means to control localized disease, over time observing fewer relapses with increasing radiation fields, leading to adoption of extended-field irradiation encompassing the mantle field (axillary, mediastinal, and axillary nodes), spleen, and para-aortic region. Simultaneously, chemotherapy evolved from single-agent nitrogen mustard to combinations resulting in the possibility of cure, even for those with advanced disease. Cure rates increased with coadministration of radiation and chemotherapy, termed combined modality therapy (CMT). By the turn of the millennium, long-term disease control was observed in 80% to 90% of those with favorable limited-stage Hodgkin lymphoma treated with CMT, and many reports described cure rates of greater than 90%. With long-term follow-up of young survivors, there was increasing recognition of the problem of late treatment-related toxicities, resulting in premature death, despite cure of Hodgkin lymphoma. Over recent decades, strategies attempting to reduce late effects and maintain or improve disease control in newly diagnosed patients have included adoption of ABVD as the standard chemotherapy regimen because it is more effective than historical regimens and is not associated with the risks of leukemogenesis or gonadal toxicity, reduction of radiation from extended to involved fields, abbreviation of the number of cycles of chemotherapy as part of CMT, and omission of RT by using ABVD chemotherapy alone.
The objective of this review is to describe the rationale and context of a treatment decision to use chemotherapy alone as treatment of patients with stage IA and IIA nonbulky Hodgkin lymphoma, building on evidence presented in a recent review. As described in that review, a focus of the current debate about treatment options has been the reporting of the final results of the NCIC Clinical Trials Group (NCIC CTG)–Eastern Cooperative Oncology Group (ECOG) HD.6 trial, in which patients with stage IA and IIA nonbulky Hodgkin lymphoma were randomized to receive 4 to 6 cycles of ABVD alone (the choice of 4 vs 6 cycles was based on disease-control assessment after 2 treatment cycles) or subtotal nodal radiation, given as a single modality to patients younger than 40 years who had no more than 3 nodal sites of disease and an erythrocyte sedimentation rate (ESR) of less than 50, or combined with 2 cycles of ABVD for patients not satisfying these 3 qualifying criteria. The main results of this trial were that chemotherapy alone was associated with superior 12-year overall survival (94% vs 87%; hazard ratio [HR] = 0.50, 95% confidence interval [CI], 0.25–0.99; P = .04) but inferior disease control as assessed by 12-year freedom from disease progression (87% vs 92%; HR = 1.91, 95% CI, 0.99–3.69; P = .05). Superior overall survival with ABVD alone was because fewer deaths were observed attributed to causes other than Hodgkin lymphoma.
Patient population and current guidelines
Several bodies have used various approaches to consider evidence and reach consensus to produce practice guideline recommendations intended to assist those managing patients with favorable-risk, limited-stage Hodgkin lymphoma ( Table 1 ). Although some details differ, these bodies have generally defined this population as those with stage IA and IIA nonbulky disease. Added factors used by some have included the number of nodal sites or areas, ESR and, less commonly in selected patients, B symptoms. Our discussion is based on a population that includes all those with stage IA and IIA, nonbulky disease excepting those rare patients with these features who also have intra-abdominal disease. We have chosen these parameters because they were used as eligibility criteria in the NCIC CTG-ECOG HD.6 trial.
| Guideline Body | Patient Population | Recommended Therapy |
|---|---|---|
| National Comprehensive Cancer Network, 2013 | Stage IA–IIA favorable disease: no mediastinal bulk, peripheral disease ≤10 cm, ESR ≤50, ≤3 sites of disease | CMT (2–4 ABVD + ISRT) (category 1) or Stanford V × 8 weeks (category 2A) or ABVD alone (category 2A) |
| European Society for Medical Oncology, 2011 | Stage IA–IIA without risk factors: GHSG – large mediastinal mass, extranodal disease, increased ESR, ≥3 nodal areas; EORTC/GELA – large mediastinal mass, age ≥50 y, increased ESR, ≥4 nodal areas | 2 cycles of ABVD + 20-Gy IFRT (grade A recommendation) |
| Italian Society of Haematology, 2009 | Stage I–II favorable disease: EORTC – ≤3 nodal involved areas, age <50 y, and M/T ratio <0.33, and ESR <50 without B symptoms, or ESR <30, with B symptoms | 3 to 4 courses of ABVD + 30-Gy IFRT (grade A recommendation) OR 2 ABVD + 20-Gy IFRT in clinical trial setting (grade B recommendation) |
| American College of Radiology, 2008 | Stage I–II favorable disease: according to GHSG or EORTC criteria above | CMT (2–4 cycles of ABVD + 20-Gy to 30-Gy IFRT) |
Allowing for slight differences in population definition, recommendations from all 4 guideline bodies include use of CMT with 2 to 4 cycles of ABVD and involved-field RT (IFRT) with 20 to 30 Gy (variations within these ranges depend on patient specifics). In addition, one body, the National Comprehensive Cancer Network (NCCN), also includes 4 to 6 cycles of ABVD alone as an acceptable alternative, but assigns this recommendation a lower level of evidence (level 2A). Of these 4 guidelines, only the NCCN guidelines included consideration of the final analysis of the HD.6 trial, which was published in 2012 and reported 12-year outcomes. The other 3 bodies published their recommendations before 2012 and considered the initial HD.6 results, which were published in 2005 and described 5-year outcomes ; compared with the final reporting of that trial, these earlier results reported similar superiority of radiation-based therapy with respect to disease control with no differences in overall survival detected. Among guideline developers who do not recommend ABVD alone as a treatment option, reasons cited include the preliminary nature of the report describing 5-year outcomes, inferior disease-control outcomes associated with this therapy, the outdated nature of subtotal nodal RT received by patients in the HD.6 control arm, and the expectation that current use of IFRT is associated with a reduced risk of adverse late treatment effects.
Components of the decision-making process
Outcome Measures and Surrogates
Although overall survival coupled with health-related quality of life are likely the outcome measures of most importance to patients, clinical trials testing treatment strategies for populations with stage IA and IIA nonbulky Hodgkin lymphoma usually prioritize end points that measure disease control. The reasons for using these end points include the historical observation that disease control was a reasonable surrogate for overall survival and that earlier availability of robust disease-control data facilitated timely reporting. However, recent data show that progressive disease accounts for few deaths in patients with stage IA and IIA nonbulky Hodgkin lymphoma, even when median follow-ups are less than a decade (and thus do yet account for the additional late effects mortality that is observed in the second and third decades after treatment). Therefore, although common, use of disease-control surrogate outcomes obtained with less than a decade of follow-up is associated with substantial risks if intended to be a proxy for long-term survival.
Compounding this generic limitation is the need for clarity around the problems associated with specific disease-control end points. For instance, a recommended end point is time to progression (TTP). In calculating TTP, events contributing to this composite end point include the first of progressive Hodgkin lymphoma or death from Hodgkin lymphoma. The end point thus fails to capture deaths attributed to causes other than progressive Hodgkin lymphoma. Furthermore, given that stage IA and IIA nonbulky Hodgkin lymphoma is not associated with a sufficient disease burden to directly affect survival unless preceded by disease progression, all events associated with this end point are caused by disease progression. Because most deaths in this population are unassociated with disease progression, TTP is not, nor can it be expected to be, an accurate surrogate for overall survival.
Another surrogate end point is progression-free survival (PFS), which is also a composite, in that events include the first of progressive Hodgkin lymphoma or death from any cause. This end point thus has the potential to account for deaths attributed to late treatment effects provided that the reporting of the clinical trial includes a duration of follow-up that is sufficient to span the period when late effects occur, and that the number (and thus relative proportion) of types of events (ie, progressive disease vs death from any cause) that comprise this composite are provided. Without such clarity, PFS has also been shown to be an unreliable proxy for long-term overall survival.
Practitioners and policy makers need to recognize the important differences that exist between debates around use of disease control end points as surrogates for overall survival in trials testing therapies for patients with metastatic carcinoma versus those evaluating patients with stage IA and IIA nonbulky Hodgkin lymphoma. In the former, any lack of surrogacy likely emphasizes the extreme precision of disease progression measurement associated with principles of conducting explanatory clinical trials, especially within the context of limited survival durations; this limitation may be exaggerated when subsequent lines of therapy provide transient benefits. In contrast, for patients with stage IA to IIA nonbulky Hodgkin lymphoma, lack of surrogacy relates to deaths from causes other than progressive disease—an observation that is further compounded by knowledge that some of these deaths are related to the therapy received. Thus, the outcomes of interest that are associated with treatment policy determinations for these patients are associated with a unique paradigm that is distinct from considerations used for most other cancers: long-term overall survival is the outcome of greatest importance, data for this end point are largely unavailable for purported new therapeutic advances, and reliable proxy outcome measures that might facilitate estimates of the relative long-term survivals associated with competing options do not exist. Contributing to current dilemmas about outcome measures is the lack of published quality-of-life data from trials that compare approaches to management using CMT and ABVD alone. Evaluations of this end point, both during and after treatment, may assist in determining patient preferences.
Therapy that Includes Radiation
The background and rationale for choosing CMT that includes modern radiation treatment practices is detailed in the article by Marc André elsewhere in this issue. Recognizing that the options of CMT and ABVD alone are associated with relative risks and benefits, and trade-offs thus exist, we recognize several parameters that require consideration when choosing an overall treatment policy and implementing this policy in individual patients.
First, adverse late effects associated with wide-field radiation strategies, such as subtotal nodal and extended-field RT (EFRT), are well documented. Examples include a population-based cancer registry study that includes 32,591 patients and identified significantly increased risks for all solid tumors; whereas chemotherapy alone resulted in an observed/expected ratio of 1.7, the ratio for patients receiving RT was 4.4 when follow-up was 20 to 24 years after diagnosis. Associations were observed between RT and risks for esophageal, gastric, rectal, female breast, bladder, thyroid, and bone/connective tissue malignancies. That these risks exist with outdated RT strategies is well accepted and includes recognition that risks increase through at least the second and third decades of follow-up. The outcomes of patients in the HD.6 control arm, who received subtotal nodal RT, are consistent with these data. With a median follow-up of 4.2 years, there were 10 second cancers and 12 cardiac events in the 203 patients allocated to receive RT-containing therapy, whereas with a median follow-up of 11.3 years, there were 23 and 26 events observed, respectively.
Second, we recognize that when compared with outdated RT, modern RT includes reduced doses and smaller target volumes and is associated with reduced risks of late effects. However, the magnitude of reduction remains uncertain and likely remains important. For instance, observations from the German Hodgkin Study Group (GHSG) HD8 trial, which randomly allocated patients to receive CMT that included EFRT or IFRT observed that with a median follow-up of 4.6 years, 24 of 532 (4.5%) patients assigned to EFRT and 15 of 532 (2.8%) assigned to IFRT had developed secondary malignancies ; with a median follow-up of 11 years, the incidence of second malignancies had increased to 58 (10.9%) and 45 (8.5%) events, respectively. Although cardiac events were not tracked, the incidence of deaths attributed to cardiac causes showed a similar trend: at a median of 4.6 years, there were 5 and 7 cardiac deaths in the EFRT and IFRT groups, respectively, whereas with a median follow-up of 11 years, there were 15 and 12 such deaths. Thus, although risks are reduced with IFRT compared with EFRT, important risks of late effects seem to remain with IFRT, especially considering that the data on which this concern is based do not yet include follow-ups well into the second decade after treatment.
Third, we have previously reported our speculation that risks of late effects first become apparent in older patients. This premise is based on the apparent increased risks of both death and nonfatal second cancers and cardiac events observed in patients allocated to receive RT in the HD.6 trial who were assigned to the unfavorable-risk cohort, a determination that included older age as a criterion, and the observation that excess absolute risks of second cancers increase with attained age after treatment of Hodgkin lymphoma. This premise is further suggested by outcomes observed in an individual patient cross-trial comparison of patients assigned to ABVD alone in HD.6 with patients assigned to the preferred arms of the GHSG HD10/11 trials; more deaths without previous progression of Hodgkin lymphoma were observed in older patients. If true, our supposition means that patients younger than 40 years at diagnosis require follow-up well into the second, if not third, decade after therapy before risks of adverse late effects can be understood.
Excellent outcomes have been achieved with CMT strategies tested in the GHSG HD10 trial (8-year freedom from treatment failure of 86% with 2 cycles of ABVD and 20-Gy IFRT) and the European Organization for Research and Treatment of Cancer (EORTC) H8 F trial (10-year event-free survival of 93% with 3 cycles of mechlorethamine, vincristine, procarbazine, and prednisone (MOPP) doxorubicin, bleomycin, vinblastine (ABV) and 36-Gy to 40-Gy IFRT). These results establish a baseline for comparing new strategies that include involved nodal RT (INRT) and involved site RT (ISRT) radiation. However, although results from these 2 trials inform current recommendations, neither report includes median follow-up into the beginning of the second decade, let alone into the third decade after treatment, and median ages were 38.7 and 30 years, respectively, meaning that extended follow-up is needed to better understand the risks of late effects in younger patients. The relative merits of CMT strategies incorporating INRT and ISRT require similar scrutiny, with long-term follow-up of comparative data evaluating disease control and occurrence of late effects; given the time frames required, reports with such details are more than a decade away.
Therapy with Chemotherapy Alone
Systematic evaluations of a strategy to treat patients with stage IA and IIA nonbulky Hodgkin lymphoma with chemotherapy alone should consider only trials that test optimum chemotherapy (ie, ABVD). Thus, previous trials evaluating regimens such as MOPP (or its equivalents), epirubicin, vinblastine, etoposide, and epirubicin, bleomycin, vinblastine, and prednisone do not provide data that contribute to current decision making. Similarly, a 2011 Cochrane review does not assist policy determination, because this analysis included 3 such trials; the other 2 trials included did test ABVD but evaluated populations that included patients with stage III and bulky disease.
We have previously summarized the specifics and context of randomized trials comparing chemotherapy alone with strategies that include RT for patients with limited-stage Hodgkin lymphoma. One of these reports described results of a MEDLINE search using PubMed from January 2002 to February 2007 and a review of abstracts from the American Society of Hematology (ASH) and American Society of Clinical Oncology (ASCO) annual meetings from 2002 to 2006. This search identified 3 randomized controlled trials (RCTs) testing ABVD alone in previously untreated adult patients with limited-stage Hodgkin lymphoma. One of these trials was the initial report of HD.6, one was a single institution comparison of ABVD with CMT in patients with stage I to IIA or B and stage IIIA disease, and one was a subset analysis of a larger trial evaluating patients with all stages of Hodgkin lymphoma. None of the trials detected differences in overall survival, and only the HD.6 trial had sufficient power to detect differences in disease control, which favored the group receiving RT. The premise that ABVD alone is associated with disease control that is approximately 7% worse than CMT approaches comes from these data.
We repeated this PubMed search in May 2013 using the MESH term Hodgkin and the limiting terms “randomized controlled trials” and “meta-analysis” to identify publications since February 2007; 231 citations were identified and reviewed. After excluding articles that were not phase III RCTs or meta-analyses testing ABVD-based chemotherapy alone as initial therapy for patients with nonbulky stage IA and IIA Hodgkin lymphoma, 3 publications remained. These publications included the final analysis of HD.6 described earlier, the Cochrane meta-analysis referred to above, and a pediatric study testing cyclophosphamide, vincristine, prednisone, and procarbazine/ABV therapy alone. Furthermore, we reviewed ASH abstracts from 2007 to 2012 and identified 2 additional RCTs: the UK National Cancer Research Institute RAPID (PET Scan in Planning Treatment in Patients Undergoing Combination Chemotherapy For Stage IA or Stage IIA Hodgkin Lymphoma) and EORTC-led H10 trials, which both tested response-adapted strategies using positron emission tomography (PET).
We have previously provided our interpretation of the important conclusions associated with HD.6 data. First, superior 12-year overall survival (94% vs 87%) was observed with ABVD alone, and this difference was caused by more deaths in the RT arm from causes other than progressive Hodgkin lymphoma or early treatment complication; it is recognized that these results are contributed to by use of an outdated RT strategy. Second, the 12-year outcomes of patients assigned to ABVD alone, overall survival of 94%, and freedom from disease progression of 87%, are considered to be excellent and comparable with those reported in randomized trials testing CMT that includes IFRT. Third, the observation that the risk of disease progression with chemotherapy alone was almost twice that associated with treatment that includes RT (HR = 1.91), whereas the risk of death was reduced by half (HR = 0.5), refutes use of disease-control measures as an accurate proxy for overall survival in this patient population. Among patients in the ABVD-alone cohort who were evaluable for response after 2 treatment cycles (90% of patients), those with a complete remission (CR) or unconfirmed complete remission (CRu) had superior 12-year disease control (94% vs 81%; HR = 0.28; P = .02) and a trend to superior 12-year overall survival (98% vs 92%; HR = 0.17; P = .06) when compared with those not achieving a CR/CRu status; these data provide a basis of support to test response-adapted therapy using PET.
The UK RAPID and EORTC-led H10 trials have been reported in preliminary abstract form and provide insights about PET-directed response-adapted therapy and comparisons of chemotherapy alone with CMT. Using a noninferiority design, the RAPID trial enrolled 602 patients with stage IA and IIA nonbulky Hodgkin lymphoma between 2003 and 2010. All patients received 3 cycles of ABVD followed by centrally reviewed PET. Those with a scan deemed to be consistent with potential persisting disease received a fourth cycle of ABVD and IFRT. Those with a negative scan were randomized to receive IFRT and no further treatment. Of the 426 patients with a negative scan (75% of those accrued), 420 were randomized and included in an intent-to-treat analysis that reported 3-year PFS of 90.7% among those receiving ABVD alone and 93.8% among those randomized to receive IFRT. This difference of –2.9% was associated with 95% CIs of –10.7% and +1.4% and thus exceeded the prespecified noninferiority boundary of –7%. The 3-year overall survivals were 99.5% among those allocated to ABVD alone and 97% among those allocated to IFRT. The investigators concluded that patients with a negative PET scan after 3 cycles of ABVD do not require IFRT, chemotherapy alone reduces treatment time and costs, improves tolerability, and removes the burden of early and late radiation-associated toxicities. Given the noninferiority design, subsequent reporting also include the results of a per-protocol analysis.
The H10 trial began accruing patients ages 15 to 70 years with newly diagnosed supradiaphragmatic stage I to II classic Hodgkin lymphoma in 2006; our synopsis is confined to the favorable-risk population who were identified as having fewer than 4 nodal areas of disease with no areas of bulky disease, ages younger than 50 years, and ESR less than 30 with B symptoms or less than 50 without B symptoms (the H10F trial). All patients underwent PET after 2 cycles of ABVD; patients randomized to a control arm received 3 cycles of ABVD and 30-Gy INRT without consideration of PET results, whereas those in the experimental arm received PET-directed therapy. Patients in the experimental arm with negative PET scan received 2 additional cycles of ABVD (total of 4 cycles) and those with a positive scan received 2 cycles of escalated-dose BEACOPP and 30-Gy INRT. Of 444 patients, 381 (86%) were deemed to be PET negative; the 1-year PFS outcomes were 94.9% with ABVD alone and 100% with CMT (HR 9.36, 95% CI 2.45, 35.73). These findings led an independent Data Safety Monitoring Committee to recommend that the trial be halted, including with an explanation that the a priori null hypothesis for noninferiority would not be rejected.
Although determination of practice policies requires results from properly designed and executed RCTs, nonrandomized comparisons may contribute additional context and allow for hypothesis generation. The NCIC CTG and GHSG have thus conducted an individual patient cross-trial comparison evaluating 588 patients who would have been mutually eligible for the HD.6 trial and either the GHSG HD10 or HD11 trial ( Fig. 1 , Table 2 ). The 8-year TTP was superior and PFS trended toward superiority in the GHSG cohort, who were assigned to CMT. No differences in 8-year overall survival were apparent (95% in both cohorts). With a median follow-up of 11.2 years, 10 deaths were observed in 182 patients in the HD.6 cohort, 5 from Hodgkin lymphoma or immediate treatment toxicity and 5 from other causes; with a median follow-up of 7.6 years, 19 deaths were observed in 406 patients in the GHSG cohort, 7 from Hodgkin lymphoma or immediate treatment toxicity and 12 from other causes. All TTP events were caused by disease progression. Of the 27 PFS events in HD.6, 23 were caused by disease progression and 4 were deaths without previous disease progression; of the 38 PFS events in the GHSG cohorts, 25 were caused by disease progression and 13 were deaths without previous disease progression. A subset analysis of those eligible for HD10 suggested that disease-control benefits associated with CMT may be limited to those not achieving CR/CRu, as assessed by physical examination and computed tomographic scanning. Among those assessed as achieving a CR/CRu after 2 cycles of ABVD, the respective 8-year TTP and overall survivals were 95% and 100% in the HD.6 cohort and 93% and 96% in the HD10 cohort. In contrast, among those assessed as not achieving a CR/CRu after 2 cycles of ABVD, the respective 8-year TTP and overall survivals were 78% and 91% in the HD.6 cohort and 92% and 95% in the HD10 cohort.
