Canadian evidence-based guideline for the first-line treatment of chronic lymphocytic leukemia

Chronic lymphocytic leukemia (cll) is the most common adult leukemia in North America. In Canada, no unified national guideline exists for the front-line treatment of cll; provincial guidelines vary and are largely based on funding. A group of clinical experts from across Canada developed a national evidence-based treatment guideline to provide health care professionals with clear guidance on the first-line management of cll. Consensus recommendations based on available evidence are presented for the first-line treatment of cll.


INTRODUCTION
Chronic lymphocytic leukemia (cll) is a clinically and biologically heterogeneous disease, and the most common adult leukemia [1][2][3][4] .According to 2016 statistics, the annual incidence of cll in Canada is about 2400 5 .
Guidelines developed by the International Working Group on CLL provide concise standardized criteria for the diagnosis of cll that include a clonal B lymphocytosis in the peripheral blood (≥5.0×10 9 /L) with a characteristic morphology and immunophenotype 6 .In most cases, examination of the bone marrow is not required for diagnosis.Heterogeneity in the clinical course of cll is attributable mainly to variations in the biology of the disease and, particularly, genetic lesions that correlate with response to therapy, the most relevant prognostic factor for overall survival (os) [7][8][9] .Two widely accepted clinical staging methods-the Binet and Rai systems-are the simplest and best-validated means for identifying patients who require treatment and for predicting survival [10][11][12] .Clinical staging relies solely on physical examination and standard laboratory tests, and does not require computed tomography imaging.Furthermore, with limited value in predicting patient outcome at diagnosis, computed tomography is not recommended outside of clinical trials 13 .
Recent advances in treatment since about 2008 have significantly improved outcomes in cll; however, the disease is still considered incurable except in rare cases of allogeneic hematopoietic stem-cell transplantation (allo-hsct) 4,14 .Consequently, the goal of treatment is to achieve effective and durable disease control [measured as progression-free survival (pfs) and os], with minimal toxicity and acceptable quality of life [14][15][16] .With the availability of several new therapeutic options, treatment decisions based on individual and disease characteristics are paramount in achieving the best outcomes for patients.
Several international guidelines for cll have been published 6,14,[17][18][19][20] ; however, no unified national cll guidelines have been developed in Canada.Although individual provinces have created guidelines, those guidelines differ in their recommendations and are based primarily on the availability of therapeutic options in the provincial formulary 16,[21][22][23] .Accordingly, an evidence-based national treatment guideline that is supported by Canadian hematologists is needed to ensure that all patients with cll in Canada have access to the best available care.In association with Lymphoma Canada, a group of Canadian cll experts therefore developed a national evidence-based consensus guideline for the first-line management of patients with cll.

METHODS
An initial literature search, plus two updates, queried 3 databases (medline, PubMed, and the Cochrane Database of Systematic Reviews) to identify meta-analyses, randomized controlled trials (rcts), and single-arm prospective studies published between January 2000 and July 2017 that investigated first-line treatment of cll.Key search terms for each question were combined with the Medical Subject Heading term "leukemia, lymphocytic, chronic, B-cell."In addition to those searches, abstracts from the proceedings of selected conferences (American Society of Hematology, European Hematology Association, American Society of Clinical Oncology) held between January 2015 and July 2017 were hand-searched.The ClinicalTrials.gov and Cochrane Central Register of Controlled Trials Web sites were also searched for trials in progress.Language of publication was restricted to English.Detailed screening of the full-text versions of all studies was performed to identify the final list of studies.Study selection was limited to those that met these criteria: confirmed diagnosis of cll; adult study population (≥18 years); prospective design; rct, comparative, or single-arm trial with 20 or more study participants; evaluation of first-line treatment for cll; and inclusion of survival outcomes (pfs, os).When rct data relating to a particular question were available, only the rcts were included.When few rcts relating to a particular question were identified, prospective single-arm studies were considered.National Comprehensive Cancer Network categories of evidence and consensus were used to grade the level of evidence supporting recommendations 24 .Details of those categories are presented in Table i.

GUIDELINE Question 1
What prognostic investigations should be performed in patients with previously untreated cll?

Background
The clinical staging systems described by Rai and Binet more than 40 years ago have proved useful as prognostic tools; however, they are not able to determine an individual patient's ongoing clinical course, particularly in the early stages.Prognostic biomarkers provide information about a patient's overall cancer outcome regardless of therapy.
Since about 2004, significant progress has been made in identifying host-and tumour-related prognostic biomarkers, including serum markers, cytogenetic abnormalities, and gene mutations-although relatively few have been prospectively validated within clinical trials 25 .The ability to predict the outcomes of newly diagnosed patients with cll has remarkably improved, but ideally, the hematology community would like to have predictive biomarkers that can help to determine which therapy will work best for a given patient.To date, however, no predictive biomarkers for cll have been validated in prospective clinical trials.

Summary of Evidence
Multivariable analyses of known prognostic biomarkers influencing pfs or os were reported in eight rcts and two meta-analyses of rcts evaluating first-line treatment of cll (Table ii).In the rcts, del(17p) or TP53 mutation (or both), del(11q), unmutated IGHV (ighv-u), β 2 -microglobulin (β2M) concentrations of 3.5 mg/L or greater, and serum thymidine kinase concentrations of 10 U/L or greater were most commonly reported as negative prognostic biomarkers for pfs.Only del(17p) or TP53 mutation (or both), ighv-u, β2M greater than 3.5 mg/L, and thymidine kinase greater than 10 U/L were independently predictive of reduced os.Either or both of TP53 mutation and del(17p) were similarly predictive of very poor pfs and os after chemotherapy or chemoimmunotherapy with purine analogs or alkylating agents 30,32,[35][36][37] .In the cll8 trial from the German CLL Study Group (gcllsg), pfs was shorter for patients with del(11q).However, in that subgroup, the 5-year os with fcr (fludarabine-cyclophosphamide-rituximab) therapy was significantly superior to that with fc (fludarabinecyclophosphamide), suggesting that, despite the shorter duration of remission conferred by del(11q), these patients respond well to first-line fcr therapy 31 .
The correlation of IGHV mutation status with response to first-line chemoimmunotherapy was evaluated in three rcts 30,35,37 .All studies reported poorer outcomes, in terms of pfs, for patients with ighv-u.In the gcllsg cll8 study, os values were not reported for the two subgroups, but Kaplan-Meier estimates suggest that os is significantly shorter in patients with ighv-u 30 .Longer follow-up in those studies and additional investigation of IGHV mutation status in randomized trials are required to determine how this prognostic biomarker should inform decisions about first-line treatment.The influence of β2M and thymidine kinase on response to treatment has not been prospectively evaluated in randomized studies to date and remains to be defined in the setting of current first-line treatments.
To develop an integrated prognostic index, the gcllsg analyzed data from three large phase iii trials that collectively included 1948 patients 38 ; however, of the three trial cohorts analyzed, none included patients treated with chemoimmunotherapy, limiting the adoption of the gcllsg score in the current era of first-line cll treatment.More recently, the cll-ipi (International Prognostic Index) Working Group used pooled data from 3472 patients participating in eight phase iii trials (including the cll8 trial cohort treated with fcr) to develop an integrated prognostic score for patients with cll, identifying 3 biomarkers independently associated with shorter os: β2M concentration greater than  3.5 mg/L, ighv-u, and P53 gene aberrations [del(17p), TP53 mutation, or both] 39 .Four risk categories with different os rates were identified, providing additional prognostic information about os beyond conventional clinical staging.The cll-ipi has been validated in unselected patient cohorts and in patients enrolled in the gcllsg cll11 randomized trial that evaluated first-line treatment of older patients with comorbidities [40][41][42] .One limitation of that study is that, at the time of the analysis, rcts of novel targeted therapies did not have sufficiently long follow-up to be included.

Recommendations
n Testing for prognostic markers should be performed when therapy is required, but evidence is insufficient to recommend routine prognostic marker testing at diagnosis in asymptomatic patients with early-stage cll.
The decision to initiate therapy should be made independently of prognostic marker results, even in the setting of high-risk disease (level of evidence: category 2A).n Patients with TP53 abnormalities have a particularly poor prognosis, including significantly reduced pfs and os after standard chemoimmunotherapy, and might benefit from treatment with a novel targeted therapy.The expert panel strongly recommends testing for del(17p) and TP53 mutation before initiation of first-line treatment (level of evidence: category 2A).n Because the cll-ipi provides valuable prognostic information, the expert panel recommends testing for IGVH mutation status and β2M concentration (in addition to del(17p) and TP53 mutation) before initiation of firstline therapy (level of evidence: category 2B).

Question 2
What criteria should be used to assess fitness in patients with cll?

Background
The advent of newer therapies has led to a greater focus on evaluating the fitness status of patients with cll.As treatment intensity increases, reliable methods are needed to identify patients who can safely tolerate and benefit from such therapy.Traditionally, fitness was classified based on age alone; it is now well recognized that chronologic age is not a reliable surrogate for physiologic age or fitness 43,44 .Clinical trials performed in the cll population are difficult to compare, because the indices used to assess fitness are not standardized, leading to heterogeneity in trial populations.In routine clinical practice, clinical judgment remains the standard of care.
The gcllsg has used a combination of the Cumulative Illness Rating Scale (cirs) and creatinine clearance to define fitness status with respect to tolerability of fcr chemoimmunotherapy 45 .Although the cirs score establishes a clinically useful division, it has not been externally validated or universally adopted outside of clinical trials.

Summary of Evidence
Few randomized prospective studies of first-line treatment for cll have evaluated the effect of patient fitness on outcomes.Prospective analyses of defined fitness factors in treatment-naïve patients with cll receiving current chemoimmunotherapy regimens is limited to subgroup analyses in two randomized studies and four single-arm prospective studies (supplemental Table 1) 31,35,[46][47][48][49][50][51] .Two meta-analyses investigating the effect of comorbidity or age on response to first-line treatments were also included (supplemental Table 1).Those studies provide insight into some of the fitness parameters that might influence patient response and tolerability to therapy; however, the evidence is currently insufficient to define an optimal method to assess fitness for patients with cll or to indicate that a fitness score is superior to clinical judgment.

Recommendations
n Patient fitness should be considered when choosing therapy for cll patients (level of evidence: category 2B).n No specific fitness assessment tool has been proved optimal for decision-making about cll treatment, but the assessment should focus on organ impairment, particularly renal function (level of evidence: category 3).

Question 3
How should asymptomatic early-stage cll be managed?

Background
Today, almost 80% of cll patients are diagnosed at an early clinical stage 52 .Considerable interest is therefore invested in determining optimal timing of treatment initiation to achieve the best outcomes for this patient cohort.All current international guidelines recommend initiation of treatment in patients with advanced (Binet C, Rai iii-iv) or active symptomatic disease; however, they recommend that newly diagnosed patients with asymptomatic early-stage disease (Binet A-B, Rai 0-ii) be monitored without therapy unless they have evidence of disease progression 4,14,[17][18][19][20] .

Summary of Evidence
No published rcts evaluating early first-line treatment of Binet A-B or Rai 0-ii asymptomatic patients were identified after the year 2000.Studies from the French Cooperative Group on CLL and the Cancer and Leukemia Group B evaluated outcomes of early treatment with chlorambucil in patients with early-stage disease (Table iii).Although disease progression and appearance of symptoms could be delayed with early treatment, the use of alkylating agents did not prolong survival.That result was confirmed by a meta-analysis (Table iii) 55 .Furthermore, an increased frequency of fatal epithelial cancers in treated compared with untreated patients was reported 54 .
Table iii also presents results from one abstract examining early treatment with chemoimmunotherapy 56 .The randomized German-French cooperative phase iii trial analyzed the efficacy of early compared with deferred fcr therapy in treatment-naïve patients with Binet A cll having a high risk of disease progression.Patients with high-risk cll were randomized to receive fcr or to be followed in a "watch and wait" strategy.Patients with low-risk cll were observed only.After a median follow-up of 46 months, event-free survival was significantly improved in patients receiving fcr compared with patients being managed as "watch and wait" (median: not reached vs. 24.5 months; p < 0.0001).However, os was not significantly different in the fcr and "watch and wait" strategy groups, with 181 highrisk patients (90%) being alive at last follow-up.
Currently, a multicentre double-blind placebocontrolled phase iii study, gcllsg cll12 (see NCT02863718 at http://ClinicalTrials.gov/), is underway to compare the efficacy and safety of ibrutinib with a "watch and wait" approach in Binet A cll with risk of disease progression defined by the comprehensive cll score 57 .Long-term follow-up of those high-risk patients could provide additional insight about outcomes of early treatment in this patient cohort, but the potential benefit of early intervention with cll drug therapy remains to be proven.

Question 4
How should advanced symptomatic cll be managed?

Background
Monotherapy with the alkylating agent chlorambucil (Clb) was the standard-of-care therapy for cll for several decades 55 .All published international guidelines currently recommend the addition of an anti-CD20 antibody to chemotherapy for the first-line treatment of cll in most patients requiring therapy 6,14,[17][18][19][20] .The chemotherapy agents recommended depend on factors such as patient age, functional status, presence of comorbidities, and organ function.For patients with del(17p) or TP53 mutation, recently updated guidelines recommend treatment with the kinase inhibitor ibrutinib 14,[17][18][19] .Evidence of progressive marrow failure as manifested by the development or worsening of anemia or thrombocytopenia, or both 2. Massive (that is, at least 6 cm below the left costal margin) or progressive or symptomatic splenomegaly

3.
Massive nodes (that is, at least 10 cm in longest diameter) or progressive or symptomatic lymphadenopathy 4.
Progressive lymphocytosis, with an increase of more than 50% over a 2-month period or a lymphocyte doubling time (LDT) of less than 6 months The LDT can be obtained by linear regression extrapolation of absolute lymphocyte counts obtained at intervals of 2 weeks over an observation period of 2-3 months.In patients with initial blood lymphocyte counts of less than 30×10 9 /L (30,000/μL), the LDT should not be used as a single parameter to define a treatment indication.In addition, factors contributing to lymphocytosis or lymphadenopathy other than chronic lymphocytic leukemia (for example, infections) should be excluded.

5.
Autoimmune anemia or thrombocytopenia (or both) that is poorly responsive to corticosteroids or other standard therapy (see section 10.2 of the guideline)

6.
Constitutional symptoms, defined as any one or more of the following disease-related symptoms or signs: established rituximab-fc (compared with fc chemotherapy alone) as the standard of care.Subgroup analysis of prognostic factors showed that the positive effect of fcr was consistent in most prognostic groups and that the benefit of fcr was most pronounced in patients with mutated IGHV 30 .However, fcr did not improve the survival of patients with del(17p) or TP53 mutation [30][31][32] .Several phase ii studies have been initiated with the intent of improving the fcr regimen (supplemental Table 3); to date, however, few rcts determining the efficacy of those treatments in comparison with fcr have been reported.Two studies investigated the addition of alemtuzumab to fc, observing greater toxicity related to infections (Table v) 34,58 .Preliminary results from a randomized phase ii study (Cancer and Leukemia Group B 10404) evaluating fr (fludarabine-rituximab), fr followed by 6 months of lenalidomide consolidation, and fcr in previously untreated patients with cll have recently been reported (Table v) and demonstrated shorter pfs with fr than with fcr 66 .
Bendamustine regimens have also been investigated as first-line therapy in prospective trials (Table v and supplemental Table 3) 35,51,59,60 .In the international randomized phase iii noninferiority study cll10, the gcllsg evaluated the efficacy and tolerability of br (bendamustine-rituximab) compared with fcr for the first-line treatment of fit patients with cll without del(17p) (Table v) 35,59,60 .Median pfs was significantly longer in the fcr arm.Physically fit subgroups derived the most benefit from fcr therapy, but the difference in pfs between treatment groups was nonsignificant for patients more than 65 years of age and for those with a cirs score of 4-6 or the presence of more than 1 cirs item.After 5 years, no difference in os was observed between the treatment arms; however, during treatment, infections were more frequent with fcr, especially in patients 65 years of age and older.

Less-Fit Patients (Without del(17p) or TP53 Mutation):
The gcllsg cll11 trial investigated Clb in combination with anti-CD20 antibodies in previously untreated patients with cll and comorbidities, demonstrating prolonged pfs and os with the addition of anti-CD20 therapy (Table v) 61,62 .Compared with ClbR (Clb-rituximab) treatment, treatment with Clb-obinutuzumab resulted in longer pfs and higher rates of complete response.
The international complement 1 study demonstrated similarly improved pfs with a combination of the anti-CD20 antibody ofatumumab and Clb compared with Clb alone; however, at the time of publication, no difference in os had been reported 64 .
The randomized phase iii mable study evaluated the efficacy and safety of br compared with ClbR in an older less-fit cll population (Table v) 65 .In previously untreated patients, pfs was longer with br than with ClbR.The magnitude of the benefit (10 months) was relatively modest; however, the Clb dose was considerably higher than in the cll11 trial.Grade 3 adverse events were more common with br than with ClbR, driven by a slightly higher rate of infection.
In the phase iii randomized resonate-2 trial (pcyc-1115), ibrutinib was compared with Clb monotherapy in previously untreated patients with cll for whom fludarabinebased therapy was considered inappropriate 63 .Compared with Clb, ibrutinib was associated with longer pfs (median: not reached vs. 18.9 months), significantly prolonged os, and an 84% reduction in the risk of disease progression or death.The study has been criticized for its use of Clb monotherapy as a comparator because Clb was not a standard-of-care treatment option at the time of the study.
Patients with del(17p) or TP53 Mutation, or Both: Patients who have del(17p) or TP53 mutation often respond poorly to standard chemotherapy regimens, including fcr [30][31][32] .Alemtuzumab, in combination with other agents, has been studied in prospective trials in this high-risk population (Table vi) 67,68 .The results from those studies suggest that treatment regimens containing alemtuzumab (compared with standard chemoimmunotherapy) might confer a modest improvement in responses for cll with del(17p) or TP53 mutation; however, confirmatory phase iii studies are required to assess the potential benefit of those therapies.
Two prospective phase ii trials have reported results for single-agent ibrutinib in previously untreated patients with high-risk cll (Table vi) 63,70,71 .In one study, 51 patients (35 untreated) with del(17p) or TP53 mutation were treated with ibrutinib, achieving impressive pfs and os results 70 .Although the experience with ibrutinib as first-line treatment for patients with cll and a del(17p) or TP53 mutation is still limited, current data suggest that this agent might provide durable disease control in treatment-naïve patients with cll having del(17p) or TP53 mutation.

Recommendations
n For fit younger patients without del(17p) or TP53 mutation, we recommend fcr as the preferred first-line treatment (level of evidence: category 1).n For fit elderly patients (more than 65 years of age) without del(17p) or TP53 mutation, br is a reasonable treatment option and could be used in preference to fcr because of lesser toxicity (level of evidence: category 2A).n For less-fit patients, for whom fludarabine therapy is considered inappropriate, and who do not have del(17p) or TP53 mutation, treatment with Clb-obinutuzumab or with ibrutinib monotherapy is recommended.In the absence of a prospective rct comparing ibrutinib therapy with Clb-obinutuzumab (a current standard chemoimmunotherapeutic option in this population in Canada), it is not possible to determine which regimen is optimal in terms of long-term survival and toxicity (level of evidence: category 1).n Patients with del(17p) or TP53 mutation should be offered ibrutinib as first-line treatment because of demonstrated high response rates and potentially long-lasting remissions in this high-risk population (level of evidence: category 2A).

Question 5
In which patients should additional treatment be considered after a response to first-line induction therapy?

Background
Although modern treatment options for cll produce high response rates, almost all patients relapse, likely because  of the persistence of minimal residual disease.Treatment strategies aimed at eradicating minimal residual disease after initial therapy could therefore have a favourable effect on outcomes for patients with cll.Consolidation and maintenance therapy is a promising concept that can further improve the quality and duration of response in patients with cll.Since about 2008, maintenance treatments mainly based on monoclonal antibodies have been explored in chronic B cell malignancies.Rituximab maintenance is now commonly recommended in other B lymphoid diseases such as follicular lymphoma and mantle-cell lymphoma, in which phase iii studies have shown prolonged pfs [72][73][74] .High-dose therapy (with or without total-body radiotherapy) with autologous hsct (auto-hsct) is a consolidation strategy that has been investigated in the first-line treatment setting for cll.
Currently, the only potentially curative treatment for cll is allo-hsct.Previously, the rationale for allo-hsct in first remission for del(17p) or TP53-mutated cll was based on the experience that, once the disease recurred after an effective first-line treatment, the likelihood of a second remission was unlikely with available therapies.Although allo-hsct remains a curative treatment option for some patients, it is not routinely recommended for the frontline treatment of cll patients in the current era of effective targeted therapies.

Summary of Evidence
Consolidation or Maintenance Drug Therapy: Two randomized phase iii trials (one published 75 , one conference abstract 76 ) have compared maintenance therapy with observation after first-line induction with chemoimmunotherapy in cll (supplemental Table 4).Additionally, seven prospective studies evaluating drug consolidation or maintenance strategies after first-line chemoimmunotherapy were identified in the literature search (supplemental Table 5) [77][78][79][80][81][82][83][84][85] .Although current evidence indicates that maintenance therapy might prolong pfs in first remission, no study has yet demonstrated an improvement in os, and further randomized trials and long-term follow-up are required to characterize the benefits of rituximab and other agents in maintenance therapy.

Auto-HSCT:
One randomized trial with rituximab-based induction chemoimmunotherapy, which compared the combination of high-dose consolidation therapy (hdt) and auto-hsct with rituximab maintenance, was identified (supplemental Table 6) 86 .The authors reported the outcomes of hdt with auto-hsct and of rituximab maintenance after fcr induction therapy.After a median follow-up of 5 years, no difference in median event-free survival was observed between the two arms (65.1 months for hdt with auto-hsct and 60.4 months for rituximab maintenance), and hdt with auto-hsct did not result in an improved rate of os (88.1% for hdt with auto-hsct and 88% for rituximab maintenance).

Allo-HSCT:
No prospective studies comparing allo-hsct with non-transplantation strategies during remission after first-line cll treatment were identified in the literature

Recommendations
n No current high-quality evidence supports the use of maintenance therapy in patients with cll after firstline therapy (level of evidence: category 2A).n Given the lack of a survival benefit, we do not recommend hdt with auto-hsct in its current form as a consolidative approach after first-line therapy (level of evidence: category 2A).n Allo-hsct is not currently recommended as part of first-line therapy for cll (level of evidence: category 2B).

TABLE II Independent prognostic factors in chronic lymphocytic leukemia Reference Treatment Patients (n) Analysis Independent prognostic factors for survival Progression-free Overall Randomized studies
e464 Current Oncology, Vol. 25, No. 5, October 2018 © 2018 Multimed Inc.

n
6or asymptomatic patients with early-stage cll who do not meet the indications for therapy established by the International Working Group on CLL guidelines (Tableiv), clinical observation only is recommended6(level of evidence: category 2A).

Fit Patients (Without del(17p) or TP53 Mutation):
Purine analogs have replaced Clb as the backbone of first-line chemotherapy for physically fit patients, based on the results of numerous rcts (supplemental Table2).Fludarabine remains the best-studied purine analog and the one most commonly prescribed for cll.Based on improved os and a 2-year improvement in median pfs, the randomized gcllsg cll8 trial of untreated physically fit patients (cirs score ≤ 6)

TABLE III
Early therapy compared with observation

TABLE IV
6riteria for initiating therapy: summary of guidelines developed by the International Working Group on CLL61.
German CLL Study Group; ECOG PS = Eastern Cooperative Oncology Group performance status; CIRS = Cumulative Illness Rating Scale; CrCl = creatinine clearance; ORR = overall response rate; CR = complete response rate; HR = hazard ratio; CI = confidence interval; MRD = minimal residual disease; pts = patients; NS = nonsignificant; WHO PS = World Health Organization performance status.

TABLE VI
Phase II studies in patients with Del(17p) or TP53 mutations response rate; CR = complete response; PR = partial response; CI = confidence interval.search.Although allo-hsct has curative potential, this treatment option has additional limitations related to age, comorbidity, and donor availability.