iwCLL guidelines for diagnosis, indications for treatment, response assessment, and supportive management of CLL

In 2008, the International Workshop on Chronic Lymphocytic Leukemia (iwCLL) published consensus guidelines for the design and conduct of clinical trials for patients with CLL that were revised from those previously published by the National Cancer Institute–sponsored Working Group.^1-3  Those guidelines provided definitions intended to standardize the assessment of patients that were adopted by the US Food and Drug Administration and European Medicines Agency for the evaluation of new drugs. Since the publication of those guidelines, there have been major advances in the biology and treatment of patients with CLL, prompting the iwCLL to evaluate and revise the 2008 criteria.

The following major changes or additions were introduced in these updated guidelines.

• The clinical relevance of the recent discoveries on the genomic alterations found in CLL, including mutations of the TP53 gene.

• The increasingly important prognostic role of the immunoglobulin variable heavy chain mutational status.

• The current use of clinical staging, novel genetic or biological prognostic markers, and prognostic scores.

• An improved assessment of splenomegaly, hepatomegaly and lymphadenopathy, which was harmonized with the relevant sections of the updated lymphoma response guidelines.

• An updated response assessment for novel targeted drugs (kinase inhibitors, Bcl2 inhibitors) that need to be evaluated during continuous therapy.

• The increasing role of assessing minimal residual disease.

• Updates regarding the baseline assessment and prophylaxis of viral diseases before and under therapy of CLL.

The World Health Organization classification of hematopoietic neoplasias describes CLL as leukemic, lymphocytic lymphoma, being only distinguishable from small lymphocytic lymphoma (SLL) by its leukemic manifestation.⁴  In the World Health Organization classification, CLL, by definition, is always a disease of neoplastic B cells, whereas the entity formerly described as T-CLL is now called T-cell prolymphocytic leukemia.⁵ 

It is important to verify that the patient has CLL and not some other lymphoproliferative disease that can masquerade as CLL, such as hairy cell leukemia or leukemic manifestations of mantle cell lymphoma, marginal zone lymphoma, splenic marginal zone lymphoma with circulating villous lymphocytes, or follicular lymphoma. To achieve this, it is necessary to evaluate the blood smear, the immunophenotype, and, in some cases, the genetic features of the circulating lymphoid cells (see sections 1.1, 1.2, and 1.3).

The diagnosis of CLL requires the presence of ≥5 × 10⁹/L B lymphocytes in the peripheral blood, sustained for at least 3 months. The clonality of these B lymphocytes needs to be confirmed by demonstrating immunoglobulin light chain restriction using flow cytometry. The leukemia cells found in the blood smear are characteristically small, mature lymphocytes with a narrow border of cytoplasm and a dense nucleus lacking discernable nucleoli and partially aggregated chromatin. Gumprecht nuclear shadows, or smudge cells, found as cellular debris, are additional morphologic features commonly associated with CLL. A small percentage of larger or atypical cells or prolymphocytes can be found admixed with morphologically typical CLL cells. Finding ≥55% prolymphocytes would favor a diagnosis of prolymphocytic leukemia; however, this diagnosis remains difficult and is solely based on morphological criteria, because no reliable immunological or genetic marker has been identified. A significant proportion of circulating prolymphocytes (≥10%) seems to indicate a more aggressive form of CLL (with NOTCH1 or genetic TP53 aberrations).⁶ 

CLL or SLL might be suspected in otherwise healthy adults who have an absolute increase in clonal B lymphocytes, but who have <5 × 10⁹/L B lymphocytes in the blood. However, in the absence of lymphadenopathy or organomegaly (as detected by physical examination or imaging studies), or of disease-related cytopenias or symptoms, the presence of <5 × 10⁹/L B lymphocytes is defined as monoclonal B lymphocytosis (MBL).⁷  The presence of a cytopenia caused by a typical marrow infiltrate establishes the diagnosis of CLL regardless of the number of peripheral blood B lymphocytes or of the lymph node involvement. MBL has been observed to progress to CLL, requiring treatment at a rate of 1% to 2% per year.^8,9  Subjects with MBL appear to share an increased risk of secondary cancers with CLL patients, in particular of the skin, and should be encouraged to participate in the appropriate screening programs (eg, for carcinomas of the skin or colon).⁹ 

The definition of SLL requires the presence of lymphadenopathy and the absence of cytopenias caused by a clonal marrow infiltrate. Additionally, the number of B lymphocytes in the peripheral blood should be <5 × 10⁹/L. In SLL, the diagnosis should be confirmed by histopathological evaluation of a lymph node biopsy or biopsy of other tissues. Some patients may present with enlarged lymph nodes that are not suspicious for solid tumors and with peripheral blood B lymphocytes <5 × 10⁹/L that carry a typical CLL immunophenotype (see section 1.2). In these cases, a tissue or lymph node biopsy to establish the diagnosis of SLL may have limited clinical consequences and be omitted.

CLL cells coexpress the surface antigen CD5 together with the B-cell antigens CD19, CD20, and CD23. The levels of surface immunoglobulin, CD20, and CD79b are characteristically low compared with those found on normal B cells.^10-12  Each clone of leukemia cells is restricted to expression of either κ or λ immunoglobulin light chains.¹⁰  The expression of CD5 can also be observed in other lymphoid malignancies, however, such as mantle cell lymphoma.¹³  A recent, large harmonization effort has confirmed that a panel of CD19, CD5, CD20, CD23, κ, and λ is usually sufficient to establish the diagnosis.¹⁴  In borderline cases, markers such as CD43, CD79b, CD81, CD200, CD10, or ROR1 may help to refine the diagnosis.¹⁴ 

The tests described in this section are not essential to diagnose CLL, but may help predict the prognosis or assess the tumor burden. Of these different tests, only a few are needed to establish a prognostic profile in addition to the clinical staging (see section 2.3). Moreover, the indication for treatment does not depend on the results of these tests, but on the patient’s clinical stage and symptoms (see section 4).

Interphase fluorescence in situ hybridization (FISH) can be performed with peripheral blood lymphocytes and identifies cytogenetic lesions in >80% of all CLL cases.¹⁵  The most common deletions are in the long arm of chromosome 13 (del(13q)). Additional, frequent chromosomal aberrations comprise trisomy of chromosome 12 and deletions in the long arm of chromosomes 11 (del(11q)) and in the short arm of chromosome 17 (del(17p)).¹⁵ 

Appropriate stimulation of CLL cells in vitro has enabled the performance of conventional karyotyping with enhanced reliability.¹⁶  With this methodology, additional chromosomal aberrations of potential prognostic significance can be identified.^16-18  Moreover, stimulated metaphase karyotyping has demonstrated that leukemia cells with a complex karyotype (ie, ≥3 chromosomal abnormalities) may have adverse prognostic significance.^19-22  However, more data from prospective trials are needed to validate the prognostic and predictive value of stimulated metaphase karyotyping before we can recommend it for routine practice (Table 1).

Furthermore, FISH and conventional karyotyping can help distinguish CLL from other lymphoproliferative diseases, which have distinct disease-associated chromosomal abnormalities (eg, t(11;14), which is usually associated with mantle cell lymphoma). So far, other technologies such as array-based assays or next-generation sequencing have not been able to completely replace FISH or conventional karyotyping.

Prospective clinical trials indicate that certain genetic abnormalities are associated with adverse outcomes in response to standard chemo(immuno)therapy regimens. Patients with leukemia cells that carry del(17p) and/or TP53 mutations (as determined by DNA sequencing, with a cutoff of 10%)²³  have an inferior prognosis and appear relatively resistant to standard chemotherapy regimens using alkylating drugs and/or purine analogs.^23-26  In a retrospective analysis of several chromosomal aberrations detected by FISH, patients who had CLL cells with chromosomal aberrations del(11q) and del(17p) had an inferior outcome compared with that of patients who had leukemia cells with a normal karyotype or del(13q) as the sole genetic abnormality.¹⁵  On the other hand, patients who have leukemia cells with del(17p) and/or TP53 mutation respond poorly to chemo(immuno)therapy but fare significantly better when treated with nonchemotherapeutic agents, such as small molecule inhibitors of BTK, phosphatidylinositol 3-kinase, or BCL2. It has been demonstrated that the progression-free survival and overall survival of CLL patients carrying a del(17p) and patients carrying a TP53 mutation, as detected by Sanger sequencing in the absence of del(17p), are similar.²⁷  Therefore, the assessment of both del(17p) and TP53 mutation has prognostic and predictive value and should guide therapeutic decisions in routine practice. For clinical trials, it is recommended that molecular genetics be performed before treating a patient on protocol. Because additional genetic abnormalities may be acquired during the course of the disease,²⁸  genetic analyses (in particular for del(17p)/TP53 mutations) should be repeated before any subsequent second- or third-line treatment.

Next-generation whole exome or whole genome sequencing have identified additional genomic abnormalities, such as mutations in NOTCH1 or SF3B1 that have pathogenic as well as prognostic significance. However, more data from prospective trials are needed to validate the prognostic and predictive value of these genomic abnormalities before we can advocate using them in routine practice.

The leukemia cells use immunoglobulin variable heavy chain (IGHV) genes that may or may not have undergone somatic mutations.^29-31  The outcome of patients with leukemia cells that use an unmutated IGHV gene (usually defined as 98% or more sequence homology to the nearest germ line gene) is inferior to that of patients with leukemia cells that use a mutated IGHV gene.^32,33  Moreover, the presence of mutated IGHV genes, in particular when combined with additional prognostic factors such as favorable cytogenetics or attainment of a minimal residual disease (MRD) negative state after therapy, characterizes a CLL patient subgroup with excellent outcome following chemoimmunotherapy with fludarabine, cyclophosphamide, and rituximab.^34-36 

The discovery of almost identical or “stereotyped” B-cell receptor immunoglobulins among unrelated CLL patients suggests that (auto)antigen selection may play a role in disease pathogenesis.³⁷  Approximately one-third of patients can be grouped into subsets based on shared sequence motifs within the IGHV region complementarity determining region 3.³⁷  It seems that some of these subgroups share a similar prognosis. For example, IGHV3-21 gene usage (of stereotype subset 2) may be associated with an unfavorable prognosis independent of the IGHV mutational status.^38,39  As of today, assessment of IGHV stereotypes is not an element of the routine prognostic work up in CLL.

Leukemia cell expression of ZAP-70 and CD38 correlates with the expression of unmutated IGHV genes and can be associated with poor prognosis.^25,40-47  The association between expression of ZAP-70 or CD38 with the presence of unmutated IGHV genes is not absolute, and discordant cases are more frequently found in patients with high-risk cytogenetics.⁴⁸  CD49d, the α chain of the alpha4beta1 integrin heterodimer, has been associated with an unfavorable prognosis in CLL and was shown to be the strongest flow cytometry-based predictor of overall survival and treatment-free survival in a large, multicenter effort.^49,50 

Several studies have found that serum markers such as levels of soluble CD23, thymidine kinase, and β₂-microglobulin are associated with overall survival or progression-free survival.^51-58  Of these, β₂-microglobulin has retained independent prognostic value in several multiparameter scores.^55,58,59  Assays for these markers should be standardized and used in prospective clinical trials to validate their relative value in the management of patients with CLL.

In CLL, typically >30% of the nucleated cells in the aspirate are mature lymphoid cells. The extent and pattern of marrow infiltration (diffuse vs nondiffuse) may reflect the tumor burden.⁶⁰  A marrow aspirate and biopsy generally are not required for the diagnosis of CLL; however, a marrow biopsy and aspirate can help clarify whether cytopenias (neutropenia, anemia, thrombocytopenia) are related or unrelated to leukemic infiltration of the marrow. In these cases, a marrow biopsy may provide important information, in particular before starting therapies with cytotoxic agents. It is recommended to repeat a marrow biopsy in patients with persisting cytopenia after treatment to clarify disease- vs therapy-related causes. A marrow biopsy is mandatory to confirm a complete remission (CR; see section 5.1).

There are 2 widely accepted staging systems for use in both patient care and clinical trials: Rai⁶¹  and Binet.⁶²  The original Rai classification was modified to reduce the number of prognostic groups from 5 to 3.⁶³  As such, both systems now describe 3 major subgroups with distinct clinical outcomes. These 2 staging systems are simple, inexpensive, and can be readily and consistently applied by physicians worldwide. Both rely solely on a physical examination and standard laboratory tests and do not require imaging studies.

The modified Rai classification defines low-risk disease as occurring in patients who have lymphocytosis with leukemia cells in the blood and/or marrow (formerly considered Rai stage 0). Patients with peripheral blood lymphocytosis, enlarged lymph nodes in any site, and splenomegaly and/or hepatomegaly (lymph nodes being palpable or not) are defined as having intermediate-risk disease (formerly considered Rai stage I or II). High-risk disease includes patients with disease-related anemia (as defined by a hemoglobin [Hb] level < 1 g/dL) (formerly stage III) or thrombocytopenia (as defined by a platelet count of <100 × 10⁹/L; formerly stage IV).

The Binet staging system is based on the number of involved lymphoid areas, as defined by the presence of enlarged lymph nodes ≥1 cm in diameter or organomegaly, and on whether there is anemia or thrombocytopenia.

Areas of involvement considered for staging include:

1. Head and neck, including the Waldeyer ring (this counts as 1 area, even if ≥1 group of nodes is enlarged).

2. Axillae (involvement of both axillae counts as just 1 area).

3. Groins, including superficial femorals (involvement of both groins counts as just 1 area).

4. Palpable spleen.

5. Palpable liver (clinically enlarged).

Stage A. Hb ≥10 g/dL and platelets ≥100 × 10⁹/L and up to 2 of these areas involved.

Stage B. Hb ≥10 g/dL and platelets ≥100 × 10⁹/L and 3 or more of the lymphoid areas involved.

Stage C. Hb <10 g/dL and/or a platelet count <100 × 10⁹/L.

In daily practice, Rai or Binet stages help stratify patients according to the disease risk; however, there are a large number of biomarkers that can provide additional prognostic information.^64-66  The most relevant prognostic parameters are IGHV mutational status, serum β₂-microglobulin, and the presence of del(17p) and/or TP53 mutations. Usually, high-risk CLL is defined, at least in part, by a genetic aberration of the TP53 gene (ie, del(17p) or TP53 mutation).

Following the identification of new prognostic parameters, several prognostic scores and stratification systems have been proposed based on multivariate analyses to extract the most significant independent prognostic information from the plethora of known prognostic markers.^55,58,59  These models are very useful to identify high-risk patient populations for experimental protocols, but also those patients with a very good prognosis even at advanced stages. One of these prognostic scores, the CLL international prognostic index (CLL-IPI) consists of a weighed score that includes the clinical stage, age, IGHV mutational status, serum β₂-microglobulin, and the presence of del(17p) and/or TP53 mutations.⁵⁹  It was originally developed using datasets of ≥4500 patients treated within or outside of clinical trials, divides 4 different prognostic subgroups, and has been validated extensively in various cohorts.^67-73  The value of prognostic markers or scores might change with the application of novel therapies.

The selection of CLL patients for clinical trials is similar to that for patients with other malignancies. Phase 1-2 clinical trials commonly, although not invariably, are intended for patients who have received prior therapy. The inclusion of patients with SLL in clinical trials for CLL is encouraged. The combination of new agents with standard therapy as part of phase 2 studies may be investigated in both untreated and previously treated patients. Phase 3 clinical trials are used to compare the clinical outcome using new treatment modalities with that attained using current standard therapy. Other requirements for eligibility with respect to age, clinical stage, performance status, biomarkers, organ function, or status of disease activity should be defined for each study.

Before inclusion in a trial, the performance status by the Eastern Cooperative Oncology Group (ECOG) should be determined. Future clinical trials involving elderly patients ideally should assess the comorbidity (fitness) and/or functional activity of patients by appropriate scores.^74-76  The cumulative illness rating scale has been used successfully in the setting of multicenter clinical trials to characterize patients with concomitant morbidity,⁷⁷  although other scores to identify unfit patients are available as well.⁷⁸ 

Most chemotherapy agents have potential toxicity to the liver, kidneys, heart, lungs, nervous system, or other organ systems; therefore, organ function requirements should be guided by the known or suspected toxicity profile of each agent based on preclinical studies or prior clinical studies. Patients enrolled on protocols evaluating agents with known or suspected toxicity for a given organ should have the specific organ function documented before therapy.

The status of specific infectious diseases, as outlined in section 3.5, should be documented. Patients with active infections requiring systemic antibiotics or antifungal or antiviral drugs should have their infection controlled before initiating therapy in a clinical trial.

Patients with active second malignancies generally are not considered candidates for entry into clinical trials. Second cancers in remission and nonmelanoma skin cancers should not necessarily be an exclusion criterion for clinical trials involving patients with CLL.

Parameters considered necessary for a complete baseline evaluation might differ depending on whether the patient is treated in a clinical protocol. A clear distinction is made in sections 3.5 and 5 between recommendations for routine practice and the requirements for clinical trials (Tables 1, 2, and 3). Unless indicated, recommendations are the same for clinical trials and routine practice. In general, baseline evaluation studies for defining these parameters should be performed within 2 weeks of clinical trial enrollment, except for molecular cytogenetics (FISH), marrow aspirate and biopsy, as well as computed tomography (CT) scans (see sections 3.5.1 and 3.5.2).

3.5.1.1. Physical examination. The bidimensional diameters of the largest palpable lymph nodes in each of the following sites should be recorded: cervical, axillary, and inguinal (Table 1). The dimensions of the liver and spleen below their respective costal margins, as assessed by palpation, should also be recorded. Note that any of these manifestations of CLL, in particular hepatomegaly, could be caused by a variety of other diseases.

3.5.1.2. Assessment of performance status (ECOG score).

3.5.1.3. A complete blood cell count (white blood cell count, Hb, hematocrit, reticulocyte, and platelet count) and differential leukocyte count, including both percent and absolute number of lymphocytes. Reporting the proportion of prolymphocytes is desirable when these are present.

3.5.1.4. Marrow biopsy. Before initiating treatment within in a clinical trial, a unilateral marrow aspirate and biopsy are recommended. Repeat marrow biopsies may be compared with the baseline marrow specimen to assess the cause of cytopenias (eg, bone marrow toxicity, disease progression).

3.5.1.5. Serum chemistry (eg, creatinine, bilirubin, lactate dehydrogenase, haptoglobin, transaminases, alkaline phosphatase, β₂-microglobulin).

3.5.1.6. Serum immunoglobulin levels.

3.5.1.7. Direct antiglobulin test.

3.5.1.8. Chest radiograph (when a CT scan is not performed).

3.5.1.9. HIV serology. Patients who are infected with HIV should be given special consideration because of the potential risks for immune suppression with most antileukemia therapies and the potential for compounded myelotoxicity of treatment with antiretroviral therapy.

3.5.1.10. CMV. Therapies associated with the potential for reactivation of infection with cytomegalovirus (CMV), such as alemtuzumab, idelalisib, or allogeneic stem cell transplantation, should include plans for monitoring for active CMV disease and/or for providing anti-CMV therapy (eg, ganciclovir, valganciclovir).^79-81  These should cover screening or early diagnosis of CMV reactivation and its subsequent management. However, positive CMV serology does not represent a contraindication to treatment of CLL. As a general recommendation, patients treated with immune-suppressive agents (eg, alemtuzumab, idelalisib), should be monitored for CMV and be considered for antiviral therapy if found to have increased levels of CMV in the blood by the polymerase chain reaction (PCR), even in the absence of clinical symptoms. Moreover, antiviral therapy is recommended for infected patients with clinical symptoms of active CMV infection.

3.5.1.11. Hepatitis B and hepatitis C. Before initiating treatment, patients should be evaluated for infection with hepatitis B virus (HBV) or hepatitis C virus (HCV) because reactivation of HBV and HCV may occur following treatment with immunosuppressive or myelosuppressive drugs, including anti-CD20 antibodies. Patients found to be chronic carriers of HBV, as defined by positive surface antigen, HB core antibody, and/or low HBV titers in serum, should receive prophylactic antiviral agents while undergoing therapy for CLL with immunosuppressive drugs.⁸²  In patients with high titers of HBV or HCV DNA, initiation of antiviral therapy before antileukemic treatment should be considered.

The following recommendations are made for clinical trials or for the assessment in specific clinical situations (Table 1).

3.5.2.1. Examination of leukemia-cell cytogenetics (eg, metaphase karyotyping, FISH [in particular for del(17p)]) and analysis for inactivating mutations in TP53 should be performed before any line of therapy. The recommended threshold for reporting of mutations detected by next-generation sequencing should reflect the Sanger-like threshold of ∼10% variant allele frequency.⁸³  Patients with genetic TP53 aberrations should not receive chemotherapy, if other options are available.

3.5.2.2. CT scans. CT scans generally are not required for initial evaluation or for follow-up. The staging of CLL does not use CT scans but relies on physical examination and blood counts. Enlarged lymph nodes, if detected only by CT scan, do not change the Binet or Rai stage. It has been shown that patients with Rai stage 0 but abdominal disease detectable by CT scans may have a more aggressive course.⁸⁴  This requires further investigation before recommending CT scans for routine initial evaluation of patients with CLL. On the other hand, it has been demonstrated that the majority of relapses or progressions in CLL are detected by physical examination and blood counts, not by imaging studies.⁸⁵  Moreover, the decision for relapse treatment was determined by imaging studies in only 1% of patients⁸⁵ ; therefore, the routine follow-up evaluation of CLL patients does not require CT scans.

In clinical trials where the treatment intent is to maximize the overall response rate, neck, chest, abdominal, and pelvic CT scans are recommended to evaluate the response to therapy. One CT scan should be performed before the start of therapy and a second CT scan at the final response assessment (usually the first restaging after the end of therapy) within a study protocol, if abnormal at baseline. For the assessment of continuous therapies, CT scans should be performed at the time point of clinically evaluated maximal response or, alternatively, at a time point defined by the protocol. Additional, repetitive CT scan monitoring is usually not clinically relevant and potentially harmful for the patient.

3.5.2.3. Other imaging methods. Except in patients with proven or suspected Richter transformation, positron emission tomography (PET) scans do not provide information that is useful in the management of CLL.⁸⁶  Similarly, nuclear magnetic resonance imaging generally does not provide useful information beyond that of CT scanning in the management of CLL and therefore is not recommended outside of clinical trials.

3.5.2.4. Ultrasound imaging. In some countries, ultrasound imaging is used to assess the extent of lymphadenopathy and organomegaly in CLL. Although ultrasound imaging may be very useful in the clinical management of individual patients, the results obtained by this methodology are investigator-dependent and difficult to centrally verify. Therefore, ultrasound imaging is currently not recommended for response evaluation in clinical trials.

3.5.2.5. A lymph node biopsy is generally not required, unless it is necessary for companion scientific studies or in rare cases in which the diagnosis is difficult. A lymph node or tissue biopsy is performed to establish the diagnosis of a transformation into an aggressive lymphoma (Richter transformation). A PET scan can be useful to identify the best lymphoid area for establishing the diagnosis by biopsy.

Criteria for initiating treatment may vary depending on whether the patient is treated in a clinical trial (Table 2). In general practice, patients with asymptomatic early-stage disease (Rai 0, Binet A), should be monitored without therapy unless they have evidence of disease progression or disease-related symptoms. Several studies have shown that treating patients with early-stage disease does not result in a survival benefit^87-90 ; therefore, an early-intervention therapy with antileukemia drugs, including signaling inhibitors or BCL2 antagonists, alone or in combination with monoclonal antibodies, currently is not indicated.

Although patients with intermediate-risk (stages I and II) and high-risk (stages III and IV) disease according to the modified Rai classification or at Binet stage B or C usually benefit from the initiation of treatment, some of these patients (in particular, Rai intermediate risk or Binet stage B) can be monitored without therapy until they have evidence for progressive or symptomatic disease (summarized as “active disease”).

Active disease should be clearly documented to initiate therapy. At least 1 of the following criteria should be met.

1. Evidence of progressive marrow failure as manifested by the development of, or worsening of, anemia and/or thrombocytopenia. Cutoff levels of Hb <10 g/dL or platelet counts <100 × 10⁹/L are generally regarded as indication for treatment. However, in some patients, platelet counts <100 × 10⁹/L may remain stable over a long period; this situation does not automatically require therapeutic intervention.

2. Massive (ie, ≥6 cm below the left costal margin) or progressive or symptomatic splenomegaly.

3. Massive nodes (ie, ≥10 cm in longest diameter) or progressive or symptomatic lymphadenopathy.

4. Progressive lymphocytosis with an increase of ≥50% over a 2-month period, or lymphocyte doubling time (LDT) <6 months. LDT can be obtained by linear regression extrapolation of absolute lymphocyte counts obtained at intervals of 2 weeks over an observation period of 2 to 3 months; patients with initial blood lymphocyte counts <30 × 10⁹/L may require a longer observation period to determine the LDT. Factors contributing to lymphocytosis other than CLL (eg, infections, steroid administration) should be excluded.

5. Autoimmune complications including anemia or thrombocytopenia poorly responsive to corticosteroids.

6. Symptomatic or functional extranodal involvement (eg, skin, kidney, lung, spine).

7. Disease-related symptoms as defined by any of the following:

   • Unintentional weight loss ≥10% within the previous 6 months.

   • Significant fatigue (ie, ECOG performance scale 2 or worse; cannot work or unable to perform usual activities).

   • Fevers ≥100.5°F or 38.0°C for 2 or more weeks without evidence of infection.

   • Night sweats for ≥1 month without evidence of infection.

Hypogammaglobinemia, or monoclonal or oligoclonal paraproteinemia does not by itself constitute a basis for initiating therapy. However, it is recommended to assess the change in these protein abnormalities, if patients are treated. Also, patients with CLL may present with a markedly elevated leukocyte count; however, leukostasis rarely occurs in patients with CLL. Therefore, the absolute lymphocyte count should not be used as the sole indicator for treatment.

Disease relapse alone is not a criterion to restart therapy unless the disease is symptomatic (see active disease criteria). Asymptomatic increases of lymphocyte counts alone without other signs of progression are generally not an indication to restart therapy. Second- and subsequent-line treatment decisions should generally follow the same indications as those used for first-line treatment. Where the original indication for treatment has not resolved with initial therapy, provided treatment-related toxicities have recovered, it is reasonable to initiate second-line treatment without waiting for formal disease progression to be manifest. Also, the rate of disease progression after some newer therapies can be rapid; in such circumstances, it can be acceptable to initiate subsequent therapy before formal progression where there is substantial persisting disease burden.

Patients with any of the following features generally do not respond to second-line chemo(immuno)therapy: primary resistance to first-line chemo(immuno)therapy; time to progression after first-line, fludarabine-based chemo(immuno)therapy of 2 to 3 years^91,92 ; or leukemia cells with del(17p)/TP53 mutations. These patients should be offered a nonchemotherapy regimen and/or entrance into clinical trials. In selected cases, allogeneic hematopoietic stem cell transplantation should be considered.^93-95 

Assessment of response should include a careful physical examination and evaluation of the blood and bone marrow (Tables 3 and 4). The timing of response assessment for therapies with a defined treatment duration (such as chemoimmunotherapeutic approaches) should be at least 2 months after completion of therapy. To define the response to therapy, 2 groups of parameters need to be assessed and documented: parameters of group A assess the lymphoid tumor load and constitutional symptoms; parameters of group B assess the hematopoietic system (Table 4).

For continued therapies or treatment strategies that contain a maintenance phase, the assessment of response should be performed at least 2 months after patients achieve their maximum response or at a time point that is predefined in the protocol; in this case, it is not necessary to interrupt therapy for response assessment. Maximum response can be defined as a treatment phase in which no additional improvement is seen during at least 2 months of therapy. In clinical trials, any response (eg, CR, partial remission) should be sustained for at least 2 months before using this response in the assessment. In addition, where appropriate, a further assessment of response (ie, marrow assessment) may be performed at least 2 months after the patient has cleared MRD from the peripheral blood.

CR requires all of the following criteria (Table 4).

5.1.1. Peripheral blood lymphocytes (evaluated by blood and differential count) <4 × 10⁹/L.

5.1.2. Absence of significant lymphadenopathy by physical examination. In clinical trials, a CT scan of the neck, abdomen, pelvis, and thorax is desirable if previously abnormal. Lymph nodes should be <1.5 cm in longest diameter. Once this is determined, further imaging should not be required until disease progression is apparent by clinical examination or on blood testing.

5.1.3. No splenomegaly or hepatomegaly by physical examination. In clinical trials, a CT scan of the abdomen should be performed at response assessment and should show no evidence for lymphadenopathy and splenomegaly. We propose to use a recent consensus response cutoff for splenomegaly of 13 cm in craniocaudal length.^96,97  However, the persistence of splenomegaly may not correlate with outcome.⁹⁶  The quantitative determination of hepatomegaly seems more difficult; changes such as focal or disseminated hepatic nodules support liver involvement.

5.1.4. Absence of disease-related constitutional symptoms.

5.1.5. Blood counts need to show the following values:

5.1.5.1. Neutrophils ≥1.5 × 10⁹/L.

5.1.5.2. Platelets ≥100 × 10⁹/L.

5.1.5.3. Hemoglobin ≥11.0 g/dL (without red blood cell transfusions).

5.1.6. MRD assessment.

In clinical trials aimed at maximizing the depth of remission, the presence of MRD after therapy should be assessed (see section 5.9). The sensitivity of the method used to evaluate for MRD should be reported, as should the tissue studied (blood or marrow). The proportion of patients achieving undetectable MRD should be reported with the total number of patients treated with the specific therapy as the denominator (not as a proportion of responders or those in CR).

5.1.7. For patients in clinical trials (Table 3). A bone marrow aspirate and biopsy should be performed if clinical and laboratory results listed in sections 5.1.1 to 5.1.5 demonstrate that a CR may have been achieved. To define a CR, the cytological or pathological evaluation of the bone marrow smear or biopsy must be at least normocellular for age, without evidence for typical CLL lymphocytes by morphological criteria. This evaluation is not based on a flow cytometry–based MRD assessment (see section 9).

In a clinical trial, the time point of marrow biopsy should be defined by the protocol. For example, in patients receiving chemo(immuno)therapy, the time point of marrow biopsy is typically 2 months posttherapy.

When performing marrow biopsies in clinical trials, lymphoid nodules can be found that may reflect residual disease.^98,99  These nodules may be recorded as “nodular partial remission.” Immunohistochemistry may be performed to define whether the nodules comprise primarily T cells, B cells other than CLL cells, or CLL cells. If nodules are not composed of CLL cells, a CR can be documented provided all other criteria are met. If the marrow is hypocellular, a repeat determination should be performed 4 weeks or later, when peripheral blood counts have recovered; however, this interval should not exceed 6 months after the last treatment. In cases in which a marrow biopsy was obtained at baseline, a comparison of pre- vs posttherapy biopsies should be performed. In general practice, the use of a marrow biopsy for evaluating a CR is at the discretion of the physician.

In clinical trials aimed at maximizing the response rate, the quality of the response should be assessed in the marrow for MRD by highly sensitive molecular-based assays or immunophenotyping (see section 5.9).

5.1.8. Some patients fulfill all the criteria for a CR (including the marrow examinations described in section 5.1.7), but have a persistent anemia, thrombocytopenia, or neutropenia apparently unrelated to CLL, but related to drug toxicity. These patients should be considered as a different category of remission, CR with incomplete marrow recovery (CRi). For the definition of this category, the marrow evaluation (see section 5.1.7) should be performed with scrutiny and not show any clonal disease infiltrate. In clinical trials, patients having CR with incomplete marrow recovery should be monitored prospectively to determine whether their outcome differs from that of patients with detectable residual disease or with noncytopenic CR.

To define a partial remission, at least 2 parameters of group A and 1 parameter of group B need to improve, if previously abnormal (Table 4; sections 5.2.1 to 5.2.5). If only 1 parameter of both groups A and B was abnormal before therapy, only 1 needs to improve. Constitutional symptoms persisting for >1 month should be recorded.

5.2.1. A decrease in the number of blood lymphocytes to 50% or less from the value before therapy.

5.2.2. Reduction in lymphadenopathy compared with baseline (by cross-sectional imaging scans in clinical trials or by palpation in general practice) as defined by:

5.2.2.1. A decrease in lymph node size by 50% or more in

• the sum of the products of the same enlarged lymph nodes selected at baseline as assessed by imaging (an established number in clinical trials of lymph nodes has been up to 6).

• and the sum of longest diameters of the same enlarged lymph nodes selected at baseline as assessed by physical examination (an established number in clinical trials of lymph nodes has been a maximum of 6).

5.2.2.2. No increase in any lymph node and no new enlarged lymph node (diameter ≥1.5 cm). For small lymph nodes (longest diameter <1.5 cm), an increase <25% is not considered significant.

5.2.3. A regression ≥50% of the extent of enlargement of the spleen below the costal margin defined by palpation, or normalization in size. When assessed by CT, scan spleen size must have regressed by ≥50% in length beyond normal.⁹⁶  A persistence of splenomegaly posttherapy may have limited influence on outcome in CLL.⁹⁶ 

5.2.4. A regression of ≥50% of the extent of enlargement of the liver below the costal margin defined by palpation, or normalization in size. Given the impact of numerous medical conditions, liver size by physical examination or CT scan is not a reliable measure of hepatic involvement by CLL and should only be counted if hepatomegaly is clearly attributable to lymphoid involvement.

5.2.5. The blood count should show 1 of the following results:

5.2.5.1. Platelet counts >100 × 10⁹/L or 50% improvement over baseline.

5.2.5.2. Hb >11.0 g/dL or 50% improvement over baseline without red blood cell transfusions or erythropoietin support.

Progressive disease (PD) during or after therapy is characterized by at least 1 of the following, when compared with nadir values (Table 4):

5.3.1. Lymphadenopathy. Progression of lymphadenopathy is often discovered by physical examination and should be recorded at regular intervals. In CLL, the use of imaging (CT scans) usually does not add much information for the detection of progression or relapse.¹⁰⁰  Disease progression occurs if 1 of the following events is observed.

• Appearance of any new lesion such as enlarged lymph nodes (≥1.5 cm), splenomegaly, hepatomegaly, or other organ infiltrates. Transient increases of lymph node size during treatment with novel inhibitors may occur and should not be counted as PD.

• An increase by ≥50% in greatest determined diameter of any previous site (≥1.5 cm).

5.3.2. An increase in the spleen size by ≥50% or the de novo appearance of splenomegaly. In the setting of splenomegaly, the splenic length must increase by ≥50% of the extent of its prior increase beyond baseline (eg, a 15-cm spleen must increase to ≥16 cm). If no prior splenomegaly was observed at baseline or if splenomegaly has resolved with treatment, the spleen must increase by at least 2 cm from baseline.⁹⁶ 

5.3.3. An increase in the liver size of ≥50% of the extent enlargement of the liver below the costal margin defined by palpation, or the de novo appearance of hepatomegaly. Given the impact of numerous medical conditions, liver size by physical examination or by CT scan is not a reliable measure of hepatic involvement by CLL and should only be counted if hepatomegaly is clearly attributable to lymphoid involvement.

5.3.4. An increase in the number of blood lymphocytes by 50% or more with at least 5 × 10⁹/L B lymphocytes. Certain therapies (eg, kinase inhibitors) may cause lymphocytosis. In the setting of therapy with such agents, an increase in blood lymphocyte count by itself does not uniformly indicate an increased tumor burden, but may reflect redistribution of leukemia cells from lymphoid tissues to the blood. This should be predefined in the protocol of clinical trials for therapies in which redistribution of disease occurs. In such cases, increased lymphocytosis alone is not a sign of treatment failure or PD.¹⁰¹ 

5.3.5. Transformation to a more aggressive histology (Richter syndrome or Richter transformation). The diagnosis of Richter transformation should be established by lymph node or other tissue biopsy.

5.3.6. Occurrence of cytopenia (neutropenia, anemia, or thrombocytopenia) directly attributable to CLL and unrelated to autoimmune cytopenias.

5.3.6.1. During therapy. Cytopenias may occur as a side effect of many therapies and should be assessed according to Table 5. During therapy, cytopenias cannot be used to define disease progression. Each protocol should define the amount of drug(s) to be administered with such cytopenias.

5.3.6.2. Posttreatment. The progression of any cytopenia (unrelated to autoimmune cytopenia), as documented by a decrease of Hb levels ≥2 g/dL or <10 g/dL, or by a decrease of platelet counts ≥50% or <100 × 10⁹/L, which occurs at least 3 months after treatment, defines disease progression, if the marrow biopsy is consistent with the cytopenia resulting from increased marrow infiltration of clonal CLL cells and is not considered a treatment related toxicity.

Patients who have not achieved a CR or a partial remission, and who have not exhibited PD, will be considered to have stable disease (which is equivalent to a nonresponse).

Responses that should be considered clinically beneficial include CR and PR; all others (eg, stable disease, nonresponse, PD, death from any cause) should be rated as a treatment failure.

Progression-free survival is defined as the interval between the first treatment day (in phase 3 trials: day of randomization for intent-to-treat analysis) to the first sign of disease progression or death from any cause. Event-free survival is defined as the interval between the first treatment day (in phase 3 trials: day of randomization for intent-to-treat analysis) to the first sign of disease progression or start of a new treatment or withdrawal from the trial because of toxicity or death (whichever occurs first). Overall survival is defined as the interval between the first treatment day (in phase 3 trials: day of randomization for intent-to-treat analysis) to death. Time to next treatment is defined as interval between the first treatment day until the patient starts an alternative therapy for progressive CLL.

Note that the response duration may be assessed during therapy for continuous treatment, in particular with oral agents, as well as after the end of treatment, in particular with chemo(immuno) therapy. Study protocols should provide detailed specifications of the planned time points for the assessment of the treatment response under continuous therapy. Response durations <6 months are not considered clinically relevant (see refractory disease, section 5.8).

Relapse is defined as evidence of disease progression (see section 5.3) in a patient who has previously achieved the above criteria of a CR or partial remission (sections 5.1-5.2) for ≥6 months.

Refractory disease is defined as treatment failure (as defined in section 5.5) or as progression within 6 months from the last dose of therapy.

The complete eradication of the leukemia is a desired end point. Use of sensitive multicolor flow cytometry, PCR, or next-generation sequencing can detect MRD in many patients who achieved a complete clinical response. Prospective clinical trials have provided substantial evidence that therapies that are able to eradicate MRD usually result in an improved clinical outcome.^97,102-106  The techniques for assessing MRD have undergone a critical evaluation and have become well standardized.^107,108  Six-color flow cytometry (MRD flow), allele-specific oligonucleotide PCR, or high-throughput sequencing using the ClonoSEQ assay are reliably sensitive down to a level of <1 CLL cell in 10 000 leukocytes.¹⁰⁸  Refinement and harmonization of these technologies has established that a typical flow cytometry–based assay comprises a core panel of 6 markers (ie, CD19, CD20, CD5, CD43, CD79b, and CD81).¹⁰⁸  As such, patients will be defined as having undetectable MRD (MRD-neg) remission if they have blood or marrow with <1 CLL cell per 10 000 leukocytes. The blood generally can be used for making this assessment because the marrow will have detectable CLL when it is also found in the peripheral blood. However, there are therapies that preferentially clear the blood but not the marrow (such as monoclonal antibodies); therefore, it may be important to confirm that the marrow aspirate also is MRD-neg when the blood is found to be MRD-neg. Clinical trials aimed at maximizing the depth of remissions should include at least 1 test to assess for MRD, because the lack of leukemia persistence using these sensitive tests has a strong, positive prognostic impact. The report should be clear as to whether blood and/or marrow have been assessed and should report the proportion of MRD-neg patients on an intent-to-treat basis using the total number of patients in that treatment arm as the denominator (not those assessed or those who responded to treatment).

6.1. Patients ideally should be stratified with regard to previous treatment vs no previous treatment, and as purine analog-sensitive vs purine analog-refractory in studies for which prior therapy is allowed.

6.2. If >1 clinical stage is allowed, patients ideally should be stratified with respect to clinical stage.

6.3. If the specific patient subgroup is not excluded, then they should be stratified or analyzed as a subgroup analysis based upon whether they have leukemia cells with del(17p) or del(11q), and for mutations of the TP53 gene.

6.4. Patients should be stratified for the mutational status of the IGHV gene locus (mutated vs unmutated). If it is not possible to stratify prospectively, then the IGHV mutated and IGHV unmutated patients should be analyzed as a planned subgroup analysis.

Evaluation of treatment-related toxicity requires careful consideration of both the manifestations of the underlying disease and the anticipated adverse reactions to the agents used in therapy. For this reason, some of the conventional criteria used for assessing toxicity are not applicable to clinical studies involving patients with hematological malignancies in general or CLL in particular. An example is hematological toxicity; patients with advanced CLL generally have cytopenias that may be caused by the underlying CLL and/or prior therapy. A few recommendations are presented to help evaluate for treatment-induced toxicity in CLL.

Evaluation of hematological toxicity in patients with CLL must take into consideration that many patients have low blood cell counts at the initiation of therapy. Therefore, the standard criteria used for solid tumors cannot be applied because many CLL patients then would be considered to have grade 2 to 4 hematological toxicity at the initiation of treatment. Furthermore, the absolute blood neutrophil counts are rarely used at the initiation of therapy to modify the treatment dose because these values typically are unreliable in CLL patients with lymphocytosis. However, the increasing use of more effective therapeutic agents, particularly those with neutropenia as a dose-limiting toxicity (eg, nucleoside analogs), can result in clinically significant myelosuppression. Therefore, the 1996 guidelines proposed a new dose-modification scheme for quantifying hematological deterioration in patients with CLL, which included alterations in the dose of myelosuppressive agents based on the absolute neutrophil count. This dose modification scheme has proven very helpful in the context of several large prospective trials in CLL and is therefore retained in the current version of the guidelines (Table 5).

Patients with CLL are at increased risk for infection because of compromised immune function, which might be related to the disease itself and/or to the consequences of therapy. Nevertheless, the rate of infection following treatment can be used in assessing the relative immune-suppressive effects of a given therapy. The etiology of the infection should be reported and categorized as bacterial, viral, or fungal, and as proven or probable. The severity of infections should be quantified as minor (requiring either oral antimicrobial therapy or symptomatic care alone), major (requiring hospitalization and systemic antimicrobial therapy), or fatal (death as a result of the infection).

Particular attention should be given to monitoring for symptoms or laboratory evidence of opportunistic infections such as Pneumocystis jirovecii or Herpesviridae (herpes simplex virus, varicella-zoster virus, cytomegalovirus, Epstein-Barr virus) in patients treated with agents such as alemtuzumab and idelalisib (alone or in combination) or with allogeneic stem cell transplantation. Patients receiving anti-CD20 antibodies may experience reactivation of HBV infections.¹⁰⁹  HBV serological status should be evaluated before treatment with such agents; therefore, appropriate antiviral prophylaxis should be initiated in patients with a history of HBV infection.¹⁰⁹  In contrast, the infection rate seems low in patients age <65 years treated with fludarabine-based first-line therapy, in which no monitoring or routine anti-infective prophylaxis is required.¹¹⁰  Progressive multifocal leukoencephalopathy has been reported in a few CLL patients treated with anti-CD20 antibodies; therefore, infections with JC virus should be ruled out in situations of unclear neurological symptoms.^111-114 

Patients with CLL rarely experience tumor lysis syndrome (TLS) after therapy with purine analog–based regimens.¹¹⁵  However, TLS might occur following treatment with drugs such as lenalidomide,¹¹⁶  venetoclax,¹¹⁷  or type II anti-CD20 monoclonal antibodies.¹¹⁸  For this reason, patients in early-phase clinical trials should be monitored for possible TLS, which should be treated appropriately. If observed, the occurrence and severity of TLS should be recorded in clinical trials using established criteria.¹¹⁹ 

Other nonhematological toxicities should be graded according to the latest version of the National Cancer Institute Common Terminology Criteria for Adverse Events. Some of the newer agents show a new spectrum of side effects that should be carefully monitored, including cardiac arrhythmias (for ibrutinib), autoimmune colitis (for idelalisib) or diarrhea (ibrutinib), bleeding (ibrutinib), and autoimmune pneumonitis (idelalisib).

Clear and careful reporting of data are essential parts of any clinical trial. In clinical studies involving previously treated patients, patients who are relapse, or are refractory should be clearly distinguished. Relapse and refractory disease are defined previously (sections 5.7 and 5.8). For those patients who have relapsed, it is also useful to describe the quality and duration of their prior response.

Given the recent increase of treatment options for CLL patients, the choice of treatment and the end points of clinical trials may depend on the fitness of the patients (see section 3.1). For example, the number of MRD-neg complete remissions or the overall survival might be appropriate end points in physically fit patients. In contrast, trials in patients with reduced physical fitness might choose the time to progression or health-related quality of life as trial end points. Moreover, the quality of life in patients with CLL may be reduced compared with that seen in the normal population and only moderately increased by some of the current treatment options.^120-122  Further studies assessing the health-related quality of life in CLL are encouraged.

While patients are under myelosuppressive chemo(immuno)therapy, growth factors such as granulocyte-colony-stimulating factor (G-CSF) should be given according to the guidelines of the American Society of Clinical Oncology.¹²³  G-CSF might also benefit patients who experience prolonged cytopenias following treatment with alemtuzumab. Similarly, some patients with anemia may benefit from erythropoiesis-stimulating factors, if used according to recently published guidelines.¹²⁴  However, CLL-related cytopenias are often efficiently corrected by an appropriate antileukemic therapy.

Infections are frequent problems during management of CLL patients. Excellent reviews regarding their prevention and therapy have been published recently.^125-127  Unfortunately, there are no randomized studies showing that vaccination may alter infection rates or outcomes from acquired infections in CLL. It is generally recommended that routine vaccinations be performed before initiation of treatment if possible. Vaccinations achieve reasonable rates of seroprotection and seroconversion in immunocompromised cancer patients, with minimal side effects.¹²⁷  Conjugate vaccines have proved to be highly immunogenic and are to be preferred, where available, in CLL patients.¹²⁸  Vaccines against seasonal influenza and H1N1 can be recommended, given the severity of the H1N1 pandemic and the highly severe flu impact in immunocompromised CLL patients.¹²⁹  Live vaccines are contraindicated in CLL patients because severe or even fatal complications have been reported.¹²⁵ 

Hypogammaglobulinemia (low serum levels of IgG and IgA with variable IgM) is a well-recognized complication associated with CLL. Regarding the substitution of CLL patients with hypogammaglobulinemia and history of infections, 6 randomized studies have shown that the prophylactic use of intravenous immunoglobulins decreases the rate of bacterial infections and prolongs the time to first infection, but does not produce differences in survival or other outcome parameters (summarized in Sánchez-Ramón et al¹²⁵ ). Therefore, the use of intravenous immunoglobulin cannot be routinely recommended, but should be reserved to individual situations of hypogammaglobulinemia and repeated infections.

The relationship between CLL and autoimmune cytopenias is well established. The potential mechanisms, particularly the role of leukemic cells in stimulating the production of polyclonal autoantibodies, are increasingly understood.¹³⁰  Autoimmune thrombocytopenia (ITP) and autoimmune hemolytic anemia (AIHA) as a single abnormality caused by CLL initially should be treated with glucocorticoids and not initially with chemotherapy or targeted agents. Second-line treatment options for AIHA include rituximab, splenectomy, intravenous immunoglobulins, and/or immunosuppressive therapy with agents such as cyclosporine A, azathioprine, low-dose cyclophosphamide, or alemtuzumab.^131-133  Some ITP patients not responding to glucocorticoids may benefit from rituximab, immunosuppressive agents (eg, mycophenolate), or thrombopoietin analogs.¹³⁴  Refractoriness of autoimmune cytopenias to therapy is an indication for treatment directed at the underlying CLL.¹³⁵  In this regard, the Binet or Rai staging systems do not distinguish between ITP/AIHA or bone marrow infiltration as the cause for anemia or thrombocytopenia that results in classifying a patient as having stage C or high-risk disease.

The production of these guidelines is the result of a permanent process of consolidation with the goal of integrating novel findings and concepts into current practice and future clinical trials; therefore, we wish to give credit to all the colleagues that have participated to create earlier versions of the guidelines.^1-3 

This work was supported by the Kompetenznetz Maligne Lymphome (Competence Network Malignant Lymphoma) and the German Research Council (DFG) (clinical research unit, CRU, 286) (M.H.).

Contribution: All authors debated on all recommendations until a consensus was reached and all authors wrote the article; and M.H. provided the first draft, organized the consensus meetings, and compiled the last version of the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Michael Hallek, Klinik I für Innere Medizin, Universität zu Köln, Joseph-Stelzmann Str 9, 50924 Köln, Germany; e-mail: michael.hallek@uni-koeln.de.