Chronic lymphocytic leukemia patients with highly stable and indolent disease show distinctive phenotypic and genotypic features

B-cell chronic lymphocytic leukemia (CLL) is the most frequent leukemia in the Western world.¹  After decades of conservative approach, over the last years the overall management of patients with CLL has changed considerably on the basis of an improvement in the understanding of the biologic properties of the disease and of the newly developed therapeutic strategies.^2,3  The disease presents a heterogeneous clinical course, with some patients surviving for many years without requiring any specific therapy and others progressing despite aggressive treatment. Although 2 clinical staging systems offer important prognostic clues at the time of CLL diagnosis,^4,5  they fail to provide indications about disease heterogeneity and treatment requirements. At the present time, management of CLL poses 2 major issues. The first issue concerns the decisions of “if ”and “when” to begin treatment since (a) more frequent early diagnoses are made due to the growing use of routine blood tests; (b) in a considerable proportion of patients (about 30%) the diagnosis is made under the age of 60 years; (c) the biologic age of “elderly” people has improved dramatically; and (d) the overall median life expectancy has also increased substantially both in males and females. The second issue concerns the choice of the therapeutic strategy within the broad armamentarium available for CLL, ranging from conservative therapy to allografting procedures.

For these reasons, and in view of the evidence that treatment would have a greater likelihood of success for patients with progressive disease if implemented earlier, the possibility of recognizing reliable and reproducible prognostic markers in CLL is of utmost importance for an optimal clinical management. A number of biologic features that appear to have prognostic implications have been identified. These include the presence of an atypical morphology or immunophenotype,^6-8  the presence of p53 abnormalities,⁹  the genetic profile,¹⁰  and the presence of somatic mutations of the immunoglobulin (Ig) heavy (H) chain variable (V) genes.^11,12  Although with a lesser degree of concordance, the presence or absence on neoplastic B cells of the CD38 antigen has also been related to outcome.^11,13,14  In particular, the presence of IgV_H somatic hypermutation and CD38 expression identifies 2 main clinico-molecular variants of CLL, one of which (IgV_H mutation positive and CD38 negative) derives from post–germinal center cells and displays a favorable outcome, whereas the other type (IgV_H mutation negative and CD38 positive) is characterized by a relatively severe prognosis.^11,13,14 

In order to evaluate the role of a number of potential prognostic features in a series of CLL patients with unequivocally indolent disease, we have studied a series of patients with very stable disease observed at our institution (Università La Sapienza, Rome, Italy) over a follow-up period of 10 to 23 years and who never required treatment. We have characterized the peripheral blood cell morphology; the immunophenotypic profile of the leukemic and nonleukemic populations; the molecular status of the IgV_H genes, the BCL-6 gene, and p53 mutations; as well as the presence of cytogenetic abnormalities. The results obtained indicate that indolent CLL shows a distinct and reproducible biologic profile that should be used to identify at presentation patients with stable disease who are likely to have a very benign clinical course without treatment.

There were 20 patients, 12 males and 8 females, median age of 56 years at the time of presentation (range, 48-69 years), studied at a median of 12 years (range, 10-23 years) from diagnosis (Table 1). All patients gave their informed consent to the blood collection and to the biologic analyses included in the present study according to the Declaration of Helsinki. The diagnosis of CLL was based on the presence of more than 5000 lymphocytes/μL in the peripheral blood that expressed a conventional CLL immunophenotype (CD5/CD20⁺, CD23⁺, weak CD22⁺, weak sIg⁺, CD10^-) and morphology. According to the Rai staging system, 16 of 20 and 4 of 20 patients were classified, respectively, as stage 0 or I, while by the Binet staging system all patients were in stage A. Following diagnosis, all patients were seen at our institution every 3 months during the first year, every 4 months during the second year, every 6 months during the third year, and, subsequently, once a year. All patients underwent a physical examination, a full blood count with a differential analysis, and routine laboratory parameters. Infectious episodes were diagnosed according to clinical and bacteriologic investigations; second tumors were checked for based on clinical symptoms. All patients were subjected to a lymphocyte count every 3 months. During the clinical follow-up, a Herpes infection was documented in 2 patients (nos. 10 and 15), and in 2 patients (nos. 2 and 10) sporadic episodes of bacterial infections were reported. A second neoplasia was diagnosed in 5 of 20 patients and a paraneoplastic disorder in 2 of 20. The levels of serum Ig remained within the normal range in 12 of 16 tested patients. Autoimmune hemolytic anemia and autoimmune disorders, suspected on clinical and hematologic grounds and confirmed serologically, were never recorded. No patient progressed to a more advanced stage during the clinical follow-up. The disease remained stable over the years and no patient has required any form of antileukemic treatment.

At the time of the study, diagnosis and clinical staging were compared with the criteria recommended by the International Workshop on CLL.¹⁵ 

All biologic studies were carried out on cells freshly drawn from patients informed of the purpose of the work-up.

The peripheral blood counts were analyzed over time using the H3 and ADVIA instruments (Bayer, Milan, Italy). At the time of the study, the morphologic observation was carried out by 3 independent observers on May-Grunwald/Giemsa–stained peripheral blood films. The presence of small lymphocytes, medium lymphocytes, prolymphocytoid cells, and cleaved nuclei cells was expressed as a percent on 500 lymphoid cells observed. A morphologically typical CLL was defined on the basis of the presence of 90% or more of small plus medium lymphocytes and 10% or less of the other cell types. In all other cases, CLL was defined as morphologically atypical.⁶ 

Fresh peripheral blood mononuclear cells (PBMCs) were obtained after gradient separation (Nycomed Pharma AS, Oslo, Norway) and phosphate-buffered saline (PBS) washings. PBMCs (5 × 10⁵) were marked with 5μL of the following fluorescent-conjugated monoclonal antibodies: anti-CD5, -CD19, -CD20, -CD23, -CD22, -CD38, -CD3, -CD4, -CD8 (all from Becton Dickinson Immunocytometry Systems, San Jose, CA), and anti-κ and -λ Ig chains (Dako, Glostrup, Denmark). The intracytoplasmic staining of the Ig light chains κ and λ was evaluated by fluorescent-conjugated monoclonal antibodies (Dako) after fixation of cells by the Fix and Perm cell permeabilization method (Caltag Laboratories, Burlingame, CA). Phycoerythrin (PE)–conjugated CD38 was assessed in association with fluorescein isothiocyanate (FITC)–conjugated CD2. Marked cells were analyzed using a fluorescence-activated cell sorting (FACS) Calibur Flow Cytometer (Becton Dickinson). The results were processed by the CellQuest software system (Becton Dickinson).

DNA sequencing analysis of IgV_Hgenes.IgV_H gene rearrangements were amplified with a set of 6 V_H gene family-specific primers that hybridize to sequences in the V_H leader region in conjunction with a J_H primer mix in separate reactions for each V_H primer.^16-18  Briefly, 300 ng genomic DNA, 15 pmol of each primer, 300 μM deoxynucleoside triphosphates (dNTPs), 1.5 mM MgCl₂, and 0.5 U AmpliTaq polymerase (Taq gold; Perkin-Elmer, Norwalk, CT) were mixed with Perkin-Elmer commercial buffer in a final volume of 30 μL. Cycles (35) of denaturation (95°C), annealing (60°C), and extension (72°C) were performed in a thermocycler (Biorad I Cycler; Biorad, Hercules, CA). The amplified polymerase chain reaction (PCR) fragment was directly sequenced with appropriate primers using a commercially available kit (ThermoSequenase; Amersham Life Sciences, Amersham, United Kingdom). [α-³³P]-labeled terminator dideoxynucleotides (Amersham Life Sciences) were included in the sequencing mixture. Sequences were compared with the V BASE sequence directory (MRC Centre for Protein Engineering, Cambridge, United Kingdom) using the MacVector 6.0.1 software (Oxford Molecular Group, Oxford, United Kingdom) for comparison of the rearranged IgV genes with the most homologous germ-line sequences. Percentage of divergence from the most homologous germ-line sequence was calculated by 2 methods, that is, by counting the number of mutations between the 5′ end of framework region 1 (FR1) and the 3′ end of FR3 and by counting the number of mutations between the 5′ end of complementarity-determining region 1 (CDR1) and the 3′ end of FR3.

DNA direct sequencing of the BCL-6 gene. For DNA sequencing of BCL-6 5′ noncoding regions, a unique PCR product encompassing nucleotides +404 to +1142 was amplified by primers E1.21B (5′-CTCTTGCCAAATGCTTTG-3′) and E1.26 (5′-CACGATACTTCATCTCATC-3′).¹⁹  Briefly, 200 ng genomic DNA, 10 pmol of each primer, 100 μM dNTPs, 1.5 mM MgCl₂, and 0.5 U AmpliTaq polymerase were mixed with Perkin-Elmer commercial buffer in a final volume of 25 μL. Cycles (35) of denaturation (95°C), annealing (52°C), and extension (72°C) were performed in a thermocycler (Biorad I Cycler). PCR products were directly sequenced according to the sequencing procedures described above.

DNA direct sequencing of the p53 gene. DNA sequencing analysis of p53 exons 5 to 8 was performed as described above using previously reported primers.²⁰ 

Conventional karyotyping. Conventional chromosome analysis was performed on PBMC samples according to methods previously described²¹  using the following mitogens: phorbol myristate acetate (50 ng/mL) and lipopolysaccaride from Escherchia coli (100 mg/mL). Whenever possible, 20 to 30 metaphases were studied and karyotypes described according to the International System for Human Cytogenetic Nomenclature (ISCN).²² 

Interphase cytogenetics. Fluorescence in-situ hybridization (FISH) was performed on cells taken from the same samples that were used for cytogenetic analysis, according to the methods previously described.²¹  The following probes were used: 2 lambda EMBL3 clones (nos. 19 and 65), spanning an area of approximately 40 kb within the middle portion of the ataxia teleangectasia (ATM) gene at 11q22-23 (the frequency of 11q22-23 deletion in CLL using this probe is, in our experience, 16%); the C21 cosmid that recognizes DNA sequences at 13q14 between the Rb gene and the D13S25 marker, a region known to be deleted in approximately 40% of CLL²¹ ; a 6q21 probe prepared by B. Schlegelberger (Institute for Human Genetics, Kiel, Germany)²³  that detects 6q deletion in 7% of unselected CLL in our experience (A.C. and G.C., personal data, 2002); and a 17p13.3 cosmid recognizing p53 gene sequences, made available by F. Birg (Institute de Cancerologie et Immunologie de Marseille, Marseille, France) in the context of the Biomed I program, the “EU concerted action for cytogenetic diagnosis of haematologic malignancies.” The frequency of 17p deletion in lymphoid neoplasias analyzed in our laboratories using this probe is in the range of 3% to 26%. A chromosome 12–specific centromeric probe was purchased (Oncor, Gaithersburg, MD) to study trisomy 12 in interphase cells.

The hybridization protocol was described in detail in previous studies.²¹  Cohybridization of the test probe with a control probe was performed by dual-color FISH. To prevent false-positive results due to inefficient hybridization, signal screening was performed on slides with a high hybridization efficiency, having more than 80% interphase cells showing 2 signals with the control probe. The evaluation was performed on a fluorescence microscope (Nikon Italia, Florence, Italy); 200 cells with well-delineated signals were observed and images were captured with a couple-charged camera device (Cytovision; Nikon Italia). Using 5 normal control samples, less than 1% of the cells showed 3 signals (false trisomy) with all probes, and the cut-off point for recognition of trisomy was set at 2%. Using the 11q22-23, 13q14, 6q21, and 17p13 probes, the mean number of cells with 1 signal (false deletion) was 2.1%, 1.9%, 2.5%, and 2%, respectively, and the cut-off point for the identification of each deletion was set at 10%.

The median lymphocyte count at the time of diagnosis and at the time of the present study was 15.5 × 10⁹/L (range, 8.0-29.0 × 10⁹/L) and 19.4 × 10⁹/L (range, 9.0-76.0 × 10⁹/L), respectively; the median hemoglobin (Hb) value was 9.37 mM (range, 8.06-11.17 mM) and 8.50 mM (range, 7.44-9.80 mM), respectively; and the median platelet count was 200 × 10⁹/L (range, 177-326 × 10⁹/L) and 161 × 10⁹/L (range, 124-253 × 10⁹/L), respectively.

In all but 2 patients (nos. 1 and 9), the morphologic observation showed a “typical” CLL. The lymphoid cells were represented by small and medium lymphocytes with a constant and homogeneous volume (≥ 90% of cells), with scanty but visible cytoplasm and a characteristically clumped chromatin without evident nucleoli. A smaller proportion (≤ 10%) of lymphoid cells was larger or presented a visible nucleolus (prolymphocytoid cells) or an indented nucleous (cleaved nuclei cells) or had cytoplasmatic granules (large granular lymphocytes). Patients nos. 1 and 9 showed a proportion of cleaved nuclei cells more than 10% (14% and 29%, respectively). The mean presence of smudge (basket) cells was 98.2‰ ± 83.7‰ (median, 68‰; range, 8‰-307‰).

The phenotype of all patients was evaluated at the time of this study to confirm the diagnosis of CLL made many years earlier. A CD5/CD20 positivity (defined as more than 20% of cells positive) was found in all cases but 1 (no. 3). CLL cells were always positive for the CD23 antigen, while CD22 was weakly expressed. The analysis of the expression of Ig light chains showed that 17 of 20 cases presented membrane or intracytoplasmic positivity for the κ chain and only 3 of 20 for the λ chain. In all cases, the Igs were weakly expressed. The CD38 antigen was expressed in only one case (no. 3).

The peripheral blood T-cell subset distribution was evaluated in all cases. The percent and absolute number of CD3⁺, CD4⁺, and CD8⁺ cells are reported in Table 2. It is worth noting that the CD4/CD8 ratio was in all cases more than 1.

Conventional karyotype. The cytogenetic features of 19 of 20 CLL patients were initially investigated by conventional chromosome analysis on mitogen-stimulated cells²²  in short-term cultures. In 12 of 19 patients, 15 or more metaphases were observed, and in 8 of 12 patients normal metaphases were obtained. Cytogenetic alterations were detected in 4 patients (nos. 2, 8, 14, and 19). In patient no. 14, the t(1;20)(p13;q2.3-2.3) translocation was detected in only 1 of 15 metaphases. In patient no. 2, the t(2;18) (p11;q2.3-2.3) translocation, associated with follicular lymphoma and CLL,^24,25  was further investigated. The translocation did not involve the Igκ chain gene (on chromosome 2) or the BCL-2 gene (on chromosome 18) (data not shown). The alteration detected in patient no. 8, t(2;16)(p12;p13) in 4 of 15 metaphases, has so far never been reported. Patient no. 19 showed, in 13 of 15 metaphases analyzed, an 11q21-22 deletion, without involvement of the ATM gene (below) on 11q22-23 chromosome, or of the ALL1 gene (data not shown).

FISH. There were 17 patients (all but nos. 14, 16, and 20) analyzed by FISH for known genomic aberrations that may be associated with the clinical course of the disease.¹⁰  No patient showed 17p or 11q deletions, which are correlated with an unfavorable prognosis. Of the patients, 2 (nos. 4 and 8) presented a 12q trisomy, and 3 (nos. 1, 4, and 19) showed a 13q14 deletion, which is associated with good prognosis. Patient no. 4 showed both 12q trisomy and 13q14 deletion. The 6q and 8q24 deletions were never found. In 13 of 17 cases, none of the studied abnormalities was detected. The absence of genomic aberrations by interphase FISH has been correlated with good prognosis.¹⁰ 

IgV_Hgene mutation status. Amplification and sequencing of the rearranged V_H genes used by CLL cases yielded an in-frame V_HDJ_H rearrangement in all samples. In all cases, the frequency of mutations was calculated by 2 methods, that is, by counting the number of mutations between the 5′ end of FR1 and the 3′ end of FR3 or, alternatively, between the 5′ end of CDR1 and the 3′ end of FR3. By setting a cut-off value of more than 2% mutation rate, all but one (no. 15) of the CLL patients were found to carry somatically mutated V_H genes by both calculation methods used (Table 3). Similarly, by setting a cut-off value of 3% or more mutation rate, all but one (no. 15) of the CLL patients were scored as mutated when counting mutations between the 5′ end of FR1 and the 3′ end of FR3. Finally, with a 3% or more cut-off value and by counting mutations between the 5′ end of CDR1 and the 3′ end of FR3, all but 2 cases (nos. 15 and 18) were scored as mutated. The discrepant result obtained for case 18 when applying the 3% or more cut-off value is due to a relative excess of IgV_H mutations falling within the FR1 in this case (Table 3).

In each individual case harboring V_H mutations, the mutation rate varied slightly depending on the method used to calculate mutations. In particular, by counting mutations between the 5′ end of FR1 and the 3′ end of FR3, the average mutation rate in patients harboring somatic hypermutation was 7.006 ± 2.928 bp (median, 6.820 bp; range, 3.34-13.4 × 10^-2/bp). Conversely, by counting mutations between the 5′ end of CDR1 and the 3′ end of FR3, the average mutation rate in patients harboring somatic hypermutations was 7.775 ± 3.496 (median, 7.690 bp; range, 2.96-14.1 × 10^-2/bp).

The V_H gene families used in the 20 respective patients were as follows: V_H3 in 7 patients (nos. 3, 4, 5, 7, 11, 16, and 19), V_H4 in 5 (nos. 6, 12, 14, 17, and 20), V_H2 in 4 (nos. 1, 2, 8, and 10), V_H1 in 3 (nos. 9, 13, and 18), and, finally, V_H5 in 1 (no. 15). The specific V_H genes used are detailed in Table 3. Notably, no case displayed the V_H3-21 or the V_H1-69 genes, which have been associated with mutated CLL with a relatively poor prognosis and with unmutated CLL, respectively.^26,27 

BCL-6 gene mutations. Mutations of the BCL-6 gene were detected in 12 of 20 CLL patients studied (60%). The average frequency of mutation was 0.18 ± 0.13 bp (median, 0.14 bp; range, 0.07-0.47 × 10^-2/bp; Table 3). All patients with mutated BCL-6 presented a frequency of IgV_H mutations more than 2 and also more than 3. The only patient (no. 15) without IgV_H mutations did not present BCL-6 mutations, in agreement with previous reports.²⁸ 

P53 gene mutations. DNA direct sequencing analysis investigated p53 exons 5 to 8. No mutations were recorded in 20 of 20 cases evaluated (Table 3).

Numerous studies have independently evaluated in CLL patients several of the biologic parameters investigated in the present study and often associated them with a favorable or unfavorable prognostic likelihood. We have taken a different approach and have identified a series of CLL patients followed at a single institution (Chair of Hematology, University La Sapienza), who over a period of 10 to 23 years from diagnosis have shown a highly stable and indolent disease that has never required treatment. The main question addressed was the following: what are the phenotypic, cytogenetic, and genotypic features of such a unique and uniform series of stable patients? Do they display univocal distinctive biologic features that can in the future be used in prospective studies to identify such patients upfront at the time of presentation? In view of the growing evidence of the clinical and biologic heterogeneity of CLL patients, and of the always greater therapeutic armamentarium available, this represents a relevant issue in order to distinguish at diagnosis patients who may benefit from early treatment from those for whom treatment would be detrimental.

Clinically, all patients maintained over the years an indolent hematologic picture and the stage remained unchanged: they were all stage A according to Binet et al⁵  and 16 patients were stage 0 and 4 stage I according to Rai et al.⁴  Immunocompetence plays a central role in CLL, which is characterized by a progressive and marked hypogammaglobulinemia,²⁹  the presence of multiple defects within the T- and cytotoxic cell compartments,^30-32  and by a diminished response to antigenic stimuli.²⁹  This has important clinical implications, since about 80% of CLL patients witness during the course of their disease infective complications.³³  It is thus of relevance that in the clinical history of our patients, only 3 (15%) developed recurrent infective episodes. This is most likely a result of normal levels of serum Ig observed in 75% of cases. In addition, all patients had a normal or near-normal T-cell subset distribution (CD4/CD8 ratio more than 1), which suggests an effective immunologic response to infections. These findings differ substantially from the low levels of serum Ig and multiple CD4, CD8 defects that normally occur in CLL patients.^30-32  It is also realistic to hypothesize that in these very stable CLL patients an important immunologic control of the disease mediated by residual T cells may be operational. It has been reported that autologus T lymphocytes are able to specifically recognize the tumor-derived V_H-CDR3 region in CLL patients who presented in the neoplastic cells somatic mutations of the IgV_H genes,³⁴  as in our case series. It should also be underlined that no patient has developed in the 10 to 23 years of clinical follow-up autoimmune episodes, which represent a frequent and often life-threatening complication of CLL patients.³⁵  Once again, the evidence of the preserved T-cell competence may help to explain the lack of autoimmune complications.

The first indications that the heterogeneous clinical picture of CLL patients could have some biologic clues came from the suggestion that cases with an atypical morphologic and/or immunophenotypic profile of the leukemic cells could be associated with a more aggressive clinical picture.^7,8,36  We have addressed this issue in our series that indeed displayed in all but 2 cases a typical morphologic picture and in all but 1 a characteristic immunophenotypic profile, the latter case being CD5^-. As expected,³⁰  the large majority of cases (17/20) was Igκ⁺.

More recently, it has been documented that given karyotypic and genetic abnormalities are strongly associated with an unfavorable prognosis. These are mainly represented by the presence of more than 3 clonal abnormalities (complex karyotype), by the presence of only abnormal metaphases (AA karyotype),³⁷  and by 17p and 11q deletions,¹⁰  as well as by p53 abnormalities.⁹  All of our cases were negative for 11q and 17p deletions, negative for p53 mutations, and did not show complex karyotype aberrations. All in all, no karyotypic anomalies were recorded in 13 of 17 cases investigated. In addition, the 13q14 deletion, which has been associated with a more favorable prognosis, was found in 3 of 4 cases with cytogenetic abnormalities.

Great attention has recently been generated by the indication that 2 markers—the mutational status of the IgV_H genes and the expression of the CD38 antigen on CLL cells—may bear important prognostic implications.^11-13  While there is good agreement on the prognostic role of the IgV_H mutation status, there is less concordance on the impact of CD38 expression and on its correlation with the IgV_H status. A recent study by Hamblin et al¹⁴  showed that CD38 expression and IgV_H mutations gave discordant results in 28.3% of patients analyzed, and that in multivariate analysis they represented independent prognostic indicators. The researchers concluded that only IgV_H mutations are intrinsic to the nature of the disease and do not change during its course. In our series, all cases but one had a mutated profile of the IgV_H genes when using the conventional cut-off value of a 2% or more IgV_H mutations calculated on the entire V_H gene sequence (ie, from the 5′ end of FR1 to the 3′ end of FR3). Also, in no case but one did the leukemic cells express CD38.

Very recently, the operational parameters for defining mutated versus unmutated CLL have become a matter of debate.^38-40  In particular, it has been suggested that a cutoff value of 3% or more mutation rate may be more predictive of outcome than the conventional cutoff value of 2% or more. Also, it has been realized that the mutation rate of an individual CLL case may vary depending on the portions of the V_H gene considered for mutational analysis. We have, therefore, assessed the IgV_H mutational status of our highly stable CLL panel by using 2 different calculation methods applied uniformly to all cases. One calculation method considered the entire V_H gene from the 5′ end of FR1 to the 3′ end of FR3, whereas the second method considered only the V_H gene portion included between the 5′ end of CDR1 and the 3′ end of FR3. Also, both cutoff values (≥ 2% and ≥ 3%) previously proposed in the literature have been applied to our CLL series. Overall, the results indicate that virtually all highly stable CLL may be scored as positive for IgV_H mutations independent of the calculation method used and of the cutoff value chosen for the analysis.

With respect to the IgV_H genes used by highly stable CLL, we failed to identify specific genes positively associated with this disease type. However, it is remarkable that the V_H3-21 gene, which identifies a subset of mutated CLL with poor outcome,²⁶  was never found in our series. Also, as expected, the V_H1-69 gene associated with unmutated CLL cases,²⁷  did not occur in our leukemias. Finally, BCL-6 mutations were found in a high proportion of cases (60%). Mutations of the BCL-6 proto-oncogene have been associated with the presence of IgV mutations and appear associated with the postgerminal nature of the mutated CLL cases,⁴¹  although the true histogenetic value of these mutations in the context of CLL is uncertain.⁴² 

The attitude toward the management of CLL patients is changing very rapidly. This is due to a number of considerations, which include, among others, the increasing frequency of cases diagnosed under the age of 60 years (young CLL); the knowledge of the clinical heterogeneity of this condition; the broad range of therapeutic approaches today available, including transplantation procedures; and the evidence that biologic features that may bear important prognostic implications can at last be identified. This is progressively leading to a reconsideration of the watchful waiting policy normally adopted in CLL. While early treatment intervention has not proved valuable as an overall approach for early-stage patients,^43,44  it may prove beneficial if implemented early in patients with well-defined adverse prognostic features. Conversely, treatment would be detrimental for patients with recognizable favorable prognostic features. Based on the results of our study, patients with a long-lived indolent disease and a clinical course similar to that of age-matched controls not suffering from CLL display a distinct biologic profile characterized by a typical morphology and immunophenotype, a normal T-cell subset distribution and frequent normal serum Ig levels, lack of CD38 expression, absence of poor prognosis genetic markers, and a mutated profile of the IgV_H regions, with no Ig V_H3-21 involvement. If broadly used at diagnosis, such extended characterization should allow the identification of patients with a very favorable prognostic likelihood who should not undergo treatment, even at an early age. Although some analyses, namely the cytogenetic and genetic investigations, are to date not routinely performed, the presence of a typical morphology and phenotype, and a normal CD4/CD8 T-cell ratio and levels of serum Ig, in the absence of CD38 expression, may help to identify indolent CLL patients. Based on the results of this study, we are currently submitting all newly diagnosed CLL patients 65 years or younger to this extended biologic work-up in order to verify prospectively the impact of this approach.

At a time when more aggressive therapeutic strategies are progressively being used in CLL, the distinct features of patients with indolent disease should be taken into account. Prospective studies aimed at investigating the impact of early aggressive intervention need to include in their pretreatment work-up an extended biologic characterization. This in turn would allow the design of an innovative scoring system based on biologic markers of prognostic relevance.