B-Cell Leukemias of Mature Lymphocytes

Morphology

The peripheral blood smear of patients with CLL is characterized by a variable lymphocytosis that may uncommonly reach over 500 × 10 ⁹ /L. The lymphocytes are typically small and round with condensed chromatin alternating with lighter areas, imparting a “soccer ball” or “cracked” chromatin pattern ( Fig. 12.1 ). The cytoplasm is usually scanty. Variation from this typical form is acceptable within the spectrum of CLL, and some cases may have substantial numbers (>10%) of cells with nuclear irregularity and/or moderate amounts of pale blue cytoplasm ( Fig. 12.2 ). Some studies have associated such variant morphology with worse outcome. Prolymphocytes are present in varying proportions. These cells are 1.5 to 2 times larger than the typical CLL cells with slightly open chromatin and a central nucleolus. Increased numbers of prolymphocytes (>5 × 10 ⁹ /L) have been associated with a poor prognosis and certain genetic alterations such as TP53 abnormalities and trisomy 12. Cases in which prolymphocytes constitute more than 10% but less than 55% of the lymphocytes have been termed mixed cell type CLL or CLL/PL in the original French-American-British classification scheme but are not distinguished from CLL in the 2016/17 updated World Health Organization classification. The bone marrow may show variable involvement. Four major patterns are recognized: nodular, interstitial, mixed, and diffuse ( Fig. 12.3 ). The nodules tend to be nonparatrabecular. The cells are similar in appearance to those seen in lymph nodes: small and round with condensed chromatin. Proliferation centers can be seen. In the interstitial pattern, the lymphocytes infiltrate around preserved fat spaces, admixed with varying amounts of residual hematopoietic elements. The diffuse pattern, with areas totally replaced by sheets of leukemia cells, has been associated with a poor prognosis and more advanced disease.

Transformation to PLL is defined by more than 55% of lymphocytes being prolymphocytes and is usually characterized by worsening symptoms, loss of response to therapy, and poor prognosis. Richter's syndrome (transformation to a diffuse aggressive B-cell lymphoma) occurs in approximately 3% to 4% of patients, and survival is short. Unusual transformations/second malignancies such as Hodgkin lymphoma (HL), acute lymphoblastic leukemia, and multiple myeloma have been reported. Of note, Reed-Sternberg (RS)-like cells have been shown in some cases to be positive for Epstein-Barr virus and clonally related to the CLL. HL transformation should be reserved for cases with tissue biopsy demonstrating areas with all the histologic features of HL and not used for cases with scattered RS-like cells in a background of SLL.

Immunophenotype

Flow cytometric immunophenotyping is required to confirm the diagnosis of CLL. CLL cells express CD5, CD19, CD20, CD22, CD23, CD200, and restricted surface immunoglobulin light chain. The expression of CD20 and immunoglobulin is usually dim compared to normal B cells ( Fig. 12.4 ). In occasional cases, the immunoglobulin light chain may be so dim as to be virtually undetectable by routine flow cytometry. Heavy chain expression is usually of the IgM and IgD type. FMC7 and CD79b are absent or only dimly expressed. Deviation from this typical immunophenotype occurs, and a scoring system has been proposed to help quantify the likelihood that the diagnosis is CLL ( Table 12.2 ). One caveat to the presence of restricted light chain expression is the existence of rare cases of biclonal CLL in which a kappa-restricted CLL clone and a lambda-restricted CLL clone are seen. Cases might seem polytypic but have the other immunophenotypic features of CLL and, on molecular analysis, show biclonal rearrangement patterns.

Expression of LEF1 is highly specific and sensitive for CLL and is useful in the differential diagnosis of small B-cell leukemias/lymphomas. A nuclear protein involved in WNT/beta-catenin signaling and expressed in normal T cells, LEF1 is most easily assessed by immunohistochemistry in bone marrow trephine or other tissue biopsies, although assessment by flow cytometry has been reported.

Two additional markers have been shown to be important in the prognosis of CLL. CD38 is a marker that initially was thought to correlate with IGH variable region (IGHV) mutational status (see following). The correlation has subsequently been proved to be imperfect, but it has been shown in numerous studies that CD38 expression (a commonly used cutoff is 30% of CLL cells) is a poor prognostic indicator in CLL. Several studies have also shown it to be independent of other clinical parameters. The intracellular nonreceptor tyrosine kinase ZAP-70 correlates fairly well with IGHV gene mutational status and can also be assessed by flow cytometry or immunohistochemistry. Expression in more than 20% of CLL cells is associated with unmutated IGHV genes and poor prognosis. Technical challenges, lack of standardization, and advances in molecular genetic markers have prevented widespread clinical use of these markers.

Molecular Genetics

Interphase fluorescence in situ hybridization (FISH) studies have revealed many recurrent genetic abnormalities that also have clinical importance ( Table 12.3 ). Deletion of 13q, the most common abnormality, is associated with typical morphology and good prognosis ( Figs. 12.5 and 12.6 ). Trisomy 12 is associated with atypical morphology and intermediate prognosis. Del6q and del11q22-23 have poor prognosis. Deletion of 17p involving TP53 is uncommon but associated with poor prognosis and poor response to conventional therapy. Assessment of CLL cases for these markers by FISH panels has become routine. More than any other biomarker in CLL, deletion of 17p is beginning to impact selection of therapy in individual CLL patients outside of clinical trials. Studies evaluating driver mutations, including these, copy number abnormalities in pretreatment and relapse samples show that del13q, trisomy 12, and del11q remain stably clonal over time, indicating that they are early genetic events.

IGHV mutational status has been shown to be an independent predictor of outcome in CLL. IGHV can be characterized as unmutated based on 98% or greater homology to the germline sequence. Patients whose CLL cells have unmutated IGHV genes generally have a poorer prognosis than patients whose CLL cells have mutated IGVH genes. Approximately 45% of CLL cases will have unmutated IGVH genes. As with FISH testing, IGHV mutational testing at diagnosis is recommended by some experts to inform clinical management decisions.

MicroRNAs (miRNAs) have become a focus of investigation in CLL. These small noncoding RNAs are important in regulation of gene expression. Two miRNA genes, miR-16 and miR-15 , have been mapped to the minimally deleted region in CLL at 13q14.3. These miRNAs control B-cell proliferation and appear to modulate BCL2 expression. Deletion in transgenic mice leads to monoclonal B-cell lymphocytosis and CLL-like disease.

The categorization of CLL cases based on shared B-cell receptor (BCR) structural patterns at the protein level, derived from analysis of complementarity determining region 3 (CDR3) amino acid sequences, is termed BCR-stereotypy . Subsetting of CLL cases into one of at least major 19 stereotypic subsets (accounting for perhaps 30% of CLL cases) provides potential information on antigens recognized by these CLL clones and on the possibility of antigen drive. Correlations between stereotypes and other genetic abnormalities, biological features such as signaling pathways, and prognosis are being identified. How these may inform prognosis and treatment decisions remains an open question.

Massively parallel sequencing studies evaluating whole exomes and genomes for CLL samples have begun to define the mutational landscape and patterns of clonal evolution in this disease. Although integrative analysis of these complex data and determination of the clinical relevance in daily practice are ongoing challenges, there are emerging themes. First, there is a limited set of mutations in CLL, and studies have suggested a mean of 15 nonsynonymous mutations per patient. Among the more commonly mutated genes are SF3B1 , ATM , TP53 , BIRC3 , and NOTCH1 . Second, although there is a large list of mutated genes, driver mutations (those that are above background mutation rates and felt to be biologically relevant to leukemogenesis and progression) may be present in only 30 or so genes. These genes can be organized into a limited set of pathways or functional groups. Most driver mutations associate with the unmutated IGHV subtype (poor prognosis) of CLL, with the exception of del(13q), MYD88 , and CHD2 , which appear to segregate more with mutated IGHV CLL. Third, clonal evolution can occur at relapse and is often influenced by treatment. When sensitive methods are used, relapse driver mutations can be detected in the pretreatment samples in approximately 50% of cases. Fourth, presence of subclonal driver mutations in pretreatment samples is associated with shorter progression-free survival independent of other risk factors. Table 12.4 shows some of the more commonly mutated genes, functional pathways, and approximate frequencies. This is a rapidly changing area, and the interested reader may refer to key references for this chapter.

TABLE 12.4

CLL Mutations

Adapted from Puente XS, Pinyol M, Quesada V, et al: Whole-genome sequencing identifies recurrent mutations in chronic lymphocytic leukaemia. Nature 475:101-105, 2011.

Open full size image

^a Approximate frequency.