The Ph chromosome abnormality is present in more than 90% of patients with typical CML (Fig. 101-1). It results from a balanced reciprocal translocation between the long arms of chromosomes 9 and 22, t(9;22)(q34;q11.2). It is found in hematopoietic cells but not in other human cells. Its origin is close to the pluripotent stem cell, and it is present in erythroid, myeloid, monocytic, and megakaryocytic cells, less commonly in B lymphocytes, rarely in T lymphocytes, but not in marrow fibroblasts. The breakpoint of chromosome 9 results in translocation of the cellular oncogene ABL1 to a region on chromosome 22 coding for the major breakpoint cluster region (BCR). ABL1 is a homolog of V-ABL, the Abelson virus that causes leukemia in mice. This juxtaposes a 5′ portion of BCR to a 3′ position of ABL1 and produces a new hybrid oncogene, BCR-ABL. This codes for a novel BCR-ABL oncoprotein of molecular weight 210 kDa (p210^BCR-ABL). The p210^BCR-ABL oncoprotein results in uncontrolled kinase activity, which causes excessive proliferation and reduced apoptosis of CML cells,^13–13 leads to a growth advantage of CML cells to normal cells, and suppresses normal hematopoiesis. The normal stem cells, although suppressed, persist and reemerge after effective therapy of CML.

Figure 101-1 The Philadelphia (Ph) chromosome, initially described as a minute chromosome, was later identified as a balanced reciprocal translocation between the long arms of chromosomes 9 and 22, t(9;22) (q34;q11.2). This results in the translocation of the *ABL1* gene from chromosome 9 in proximity to the breakpoint cluster region (*BCR*) on chromosome 22. Depending on the breakpoint site on chromosome 22, three resultant BCR-ABL oncoproteins are generated: (1) P210^BCR-ABL, which occurs in the large majority of patients with Ph-positive chronic myeloid leukemia (CML); (2) P190^BCR-ABL, present in two thirds of patients with Ph-positive acute lymphocytic leukemia (the other one third have the P210^BCR-ABL product); and (3) P230^BCR-ABL, which is found rarely (1% to 2%) in patients with an indolent course of Ph-positive CML. The three breakpoints on chromosome 22, M-Bcr (P210), m-bcr (P190), µ-bcr (P230), which correspond to the three resultant BCR-ABL oncoproteins, are shown.

In Ph-positive acute lymphocytic leukemia, the breakpoint in BCR often occurs in a more centromeric region called the minor BCR region (mBCR). This produces a smaller BCR gene opposing ABL1, and therefore a fusion gene, messenger RNA, and BCR-ABL oncoprotein (p190^BCR-ABL) of smaller sizes. A third rare breakpoint distal to the major BCR region, called µ-BCR, produces a p230^BCR-ABL hybrid oncoprotein associated with a very indolent course of CML (see Fig. 101-1).¹⁴

The constitutive activation of BCR-ABL results in autophosphorylation and activation of downstream pathways that alter gene transcription, apoptosis, cytoskeletal organization, cytoadhesions, and degradation of inhibitory proteins. These signal transduction pathways involve RAS, mitogen-activated protein (MAP) kinases, signal transducers and activators of transcription (STAT), phosphatidylinositol-3-kinase (PI3K), MYC, and others. Many of these interactions are mediated through tyrosine phosphorylation and require binding of the BCR-ABL to adapter proteins such as GRB-2, CRK, CRK-like protein (CRKL), and Src homology–containing proteins (SHC). Understanding the pathophysiology of these events could result in the successful development of targeted therapies that may synergize with TKIs to improve further the prognosis in CML.

What causes the BCR-ABL molecular rearrangement is unknown. Molecular techniques that amplify detection of BCR-ABL to 1 in 10⁸ detect it in marrow cells of 25% to 30% of normal adults, in 5% of infants, but not in cord blood.¹⁵ Because CML develops in only 1.5 of 100,000 individuals (i.e., 1 to 2 per 25,000 to 30,000 individuals who express BCR-ABL in their bone marrow), additional molecular events or lack of immune recognition of the clonal cells may contribute to the development of CML.

The fusion BCR-ABL gene and the p210 oncoprotein can be found in some patients with typical morphologic CML, in whom the t(9;22)(q34;q11.2) is not identified. These patients have a response to therapy and a survival rate similar to those with Ph-positive cases. Patients with Ph-negative and BCR-ABL-negative CML (see later discussion) constitute a different entity (atypical CML) and have a worse prognosis.16,17 The molecular pathophysiology of CML transformation is poorly understood. Several molecular events have been associated with CML transformation, including point mutations or deletions in the TP53 tumor suppression gene, MYC amplification, deletions in the CDKN2A tumor suppressor gene, alteration of the RB retinoblastoma gene, and others. The causal relation of these events to transformation is not clear.

Animal Models of Chronic Myeloid Leukemia

Experimental models have established a causal relationship between the BCR-ABL molecular events and the development of CML. Transgenic mice that express Bcr-Abl developed acute leukemia. A Bcr-Abl–expressing retrovirus used to infect murine bone marrow cells that later repopulate irradiated mice resulted in myeloproliferative disorders, including a CML-like syndrome.20–20 Bcr-Abl, under the control of a tetracycline-repressible promoter, expressed in mice, resulted in a lymphoid leukemia that reversed in the presence of tetracycline,²¹ attesting to the leukemic potential of Bcr-Abl as a sole oncogenic abnormality. The recapitulation of CML-like disorders in animal models mediated by the Bcr-Abl molecular events established them as a root cause for CML and a legitimate molecular target for therapeutic interventions with TKIs.

Clinical Presentation

Chronic Phase

About 50% to 60% of patients with CML diagnosed in the United States are asymptomatic. The disease is found on routine physical examination or blood tests.

Common signs and symptoms of CML, when present, result from anemia and splenomegaly. These include fatigue, weight loss, malaise, early satiety, and left upper quadrant fullness or pain (Table 101-1). Rare manifestations include bleeding (associated with a low platelet count and/or platelet dysfunction), thrombosis (associated with thrombocytosis and/or marked leukocytosis), gouty arthritis (from elevated uric acid levels), priapism (usually with marked leukocytosis or thrombocytosis), retinal hemorrhages, and upper gastrointestinal ulceration and bleeding (from elevated histamine levels due to basophilia). Leukostatic symptoms (dyspnea, drowsiness, loss of coordination, confusion) caused by sludging in the pulmonary or cerebral vessels, are uncommon in chronic-phase CML despite white blood cell (WBC) counts exceeding 100 × 10⁹ cells/L.

Table 101-1

Features of Patients with Newly Diagnosed Philadelphia Chromosome–Positive Chronic Myelogenous Leukemia in Chronic Phase Referred to MD Anderson Cancer Center (1990–present; N = 1469)

Parameter	Percentage
Age ≥ 60 years (median)	18 (46)
Female gender	41
Splenomegaly	32
Hepatomegaly	5
Lymphadenopathy	5
Other extramedullary disease	2
Hemoglobin < 10 g/dL	12
Platelets
>450 × 10⁹ cells/L	31
<100 × 10⁹ cells/L	3
WBC ≥ 50 × 10⁹ cells/L	38
Marrow
≥5% blasts	6
≥5% basophils	14
Peripheral blood
≥3% blasts	9
≥7% basophils	10
Cytogenetic clonal evolution other than the Philadelphia chromosome	4
Sokal risk
Low	62
Intermediate	27
High	10

Splenomegaly is the most consistent physical sign in CML and is detected in 30% to 50% of cases. Hepatomegaly is less common (10% to 20%). Lymphadenopathy, infiltration of skin or other tissues, is uncommon. When present, these findings may suggest accelerated or blastic phases of CML or Ph-negative CML. Headaches, bone pain, arthralgias, pain from splenic infarction, and fever are more frequent with CML transformation.

Laboratory features of untreated CML include leukocytosis with predominance of neutrophils and a left shift extending to blast cells. Basophils and eosinophils might be increased. Thrombocytosis is common; thrombocytopenia is rare and, if present, suggests a worse prognosis. Anemia (hemoglobin < 11 g/dL) is present in one third of patients. Biochemical abnormalities include a low leukocyte alkaline phosphatase score, which also occurs in some patients with agnogenic myeloid metaplasia. Serum levels of vitamin B₁₂, lactate dehydrogenase, uric acid, and lysozyme are often increased. Some patients demonstrate a cyclic oscillation of the WBC count.

The bone marrow is hypercellular with marked myeloid hyperplasia. The myeloid-to-erythroid ratio is usually 15 : 1 to 20 : 1. About 15% of patients have 5% or more blast cells in the peripheral blood or bone marrow at diagnosis. Increased reticulin fibrosis is common (30% to 40% grade 3/4 reticulin fibrosis by silver staining), but collagen fibrosis is rare at presentation.²² Interestingly, whereas the “spent phase” of myelofibrotic CML was commonly reported in the past (with busulfan therapy) as an end-stage CML event, it has become uncommon in the eras of IFN-α and imatinib therapy. Imatinib effectively reduces myelofibrosis in CML.^23,24

Accelerated and Blastic Phases

The definitions of accelerated and blastic phases of CML are shown in Table 101-2. In most patients the transformation from chronic to advanced phase is insidious, with the disease becoming more difficult to control. This phase is referred to as the accelerated phase. The criteria of accelerated phase are variable. In one study, features that correlated with a median survival of 18 months or less included a blast percentage equal or greater than 15%, blasts plus promyelocytes equal or greater than 30%, basophils equal or greater than 20%, a platelet count less than 100 × 10⁹ cells/L unrelated to therapy, and cytogenetic clonal evolution.²⁵ Features of accelerated-phase CML should be revisited because some (e.g., clonal evolution more favorable, blasts >15% less favorable) have different prognostic implications from recent studies (Fig. 101-2).²⁶ Five to 10 percent of patients are seen in the accelerated phase; these patients have an estimated 8-year survival rate of 75% when treated frontline with TKIs.²⁷ The accelerated phase is also associated with worsening anemia and splenomegaly, organ infiltration (liver, lymph nodes, skin, bones, or other tissues), and constitutional symptoms (aches, fever, malaise, weight loss).

Table 101-2

Definitions of the Accelerated and Blastic Phases of Chronic Myelogenous Leukemia

Criteria	MD Anderson Cancer Center*	International Bone Marrow Transplant Registry	World Health Organization
ACCELERATED PHASE
Percent blasts	15-29	10-29	10-19
Percent blasts + promyelocytes	≥30	≥20	NA
Percent basophils	≥20	≥20% (basophils + eosinophils)	≥20
Platelets (×10⁹/L)	<100	Unresponsive ↑, persistent ↓	<100 or > 1000 unresponsive
Cytogenetics	CE	CE	CE not at diagnosis
WBC	NA	Difficult to control, or doubling in < 5 days	NA
Anemia	NA	Unresponsive	NA
Splenomegaly	NA	Increasing	NA
Other		Chloromas, myelofibrosis	Megakaryocyte proliferation, fibrosis
BLASTIC PHASE
30% or more blasts; or extramedullary blastic disease, except for the World Health Organization classification, which requires 20% or more.

CE, clonal evolution; NA, not applicable; WBC, white blood cells.

^*These criteria were also used in all the IFN-α and tyrosine kinase inhibitors studies.

Figure 101-2 Progression-free survival (PFS) in CML by different characteristics, highlighting that clonal evolution (CE) at diagnosis is not associated with adverse prognosis. CE during the course of CML has a slight adverse prognostic impact but cannot be considered an accelerated phase feature because the associated estimated 4-year PFS rate is 68%, similar to that of chronic phase. Similarly, thrombocytosis (>1000 × 10⁹ cells/L) is associated with an estimated 4-year PFS rate of 84%. The estimated 4-year PFS rate in accelerated phase (classic standard criteria) is 45%. The figure also highlights that although a blast percentage of 20% to 29% is associated with an adverse accelerated phase like prognosis, it cannot be considered a blastic-phase feature (as proposed by the World Health Organization) because it seems to be closer in outcome to accelerated than to blastic phase. Patients who are seen with 10% to 14% blasts have an adverse prognosis similar to those with accelerated-phase disease. Therefore, the proposed new criteria of accelerated phase may exclude clonal evolution but include patients with blast percentages greater than 10%.

Blastic-phase CML is diagnosed by the presence of 30% or more blasts in the bone marrow and/or peripheral blood or by the presence of extramedullary blastic disease. Blastic-phase CML resembles acute leukemia. Patients may experience fever, bone aches, bleeding, infections, weight loss, and increasing splenomegaly. Approximately 70% of patients have a myeloid or undifferentiated blastic phase and 30% have a B-cell lymphoid blastic phase.

In most patients the CML evolves into the accelerated phase before the blastic phase, but in 20% of patients transformation to a blastic phase may occur without warning. Most patients in accelerated- or blastic-phase CML have additional chromosomal abnormalities, such as a double Ph, trisomy 8, or isochromosome 17. Extramedullary blastic-phase CML can occur in the lymph nodes, skin, meninges (especially in lymphoid blastic-phase CML), bone, and other sites.

Diagnosis

The diagnosis of typical CML is simple and consists of documenting, in the setting of persistent unexplained leukocytosis (or occasionally thrombocytosis), the presence of the Ph chromosome t(9;22)(q34;q11.2) by routine cytogenetics or of the corresponding BCR-ABL rearrangement by fluorescent in situ hybridization (FISH) analysis or molecular studies.

A FISH analysis relies on the colocalization of large genomic probes specific to the BCR and ABL1 genes. Comparison of simultaneous marrow and blood samples by FISH analysis shows high concordance. FISH studies may have a false-positive range of 1% to 5%, depending on the probes used.

Reverse-transcriptase polymerase chain reaction (RT-PCR) amplifies the region around the junction between BCR and ABL1. It is highly sensitive for the detection of minimal residual disease. PCR testing can either be qualitative, providing information about the presence of the BCR-ABL transcript, or quantitative, assessing the amount of BCR-ABL transcripts. Qualitative PCR may be useful for diagnosis of CML; quantitative PCR (qPCR) is ideal for monitoring response to therapy and is usually performed by real-time PCR. Simultaneous peripheral blood and marrow PCR studies show a high level of concordance. False-positive and false-negative results can happen with PCR. False-negative results may be from poor-quality RNA or failure of the reaction or from atypical transcripts (e.g., b₃a₃); false-positive results can be due to contamination. A 0.5- to 1-log coefficient of variability in some samples can occur depending on testing procedures, sample handling, and laboratory experience.28,29

Because of the increasing importance of monitoring patient response to TKI by various methods including qPCR and FISH studies, it is important to measure these at diagnosis to ensure their positivity and quantitation. In 5% of patients with CML, variant breakpoints result in a false-negative qPCR test because the traditional probes used do not detect these variants. If the test is not performed at diagnosis in these patients, subsequent qPCR studies will suggest a “false” complete molecular response.

The Ph chromosome is usually present at diagnosis in 100% of metaphases, often as the sole abnormality. Between 10% and 15% of patients have additional chromosomal changes (clonal evolution) involving trisomy 8, isochromosome 17, additional loss of material from the second chromosome 22 (double Ph), or others.

Eighty-five percent of patients have a typical t(9;22) translocation; 5% have variant translocations, which can be simple (involving chromosome 22 and a chromosome other than chromosome 9) or complex (involving one or more chromosomes in addition to chromosomes 9 and 22). With imatinib, patients with Ph variants have a response to therapy and prognosis similar to those with Ph-positive CML.

Diagnostic and Monitoring Procedures

The improved rates of complete cytogenetic response and of molecular response with imatinib now require new techniques that measure these responses more accurately (rather than relying on evaluation of only 20 metaphases by routine cytogenetic studies), with less painful procedures (peripheral blood rather than marrow samples), and with methods that measure minimal disease below the level of detection by routine karyotypic analysis (molecular studies). FISH studies can assess rapidly the disease status in 200 cells; this assessment may be done in cells in interphase and can be performed on blood specimens. Real-time PCR studies usually measure the ratio of the abnormal message, BCR-ABL, to a control gene (e.g., ABL). Because this is variable by laboratory and with time, efforts are ongoing to standardize the measurement of BCR-ABL transcripts in relation to an international standard (IS).³⁰ Recent monitoring procedures in CML have emphasized replacing bone marrow studies (cytogenetics) with peripheral blood studies (FISH, qPCR) as more reliable and measuring deeper levels of response to TKIs. FISH studies are reliable in patients in complete cytogenetic response with a strong concordance between negative FISH studies and complete cytogenetic response. However, the concordance between the percent of Ph positivity by FISH and cytogenetic studies is not perfect. Recently, landmark studies have emphasized the importance of BCR-ABL transcripts of 10% or less as early indicators for response and long-term prognosis and that the finding of BCR-ABL transcripts of 1% or less (IS) (grossly equivalent to a complete cytogenetic response) is an important response criterion at 1 year or later. In general, BCR-ABL transcripts less than 10% (IS) are equivalent to a partial cytogenetic response with Ph-positive metaphases less than or equal to 35%; and BCR-ABL transcripts less than or equal to 1% (IS) are equivalent to a complete cytogenetic response.^31–34 A BCR-ABL transcript of less than or equal to 0.1% (IS) (approximately a 3-log reduction of disease from a standardized baseline) has been associated with a low risk of relapse and with favorable PFS. A negative qPCR analysis (undetectable BCR-ABL transcripts; 4.5-log reduction or more) may be technique dependent and is referred to as a complete molecular response.