• ~150 new cases of all MPN per year in the entire Hong Kong population ^[6]
• Disease incidence increases with age; median age 65–70 for MPN overall ^[6]
• CML is not a common disease but has a stable incidence

An important epidemiological point: since the introduction of tyrosine kinase inhibitors (TKIs), patients with CML now have near-normal life expectancy. This means that while the incidence remains stable, the prevalence is steadily rising — there are more and more patients living with CML in the community.

To understand CML, you need to understand where the disease originates:

1. Pluripotent haematopoietic stem cell (HSC) in the bone marrow gives rise to:

   • Common myeloid progenitor (CMP) → granulocytes (neutrophils, eosinophils, basophils), monocytes, erythrocytes, megakaryocytes/platelets
   • Common lymphoid progenitor (CLP) → B cells, T cells, NK cells

2. Granulopoiesis — the normal maturation sequence:

   • Myeloblast → Promyelocyte → Myelocyte → Metamyelocyte → Band → Segmented neutrophil
   • This process takes ~10–14 days and is tightly regulated by growth factors (G-CSF, GM-CSF) and intracellular signalling cascades

3. The spleen — relevant because:

   • Acts as a filter for senescent/abnormal blood cells
   • Can become a site of extramedullary haematopoiesis when the bone marrow is overwhelmed or fibrosed
   • CML cells infiltrate the spleen → massive splenomegaly ^[3]

• ABL1 (Abelson murine leukaemia) is a ubiquitously expressed non-receptor tyrosine kinase normally involved in:
  • Cell differentiation
  • Cell division
  • Cell adhesion
  • DNA damage response/apoptosis
• Under normal conditions, ABL1 activity is tightly regulated by intramolecular autoinhibition
• The key point: when ABL1 is fused with BCR, this autoinhibition is lost → constitutive (always-on) kinase activity

• CML arises from a single pluripotent HSC → therefore the Philadelphia chromosome is found in cells of all myeloid lineages (granulocytes, erythrocytes, megakaryocytes, monocytes) and sometimes even in B lymphocytes
• However, the clinical manifestation is predominantly granulocytic proliferation
• This explains why CML can transform into either AML (~2/3 of blast crises) or ALL (~1/3 of blast crises) — because the original transformed cell is pluripotent

CML is considered a type of MPN but is uniquely defined by the presence of the BCR-ABL1 fusion gene (which is not present in ANY other type of MPN) ^[4][6].

The other BCR-ABL1 negative MPNs include:

• Polycythaemia vera (PV) → JAK2 V617F (100%)
• Essential thrombocythaemia (ET) → JAK2 V617F / CALR / MPL
• Primary myelofibrosis (PMF) → JAK2 V617F / CALR / MPL

"So unlike BCR-ABL and CML, there are no defining markers for [the other MPNs]" ^[6]

All forms of MPN share the potential to progress to myelofibrosis and blastic transformation ^[6].

Chronic leukaemia will transform into acute leukaemia if left untreated, since leukaemic cells will acquire more and more mutations over time ^[3].

The concept of clonal evolution:

• The original BCR-ABL1+ clone is genetically unstable
• Over time, additional mutations accumulate (e.g., TP53 loss, RB1 deletion, additional copies of Ph, trisomy 8, isochromosome 17q)
• These secondary mutations impair differentiation (the hallmark lost in blast crisis)
• The risk of AML transformation in untreated CML is > 90% ^[4]

The clinical presentation of CML depends on the phase of disease. Most patients present in the chronic phase.

Up to 50% of patients are asymptomatic at diagnosis — picked up incidentally on a routine complete blood count (CBC) showing leukocytosis ^[4].

"CBC d/c: ↓Hb, ↑WBC (basophilia, myelocyte > metamyelocyte), N/↑ Plt" ^[2]

• Complete spectrum of myeloid cells (left shift) ^[2] — you see the entire granulocytic maturation series from myeloblasts through to mature segmented neutrophils
• Bimodal distribution: two peaks — one at the myelocyte stage and one at the mature neutrophil stage (reflecting both proliferation at the myelocyte level and retained ability to mature)
• Basophilia and eosinophilia
• Myelocyte > metamyelocyte ratio (this is characteristic — in reactive leukocytosis, the metamyelocytes would outnumber myelocytes)

• Hypercellular for age ^[3]
• Marked leukocytosis with moderately reduced erythropoiesis → full maturation is attained ^[3]
• Megakaryopoiesis is markedly increased, with frequent small hypolobated forms ^[3]
• Reduced fat spaces
• May show reticulin fibrosis (particularly in later phases)

This is a commonly tested distinction:

The approach to diagnosis of haematological malignancy follows 5 steps (MCICM) ^[5]:

1. Morphology: PBS (number, morphology, blasts), BM aspirate and trephine
2. Cytochemistry: MPO/Sudan Black B for myeloid lineage
3. Immunophenotyping: flow cytometry for surface markers
4. Cytogenetics: karyotyping (for Ph chromosome), FISH
5. Molecular genetics: PCR (RT-PCR for BCR-ABL1 fusion transcript), sequencing

CML represents one of the greatest success stories in targeted cancer therapy:

• The BCR-ABL1 fusion protein is a constitutively active tyrosine kinase ^[1][3]
• Specific targeted treatment against the tyrosine kinase → tyrosine kinase inhibitors (TKIs) ^[3]
• Imatinib ("i-MATIN-ib" → "matin" relates to mesenchymal/matrix, "-ib" = inhibitor) was the first TKI designed to fit precisely into the ATP-binding pocket of the BCR-ABL1 kinase, blocking its activity
• This transformed CML from a fatal disease (median survival 3–5 years) to a chronic manageable condition with near-normal life expectancy

This is the most important differential to exclude in everyday clinical practice, because reactive causes of leukocytosis are far more common than CML.

"Leukaemoid reaction to infection: WBC can be as high as 50, but should have ↑LAP score, toxic granulation in neutrophils and a readily identifiable precipitating cause" ^[4]

A leukaemoid reaction is a non-neoplastic, exaggerated granulocytic response to a severe stimulus — typically a severe infection (e.g. pneumonia, sepsis, abscess), severe haemolysis, or metastatic cancer to bone marrow.

Why it mimics CML: Both produce marked leukocytosis with immature myeloid precursors visible on the peripheral blood film (a "left shift"). Both can cause WBC counts in the tens of thousands.

How to distinguish from CML:

Neutrophils in CML are morphologically normal but cytochemically abnormal with ↓ leukocyte alkaline phosphatase (LAP) activity, which helps to exclude reactive leukocytosis (leukemoid reaction) due to infection ^[1]

Why is the LAP score low in CML? Because the neoplastic neutrophils, although they look morphologically mature under the microscope, are derived from the BCR-ABL1+ clone and have abnormal enzyme content. In contrast, reactive neutrophils are normal cells that have been activated and degranulated — they are packed with enzymes.

CML is the only MPN defined by BCR-ABL1; the other MPNs (PV, ET, PMF) are BCR-ABL1 negative ^[4][6].

However, some features overlap:

"Other chronic myeloid neoplasm: MDS, MPN, MDS/MPN, distinguished by genetic testing" ^[4]

Why do we need to distinguish these? Because treatment is completely different. CML is treated with TKIs targeting BCR-ABL1; the other MPNs are treated with hydroxyurea, JAK2 inhibitors (ruxolitinib), aspirin, phlebotomy (PV), etc.

This is a distinct entity and is a common source of confusion.

"Atypical CML": so-called BCR-ABL1-negative CML but is by definition NOT a CML ^[4]

• These patients have a clinical picture that resembles CML (leukocytosis with granulocytic left shift) but are BCR-ABL1 negative by all methods (karyotype, FISH, RT-PCR)
• They are classified under MDS/MPN overlap syndromes (specifically "atypical CML, BCR-ABL1 negative" in the WHO classification)
• They have a distinct and generally poorer clinical course with poor response to TKIs ^[4]
• Common mutations include SETBP1, ETNK1, and ASXL1

• Falls under the MDS/MPN overlap category
• Presents with persistent monocytosis (≥ 1 × 10⁹/L and ≥ 10% of WBC differential) + dysplastic features
• Can have leukocytosis and splenomegaly, mimicking CML
• Distinguishing features: prominent monocytosis (unlike CML where monocytosis is not the dominant feature), dysplastic morphology, BCR-ABL1 negative
• Often has mutations in TET2, SRSF2, ASXL1

MDS is characterised by ineffective haematopoiesis leading to peripheral blood cytopenia ^[10]

• MDS typically presents with cytopenias (low counts), not cytoses (high counts) → this is the key difference from CML
• However, some MDS subtypes can have a modestly elevated WBC, and the bone marrow can be hypercellular
• Distinguishing features:
  • Pancytopenia (not leukocytosis) is the hallmark ^[10]
  • Dysplastic cells visible on PBS (pseudo-Pelger-Huët cells, hypogranular neutrophils, ring sideroblasts) ^[10]
  • Macrocytic anaemia (↑ MCV) is typical ^[10]
  • BCR-ABL1 negative
  • Different cytogenetic abnormalities: del(5q), del(7q), del(20q), +8 ^[10]

• AML is characterised by ≥ 20% blasts in peripheral blood or bone marrow ^[4][11]
• It typically presents with pancytopenia (cytopenias due to BM failure), NOT the marked leukocytosis with mature cells seen in CML ^[4][11]
• Features of BM failure are prominent: profound anaemia, thrombocytopenia, neutropenia → infections and bleeding ^[4]

When confusion arises: CML in blast crisis can look identical to de novo AML. The distinction matters because:

• CML in blast crisis may still respond to TKIs (targeting the underlying BCR-ABL1)
• A history of prior CML, presence of the characteristic granulocyte series changes, and BCR-ABL1 positivity point to CML blast crisis rather than de novo AML

"CML in blastic crisis: essential to identify underlying condition as BCR-ABL TKIs may still have a role" ^[4]

"Other Ph-positive malignancies: a minority of ALL and AML may be Ph-positive and can be confused with CML in blastic crisis, but often lack the characteristic granulocyte series changes in blood picture" ^[4]

• Ph-positive ALL: ~25% of adult ALL cases carry the Philadelphia chromosome. This is a distinct entity from CML blast crisis with lymphoid phenotype.

  • Distinguishing clue: Ph+ ALL presents de novo with blast morphology and NO antecedent history of a chronic myeloproliferative phase; PBS does not show the full spectrum of myeloid maturation
  • Treatment approach differs (combination of TKI + chemotherapy for Ph+ ALL vs TKI-based approach for CML)

• Ph-positive AML: Rare, but can occur. Again, lacks the chronic-phase blood picture of CML.

CLL is unlikely to be confused with CML if the PBS is examined carefully, but both are "chronic leukaemias" and may appear in differential lists:

These are rare MPNs that can present with leukocytosis:

• CNL: Persistent neutrophilia (≥ 25 × 10⁹/L) with mature segmented neutrophils predominating, BCR-ABL1 negative, often has CSF3R mutations
• CEL: Persistent eosinophilia with clonal evidence; must exclude reactive causes and other myeloid neoplasms

Both are far less common than CML and are distinguished by the specific cell type that predominates and by molecular testing.

Massive splenomegaly differential ^[6][12]:

"The gold standard diagnosis of CML is based on cytogenetics — defined by the presence of the Philadelphia chromosome t(9;22)(q34.1;q11.2)" ^[3]

"No t(9;22), not CML" ^[3]

The diagnostic criteria can be summarised as:

Both criteria must be met. Let's unpack why each matters:

• Why do we need the blood/BM picture? Because other malignancies can carry the Philadelphia chromosome (Ph+ ALL, rarely Ph+ AML) — a compatible chronic-phase myeloid picture distinguishes CML from these.
• Why do we need BCR-ABL1? Because other MPNs and reactive conditions can cause leukocytosis with a left shift — BCR-ABL1 is the definitive and sine qua non marker of CML.

Once CML is confirmed, you must determine the phase — this dictates prognosis and treatment urgency.

Why does phase matter? Because:

• CP patients → excellent long-term outcomes with TKI alone (> 90% 5-year survival)
• AP patients → need more aggressive TKI (often 2nd-gen) ± consideration of allogeneic HSCT
• BC patients → treated as acute leukaemia with TKI + induction chemotherapy ± HSCT; prognosis is poor

This is the first and most accessible investigation. Let's go through each parameter systematically.

The PBS consists of morphologically normal but cytochemically abnormal leukaemic cells ^[4].

Key PBS findings:

• Complete spectrum of myeloid cells (left shift) ^[2] — from rare myeloblasts through to mature segmented neutrophils
• Bimodal distribution — two peaks at the myelocyte and segmented neutrophil stages
• Basophilia — increased basophils scattered among the granulocytes
• No toxic granulation, no Döhle bodies — these are features of reactive neutrophilia, NOT CML
• No Auer rods (unless in blast crisis when it transforms to AML)
• No smear/smudge cells (those suggest CLL ^[8])
• No significant dysplasia (that would suggest MDS)

Left shift: abundant immature forms of granulocyte series (band cells, metamyelocytes, myelocytes) → indicates severe infection/sepsis or CML ('reticulocytosis' of neutrophils) ^[14]

"Advise bone marrow examination with cytogenetics" ^[15] — this is the standard recommendation when the PBS is suggestive of CML.

Why do we do BM if blood is enough to suspect CML?

Three reasons:

1. Phase determination — the blast % in BM determines CP vs AP vs BC ^[4]
2. Cytogenetics — karyotyping requires dividing cells from the marrow aspirate; this is where you detect the Ph chromosome AND any additional chromosomal abnormalities (cytogenetic evolution) ^[2][3]
3. Baseline assessment — for monitoring treatment response later

BM site: typically posterior iliac crest; alternative includes anterior iliac crest and sternum (aspiration only) ^[14]

Why is the LAP score low in CML? The LAP enzyme is a marker of neutrophil maturity and activation. In CML, the neutrophils are derived from the BCR-ABL1+ clone and, despite looking mature, have deficient enzymatic content. In reactive states, normal neutrophils are activated by infection/stress → enzyme content increases → LAP score is high. In PV (another MPN), the neutrophils are activated and LAP is normal-to-high, which helps distinguish PV from CML.

Immunophenotyping is not the primary diagnostic tool for CML in chronic phase — the diagnosis is made on morphology + cytogenetics/molecular genetics.

However, immunophenotyping becomes critical in blast crisis to determine the lineage of the blasts:

Flow cytometry markers:

• Myeloid blasts: CD34+, CD117+, CD13+, CD33+, HLA-DR+, MPO+
• Lymphoid (B-ALL type) blasts: CD19+, CD10+, CD22+, TdT+

These are not diagnostic for CML per se, but are important parts of the initial work-up:

"CBP: Hb 9.8, MCV 98.2 (NcNc anaemia), WBC 150.0, Plt 160. Blood film reviewed by pathologist: Features consistent with chronic myeloid leukaemia. Advise bone marrow examination with cytogenetics" ^[15]

This clinical vignette from the molecular pathology lecture illustrates the real-life diagnostic pathway:

1. Incidental finding of WBC 150 on CBP
2. Pathologist reviews the film → identifies the characteristic full myeloid spectrum with basophilia
3. Reports "features consistent with CML"
4. Recommends BM with cytogenetics → to confirm Ph chromosome and determine phase

"HSCT very good, but no longer first line → has a very good graft versus tumour effect" ^[3]

Why is HSCT no longer first-line?

• TKIs achieve > 90% 5-year survival in CP with an oral medication and minimal toxicity
• HSCT carries ~10–20% transplant-related mortality and significant morbidity
• The risk-benefit ratio favours TKI in the vast majority of CP patients
• HSCT is reserved for patients who have exhausted TKI options or who present in advanced phases

This represents a potential "functional cure" — stopping TKI in carefully selected patients.

Why does this work? It is hypothesised that in patients with deep molecular response, the residual BCR-ABL1+ cells are held in check by the patient's immune system (particularly NK cells and T cells). When TKI is stopped, immune surveillance maintains control in some patients but not all.

"Complete molecular response DOES NOT MEAN zero → just means lower than the detection limit of the test" ^[3]

This is why TFR is possible — even with "undetectable" BCR-ABL1, residual leukaemic stem cells likely persist, but immune control prevents expansion.

Busulfan is mentioned in older notes ^[2] but is now largely of historical interest in CML management:

• An alkylating agent previously used as the first-line cytoreductive agent before TKIs were available
• Now superseded by TKIs and hydroxyurea
• Occasionally still used in conditioning regimens for HSCT (myeloablative conditioning)
• Hydroxyurea ± busulfan: cytoreduction for symptomatic relief ^[2] — busulfan is largely replaced by hydroxyurea for this purpose due to busulfan's significant toxicity (pulmonary fibrosis, skin pigmentation, marrow aplasia)

• All TKIs are teratogenic — must be discontinued
• Interferon-α (IFN-α) is the treatment of choice during pregnancy — it does not cross the placenta and has a proven track record of safety
• Hydroxyurea may be used short-term for cytoreduction but is also potentially teratogenic
• Leukapheresis is a safe temporising option

• Hong Kong has a high prevalence of chronic hepatitis B (~8% of population)
• TKIs can cause immunosuppression-related HBV reactivation
• HBV serology (HBsAg, anti-HBc, anti-HBs) must be checked before starting TKI
• If HBsAg positive → start antiviral prophylaxis (entecavir or tenofovir) before TKI

The spleen is the organ most affected by CML. Understanding the complications requires recalling why the spleen enlarges: massive splenomegaly results from (1) infiltration of splenic red pulp by leukaemic granulocytes, (2) extramedullary haematopoiesis, and (3) sequestration of blood cells ^[3].

When WBC counts are extremely high (typically > 100 × 10⁹/L, more common in blast crisis where blasts are larger and stickier), the blood viscosity rises and cells physically sludge in the microcirculation.

Why is leukostasis worse in blast crisis than chronic phase? Because blasts are larger, more rigid, and express more adhesion molecules compared to mature granulocytes — they physically obstruct small vessels more easily.

CML (like all MPNs) predisposes to both thrombosis and bleeding — a seemingly paradoxical combination that makes sense from first principles.

Thrombotic complications:

Haemorrhagic complications:

"High platelet consumes all the vWF, results in acquired vWD… usually > 1000" ^[6]

• Anaemia occurs in 45–60% of CML patients ^[1][4] — caused by the expanded granulocytic clone crowding out erythropoiesis in the marrow
• Chronic anaemia → fatigue, exercise intolerance, cardiac complications (high-output cardiac failure if severe and prolonged)
• Compounded by hypersplenism (splenic sequestration destroys RBCs faster)

• In chronic phase, neutrophil numbers are actually elevated — but the neoplastic neutrophils are functionally deficient (low LAP score reflects their poor enzymatic/bactericidal capacity)
• Clinical paradox: many neutrophils on the CBC but still susceptible to infection because they don't work properly
• In advanced phases (AP/BC), effective neutropenia from marrow failure → severe infection risk
• Febrile neutropenia is a medical emergency requiring blood cultures and broad-spectrum antibiotics within one hour ^[17]

Chronic leukaemia will transform into acute leukaemia if left untreated, since leukaemic cells will acquire more and more mutations over time ^[3].

Why can it transform into either AML or ALL? Because CML originates from a pluripotent haematopoietic stem cell — the original BCR-ABL1+ clone has the potential to give rise to both myeloid and lymphoid lineages. When blast crisis occurs, the maturation arrest can happen along either path.

These are not just "side effects" — some are clinically significant and dictate treatment changes.

Long-term TKI complications (relevant for lifelong therapy):

For patients who require HSCT (TKI failure, blast crisis), the transplant carries its own set of significant complications ^[18]:

Early complications (< 1 year):

Late complications ( > 1 year):

Allogeneic HSCT: 75% long-term survival, 9.9× mortality of general population (highest years 2–5 post-HSCT) ^[18]