When CML has the typical picture of a WBC count more than 100,000/mm3 (100 x 109/L) with myelocytic predominance, increased platelets, and basophilia, the diagnosis is reasonably safe. Otherwise, the two conditions that most frequently enter the differential diagnosis are agnogenic myeloid metaplasia and leukemoid reaction. CML tends to have a greater degree of leukocytosis than AMM or leukemoid reactions, so that WBC counts more than 100,000/mm3 are more likely to be CML. An increased basophil count is more likely to occur in CML than in AMM and is not expected in leukemoid reactions. More than an occasional teardrop cell is more often seen in AMM than in CML, whereas teardrop cells are not expected in leukemoid reactions.

Bone marrow examination is a valuable differentiating procedure for CML, AMM, and leukemoid or leukoerythroblastic reactions. CML has a hypercellular marrow with a moderate degree of immaturity; AMM most often has a marrow with varying degrees of fibrosis (although a hypercellular marrow may occur); and bone marrow in leukoerythroblastic reactions due to metastatic tumor frequently contains tumor cells. Clot sections as well as smears are desirable to increase diagnostic accuracy. In some patients with AMM and occasionally in patients with bone marrow tumor metastasis, no marrow can be obtained by aspiration, and a bone biopsy is necessary to be certain that the problem was not due to technical factors but to actual absence of marrow.

Although these distinguishing features suggest that differentiation between CML and AMM should be easy, the differences may be slight and the differentiating elements may at times appear in either disease. In fact, CML and AMM have been classified together under the term “myeloproliferative syndrome.” Some workers include the whole spectrum of leukemic-type proliferations of granulocytes, RBCs, and platelets within this term.

Leukocyte alkaline phosphatase (LAP) stain is the second useful test for differentiating leukemoid reaction, CML, and AMM. A fresh peripheral blood smear is stained with a reagent that colors the alkaline phosphatase granules normally found in the cytoplasm of mature and moderately immature neutrophils. One hundred neutrophils are counted, each neutrophil is graded 0 to 4+, depending on the amount of alkaline phosphatase it possesses, and the total count (score) for the 100 cells is added up. In most patients with leukemoid reaction or simple leukocytosis due to infection, or leukocytosis of pregnancy or estrogen therapy (birth control pills), and in 80%-90% of patients with polycythemia vera, the score is higher than reference range. About two thirds of patients with AMM have elevated values, about 25% have values within normal limits, and about 10% have low values. In CML, about 90% of patients have below-normal values, but 5%-10% reportedly have normal values. Values may be normal in CML during remission, blast crisis, or superimposed infection. In acute leukemia, each cell type differs in percentage of LAP values that are low, normal, or high, but results are not sufficiently clear cut to provide adequate cell type diagnosis. Overlap and borderline cases limit the usefulness of LAP in establishing a definitive diagnosis; however, values that are elevated are substantial evidence against CML, whereas values that are definitely low suggest CML rather than AMM. An experienced technician is needed to make the test reliable because the test reagents often give trouble, and the reading of the results is subjective and sometimes is not easy. Therefore, diagnosis should not be based on this test alone. In obvious cases, there is no need to do this test.

The LAP value is elevated in many patients with “active” Hodgkin’s disease. Infectious mononucleosis in early stages is associated with low or normal values in 95% of cases. In sickle cell anemia, LAP values are decreased even though WBCs are increased; however, if infection is superimposed, there may be an LAP increase (although a normal LAP value does not rule out infection). Low values are found in paroxysmal nocturnal hemoglobinuria.

A cytogenetic or DNA probe study for the Philadelphia chromosome is a third test that may be helpful in occasional diagnostic problems. Presence of the Philadelphia chromosome in a clinical and laboratory setting suspicious for chronic leukemia would be strong evidence in favor of CML.