Part VI

Diagnosis

From the blood smear to NGS-MRD: the diagnostic stack that classifies a leukaemia, stratifies its risk, and tracks its response. Modern haematology hangs on multiparametric integration — morphology, immunophenotype, cytogenetics, and molecular — not on any one assay.

1. The Diagnostic Workflow

A patient with suspected leukaemia is worked up in stereotyped order:

  1. Complete blood count and differential — cytopenias, leucocytosis, percentage and morphology of blasts.
  2. Peripheral blood smear — Wright-Giemsa stain examined by haematologist or pathologist.
  3. Comprehensive metabolic panel + LDH + uric acid + coagulation — baseline, evaluate for tumour lysis, DIC.
  4. Bone-marrow aspirate and trephine biopsy — morphology, blast count, cellularity, dysplasia, fibrosis.
  5. Flow cytometry — on aspirate or peripheral blood; immunophenotype assignment.
  6. Cytogenetics — karyotype + targeted FISH panel (myeloid panel, ALL panel).
  7. Molecular — NGS panel (~30–500 genes); RT-PCR for fusion transcripts; FLT3-ITD analysis; for ALL, MRD-target identification.
  8. CSF examination — for ALL and AML M4/M5; lumbar puncture with cytology and flow.
  9. Imaging — CT of chest/abdomen/pelvis (CLL staging, mass evaluation); MRI for extramedullary myeloid sarcoma.

Most of these results return within 24–72 hours; cytogenetic karyotype takes longer (~5–7 days). A WHO/ICC-conformant diagnosis requires integration of all the strands — a workflow that is impossible at low-resource sites and a major contributor to global outcome disparities.

2. Peripheral Blood Smear

Despite a century of technology, a well-prepared, well-stained Wright-Giemsa smear read by an experienced morphologist remains the cheapest, fastest, most diagnostic single test in haematology. Pattern-recognition cues:

Smear findingSuggests
Auer rods (single, multiple, “faggot” bundles)AML; faggots → APL
Hypergranular promyelocytes with azurophilic granules + abundant Auer rodsAPL (hypergranular variant)
Microgranular / bilobed promyelocytesMicrogranular APL (M3v) — high WBC, FLT3-ITD common
Left-shift with full granulocytic spectrum, basophiliaCML chronic phase
Blasts ~10–19% with otherwise CML patternCML accelerated phase
Lymphocytosis with smudge cells, small mature lymphocytesCLL
Lymphoblasts (TdT+ on later flow)ALL
“Hairy” cells with cytoplasmic projectionsHairy-cell leukaemia
Blasts with vacuoles, deep blue cytoplasm (FAB L3)Burkitt / mature B-ALL
Pelger-Huët, dysgranulopoiesis, hypogranular neutrophilsMDS or MDS-related AML
Tear-drop poikilocytes, leucoerythroblastic filmMarrow infiltration, myelofibrosis

The differential is performed on at least 200–500 cells in the “feathered edge” of a peripheral smear, and on at least 500 nucleated marrow cells in a 200x oil immersion field for a marrow blast count.

3. Bone-Marrow Aspirate & Biopsy

The marrow examination remains foundational. The sample is taken from the posterior iliac crest under local anaesthesia (or sternum aspirate in select circumstances). Two specimens are produced:

  • Aspirate — liquid, smeared and stained for cytology; sent in EDTA for flow; in heparin for cytogenetics; in fixative or extra slides for FISH; in EDTA for DNA / RNA extraction.
  • Trephine biopsy (core) — ~1.5–2 cm of cortical/medullary bone; decalcified and embedded for histology; assesses cellularity, architecture, fibrosis (reticulin / Masson), megakaryocyte morphology.

The marrow report integrates:

  • Cellularity — normal-for-age (~100−age% in young adults; ~30–50% in older).
  • Blast % — the diagnostic gate for “acute” (≥20%) leukaemia.
  • M:E (myeloid:erythroid) ratio — normally 2–4:1; reversed in erythroid hyperplasia, increased in myeloid expansion (CML).
  • Dysplasia — multilineage dysplasia is required for some MDS categories; ≥10% of cells of a given lineage.
  • Fibrosis — reticulin grading 0–3; MF-3 is overt myelofibrosis.
  • Other — iron stores, ring sideroblasts (≥5 perinuclear iron granules), plasma cells, lymphoid aggregates.

The trephine is irreplaceable for assessment of fibrotic marrows, “dry tap,” and detection of focal infiltration that aspirate may miss.

4. Flow-Cytometry Immunophenotyping

Flow cytometry passes single cells through a focused laser, measures forward and side scatter (size, granularity), and detects fluorescence from antibodies bound to surface and intracellular antigens. Modern multiparametric flow uses 8–14 (and now up to 30+ in spectral) markers per tube, with software gating defining cell populations.

LineageDefining markers
B-lineageCD19, CD20, CD22, CD79a, surface Ig (κ/λ)
T-lineagecytoplasmic CD3, CD2, CD5, CD7, CD4, CD8
MyeloidMPO, CD13, CD33, CD117, CD15, CD11b
MonocyticCD14, CD64, CD11c, lysozyme
ErythroidCD235a (glycophorin A), CD71
MegakaryocyticCD41, CD61, CD42b
Stem / progenitorCD34, CD117 (KIT)
Lymphoblast (immaturity)nuclear TdT, CD34, CD10 (B), CD1a (T)

Pattern recognition makes lineage assignment usually straightforward:

  • CLL phenotype — CD5+ CD19+ CD23+ CD20dim CD79bdim smIgdim light-chain restricted (Matutes score 4–5/5).
  • B-ALL — CD19+ CD10+ CD34+ TdT+ smIg− (in common B-ALL) or cyµ+ smIg− (pre-B).
  • T-ALL — cyCD3+ surface CD3+/− CD7+ TdT+ (CD1a± defines cortical vs ETP).
  • AML — MPO+ CD117+ CD13+ CD33+, with subset markers for monocytic (CD14+, CD64+), erythroid (CD235a+), megakaryoblastic (CD41+ CD61+), or APL (HLA-DR−, CD34−, hypergranular CD13+ CD33+).
  • Mixed-phenotype acute leukaemia (MPAL) — satisfies both myeloid (MPO or monocytic markers) and lymphoid (cyCD3 for T or CD19 + 1 of CD79a/cyµ/CD22/CD10 for B) criteria; ~3% of acute leukaemias.

Aberrant immunophenotypes — cross-lineage marker expression, asynchronous expression, abnormally high/low intensities — allow tracking of an individual patient’s leukaemia by flow MRD even when fusion transcripts are absent.

5. Cytogenetics & FISH

Standard panels:

  • AML/MDS panel — BCR-ABL1, RUNX1::RUNX1T1 t(8;21), CBFB::MYH11 inv(16)/t(16;16), PML::RARA t(15;17), KMT2A 11q23 break-apart, −5/del(5q), −7/del(7q), trisomy 8, MECOM 3q26, NUP98, TP53 / 17p13.
  • B-ALL panel — BCR-ABL1, KMT2A break-apart, ETV6-RUNX1 t(12;21), TCF3-PBX1 t(1;19), MYC t(8;14), iAMP21 (RUNX1 amplification), CRLF2 break-apart for Ph-like.
  • T-ALL panel — TCR break-apart probes; STIL-TAL1; KMT2A.
  • CLL panel — del(13q14), del(17p13)/TP53, del(11q22.3)/ATM, trisomy 12.

The limit of detection of a typical interphase FISH assay is on the order of $10^-2$ (one abnormal cell among a hundred). In a disease where MRD operationally requires sensitivities of $10^-4$ to $10^-6$, FISH is therefore a diagnostic and not a monitoring tool.

6. Molecular Diagnostics

Three main platforms in routine use:

  • Targeted NGS panels — 30–500 genes covering the recurrent leukaemia driver list (FLT3, NPM1, IDH1/2, DNMT3A, TET2, ASXL1, RUNX1, TP53, NRAS, KRAS, KIT, CEBPA, WT1, NOTCH1, FBXW7, IKZF1, JAK1/2/3, etc.). Sequencing depth typically 500–2000×, with variant-calling sensitive to ~2–5% VAF for diagnosis, <0.1% for MRD.
  • RT-qPCR on RNA — quantification of fusion transcripts (BCR-ABL1, PML-RARA, RUNX1-RUNX1T1, CBFB-MYH11). For BCR-ABL1, results are reported on the International Scale (BCR-ABL1IS) so labs anywhere can be compared.
  • Digital droplet PCR (ddPCR) — absolute quantification by Poisson partitioning across thousands of nanolitre droplets; sensitivity ~10⁻⁵ for fusion transcripts and BCL2 mutations.

The relationship between input material and limit of detection is straightforward:

$$\;\text{LOD} \;\approx\; \frac{1}{N_{\text{cells}}\;\text{or}\;N_{\text{transcripts}}}\;.\;$$

For a 10 µg DNA input (~1.5 million cells), a 0.01% MRD (10⁻⁴) means ~150 mutant copies in the assay — comfortably above the Poisson noise floor. Pushing to 10⁻⁵ requires proportionally more input or molecular “bar-coding” (UMI consensus, error-corrected sequencing) to suppress polymerase error.

7. CSF & Extramedullary Disease

The CNS is a sanctuary site that systemic chemotherapy reaches poorly. ALL is the archetype:

  • CNS-1 — no blasts in CSF.
  • CNS-2 — <5 WBC/µL with blasts present; no clinical deficit.
  • CNS-3 — ≥5 WBC/µL with blasts, or cranial-nerve palsy / overt clinical disease.

Modern paediatric ALL achieves CNS protection with intrathecal methotrexate (with or without ara-C/hydrocortisone) plus high-dose systemic methotrexate, abolishing the need for cranial radiation in nearly all standard-risk patients.

Extramedullary myeloid sarcoma (granulocytic sarcoma, “chloroma”) is a tumefactive infiltrate of myeloblasts; classically green when fresh (myeloperoxidase). Sites: skin (CD56+ monocytic AML, CMML), gum (M5), bone, paraspinal masses (M2 with t(8;21)), testis (ALL), ovary, breast.

Skin involvement of CLL is “leukaemia cutis”; in CTCL the leukaemic phase is Sézary syndrome; in T-cell prolymphocytic leukaemia, marked splenomegaly and skin involvement are common.

8. MRD — The New Endpoint

Morphological complete remission — <5% blasts on aspirate, recovery of counts — can mask a 10⁹–10¹⁰ residual leukaemia burden. Measurable (or minimal) residual disease (MRD) is the quantification of that residue.

The clinical observation:

$$\;P(\text{relapse} \mid \text{MRD}^+) \;\gg\; P(\text{relapse} \mid \text{MRD}^-)\,.\;$$

In B-ALL, the relapse rate among MRD-positive patients (≥10⁻⁴ at the end of induction) is 5–10× that of MRD-negative patients (Borowitz, Blood 2008; Conter, Blood 2010). MRD now governs risk stratification in every modern paediatric and adult ALL trial.

Three platforms dominate MRD:

  • Multiparametric flow MRD — tracks an individual patient’s leukaemia-associated immunophenotype. Sensitivity ~10⁻⁴ standard, ~10⁻⁵ next-generation flow (EuroFlow). Universally applicable, fast (~24 h), but operator-dependent.
  • RT-qPCR / ddPCR for fusion transcripts — BCR-ABL1, PML-RARA, RUNX1-RUNX1T1, CBFB-MYH11 in the corresponding leukaemias. Highly sensitive (~10⁻⁵–10⁻⁶); only applicable when a fusion is present.
  • NGS-MRD — Adaptive’s clonoSEQ assay (FDA-cleared) sequences the unique IGH/TCR rearrangement of a B- or T-ALL clone with sensitivity ~10⁻⁶. ClonoSEQ is the standard MRD platform in many adult ALL and CLL trials. NGS panels for AML mutations (NPM1, FLT3, IDH1/2) achieve similar sensitivities with error-corrected sequencing.

MRD is now a regulatory endpoint. The FDA has accepted MRD-negative CR rates as a surrogate for survival in ALL trials supporting drug approvals (e.g. blinatumomab for MRD+ ALL, FDA 2018, based on BLAST trial). MRD-guided therapy — escalation, de-escalation, transplant timing — is the operational reality of the modern clinic.

9. Putting It Together

A worked example. A 67-year-old presents with fatigue, easy bruising, WBC 35×10⁹/L, Hb 78 g/L, platelets 22×10⁹/L. Smear shows ~80% blasts with Auer rods. The diagnostic stack:

  1. Day 0 — admission; CBC + smear → suspected AML; coagulation including fibrinogen, D-dimer (rule out APL/DIC); start hydration, allopurinol; obtain CSF baseline if monocytic features.
  2. Day 0 — bone-marrow aspirate & biopsy. Aspirate sent for: morphology, flow, cytogenetics, FISH, NGS panel, FLT3-ITD allelic ratio, NPM1, BCR-ABL by RT-PCR.
  3. Day 1–2 — flow returns: CD34+ CD117+ MPO+ CD13+ CD33+ HLA-DR+ — AML, not APL.
  4. Day 2–4 — FISH: PML-RARA negative; KMT2A negative; CBF normal; no monosomy 7 or 5q-. NPM1 mutation positive (immuno-cytochemical NPM1c+ rapid result).
  5. Day 5–7 — Karyotype: normal 46,XY. NGS: NPM1 c.863_864insTCTG, FLT3-ITD positive (allelic ratio 0.65, “high”), DNMT3A R882H, no TP53. ELN 2022 risk: intermediate (NPM1+ with high FLT3-ITD ratio).
  6. Treatment — induction with 7+3 (cytarabine + daunorubicin) + midostaurin (FLT3 inhibitor; RATIFY trial). HLA-typing initiated.
  7. Day 28 marrow — morphologic CR; flow MRD 0.05% NPM1c-positive cells. Plan: consolidation with high-dose cytarabine + midostaurin; allogeneic transplant in CR1 if a donor is identified, given high FLT3-ITD ratio.
  8. Post-induction NPM1 RT-qPCR — for serial MRD monitoring; rising NPM1 transcript by >1 log triggers pre-emptive therapy or transplant.

The next part — therapy — takes up these threads: 7+3, FLT3 and IDH inhibitors, venetoclax + HMA, ALL induction regimens, TKIs, BTK and BCL2 inhibitors, blinatumomab, inotuzumab, and CAR-T.

Key references for further reading. Borowitz et al., MRD in childhood B-precursor ALL, Blood 2008; van Dongen et al., EuroFlow next-generation flow for MRD, Leukemia 2017; Cross et al., Laboratory recommendations for scoring deep molecular responses following CML treatment, Leukemia 2015; Wood et al., FDA-cleared NGS-MRD assays in ALL, Blood 2019; Heuser et al., 2021 update on MRD in AML, Blood 2021; Khoury et al., WHO 5th edition Classification of Haematolymphoid Tumours, Leukemia 2022.
Share:XRedditLinkedIn