EEHLSS · Computational Haematology Curriculum · Week 4 of 72

Platelet Analysis & Thrombocyte Algorithm Foundations

Counting Methods · Platelet Indices · AI Discrimination · Nigeria Clinical Context

Module 4 of 8FRCPath Part I
ICSH AlignedMedLabAI-LISeehlss.io

Section 1

Platelet Counting Methods

Impedance — Coulter Principle

DC resistance pulse ∝ cell volume
Simple, robust, cost-effective
Direct volume measurement
High throughput; Nigeria primary method
Cannot distinguish microcytic RBCs
Sensitive to electrical interference

Optical — Light scatter / fluorescence

FSC → cell size; SSC → granularity
Fluorescence → IPF (RNA-binding dye)
Superior discrimination from particles
Multi-parameter: MPV, PDW, IPF
Complex calibration required
Sample prep artefact sensitive

Immunological — Flow cytometry

CD41/CD61 antibody detection
Highest platelet specificity
Enables activation/subset analysis
Gold standard: research / refractory TCP
Expensive; specialist expertise
Not high-throughput routine

Clinical selection guide

Method	Throughput	Cost	Best Use Case	Nigeria Tier
Impedance	High (60–100/hr)	Low	Routine; resource-limited settings	Primary
Optical	High	Moderate	Indices required; flag investigation	Secondary review
Immunological	Low	High	Refractory TCP; ITP subtyping; research	Tertiary referral

Section 2

Platelet Indices: MPV, PDW, PCT

Mean Platelet Volume (MPV)

MPV (fL) = Total platelet volume / Platelet count
Reference range: 7.5–11.5 fL (varies by analyser)

MPV reflects platelet production rate from bone marrow. Larger platelets are generally younger and more metabolically reactive. MPV is inversely related to platelet count in many thrombocytopenic conditions — a key AI pattern-recognition signal.

MPV Pattern	Interpretation	Conditions
>11.5 fL (High)	Increased thrombopoiesis	ITP, myeloproliferative neoplasms, acute infection
<7.5 fL (Low)	BM suppression	Aplastic anaemia, post-chemotherapy, radiation
Rising with falling PLT	Consumptive thrombocytopenia	Dengue nadir, severe malaria, DIC

Platelet Distribution Width (PDW)

PDW (%) = (SD of platelet volume distribution / MPV) × 100
Reference range: 9–17%

PDW reflects heterogeneity (anisocytosis) of the platelet population. High PDW indicates a mixed young/old platelet population; low PDW indicates a uniform but depleted population, typical of bone marrow failure states. The combination of PDW + MPV is particularly useful in differentiating central vs peripheral thrombocytopenia.

Platelet Crit (PCT)

PCT (%) = (Platelet count × MPV) / 10,000
Reference range: 0.15–0.40%

PCT represents total platelet mass in circulation — the haematocrit analogue for platelets. Confirms true low platelet mass versus artefactually low counts (e.g., pseudothrombocytopenia). Used to track platelet mass recovery post-transfusion.

Key clinical principle: The three indices together — MPV + PDW + PCT — tell a richer story than platelet count alone. In dengue and malaria, the rising MPV pattern during the thrombocytopenic nadir reflects bone marrow releasing larger, younger platelets in response to depletion — an active production signal.

Section 3

Giant Platelets & AI Discrimination Pipeline

Giant Platelets

Defined as platelets >7 fL diameter (some systems: >12 fL). Seen in myeloproliferative neoplasms, hereditary thrombocytopenias (Bernard-Soulier syndrome, MYH9-related disorders), and ITP with high thrombopoietic drive. Technical challenge: misclassified as RBCs/WBCs on impedance systems; trigger abnormal scatter flags on optical platforms.

Pseudothrombocytopenia (PTCP) — Platelet Clumps

Feature	Detail
Prevalence	0.1–0.2% general population; higher in elderly
Mechanism	IgG autoantibodies against GPIIb/IIIa complex activated by EDTA
Consequence	Spuriously low automated count — may trigger inappropriate transfusion
AI signals	Abnormal scatter patterns + fluorescence quenching + channel inconsistency
Resolution	Repeat in citrate or heparin tube; peripheral blood smear review

AI Multi-Stage Decision Pipeline

Stage 1: Primary classification (baseline size-gate identification)
↓
Stage 2: Giant platelet detection (size/shape/circularity/aspect ratio analysis)
↓
Stage 3: Clump detection (scatter anomaly + channel inconsistency + fluorescence quenching)
↓
Flag exceeds threshold?
↓ ↓
Clump suspected: True platelet count:
→ citrate tube → report + CDS output
→ PBS smear review

ML Approach	Algorithm	Application
Supervised learning	Random Forest, SVM, Neural Network	True platelet vs artefact probability score
Unsupervised / anomaly	Isolation Forest, One-Class SVM	Novel interference pattern detection
Clustering	k-means, DBSCAN	Identification of abnormal cell populations

Section 4

Clinical Utility in Nigeria: Dengue & Malaria

Disease	TCP rate	PLT nadir timing	Key context
Dengue	60–80% of symptomatic cases	Days 3–7 of illness	All 4 serotypes circulating in Nigeria; peak in rainy season
Malaria	70–90% of acute cases	Days 2–5 of illness	P. falciparum >90% of Nigerian cases; correlates with parasite density

Pathophysiology

Mechanism	Dengue	Malaria (P. falciparum)
Bone marrow	Direct viral suppression + cytokines	Inhibition by IL-10, TNF-α
Peripheral destruction	Immune-mediated; DIC; platelet activation	Antibody-mediated destruction
Sequestration	Platelet aggregation in microvasculature	Splenic and hepatic trapping
Direct effect	NS1 protein-induced apoptosis	Direct parasite-induced platelet activation
Critical threshold	PLT <20×10³/µL → plasma leakage risk	PLT recovery precedes parasite clearance

AI-Assisted Monitoring Systems

Point-of-care applications: Smartphone-based image analysis for platelet identification from PBS photographs. Integration with clinical decision support for dengue/malaria severity stratification. AI-driven early warning systems reduce unnecessary hospital admissions by detecting thrombocytopenia trends before clinical deterioration.

Resource-adapted solutions: Offline capabilities without constant internet; low-bandwidth compressed data transmission; AI-generated training cases for technician education; maintenance prediction for reagent scheduling. All critical requirements for Nigerian district hospital deployment.

Section 5

Future Directions

Technology	Application	Relevance to Africa
Imaging flow cytometry	Morphology + fluorescence at single-cell level	Tropical platelet morphology characterisation
Raman spectroscopy	Label-free biochemical characterisation	No reagent dependency; low-resource potential
Microfluidic platforms	Point-of-care functional platelet assays	District hospital deployment without large analysers
Federated Learning	Multi-centre training without data sharing	LUTH/UNTH/LAUTECH collaboration compliant with Africa CDC 2025
Explainable AI (XAI)	Grad-CAM/SHAP for decision transparency	Clinician trust and regulatory compliance

Section 6

Practical Laboratory Considerations

Phase	Variable	Requirement	Consequence of Error
Pre-analytical	Anticoagulant ratio	EDTA at manufacturer-recommended concentration	EDTA excess → pseudothrombocytopenia
	Mixing	Gentle inversion × 8–10 immediately after collection	Inadequate → platelet clumping → false low count
	Timing	Count within 4–6 hours of collection	Storage → MPV drift; platelet swelling
	Temperature	Room temperature; avoid cold	Cold → platelet activation and aggregation
Analytical	Calibration	Daily tri-level QC controls	Systematic bias in platelet count
	Interference	Monitor lipemia, icterus, haemolysis	Optical channel interference → spurious results
	Reference intervals	Population-specific Nigerian adult normals	Misclassification using non-applicable Western ranges
Post-analytical	Critical values	PLT <50×10³/µL (most protocols)	Delayed notification → preventable haemorrhage

Section 7

Case Studies: Nigerian Clinical Scenarios

Case 1 — Dengue

28M · febrile illness day 4

PLT (day 0)85 ×10³/µL

PLT (day 2)42 ×10³/µL ↓↓

PLT (day 4)18 ×10³/µL ↓↓↓

MPV trend9.1 → 10.2 → 11.8 fL ↑

WBC3.2 ×10³/µL (↓)

AI insight: Rising MPV + falling PLT = consumptive thrombocytopenia. PLT <20 → activate plasma leakage monitoring.

Dx: Dengue progression — nadir pattern days 4–5

Case 2 — Severe Malaria

5F · coma · 8% parasitaemia

PLT (admission)18 ×10³/µL ↓↓↓

PLT (24h ACT)45 ×10³/µL ↑

Hgb (admission)6.8 g/dL ↓

Hgb (24h ACT)8.1 g/dL ↑

MPV trend8.9 → 9.5 → 10.1 fL ↑

AI insight: PLT + MPV recovery trend predicts clinical improvement. Discharged day 5.

Dx: Severe P. falciparum malaria — responding to ACT

Case 3 — Pseudo-TCP

65M · hypertensive · routine

Auto PLT (EDTA)12 ×10³/µL ↓↓↓ CRITICAL

PBS smearMarked clumping

Citrate repeat~220 ×10³/µL — NORMAL

MechanismEDTA IgG → GPIIb/IIIa

AI flags: Abnormal scatter + fluorescence quenching + channel inconsistency → clump pattern. PTCP suspected.

Outcome: AI flagging prevented inappropriate transfusion

Section 8

Assignment & References

Week 4 Assignment

Task (eehlss.io student portal / MedLabAI-LIS):

1. A 7-year-old Nigerian child presents with fever and PLT 35×10³/µL; MPV 12.1 fL; PDW 19%. Using the three-index framework, classify the likely thrombocytopenia mechanism and propose the next two diagnostic steps.
2. An automated analyser flags PLT 8×10³/µL in an asymptomatic adult. Describe the AI signal pattern distinguishing EDTA-induced pseudothrombocytopenia from true thrombocytopenia. What is the resolution protocol?
3. Describe how a federated learning approach would improve platelet AI model performance across Nigerian tertiary hospitals without violating patient data sovereignty (Africa CDC 2025 framework).

300 words per answer · Minimum 2 Vancouver references each · PDF via MedLabAI-LIS portal

References

#	Type	Citation & DOI
1	Review	Jurk K, Shiravand Y. Platelet phenotyping and function testing in thrombocytopenia. J Clin Med. 2021;10(6):1114. DOI: 10.3390/jcm10061114
2	Review	Cattaneo M. Pseudothrombocytopenia. Haematologica. 2021.
3	Primary	Woo J. Definition of significant platelet clumping. Int J Lab Hematol. 2025. DOI: 10.1111/ijlh.xxxxx
4	Primary	Mayrose H, et al. ML-based detection of dengue from blood smear images. Diagnostics. 2023. DOI: 10.3390/diagnostics13xxxxx
5	Guideline	ICSH Working Group. Reference method for platelet counting. Int J Lab Hematol. 2015;37(3):287–303.
6	Textbook	Bain BJ. Blood Cells: A Practical Guide. 6th ed. Wiley-Blackwell; 2015. ISBN: 9781118408889
7	Africa	Maruta T. AI in diagnostics: laboratory medicine in Africa. Afr J Lab Med. 2025;14(1), a2952. DOI: 10.4102/ajlm.v14i1.2952

WK 4 — Platelet Analysis & Thrombocyte Algorithm Foundations