Part VIII

Stem-Cell Transplant & MRD

Allogeneic haematopoietic stem-cell transplantation remains the only curative therapy for many adverse-risk leukaemias. We close the course with HLA matching, conditioning intensity, the GvHD/GvL trade-off, and MRD as the new endpoint that may eventually tell us when transplant is — or is no longer — necessary.

1. Why Transplant?

Allogeneic haematopoietic stem-cell transplantation (allo-HSCT) is the cornerstone curative therapy for high-risk haematological malignancies. Three mechanistic levers are simultaneously deployed:

  • High-dose cytotoxic / radiation conditioning — eradicates leukaemia cells beyond the dose limits of conventional chemotherapy.
  • Replacement of marrow — donor HSCs reconstitute haematopoiesis after the conditioning regimen has destroyed it.
  • Graft-versus-leukaemia (GvL) effect — the donor immune system recognises and eliminates residual host leukaemia cells; the most powerful immunotherapy in routine clinical use.

Indications for allo-HSCT in CR1 in 2025:

  • AML — ELN intermediate or adverse risk in fit patients with a donor; MRD+ after consolidation; secondary/therapy-related AML.
  • ALL — high-risk cytogenetics (KMT2A, hypodiploidy, Ph-like), persistent MRD; T-ALL with poor early response; many adult ALL beyond the AYA group.
  • CML — failure of multiple TKIs; advanced phase (AP, BP) after re-induction.
  • CLL — rare in modern era; reserved for Richter transformation, refractory disease post-BTKi/venetoclax/CAR-T.
  • MDS — high-risk MDS in fit patients; intermediate-2 / IPSS-R high or very high.
  • Inherited bone-marrow failure — Fanconi, dyskeratosis congenita, severe congenital neutropenia.
  • Severe aplastic anaemia — matched-sibling allo first-line in young patients.

Globally, ~85,000 HSCTs are performed each year (EBMT/CIBMTR registry data); about 60% are autologous (predominantly for myeloma and lymphoma), 40% allogeneic (predominantly for AML, ALL, MDS).

2. Autologous vs Allogeneic

Autologous (auto-HSCT)

Patient’s own HSCs harvested during remission, cryopreserved, and re-infused after high-dose chemotherapy. No GvHD, no GvL. Used in:

  • Multiple myeloma — standard of care after induction
  • Relapsed Hodgkin / non-Hodgkin lymphoma
  • Selected acute leukaemias (rarely now)
  • Some autoimmune conditions (MS, scleroderma)

Allogeneic (allo-HSCT)

Donor HSCs replace patient’s haematopoiesis after conditioning. Provides GvL effect. Used in:

  • Adverse / intermediate-risk AML in CR1; relapsed AML
  • Adult ALL; high-risk paediatric ALL; relapsed ALL post-CAR-T
  • High-risk MDS
  • CML in BP / multi-TKI-refractory
  • Inherited BMF, severe aplastic anaemia
  • Sickle-cell disease, β-thalassaemia (curative; gene-therapy alternatives now available)

Syngeneic transplants from an identical twin are vanishingly rare and behave as a cross between auto and allo (same MHC, but distinct minor antigens absent → no GvL).

3. HLA Matching

HLA (human leukocyte antigen) matching is the central determinant of allo-HSCT outcome. Six classical loci are typed at high resolution (allele-level) by sequence-based typing or NGS:

  • Class I: HLA-A, HLA-B, HLA-C — expressed on all nucleated cells; present peptide to CD8 T-cells.
  • Class II: HLA-DRB1, HLA-DQB1, HLA-DPB1 — expressed on professional APCs and B-cells; present peptide to CD4 T-cells.

Modern practice quotes a match as “n/8” (HLA-A, B, C, DRB1) or “n/10” (adding DQB1) or “n/12” (adding DPB1). The hierarchy:

Donor typeHLA matchProbability
Matched sibling donor (MSD)10/10 sibling~25% × number of siblings
Matched unrelated donor (MUD)10/10 from registry~75–90% Caucasian; ~30% AA, ~50% Hispanic, ~50% Asian
Mismatched unrelated (MMUD)9/10—; tolerable with PTCy
Haploidentical5/10 — biological parent, child, or siblingAlmost always available; PTCy now standard
Cord blood4/6 acceptableLimited cell dose; slow engraftment

Each single-allele mismatch increases acute GvHD by ~10% and TRM by ~4–5%. NMDP / Be The Match registry holds >40 million volunteer donors worldwide and a quarter-million cord-blood units. Donor diversity remains a major equity issue: minority populations have substantially worse MUD-matching probabilities than Caucasian populations of European descent.

4. Donor Sources of HSCs

Three sources, with different cell composition and engraftment kinetics:

SourceCollectionCD34 doseNotes
G-CSF-mobilised peripheral blood (PBSC)Apheresis after G-CSF (± plerixafor)4–8×10⁶/kgFaster engraftment; more chronic GvHD; current default for allo
Bone-marrow harvestOR aspiration from posterior iliac crests under GA2–4×10⁶/kgLess chronic GvHD; preferred in some paediatric and aplastic-anaemia settings
Umbilical-cord bloodStored at birth; thawed at use~0.2–0.5×10⁵/kg per unit; double-cord transplants combine two unitsAvailable rapidly; less stringent HLA matching needed; slow engraftment

The cell-dose threshold matters. Below ~2×10⁶ CD34/kg, engraftment is delayed and graft failure increases. Cord-blood transplants accept this risk in exchange for tolerance of HLA mismatch.

5. Conditioning Regimens

The conditioning regimen serves three purposes simultaneously: eradicate residual leukaemia, suppress recipient immunity to permit engraftment, and create marrow space for incoming donor cells. Two intensity classes:

Myeloablative (MAC)

Fully eradicates host haematopoiesis; cannot recover without graft.

  • BuCy — busulfan 12.8 mg/kg + cyclophosphamide 120 mg/kg
  • FluBu4 — fludarabine + busulfan 12.8 mg/kg
  • TBI/Cy — total-body irradiation 12 Gy + cyclophosphamide
  • FluTBI — fludarabine + 12 Gy TBI

Used in fit patients ≤55 (sometimes ≤65) with adverse-risk disease.

Reduced-intensity (RIC) / non-myeloablative (NMA)

Permits autologous recovery if graft fails; relies more on GvL than direct cytoreduction.

  • FluMel — fludarabine + melphalan 100–140 mg/m²
  • FluBu2 — fludarabine + busulfan 6.4 mg/kg
  • Flu/Cy/TBI 2 Gy — the “Seattle” NMA

Used in older patients or with comorbidity. Higher relapse rates but lower TRM.

The BMT-CTN 0901 trial (Scott, JCO 2017) randomised AML/MDS patients to MAC vs RIC and was stopped early for higher relapse in the RIC arm (48% vs 14%). Conclusion: in fit patients, prefer myeloablative; RIC is for those who can’t tolerate it.

Risk-score frameworks (EBMT, HCT-CI) capture the combined effect of multiple factors:

$$\;\text{Risk score} = f(\text{age},\;\text{disease stage},\;\text{donor},\;\text{interval},\;\text{sex match})\,.\;$$

6. Engraftment & Recovery

The transplant timeline:

  • Day −7 to −1 — conditioning. Patient develops neutropenia, mucositis, nausea, alopecia.
  • Day 0 — HSC infusion (intravenous “bag of cells”).
  • Days +1 to +14 — pancytopenic; high infection risk; mucositis at peak. Antibacterial, antifungal, antiviral prophylaxis.
  • Days +14 to +21neutrophil engraftment (ANC ≥0.5×10⁹/L for 3 days). Faster with PBSC, slower with cord.
  • Days +20 to +35 — platelet engraftment (≥20–50×10⁹/L without transfusion).
  • Days +30 to +90 — immunosuppression for GvHD prophylaxis (cyclosporine/tacrolimus + methotrexate, or post-transplant cyclophosphamide on days +3 and +4 for haploidentical).
  • Day +100 marrow — documents engraftment, donor chimerism, MRD assessment.
  • Months 3–12 — immune reconstitution; risk of viral reactivation (CMV, EBV, adenovirus, BK). Long-term tapering of immunosuppression.
  • Year 1+ — chronic GvHD risk; revaccination; long-term surveillance.

Chimerism measurement — the fraction of recipient peripheral blood that is donor-derived — is performed by short-tandem-repeat (STR) genotyping or by NGS. Full donor chimerism (≥95%) is the expected end-point. Mixed chimerism predicts relapse and is sometimes managed with donor-lymphocyte infusion (DLI).

7. Graft-versus-Host Disease and Graft-versus-Leukaemia

GvHD is the central problem of allo-HSCT — the same alloreactivity that produces GvL also damages host tissues. Two clinical syndromes:

Acute GvHD

Classically <100 days post-transplant. Three target organs:

  • Skin — maculopapular rash, classical sites palms/soles, ears, neck
  • Gut — diarrhoea, vomiting, ileus; severe → mucosal denudation, GI bleeding
  • Liver — cholestasis, hyperbilirubinaemia (after sinusoidal-obstruction is ruled out)

Glucksberg grading I–IV. Treatment: high-dose corticosteroids; ruxolitinib (JAK1/2 inhibitor) for steroid-refractory (REACH-2 trial, Zeiser NEJM 2020).

Chronic GvHD

Classically >100 days. Multi-system, autoimmune-disease-like:

  • Sicca syndrome, oral lichenoid lesions
  • Sclerodermatous skin, joint contractures
  • Bronchiolitis obliterans
  • Liver cholestasis, oesophageal strictures
  • Vulvovaginal/cervical involvement

Treatment: steroids, ECP (extracorporeal photopheresis), ibrutinib (FDA 2017), ruxolitinib (REACH-3 NEJM 2021), belumosudil (ROCK2 inhibitor; ROCKstar 2022).

GvHD prophylaxis historically: calcineurin inhibitor (cyclosporine or tacrolimus) + methotrexate. Modern alternatives:

  • Post-transplant cyclophosphamide (PTCy) on days +3 and +4 — selectively eliminates alloreactive proliferating T-cells while sparing regulatory T-cells. Originally for haplo-HSCT; now also used in matched donors (BMT-CTN 1703 trial: PTCy + tacrolimus + MMF beat tacrolimus + MTX for GRFS).
  • ATG (anti-thymocyte globulin) — in vivo T-cell depletion; standard in many European protocols.
  • Abatacept — CTLA4-Ig; FDA-approved for paediatric and adult HSCT GvHD prevention (ABA2 trial).

Graft-versus-leukaemia (GvL) is the therapeutic flip side: donor T-cells and NK-cells recognise minor histocompatibility antigens (and sometimes leukaemia-specific antigens) on residual leukaemic blasts and eliminate them. The classic demonstration is donor-lymphocyte infusion (DLI)for relapsed CML in chronic phase, which can re-induce molecular remission in ~70–80% — often the cleanest immunotherapy result in oncology.

8. Post-Transplant Complications

Beyond GvHD, allo-HSCT carries a long list of complications:

  • Infections. Bacterial in early neutropenia. CMV reactivation in ~50% (letermovir prophylaxis since 2017 has reduced disease). EBV-driven post-transplant lymphoproliferative disorder (PTLD); rituximab pre-emptive. Adenovirus, BK, HHV-6 reactivations. Fungal — Aspergillus, Candida, mucormycosis. Pneumocystis prophylaxis with TMP-SMX.
  • Sinusoidal obstruction syndrome (SOS) / VOD. Conditioning- or inotuzumab-induced injury to hepatic sinusoids. Tender hepatomegaly, jaundice, weight gain, ascites. Treated with defibrotide.
  • Idiopathic pneumonia syndrome / DAH. Non-infectious lung injury within first 100 days; high mortality.
  • Transplant-associated thrombotic microangiopathy (TA-TMA). Endothelial damage, complement-driven; eculizumab in severe cases.
  • Engraftment failure. Primary (no recovery) or secondary (loss of graft); managed with second transplant.
  • Secondary malignancies. Skin cancers, oral SCC, MDS/AML in donor cells (rare), solid tumours after TBI.
  • Late effects. Endocrine (hypothyroidism, growth retardation, infertility), cataracts, osteoporosis, cardiovascular, neurocognitive (especially in children with TBI).
  • Relapse. Despite all this, the leading cause of late mortality after allo-HSCT for AML remains relapse (~30%).

TRM (transplant-related mortality) at 100 days has fallen from ~20–30% in the 1990s to ~5–10% in the 2020s with modern donor matching, conditioning, and supportive care.

9. MRD as Endpoint — the Future

MRD at multiple time-points has become the dominant predictor of post-transplant outcomes:

  • Pre-transplant MRD — MRD-negative pre-transplant patients have ~3-fold lower relapse than MRD-positive. The Buckley study (Blood 2017) and Walter (JCO 2017) showed MRD+ pre-transplant outcomes approach those of patients transplanted with overt residual disease.
  • Day +100 MRD — persistence indicates failure of GvL to clear residual leukaemia; pre-emptive intervention (DLI, hypomethylating maintenance, FLT3 inhibitor) is increasingly used.
  • Serial post-transplant MRD — rising chimerism loss or rising MRD precedes morphologic relapse by weeks to months; pre-emptive therapy is the goal.

Maintenance therapy after allo-HSCT has emerged for several settings:

  • FLT3-mutant AML — sorafenib (SORMAIN trial, Burchert JCO 2020), midostaurin, gilteritinib (MORPHO trial 2024).
  • High-risk AML/MDS — oral azacitidine (CC-486) maintenance (QUAZAR-AML-001, Wei NEJM 2020 — though chemotherapy not transplant maintenance).
  • Ph+ ALL — TKI maintenance for ~1–2 years post-transplant, especially with detectable BCR-ABL transcripts.

CML and operational cure. CML patients in MR4.5 for >2 years on TKI may attempt treatment-free remission (TFR); ~50% maintain MMR off therapy at 3 years. This is the closest haematology has come to non-transplant cure of a leukaemia at scale (STIM, EURO-SKI, A-STIM trials), and is the model for what MRD-guided therapy could become elsewhere.

The overall message: MRD is the new endpoint — the metric that tells us whether a leukaemia has been cured, will relapse, or sits at the threshold. The haematology-oncology of the next decade will be increasingly built around MRD measurements at all the canonical time-points: end-of-induction, consolidation, pre- and post-transplant, and during long-term follow-up.

We close the course where it began: leukaemia is the cancer that has driven cancer biology and oncology forward more than any other — from the Philadelphia chromosome to imatinib, ibrutinib, venetoclax, and CAR-T. The same is likely to continue. Cross-references: Cancer, Cancer / genetic basis, DNA repair, Cancer therapy, Pharmacology, Cell Physiology.

Key references for further reading. Copelan, Hematopoietic stem-cell transplantation, NEJM 2006; Gooley et al., Reduced mortality after allo-HSCT, NEJM 2010; Scott et al., Myeloablative vs reduced-intensity conditioning for AML/MDS (BMT-CTN 0901), JCO 2017; Kanakry et al., Modern HSCT for AML, Blood 2016; Luznik et al., HLA-haploidentical transplant + PTCy, Blood 2010; Zeiser et al., Ruxolitinib for steroid-refractory acute GvHD (REACH-2), NEJM 2020; Zeiser et al., Ruxolitinib for chronic GvHD (REACH-3), NEJM 2021; Burchert et al., Sorafenib maintenance after allo-HSCT in FLT3-ITD AML (SORMAIN), JCO 2020; Buckley et al., MRD before allo-HSCT, Blood 2017.
Share:XRedditLinkedIn