Total marrow and lymphoid irradiation (TMLI) with fludarabine-melphalan (FM) conditioning for matched donor hematopoietic cell transplant (HCT) in older patients with Relapsed/Refractory (R/R) disease

Abstract

Abstract Older adults with R/R acute leukemia or MDS have dismal treatment outcomes. TMLI is a targeted radiotherapy platform that enables radiation dose intensification to disease sites with minimized exposure to other organs. We have shown safety and efficacy of TMLI-based conditioning, up to 20Gy, in younger patients (pts) with active disease. Here, we present final results of a phase 1 trial evaluating TMLI with FM as conditioning for matched donor HCT in older adults with high-risk disease (NCT03494569). Our study incorporates comprehensive correlative assessments. Eligibility was age ≥55 year or less if not eligible for myeloablative conditioning, r/r acute leukemia (morphologically or MRD flow) or MDS, and available matched donor. Conditioning regimen was TMLI (days -8 to -5), fludarabine (30 mg/m2/day; days -4 to -2), and melphalan (100 mg/m2; day -2). TMLI dose was escalated from 12 to 18 Gy. All pts received peripheral blood stem cells and GVHD prophylaxis was tacrolimus/sirolimus. Primary endpoints were dose-limiting toxicity (DLT) and TMLI RP2D; secondary endpoints were engraftment, GVHD incidence, and survival. Correlative analyses were serial plasma cytokines, longitudinal flow cytometry on blood, and BM analysis. We treated 30 pts: 12Gy (DL1; n=6), 14Gy (DL2; n=6), 16Gy (DL3; n=12), and 18Gy (DL4; n=6). Median age was 63 years (range 53-73), 67% had KPS≥ 90%, and 47% had HCT-CI of 0. pts had AML (n=26; 86%, 21 had adverse risk), MDS (n=3) or ALL (n=1). Almost half the patients had >5% blasts (n=12/26, 46%) with the remaining being in CR1/2 with +MRD. Donors were unrelated in 80%. The median CD34 cell dose was 6.09×10⁶/kg. Median off-target organ dose was 6-7Gy (lungs, upper GI, heart, testes, parotids, thyroid), 7-8Gy (lower GI, liver, bladder, brain), or 3.5Gy (Oral cavity). Almost all pts had ≥ 1 G3 GI toxicity by CTCEA: abdominal pain, diarrhea, nausea, Ileus, colitis (DL1 n=5; DL2 n=6; DL3 n=10; DL4 n=3) and 1 G4 diarrhea at DL3. Mucositis in mouth, throat or esophagitis was common at lower grades (DL1 n=1; DL2=3; DL4 n=8; DL4 n=5). Other G3 non-heme AEs were fatigue (DL3 n=2; DL4 n=1) or bone pain (DL2 n=1; DL3 n=5; DL4 n=2). G3 Febrile neutropenia was common with a trend to increase at DL4 (DL1 n=3; DL2=3; DL3 n=9; DL4 n=6). Three pts had DLTs mostly renal, GI, and pulmonary (DL3 n=2; and DL4 n=1), all were non-related to TMLI. For the expansion cohort (n=6), 16Gy was selected as RP2D. All pts engrafted with full donor chimerism: median time to neutrophil and platelet engraftment were 14 days (range: 13-17) and 18 days (range: 16-29). All evaluable pts achieved CR with negative MRD on day 30. At a median 24 months (range=4-25) follow-up duration, 2-year overall and progression-free survival were both 55% (95% CI: 35-71 and 95% CI: 0.35-0.71, respectively). CI of 2-year relapse and non-relapse mortality (NRM) were 17% (95% CI: 6-33) and 28% (95% CI: 13-45), respectively. Day-100 CI of grade 2-4, 3-4 aGvHD and 2-year cGvHD were 30% (95% CI: 15-47),13% (95% CI: 4-28), and 29% (95% CI: 13-46), respectively. Dose-dependent increase was seen in IL-6, ST2, Reg3A, and TNFα, peaking at day +7. Day 14 ST2 levels correlated with GI dose (Spearman correlation; 0.48; p<0.01); no correlation was seen with Reg3A (p=0.15). No significant association was seen between ST2 or Reg3A levels and GVHD. GVHD CI was similar in pts receiving 16Gy vs. lower DLs (33% vs. 30%), or between those with lower GI dose above vs below the median 7.17Gy (range 4.08-10.92). Immune reconstitution was robust across all DLs: ALC increased from ~0.6 × 10⁹/L at day +30 to >1.5 × 10⁹/L by 1-year; CD4⁺ T cells reached ~500–600/µL by 1-year despite delayed early recovery at higher doses. CD8⁺ T cells, CD56+ NK cells, and Tregs recovered steadily, while CD19+ B cells had later normalization. Immune reconstitution was not different among pts receiving 16Gy TMLI vs lower DLs. Day 30 Treg-recovery above the median (6.84/µL) was associated with better OS (p=0.01), PFS (p=0.01), relapse (p=0.08), and grade 2-4 aGvHD (p= 0.079). BM cellularity at 1-year was normal for age across all DLs. BM mesenchymal cells relative frequencies were not altered across DLs. TMLI combined with FM was feasible and safe in older pts with high-risk r/r AL and MDS with acceptable NRM. Despite transient cytokine elevations and early GI toxicity, GvHD rates and immune recovery were preserved even at higher DLs. Disease control was promising in this advanced and high-risk population.