AUTHORS
Matthew P Dandan, DO, Jeet Patel, DO2, Gregory Collins3
AFFILIATIONS
1 Medical College of Wisconsin, Milwaukee, United States
2 University of Kansas-Olathe, Kansas City, United States
3 University of Florida College of Medicine-Jacksonville, Jacksonville, United States
ABSTRACT
Background: High-dose melphalan with autologous hematopoietic cell transplantation (auto-HCT) is the most effective consolidation for transplant-eligible AL amyloidosis. Cardiac involvement is the dominant determinant of transplant risk, and melphalan dose reduction is common in these patients specifically to mitigate early treatment-related mortality (TRM). Whether this achieves its intended safety benefit, and at what cost to disease control, has not been examined in the cardiac population.
Methods: Secondary analysis of CIBMTR public dataset MM17-021 (440 patients, first auto-HCT for AL amyloidosis, US, 2014-2018). Sixty-six (15%) with unknown cardiac status were excluded, leaving 225 with cardiac involvement, classified as preserved intensity (MEL180/200; n=127) or marked de-escalation (MEL100/140; n=98). The primary endpoint was relapse/progression by cumulative incidence with death without relapse as a competing risk and Fine-Gray modeling, with cause-specific Cox regression as a check. Secondary endpoints were disease-free survival (DFS), any hematologic response (CR or PR) and CR at day 100, day-100 TRM, and overall survival (OS). Models adjusted for age, Karnofsky performance status (KPS), HCT-CI, and bortezomib induction. Prespecified sensitivity analyses added renal involvement (25% missing), excluded MEL100, and modeled dose ordinally.
Results: Median follow-up was 24.7 and 24.3 months. De-escalated patients were older (64 vs 60 years, p=0.004) and more often had KPS <90 (69.5% vs 51.2%, p=0.010) and HCT-CI ≥3 (73.5% vs 51.2%, p=0.001); bortezomib induction was more frequent in that group (73.5% vs 64.6%, p=0.20). The 24-month cumulative incidence of relapse was 8.0% with preserved intensity versus 27.2% with marked de-escalation (Gray's test p=0.017; 12-month 2.5% vs 19.4%). Preserved intensity was associated with lower relapse (Fine-Gray sHR 0.50, 95% CI 0.29-0.87, p=0.014), without increased early-toxicity events: the incidence of death without relapse did not differ (Gray's p=0.71), day-100 TRM was 3.9% versus 4.1% (p=1.00), and OS did not differ (HR 0.86, 0.41-1.84, p=0.71; 28 deaths). Day-100 response was higher with preserved intensity (81.6% vs 65.5%; adjusted OR 2.66, 1.29-5.50, p=0.008); CR did not differ (25.4% vs 16.7%; OR 1.69, 0.79-3.65). DFS was directionally consistent but not significant (HR 0.62, 0.37-1.03, p=0.062). Bortezomib induction was independently associated with lower relapse (HR 0.40, 0.22-0.72, p=0.002), reproducing the parent analysis. Excluding MEL100 (n=25), the relapse association persisted (sHR 0.51, p=0.033; 24-month incidence 8.0% vs 26.1%, Gray's p=0.033) whereas DFS attenuated (HR 0.69, 0.39-1.21, p=0.193). Modeled ordinally, each higher dose category was associated with lower relapse (HR 0.73 per level, 0.56-0.95, p=0.022); adding renal involvement strengthened it (HR 0.43, 0.22-0.85, p=0.015). By dose level, 24-month relapse incidence was 30.0% (MEL100), 24.9% (MEL140), 3.9% (MEL180), and 9.2% (MEL200); MEL180 (n=28) did not differ detectably from MEL200 (p=0.68).
Conclusions: In cardiac AL amyloidosis undergoing auto-HCT, marked melphalan de-escalation to ≤140 mg/m2 was not associated with lower day-100 TRM or improved OS, but with an approximately threefold higher 24-month relapse incidence and lower day-100 response. The relapse association persisted after excluding MEL100 anddespite more frequent bortezomib induction in the de-escalated group, an imbalance expected to favor that group, arguing against frailty alone as the explanation. Confounding by indication cannot be excluded, DFS was only directionally supportive, and the MEL180 subgroup is small. These results suggest melphalan dosage of 180 mg/m2 as a clinically relevant threshold and motivate prospective evaluation of risk-adapted conditioning in cardiac AL amyloidosis.
DISCLOSURES
Jeet Patel, DO: No relevant conflicts of interest to declare.