CT-Based Muscle Quality and Visceral Obesity Prior to Allogeneic Hematopoietic Stem Cell Transplantation Is of Prognostic Relevance in Patients with Acute Myeloid Leukemia

Introduction

Non-relapse mortality (NRM) remains a relevant risk for acute myeloid leukemia (AML) patients (pts) undergoing allogeneic hematopoietic stem cell transplantation (HSCT). CT-defined body composition parameters are increasingly recognized as modifiable risk factors in cancer pts. Here, quantity and quality of skeletal muscle areas (SMA) as surrogate parameters for sarcopenia, and the amount of adipose tissue are associated with NRM. Large analyses in pts with hematological malignancies are scarce. We addressed this topic here in AML pts receiving HSCT.

Methods

We retrospectively analyzed 429 AML pts who received routine abdominal CT scans to exclude infectious foci up to 7 weeks prior to myeloablative (25%), reduced intensity (15%), or non-myeloablative (NMA, 60%) HSCT. Median age at HSCT was 59 (range 17-76) years (y) and pts were in 1^st complete remission with/without blood count recovery (CRc, 65%), later CRc (17%), or relapsed/refractory (18%). European LeukemiaNet (ELN22) risk at diagnosis was favorable 22%, intermediate 27% and adverse 51%.
Body composition parameters were determined in CT on the level of the third lumbar vertebra. SMA and visceral adipose tissue (VAT) were measured on a predetermined range of radiodensity expressed in Hounsfield Units using ImageJ software. SMA was divided by the body height squared to calculate skeletal muscle index (SMI). Muscle quality was defined by intramuscular adipose tissue content (IMAC), measured using the muscle radiodensity, divided by that of the subcutaneous adipose tissue – with a higher IMAC indicating lower muscle quality due to a higher adipose tissue portion. Cut-offs were determined using ROC curves in a test set (1/3 of pts) and validated in a confirmation set (2/3 of pts) randomly partitioned by sex.

Results

Pt characteristics and outcomes did not differ between the test and confirmation set (cumulative incidence of relapse [CIR] P=.52, NRM P=.79, OS P=.20).
Of all pts, 69% were sarcopenic, defined by lower SMI (cut-off male [m] 52.36cm²/m²; female [f] 40.38cm²/m²). Pts with sarcopenia had a lower body mass index (BMI) at HSCT (P<.01), a lower IMAC (P<.01) and a lower VAT (P<.01).
13% of pts were viscerally obese, defined by higher VAT (cut off m 159.85cm², f 106.26cm²). Pts with visceral obesity were older (P=.03), had a higher BMI at HSCT (P<.01), a higher SMI (P<.01) and higher IMAC (P<.01). They also had a higher incidence of a complex karyotype (CKT, P=.01), more SRSF2 mut (P=.03) and more often developed a chronic graft versus host disease (GvHD, P=.01) after HSCT.
69% of pts had low muscle quality, defined by higher IMAC. They were older (P<.01), had a higher BMI at HSCT (P<.01), a higher VAT (P<.01) and a higher SMI (P<.01). They also less frequently had de novo AML (P=.01), a higher incidence of a CKT (P=.03), more SRSF2 mut (P=.03) and worse ELN22 risk (P=.03). Pts with a high IMAC had a higher HCT-CI score (P=.02), more often received NMA HSCT (P<.01) and more often received HSCT from unrelated (P<.01) or mismatch donors (P<.01). After HSCT, they more often developed acute (P<.01) and chronic GvHD (P=.04).
Outcomes did not differ significantly according to the presence of sarcopenia (CIR P=.14, NRM P=.08, OS P=.20). Viscerally obese pts had similar CIR (P=.29), but a higher NRM (P=.03) which translated in a trend for shorter OS (P=.06). Pts with high IMAC had similar CIR (P=.25), but higher NRM (P<.01) and shorter OS (P<.01).
In multivariate analysis, low muscle quality (i.e. higher IMAC) and visceral obesity (i.e. higher VAT) remained prognostic for higher NRM (Hazard ratio [HR] 3.40, 95% confidence interval [CI] 1.67-6.94, P<.01 and HR 1.85, 95% CI 1.05-3.28, P=.04) after adjustment for ELN22 risk and a higher IMAC for shorter OS (HR 0.53, 95% CI 0.35-0.8, P<.01) after adjustment for remission status at HSCT.
As the IMAC significantly associated with OS, but differed according to age and BMI at HSCT, we performed additional subgroup analyses. Here, a high IMAC had the highest impact on OS in younger pts (≤60y P<.01, less in pts >60y [P=.04]) and pts with a BMI ≤ 25kg/m² at HSCT (OS P<.01), with a lower impact in overweight (P=.07) and none in obese pts (P=.40).

Conclusions

In AML pts undergoing HSCT, visceral obesity associated with higher NRM and low muscle quality with higher NRM and shorter OS. Strategies to improve mobilization, physical activity and nutrition during treatment should be investigated to improve outcomes of these patients.