Application and Modification of Nutritional Assessment Tools in Hematologic Malignancies

Area of Science:

• Oncology
• Nutrition Science
• Biostatistics

Background:

• Hematologic malignancies are linked to metabolic dysregulation and malnutrition, worsening outcomes and survival.
• Existing nutritional assessment tools lack validation in hematologic cancer patients, creating an evidence gap.
• This study assesses the prognostic performance of current nutritional tools in this population.

Purpose of the Study:

• To systematically evaluate the prognostic performance of eight nutritional assessment instruments in patients with hematologic malignancies.
• To explore a novel composite assessment framework reflecting metabolic characteristics specific to this patient group.

Main Methods:

• Prospective cohort study of 1067 hematologic malignancy patients from the INSCOC registry.
• Analysis of eight nutritional assessment systems: PG-SGA, mPG-SGA, PG-SGA SF, abPG-SGA, NRS-2002, GLIM, Scored-GLIM, and KPS.
• Kaplan-Meier, Cox regression, linear regression, LASSO regression, AUC, and C-index analyses were used to assess prognostic value and identify biomarkers.

Main Results:

• All eight nutritional assessment tools showed statistically significant associations with survival but limited prognostic discrimination (AUCs ranging from 0.473 to 0.561).
• A novel composite tool, Hematologic Marker-Patient Generated Subjective Global Assessment (HMPG-SGA), incorporating the albumin-globulin ratio (AGR), demonstrated improved prognostic value (AUC 0.616, C-index 0.605).
• HMPG-SGA was significantly associated with overall survival (p < 0.001).

Conclusions:

• Conventional nutritional assessment tools have limited prognostic ability for overall survival in hematologic malignancy patients.
• The HMPG-SGA, integrating nutritional assessment with hematologic indicators like AGR, shows preliminary promise as a prognostic tool.
• Further validation of HMPG-SGA is warranted for clinical application in managing hematologic malignancies.