Abstract
Background:
Patients with aggressive cancer (metastatic solid tumor (mST) or aggressive lymphoma (AL)) are at an increased risk of both thrombosis and death. The Khorana Score (KS) predicts cancer-associated thrombosis using hematologic parameters and the body mass index (BMI). In this study, we assessed the relationship of routine hemogram (complete blood count, CBC) parameters, ABO blood type and BMI with overall survival in patients with aggressive lymphoma and metastatic solid tumors.
Methods:
Utilizing the electronic medical record (EMR) and tumor registry from the University of Miami/Sylvester Comprehensive Cancer Center, patients were identified and data extracted, from 2018 to 2023. We analyzed patients with metastatic lung (ML), metastatic colorectal (MC) and metastatic prostate (MP) cancer, and aggressive lymphoma (AL). Prostate cancer tends to be more indolent, while ML and MC are more aggressive. This study was approved by the Institutional Review Board. Baseline demographics, CBC, and BMI were extracted from the EMR prior to initial treatment. We analyzed the association of hemoglobin (Hgb), white blood cell count (WBC), platelet count, absolute neutrophil count (ANC), absolute lymphocyte count (ALC), mean corpuscular volume (MCV), ABO blood group, and BMI with overall survival (OS), measured in days from cancer diagnosis. mST and AL were analyzed separately. Continuous variables were divided into quintiles (Q1-Q5) and analyzed using ANOVA to determine association with OS. ABO blood type was stratified into type O and non-type O and analyzed using t test. Data analysis was performed using R (version 4.3).
Results:
Of 1709 patients, 1105 (64%) had mST (621 ML, 281 MC, 203 MP), and 604 (36%) had AL. 59% were male, and the mean age at diagnosis was 62 years. For mST patients, pre-treatment Hgb levels, WBC, platelet counts, ANC, and ALC were significantly associated with OS. The effect was parallel to the known thrombosis effect of the KS, i.e. lower Hgb, higher WBC, and higher platelet count were associated with reduced OS. Higher ANC and lower ALC were associated with reduced OS. Both low BMI (16-25) and elevated BMI (30-54) were associated with reduced OS, compared with Q3, BMI 25-27. MCV and ABO blood group were not associated with OS. None of these relationships were observed in AL.
The relationships of hematologic parameters and OS in mST tended to be continuous across the quintiles, even within normal, or minimally abnormal values. As one example, from minimal anemia at baseline (Q2, Hgb 10.7-12.2 gm/dL) to normal (Q5, 14.3-17.5), OS increased from 425 to 622 days, a 46% increase, p<0.001. Similarly, from baseline WBC of 6.2-7.7 X 10^9/L (Q2) to WBC 9.4-12.0 (Q4), OS decreased by 34% (640 to 420 days) (p<0.001).
To assess if these associations simply reflected differences in baseline parameters with different cancer types, we analyzed the three ST categories separately. In ML, the relationship of Hgb, WBC, ANC, ALC with OS maintained significance, while platelet count showed a similar, but not significant trend. In MC, the relationship of WBC, platelets, ANC with OS were significant. In MC, the effect of anemia could not be meaningfully assessed as >80% of the patients were anemic at baseline. In MP, none of the relationship of CBC parameters with OS were significant.
Conclusions:
Several baseline hematology parameters significantly correlate with OS in solid tumor patients. In addition to the parameters of the KS, associated with thrombosis, we also assessed ANC, ALC, MCV, and ABO. Subgroup analysis, by cancer type, indicated that the relationships were not significant in MP or AL, however, in those cancer types there was no conflicting trends.
The magnitude of the survival effects was much larger than one would expect from a possible contribution of cancer associated thrombosis. The adverse association with elevated WBC, ANC, and platelet counts argue against a possible adverse impact on ability to administer treatments. We hypothesize that the differences of hematologic values may reflect levels of cytokines or other markers of inflammation. The absence of correlation of MCV with OS, despite the adverse association of anemia, suggests the effect is due to anemia of chronic inflammation and not iron deficiency.
Our study is limited by its single institution data and retrospective nature. Ultimately, one will need to confirm these findings and determine the mechanism(s) of the associations.
Byrnes:Sanofi: Consultancy; Anthos Therapeutics: Research Funding. Soff:Amgen: Research Funding; Janssen Scientific Affairs: Consultancy; Sobi/Dova Pharmaceuticals: Consultancy, Research Funding; Anthos Therapeutics: Research Funding; Alpine Immune Sciences, a Vertex Company: Research Funding; Sanofi: Consultancy; Agios Pharmaceuticals: Consultancy.