Prognostic value of lesion dissemination in doxorubicin, bleomycin, vinblastine, and dacarbazine-treated, interimPET-negative classical Hodgkin Lymphoma patients: A radio-genomic study

Abstract

We evaluated the prognostic role of the largest distance between two lesions (Dmax), defined by positron emission tomography (PET) in a retrospective cohort of newly diagnosed classical Hodgkin Lymphoma (cHL) patients. We also explored the molecular bases underlying Dmax through a gene expression analysis of diagnostic biopsies. We included patients diagnosed with cHL from 2007 to 2020, initially treated with ABVD, with available baseline PET for review, and with at least two FDG avid lesions. Patients with available RNA from diagnostic biopsy were eligible for gene expression analysis. Dmax was deduced from the three-dimensional coordinates of the baseline metabolic tumor volume (MTV) and its effect on progression free survival (PFS) was evaluated. Gene expression profiles were correlated with Dmax and analyzed using CIBERSORTx algorithm to perform deconvolution. The study was conducted on 155 eligible cHL patients. Using its median value of 20 cm, Dmax was the only variable independently associated with PFS (HR = 2.70, 95% CI 1.1-6.63, pValue = 0.03) in multivariate analysis of PFS for all patients and for those with early complete metabolic response (iPET-). Among patients with iPET-low Dmax was associated with a 4-year PFS of 90% (95% CI 82.0-98.9) significantly better compared to high Dmax (4-year PFS 72.4%, 95% CI 61.9-84.6). From the analysis of gene expression profiles differences in Dmax were mostly associated with variations in the expression of microenvironmental components. In conclusion our results support tumor dissemination measured through Dmax as novel prognostic factor for cHL patients treated with ABVD.

Keywords: Dmax; Hodgkin's lymphoma; PET; gene expression profiling; tumor dissemination; tumor microenvironment.

FIGURE 1

Association of Dmax with progression free survival (PFS) in classical Hodgkin Lymphoma (cHL). (A) Outline of the study (B) PET/CT scan representative of the largest lesions distance defined as low‐Dmax (left) and high‐Dmax (right) calculated in cHL patients having at least 2 detectable lesions. Red arrows indicate the lesions taken into account (C) Kaplan‐Meier curves for PFS among cHL patients with high‐ and low‐Dmax defined on the basis of the median value (20 cm) (D) Kaplan‐Meier curves for PFS in the overall cHL cohort among iPET+ and iPET‐patients with high‐ and low‐Dmax

FIGURE 2

Validation of Dmax cut‐off (A) Definition of Dmax Cut‐off by means of ROC curve. methods and Youden index considering 36 months of follow up (B) Definition of Dmax Cut‐off by means of maximally selected log‐rank test. (C) Histogram of max.star representing 1000 bootstrapped resamples

FIGURE 3

Gene expression analysis associated to tumor dissemination. (A) Volcano plot displaying differential expressed genes between classical Hodgkin Lymphoma (cHL) patients with high‐ and low‐Dmax. Black dots represent genes significantly deregulated (pValue ≤ 0.05) and dotted lines indicate absolute FC ≥ 2. (B) Principal Component Analysis (PCA) shows the variance between high‐ (red dots) and low‐ (gray dots) Dmax samples explained by the 61 genes differentially expressed with FDR ≤0.05. (C) Histogram representing Fold Change (FC) of the 61 genes differentially expressed (FDR ≤0.05) in high‐versus low‐ Dmax samples

FIGURE 4

Cell populations associated to tumor dissemination (A) GO analysis of the 40 genes down‐regulated in high‐versus low‐Dmax samples. B‐cells related and T‐cells related pathways are highlighted in red and blue respectively (B) Box plot representing the distribution of cell populations analyzed by CIBERSORTx among high‐ and low‐ Dmax samples