High Metabolic Heterogeneity on Baseline 18F-FDG PET/CT Predicts Worse Prognosis of Newly Diagnosed Diffuse Large B-Cell Lymphoma

Introduction.

Diffuse large B-cell lymphoma (DLBCL) represents the largest entity of non-Hodgkin lymphoma. In spite of a remarkable improvement in the treatment of DLBCL patients, considerable proportion of the patients fail to cure. 18F-FDG PET/CT is routinely performed for staging and monitoring of DLBCL. Metabolic heterogeneity (MH) calculated using the PET images potentially reflects heterogeneities of glucose metabolism, blood flow, fibrosis, and hypoxia. While high MH at diagnosis predicts poor prognosis of primary mediastinal large B-cell lymphoma (PMBCL) and other malignancies (Ceriani L, Blood. 2018), the prognostic significance of MH in newly diagnosed DLBCL remains to be clarified. Recently, we reported that high total metabolic tumor volume (TMTV) 150cm³ or more predicts poor prognosis of newly diagnosed DLBCL after R-CHOP-like treatment (Senjo H, Cancer Med. 2019). In the current study, we explored the impact of MH in baseline PET-CT on prognosis of newly diagnosed DLBCL and also tested if MH could be correlated with TMTV in baseline PET-CT.

Methods.

We retrospectively evaluated the impacts of MH at diagnosis on overall survival (OS) and event free survival (EFS) in 86 patients with newly diagnosed DLBCL treated with R-CHOP-like regimens at Sapporo Hokuyu Hospital (training cohort). MH was determined using the area under curve of cumulative standardized uptake value-volume histogram (AUC-CSH) method, as previously described (Ceriani L, Blood. 2018). In patients with multiple lesions of lymphoma, the lesion with the highest TMTV was selected for MH evaluation. The results were verified in the independent validation cohort of 64 patients treated at Aiiku Hospital. The study procedures were in accordance with the Helsinki Declaration and institutional ethical guidelines, and were approved by the institutional review boards.

Results.

ROC curve analysis determined the optimal cutoff value of AUC-CSH as 0.481 for separating patients with EFS failure at 24 months, and therefore we defined AUC-CSH < 0.481 and ≥ 0.481 as MH high and MH low, respectively. In the training cohort (n=86), there was no significant difference in the patient characteristics between the MH low and high groups except the significantly higher incidence of bone marrow involvement in the MH low group compared to the MH high group (23.3% vs 0%, P=0.00108). Importantly, both OS and EFS were significantly lower in patients with high MH than in those with low MH [5-year OS (5-yr OS); 89.5% vs 61.2%, P=0.0122, 5-year EFS (5-yr EFS); 73.1% vs 51.1%, P=0.0327] (Figure A). In a univariate analysis, TMTV ≥150cm³ and high MH were associated with poor 5-yr OS and EFS. Pearson's correlation tests demonstrated no correlation between MH and TMTV [R²=0.137 P=0.208] (Figure B). A multivariate analysis that included MH and all factors in NCCN-IPI; age, LDH, clinical stage, ECOG PS and major organ involvement, demonstrated that age > 70 and high MH were independently associated with poor 5-yr OS (age; HR, 6.03; 95% CI, 1.63 to 22.3, P=0.00699, MH; HR, 5.68; 95% CI, 1.46 to 22.1, P=0.0121). We performed an additional multivariate analysis including both MH and TMTV. We found that both MH and TMTV persisted as independent prognostic factors in this multivariate analysis (MH; HR, 7.20; 95% CI, 1.49 to 34.7, P=0.014, TMTV; HR, 29.60; 95% CI, 2.93 to 300.0, P=0.00411; log-rank, Table). Combined with TMTV, MH stratified patients into three distinguishable prognostic groups; MH low/TMTV low with 5-yr OS 94.7% and 5-yr EFS 87.9%, MH low/TMTV high or MH high/TMTV low with 5-yr OS 77.7% and 5-yr EFS 61.0%, and MH high/TMTV high with 5-yr OS 45.7% and 5-yr EFS 31.5% (Figure C). In the validation cohort (n=64), we confirmed high MH predicted worse prognosis [5-yr OS; 68.6% vs 37.1%, P = 0.0254, 5-yr EFS; 57.3% vs 32.6%, P=0.0375] (Figure D). MH was not correlated with TMTV in the validation cohort either, and a combination of MH and TMTV again stratified the patients into three distinctive prognostic groups (Figures E and F). In the univariate analysis, MH was again associated with poor 5-yr OS and EFS. In the multivariate analysis, MH was associated with poor 5-yr OS. Altogether, we validated that prognostic values of MH in the patients with newly diagnosed DLBCL.

Conclusion.

High MH predicts worse prognosis in the patients with newly diagnosed DLBCL independently of NCCN-IPI and was not correlated with TMTV. Baseline MH is a novel prognostic biomarker in DLBCL.

Figure

View largeDownload slide

Figure

View largeDownload slide

Close modal