Molecular Profile and FDG-PET Metabolic Volume at Staging in DLBCL—Letter

We read with interest the study by Cottereau and colleagues (1) that was recently published in Clinical Cancer Research. Cottereau and colleagues (1) claim that early and end-of-treatment ¹⁸F-fluoro-2-deoxy-D-glucose PET (FDG-PET) can be used as good prognostic markers of outcome in diffuse large B-cell lymphoma (DLBCL), and propose to enhance prognostication by adding additional prognostic biomarkers by using baseline FDG-PET tumor volume metrics and gene sequence analyses. In their retrospective study that included 81 DLBCL patients (with gene expression profiling performed in 53 patients), high baseline total metabolic tumor volume at FDG-PET, ABC subtype, MYC positivity, and BCL2 positivity were reported to be associated with both a worse progression-free survival and overall survival, and the majority of these variables retained significant prognostic value after adjusting for the age-adjusted International Prognostic Index (aaIPI; ref. 1). Cottereau and colleagues (1) concluded that next-generation prognostic models will probably incorporate FDG-PET scan metrics, other imaging results, and sequencing data in conjunction with traditional IPI risk factors.

Although these results sound very promising, several comments have to be made. First, we do not agree with their claim that early FDG-PET and end-of-treatment FDG-PET can be used as reliable prognosticators. Multiple studies have demonstrated that FDG-PET fails to predict outcome in a huge proportion of patients in these settings (2, 3). Second, although several prognostic biomarkers retained significant prognostic value after adjusting for the aaIPI, adjustment for the newly developed (prognostically superior) National Comprehensive Cancer Network (NCCN)-IPI was not performed. In previous work, it was reported that baseline FDG-PET tumor volume metrics do not provide any prognostic information in DLBCL beyond that which can already be obtained by the NCCN-IPI (4). Third, FDG-PET scans and gene sequence analyses are very expensive tests, whereas other cheaper prognostic factors are available. For example, it has been proven that cheap and easily available parameters such as the hemoglobin level and C-reactive protein level provide additional prognostic value to the recently developed NCCN-IPI (5).

In conclusion, although Cottereau and colleagues reported baseline FDG-PET metrics and gene sequence analyses to provide additional prognostic value to the aaIPI score, such an approach cannot be recommended yet for use in clinical practice, considering the facts that no comparison was made with the superior NCCN-IPI and that there are much cheaper and easily available laboratory tests that provide at least equal prognostic value.

See the Response, p. 3414