Molecular Methods for Prognostication
Gene expression profiling can be used to classify patients into 2 unique molecular subtypes: activated B-cell-like (ABC) and germinal center B-cell-like (GCB) DLBCL. Compared with GCB-DLBCL, survival outcomes for patients with ABC-DLBCL are poor, with estimated 5-year OS rates of 35% for ABC-DLBCL and 60% for GCB-DLBCL. Gene expression profiling has only been used in laboratory research and clinical trials thus far; limitations to widespread use of this technique include accessibility, time, and costs. Currently, an alternate method, immunohistochemical staining, is being used in place of gene expression profiling in clinical settings.
Other poor prognostic subtypes of DLBCL include double- and triple-hit lymphomas. At present, these subtypes require fluorescence in situ hybridization for identification. Newer methods such as the double-hit signature gene expression-based assay may allow for improved detection of these subtypes.
Next-generation sequencing is another genomic technology used to detect driver mutations among DLBCL subtypes. Several of these mutations are being investigated as potential biomarkers for upcoming personalized medicine approaches. Various clinical trials are under way, with the aim of detecting novel genome-directed therapies.
Given the recent expansion of mutational and gene expression data available for prognostic modeling, machine learning techniques are being developed to help translate these complex data into clinically meaningful tools. Advanced regression techniques and informative feature extraction methods such as principal component analysis and clustering methods are being studied. At present, whether these methods will help predict clinical outcomes remains largely unknown.
Researchers are investigating new methods of measuring treatment effectiveness in patients with DLBCL. Recent studies have suggested interim positron emission tomography-computed tomography imaging may be useful to help restage patients following the treatment completion. The technique has been adopted as the standard of care in patients with Hodgkin lymphoma, as it has been shown to improve clinical outcomes. Further studies are ongoing to examine whether this technique can be applied in other settings, such as tumor genetics. Circulating tumor DNA is another novel technique that could soon be used in this domain.
“Having reliable and reproducible predictive markers is extremely important to [identify the] patients most likely to benefit from a particular therapy,” Dr Zeidan said. He added that additional benefits may include “avoiding potential toxicities, associated costs, [and] time wasted on ineffective therapies.”
1. Harkins RA, Chang A, Patel SP, et al. Remaining challenges in predicting patient outcomes for diffuse large B-cell lymphoma [published online September 12, 2019]. Expert Rev Hematol. doi:10.1080/17474086.2019.1660159