Statistical controversies in clinical research: prognostic gene signatures are not (yet) useful in clinical practice

Abstract

With the genomic revolution and the era of targeted therapy, prognostic and predictive gene signatures are becoming increasingly important in clinical research. They are expected to assist prognosis assessment and therapeutic decision making. Notwithstanding, an evidence-based approach is needed to bring gene signatures from the laboratory to clinical practice. In early breast cancer, multiple prognostic gene signatures are commercially available without having formally reached the highest levels of evidence-based criteria. We discuss specific concepts for developing and validating a prognostic signature and illustrate them with contemporary examples in breast cancer. When a prognostic signature has not been developed for predicting the magnitude of relative treatment benefit through an interaction effect, it may be wishful thinking to test its predictive value. We propose that new gene signatures be built specifically for predicting treatment effects for future patients and outline an approach for this using a cross-validation scheme in a standard phase III trial. Replication in an independent trial remains essential.

Weighted gene expression profiles identify diagnostic and prognostic genes for lung adenocarcinoma and squamous cell carcinoma

“…10 In contrast, a prognostic signature may offer insights into the course of a disease, the prediction of survival rates and the response to a specific treatment. 10,11 As far as NSCLC is concerned, many diagnostic signatures [12][13][14][15] and prognostic signatures [16][17][18][19][20] have been identified. For example, a recent study used 183 AC and 80 SCC patients as a training set and obtained a 42-gene signature to discriminate these two subtypes.…”

Bringing 3D tumor models to the clinic – predictive value for personalized medicine

“…What initially appeared to be a very promising risk prediction method has since been somewhat downgraded by results from the OPTI-MA Study. The concordance between test results was moderate with 40.7% of patients receiving varying results regarding risk and subtype categorization [55,56]. Genetic profiling similar to those described above for breast cancer has also been done for several nongynecologic cancers with comparable results [57][58][59].…”

“…Using the same principles that are already approved for predictive biomarkers in prospective studies with randomization procedures would improve the value, reliability, and most importantly the level of evidence for chemoresponse assays. This process is similar to the evidence‐based clinical evaluation proposed for multi‐gene assays . Clinical implementation would add a functional component to the tumor characterization prior to drug treatment begin similar to the multimarker omics technologies (Fig.…”

Making Meaningful Clinical Use of Biomarkers

“…Numerous newer disease signatures have been identified as mentioned earlier, but none has been clinically validated for routine use. Several additional reviews have addressed the misconceptions and shortcomings that contribute to the void between discovery and meaningful clinical use to replace current reference standards 1,24,25…”