Can functional tumor profiling lead to superior care within standard of care?: A meta-analysis of randomized controlled trials.

Abstract

e23092 Background: The adoption of novel drugs or therapies is generally dependent on randomized controlled trials (RCTs), as they provide the highest level of evidence. In contrast, the adoption of diagnostic tests, which do not have a therapeutic effect on their own, is typically based on whether they can meaningfully inform treatment decisions. However, while dozens of studies have repeatedly shown functional tumor profiling predicting a patient’s tumor response with 80% sensitivities and specificities or more, functional profiling has been held to a higher standard, with professional organizations and insurers often opposing its use due to the lack of RCT-level evidence for improved outcomes. Methods: We performed a systematic review and meta-analysis of randomized controlled trials (RCTs) comparing functional profiling–guided therapy with physician-directed standard care in patients with solid tumors. The primary outcome was defined as 1-year overall survival (OS); with secondary outcomes as 1-year progression-free survival (PFS) and objective response rates (ORR). Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated, and pooled estimates were obtained using inverse-variance weighting. Results: We identified four RCTs enrolling 472 heavily pretreated cancer patients with analyzable clinical outcomes. One RCT in recurrent glioblastoma (n = 78) reported 1-year overall survival, with 49% (21/43) of patients alive in the assay-guided group versus 23% (8/35) in the clinician’s choice group (OR 3.37; 95% CI 1.16–9.82, p = 0.026). Two RCTs (n = 159) reported 1-year progression-free survival, with 18% (15/83) of patients being progression-free in the assay-guided arms compared with 1% (1/76) in the clinician’s choice arms (OR 11.16; 95% CI 2.02–61.81, p = 0.006). Three RCTs (n = 394) reported objective response rates of 26.7% (54/202) in the assay-guided arms versus 13.5% (26/192) in control arms; however, substantial between-study heterogeneity was observed (I² = 81%), so the random-effects pooled estimate failed statistical significance (OR 4.25; 95% CI 0.66–27.38, p = 0.128). Of note, earlier studies reported no difference in median survival but did not report 12-month survival outcomes and therefore could not be included in landmark survival analyses. Conclusions: Earlier RCTs of assay-guided therapies did not demonstrate increased survivals, supporting the official position of guidelines that such tests should not override standard-of-care recommendations. However, more recent RCTs of more advanced tests, in which assay-guided therapy was used to choose among guideline-recommended options, have shown unprecedented improvements in patient outcomes and suggesting that assay-guided therapy might help clinicians to provide “superior care within standard of care.”