Read on Twitter
#OncoAlert Results from the phase III TAHOE trial now available @JTOonline - compared 2L rovalpituzumab-teserine (Rova-T) versus topotecan in DLL3-high #SCLC. Stopped early for futility, not meeting its OS endpoint. #LCSM @IASLC https://www.jto.org/article/S1556-0864(21)01709-3/fulltext
Rova-T is an antibody-drug conjugate directed to DLL3, part of the Notch receptor ligand family, highly expressed on #SCLC and LCNE but not on normal tissue. The phase I study by @charlesrudin was genuinely exciting with clear activity, especially in DLL3-high SCLC (RR 38%)
TAHOE was a pure 2nd line #SCLC trial. Enriched population of only high DLL3 expression by IHC (75% by SP347). Allowed treated stable CNS metastases. Primary endpoint OS. 2:1 randomization to Rova-T vs standard q3w topotecan. Included stratification by platinum sensitivity.
TAHOE was stopped early by the IDMC when OS noted to be shorter with Rova-T than topotecan. Median OS 6.3m with Rova-T and 8.6m with topotecan (HR 1.46). Did control overperform? In the topo vs CAV, median OS ~6m (1999) but in more recent studies (CM-331), chemo median OS 8.4m.
Other efficacy outcomes also favored topotecan over Rova-T including response rate (15% Rova-T vs 21% topo) and PFS (mPFS 3m Rova-T vs 4.3m topo, HR 1.51). DOR 3.5m Rova-T and 4.9m topo. Global health scores at final assessment also worse with Rova-T. #LCSM
In addition, Rova-T significantly more toxic. Notable adverse events include pleural effusions (29% Rova-T vs 4% topo), pericardial effusions (20% vs 2%), photosensitivity reactions (16% vs 0%). Fatal drug related adverse events seen in 5 patients receiving Rova-T (none on topo).
Solid design - randomized phase 3 with biomarker enriched population but another example of a promising agent that could not beat topotecan for OS. Disappointing results but DLL3 remains a potential therapeutic target (see AMG-757, a DLL3-BiTE) https://clinicaltrials.gov/ct2/show/NCT03319940
Congratulations to the team for a well conducted study and for stopping the study when it was clear that the experimental arm could not be better than SOC. @fblackhall @LGreillier @NReguart @fedcappuzzo @trfmkt
