Total tumor volume and growth rate at 3 months on MRI outperform baseline imaging and circulating tumor DNA for predicting survival of patients with metastatic uveal melanoma receiving tebentafusp - 03/03/26
, Xavier-Louis Jasawant a, José Luis Sandoval b, Léah Mailly-Giacchetti c, Gaëlle Pierron d, Nathalie Cassoux e, Pascale Mariani f, Marc-Henri Stern c, Shufang Renault b, Aurore Rampanou b, Raphael Sanchez g, Sophie Piperno-Neumann g, Vincent Servois a, 1, Manuel Rodrigues c, g, 1Highlights |
• | Month-3 liver MRI tumor volume and growth rate independently predict overall survival in metastatic uveal melanoma patients. |
• | An MRI-based risk score yields three clearly distinct prognostic groups with a high C-index (0.79). |
• | Dynamic, treatment-informed MRI biomarkers outperform baseline imaging, circulating tumor DNA, and RECIST objective response. |
• | MRI-guided risk supports stratification in trials for patients with metastatic uveal melanoma treated with tebentafusp. |
Abstract |
Purpose |
The purpose of this study was to determine whether liver magnetic resonance imaging (MRI)-derived tumor burden and growth rate at 3 months better predict overall survival (OS) than baseline imaging, circulating tumor DNA (ctDNA), and RECIST objective response in patients with hepatic metastases from metastatic uveal melanoma (mUM) treated with tebentafusp.
Materials and methods |
A total of 88 patients with mUM treated with tebentafusp between 2018 and 2022 who underwent MRI examination at baseline and at three months following initiation of tebentafusp were retrospectively included. There were 52 women and 36 men, with a median age of 59.0 years (Q1, 52.0; Q3, 66.5; range: 30–77 years). Hepatic metastases were segmented to compute total tumor volumes at baseline (TTV 0 ) and at 3 months (TTV m3 ) and the exponential tumor growth rate between scans (TGR m3 , %/month). When available, ctDNA was quantified at the same time points. Associations with OS were assessed using Cox models and Harrell’s concordance index (C-index) with 95 % confidence intervals obtained by 1,000-iteration bootstrap resampling. A three-tier MRI risk score combined median-based TTV m3 and TGR m3 cutoffs.
Results |
In multivariable analysis, greater TTV m3 and TGR m3 were independently associated with shorter OS (hazard ratio [HR] per doubling of TTV m3 , 1.25; 95 % confidence interval [CI]: 1.15–1.35; HR per +10 %/month TGR m3 , 1.18; 95 % CI; 1.08–1.29; both P < 0.001), whereas baseline tumor volume and RECIST objective response were not. The MRI risk score yielded three well-separated OS curves (log-rank P < 0.001) and a C-index of 0.79 (bootstrap 95 % CI: 0.71–0.86), outperforming ctDNA-based models (C-index, 0.73; P = 0.01) and RECIST objective response (C-index, 0.65; P < 0.001).
Conclusion |
Dynamic liver MRI biomarkers in patients with mUM under tebentafusp, particularly month-3 total tumor volume and growth rate, provide independent prognostic information beyond baseline imaging, ctDNA, and RECIST and support MRI-based risk stratification as a central decision tool in mUM.
Le texte complet de cet article est disponible en PDF.Keywords : Circulating tumor DNA, Magnetic resonance imaging, Metastatic uveal melanoma, Tebentafusp, Tumor growth rate, Tumor volumetry
Abbreviations : 3D, AJCC, CD3, CI, C-index, ctDNA, DWI, gp100, HR, HyperT2, ICC, mUM, NR, OS, Q1, Q3, RECIST, SD, SF3B1, TCR, TGR, TGR m3 , TR, TTV 0 , TTV m3
Plan
Bienvenue sur EM-consulte, la référence des professionnels de santé.
L’accès au texte intégral de cet article nécessite un abonnement.
Déjà abonné à cette revue ?