Diagnostic Accuracy of Imaging in Assessing Nonviability of Disappearing Colorectal Liver Metastasis

In patients with colorectal liver metastases (CLMs), the optimal treatment of disappearing liver metastases (DLMs) diagnosed on postchemotherapy computed tomography (CT) is controversial. To examine the diagnostic value of magnetic resonance imaging (MRI) (diffusion weighted, T1/T2, and contrast enhanced) and CT for accurate assessment of the nonviability of DLMs. This was a prospective international study including patients with initially unresectable CLMs downstaged to liver resection after chemotherapy at 21 centers in France, Austria, Belgium, the US, and Japan. A total of 233 patients were registered and 112 were enrolled between November 2016 and March 2021 with a minimum 2-year follow-up. Clinical cutoff was in September 2023, and data were analyzed from August 2024 to May 2025. Postchemotherapy evaluation with both CT and MRI was performed. DLMs were defined as lesions that had disappeared on CT. Confirmed DLMs (cDLMs) were defined as those that had disappeared on both CT and MRI. The primary end point was the negative predictive value (NPV) of MRI and CT in assessing the nonviability of cDLMs using either pathological complete response (for resected lesions) or the absence of recurrence at the site of cDLMs during the 2-year follow-up (for lesions left behind) to confirm the true lesion status. The planned sample size was 149 evaluable cDLMs, aiming at excluding an NPV of 0.85 or lower with a 1-sided α of 5% and a power of 90%. Among 112 total patients (mean [SD] age, 60.0 [10.4] years; 67 [59.8%] male) a total of 152 cDLMs and 227 DLMs were evaluable. The NPV of all evaluable cDLMs, either resected or left behind, was 62.5% (95/152; 90% CI,50.8-74.2), which was lower than the prespecified threshold. The NPV of DLMs was 52.9%. The NPVs of resected cDLMs vs those left behind were 56.8% (50/88; 90% CI, 44.2-69.5) and 70.3% (45/64; 90% CI, 48.6-92.0), respectively. For patients without extrahepatic metastases who had R0/1 resection, there was no significant difference in disease-free survival and overall survival between those with all cDLMs removed vs those with at least 1 cDLM left behind. Although the combination of MRI and CT was more accurate in detection of nonviable DLMs compared to CT alone, cDLMs did not correspond to nonviability in patients with initially unresectable CLM. Survival benefit associated with removal of cDLMs is still unclear in this setting. ClinicalTrials.gov Identifier: NCT02781935.