Name of the study
A randomized controlled trial comparing CT chest/abdomen and liver-MRI versus CT chest/abdomen would neither be efficient nor practical. Any beneficial effect on patient outcomes would only be observed in those with additional CS-CRLM on MRI; the existence and size of that subgroup is the main target of this accuracy study. In addition, routine liver-MRI is already integrated in daily practice in many centers. Within the study design proposed here, liver-MRI will not be withheld from any patient.
Sample size calculation indicate that 298 patients will have to be included to obtain estimates of change in management of 10% with 95% confidence intervals (CIs) that do not exceed 7% (95%CIs: 6.5%-13.5%). We plan to recruit these patients during a 16-month period.
Only patients who are considered candidates for curative surgery (+/- per-operative ablation) based on CT findings are eligible.
First, based on the CT findings, the liver surgery plan (+/- per-operative ablation) will be documented. Thereafter, findings at MRI will be discussed.
Researchers and contact information
Prof.dr. J. Stoker, Amsterdam UMC, loc. AMC
Colorectal cancer (CRC) is the third most common cancer worldwide. Approximately 50% of patients with CRC present with or will develop liver metastases (CRLM). Surgery is the cornerstone of curative intent therapy for these patients with 5-year survival exceeding 30%. Increasingly, minimally invasive surgery is performed (combined with local ablation) for which optimal preoperative CRLM localization is of utmost importance.
Current routine preoperative workup of CRLM consists of CT chest/abdomen. Recently, it has become clear that the diagnostic accuracy of liver-MRI for (small) CRLM is superior to CT. Whether liver-MRI should be added routinely or selectively (problem-solving) in the work-up of CRLM is, however, unclear as previous studies did not address this point. A recent survey by our group identified a large (15%-90%) inter-hospital variability in the use of liver-MRI in the Netherlands.
Additional liver-MRI can detect clinically significant (CS)-CRLM: lesions in liver part(s) not incorporated in the initially planned resection/ablation based on CT. Detection of CS-CRLM will lead to either more extensive resection/ablation or to refraining from resection/ablation altogether (impact on management and informed decision making).
As adding routine liver-MRI to the work-up of (presumed) CRLM may cause additional pre-surgery delays, any recommendation about its use should be guided by the relative frequency with which additional CS-CRLM are detected and hereby considering the role of routinely performed intraoperative ultrasound (IOUS).
The current large inter-hospital variability is caused by a lack of evidence on the incremental diagnostic accuracy of routine liver-MRI, potentially leading to suboptimal care. Routine liver-MRI in CRLM patients scheduled for surgery based on CT chest/abdomen would only be justified if it leads to a sufficient proportion of patients in whom additional CS-CRLM are found.
Aim & hypothesis
Primary aim is to perform a prospective, multicenter incremental diagnostic accuracy study in patients with (recurrent) CRLM to estimate the proportion of patients in which liver-MRI finds CS-CRLM in addition to CT chest/abdomen.
We hypothesize that liver-MRI will lead to the detection of additional CS-CRLM in at least 10% of patients with CRLM scheduled for surgery based on CT chest/abdomen.
Our secondary aim is to develop a prediction rule to identify specific subgroups (e.g. post-chemotherapy, bilobar CRLM) that would benefit most of liver-MRI. Furthermore, the additional cost of liver-MRI will be determined.
Multicenter, incremental diagnostic accuracy study in 15 Dutch hospitals.