Keyword: Liver Metastases

ABSTRACT: Introduction: The main method for preoperative assessment of colorectal cancer (CRC) liver metastases remains intravenous contrast-enhanced computed tomography (CT); however, its sensitivity is limited, especially for lesions smaller than 10 mm, and neoadjuvant chemotherapy further reduces it. The characteristic CT appearance of a metastasis is hypodensity in the portal phase and a rim of enhancement at the lesion edge, which can be enhanced by injecting contrast material directly into the hepatic arteries. CT-arteriohepaticography (CT-AHG) is potentially more effective, including during drug treatment, but data on its sensitivity after chemotherapy are virtually absent.
Objective: To compare the diagnostic efficacy of CT-AHG and intravenous contrast-enhanced CT in the detection and characterization of CRC liver metastases in patients after neoadjuvant chemotherapy.
Materials and Methods: This single-center prospective study included 22 patients with CRC and liver metastases who underwent both imaging methods prior to liver resection at the N.N. Petrov National Medical Research Center of Oncology. The unit of analysis was the lesion. The reference standard was targeted histological examination of each resected lesion. CT and CT-AHG were evaluated independently, using animated images, by radiologists with more than 5 years of experience, one for each "patient–method" pair. For each method, sensitivity, specificity, and predictive values with 95% Clopper–Pearson confidence intervals (CIs) were calculated for 129 included lesions. The difference in sensitivities was reported with 95% CI (patient-level bootstrap) and tested using McNemar's exact test.
Results: With combined use of CT and CT-AHG (a lesion was considered positive if detected by at least one method), sensitivity was 80.8% [72.6–87.4], which was 10.0 percentage points (95% CI: +2.4 to +17.9) higher than the sensitivity of CT alone. Individually, sensitivity was 70.8% [61.8–78.8] for CT and 73.3% [64.5–81.0] for CT-AHG. The difference between them was −2.5 percentage points (95% CI: −9.5 to +6.8) and was not statistically significant (p = 0.66). For lesions smaller than 10 mm, sensitivity was 35.3% for CT versus 50.0% for CT-AHG; for lesions 10 mm or larger, sensitivity was 84.9% and 82.6%, respectively.
Conclusion: The sensitivity of CT-AHG does not differ significantly from that of intravenous contrast-enhanced CT; however, the combination of methods improves sensitivity by 10.0 percentage points. Therefore, CT-AHG should be considered as a complement to CT, rather than a replacement for it.

ABSTRACT: Introduction: The extent of surgery for colorectal cancer (CRC) liver metastases is currently planned on a lesion-by-lesion basis, making lesion-by-lesion assessment of the diagnostic efficacy of radiological methods important. Such studies are few in number: they are labor-intensive, and lesions detected by different methods are difficult to correlate with each other. Typically, the correspondence between radiological findings and histological conclusions is established by liver segment or anatomical landmarks, which is unreliable in cases of multiple lesions.
Objective: To develop a methodology for lesion-by-lesion correlation of radiological data, intraoperative findings, and histological examination based on a three-dimensional liver master model.
Materials and Methods: A single-center study at the N.N. Petrov National Medical Research Center of Oncology included 18 patients with CRC liver metastases who underwent intravenous contrast-enhanced CT, CT-arteriohepaticography (CT-AHG), MRI with hepatobiliary contrast agent, and MR-arteriohepaticography (MR-AHG) 1–7 days prior to liver resection. Using 3D Slicer software, a master model—a three-dimensional reconstruction of the liver with numbered lesions—was constructed from CT data. Lesions detected by at least one imaging method, including those only visible prior to chemotherapy, were mapped onto the model. The lesion number linked the label on images, the surgeon's finding on intraoperative ultrasound (IOUS), and the gross specimen. Correspondence between model lesions and the gross specimen was established by a consensus of the radiologist, surgeon, and pathologist based on location relative to the capsule, resection margin, and vessels.
Results: In 18 patients, 147 lesions were mapped on the master models: median 5.5 per patient (range 1–27). Sixteen patients had more than one lesion. A total of 111 lesions were histologically verified (median 3.5 per patient), of which 103 were metastases; 36 non-resected lesions were excluded from accuracy calculations. The model enabled independent correlation of the four imaging methods: sensitivity was 77.7% for CT, 69.9% for CT-AHG, 56.3% for MRI, and 50.5% for MR-AHG; with combined assessment (at least one method), sensitivity was 90.3%: 93 of 103 metastases were detected, while 10 were missed by all methods. No additional lesions were identified on the gross specimen outside the master model.
Conclusion: When performing liver resection for CRC metastases, the master model provides a unified coordinate system from preoperative images to the gross specimen and continuous lesion numbering. This enables rigorous lesion-by-lesion verification in cases of multiple lesions, where segmental correspondence alone is insufficient.