AJR: CT & MR more accurate than ultrasound for liver malignancy resection
Multidetector CT (MDCT) is more sensitive than intraoperative ultrasound at detecting metastatic liver segments in patients undergoing liver resection, while both MDCT and MRI demonstrated superior value in identifying disease-free hepatic segments, according to a study published in the March issue of the American Journal of Roentgenology.
“Historically, intraoperative ultrasound has been considered the reference standard for liver imaging, with improved lesion detection compared with CT or MRI,” wrote Ute Wagnetz, MD, and colleagues from Princess Margaret Hospital in Toronto. “Accurate preoperative identification of the number and location of hepatic tumors is paramount in choosing suitable patients for surgery and in planning the extent of hepatic resection,” the authors continued.
Three-quarters of liver resections with curative intent are followed by recurrences, and despite 20 to 30 percent of colorectal cancer patients harboring occult liver metastases, ultrasound has remained the reference standard for surgeons. In addition, reports indicate the variable impact of the modality on surgical planning. Wagnetz and co-authors sought to compare the depiction of malignant lesions and to predict positive and negative metastatic segments using CT, MRI and intraoperative ultrasound in patients undergoing partial hepatic resection for metastatic liver disease or primary hepatic tumors.
The retrospective study included 292 patients, 168 of whom had metastatic colorectal cancer, 70 had hepatocellular carcinoma and 54 presented with other hepatic malignancies. Among this cohort, 247 underwent 64-slice MDCT, 64 underwent MRI and 19 were imaged by both modalities. Intraoperative ultrasound was utilized in all cases and the average cross-sectional imaging exam took place an average of 38 days prior to surgery.
MDCT displayed a significantly higher sensitivity to lesions, 96.8 percent, than intraoperative ultrasound, at 95.1 percent. When controlling for cancer type and time interval between cross-sectional imaging and surgery, MDCT maintained its significantly improved sensitivity compared to ultrasound, while MRI proved equivalent to the reference standard.
Out of a total of 663 resected segments, intraoperative ultrasound detected 95.2 percent of segments harboring malignancy, compared with 97.1 percent for MDCT and 94.6 percent for MRI. Once again, controlling for patient group and pre-operative imaging time intervals, MDCT was significantly more sensitive than ultrasound, while MRI did not differ significantly.
Both MDCT and MRI demonstrated superior negative predictive values to ultrasound, with MDCT accurately identifying 96.4 percent of negative segments, MRI 95.4 percent and ultrasound 90.5 percent.
“The current study, which was performed at a high-volume hepatopancreatobiliary center using current CT and MRI scanners, shows that the sensitivity of 64-MDCT is higher than that of intraoperative ultrasound for lesion detection and for identification of segments harboring malignancy for patients for the study group as a whole and specifically for patients with metastatic colorectal cancer,” the authors explained.
“In this study population, surgical management was altered after intraoperative ultrasound in less than 3 percent of patients,” Wagnetz and co-authors noted.
The researchers pointed out that all modalities “performed poorly” in correctly classifying the few benign resected lesions.
Wagnetz and colleagues drew attention to the fact that radiologists were necessarily aware of the findings from pre-operative cross-sectional imaging when evaluating intraoperative ultrasound, introducing “significant bias” in favor of this modality. Additionally, the authors did not assess contrast-enhanced ultrasound, which Wagnetz and colleagues acknowledged to have shown improved sensitivity in preliminary reports both to ultrasound and pre-operative cross-sectional imaging.
The researchers maintained the importance of intraoperative ultrasound for hepatopancreatobiliary surgeons, while concluding that “64-MDCT and MRI have an equivalent or higher sensitivity in identifying hepatic segments with malignancy, and both 64-MDCT and MRI appear to have a higher predictive value for identifying disease-free segments than does intraoperative ultrasound.”
“Historically, intraoperative ultrasound has been considered the reference standard for liver imaging, with improved lesion detection compared with CT or MRI,” wrote Ute Wagnetz, MD, and colleagues from Princess Margaret Hospital in Toronto. “Accurate preoperative identification of the number and location of hepatic tumors is paramount in choosing suitable patients for surgery and in planning the extent of hepatic resection,” the authors continued.
Three-quarters of liver resections with curative intent are followed by recurrences, and despite 20 to 30 percent of colorectal cancer patients harboring occult liver metastases, ultrasound has remained the reference standard for surgeons. In addition, reports indicate the variable impact of the modality on surgical planning. Wagnetz and co-authors sought to compare the depiction of malignant lesions and to predict positive and negative metastatic segments using CT, MRI and intraoperative ultrasound in patients undergoing partial hepatic resection for metastatic liver disease or primary hepatic tumors.
The retrospective study included 292 patients, 168 of whom had metastatic colorectal cancer, 70 had hepatocellular carcinoma and 54 presented with other hepatic malignancies. Among this cohort, 247 underwent 64-slice MDCT, 64 underwent MRI and 19 were imaged by both modalities. Intraoperative ultrasound was utilized in all cases and the average cross-sectional imaging exam took place an average of 38 days prior to surgery.
MDCT displayed a significantly higher sensitivity to lesions, 96.8 percent, than intraoperative ultrasound, at 95.1 percent. When controlling for cancer type and time interval between cross-sectional imaging and surgery, MDCT maintained its significantly improved sensitivity compared to ultrasound, while MRI proved equivalent to the reference standard.
Out of a total of 663 resected segments, intraoperative ultrasound detected 95.2 percent of segments harboring malignancy, compared with 97.1 percent for MDCT and 94.6 percent for MRI. Once again, controlling for patient group and pre-operative imaging time intervals, MDCT was significantly more sensitive than ultrasound, while MRI did not differ significantly.
Both MDCT and MRI demonstrated superior negative predictive values to ultrasound, with MDCT accurately identifying 96.4 percent of negative segments, MRI 95.4 percent and ultrasound 90.5 percent.
“The current study, which was performed at a high-volume hepatopancreatobiliary center using current CT and MRI scanners, shows that the sensitivity of 64-MDCT is higher than that of intraoperative ultrasound for lesion detection and for identification of segments harboring malignancy for patients for the study group as a whole and specifically for patients with metastatic colorectal cancer,” the authors explained.
“In this study population, surgical management was altered after intraoperative ultrasound in less than 3 percent of patients,” Wagnetz and co-authors noted.
The researchers pointed out that all modalities “performed poorly” in correctly classifying the few benign resected lesions.
Wagnetz and colleagues drew attention to the fact that radiologists were necessarily aware of the findings from pre-operative cross-sectional imaging when evaluating intraoperative ultrasound, introducing “significant bias” in favor of this modality. Additionally, the authors did not assess contrast-enhanced ultrasound, which Wagnetz and colleagues acknowledged to have shown improved sensitivity in preliminary reports both to ultrasound and pre-operative cross-sectional imaging.
The researchers maintained the importance of intraoperative ultrasound for hepatopancreatobiliary surgeons, while concluding that “64-MDCT and MRI have an equivalent or higher sensitivity in identifying hepatic segments with malignancy, and both 64-MDCT and MRI appear to have a higher predictive value for identifying disease-free segments than does intraoperative ultrasound.”