Liver resection by Ultrasonic Dissection and lntraoperative Ultrasonography

Ultrasonic dissetion (USD) and intraoperative ultrasonography (IOUS) have shown encouraging results in a retrospective analysis of 109 patients with benign or malignant liver disease. Of 109 patients assessed between 1980 and 1993, 84 were resected: 27 by finger fracture technique (FFT) and 57 by USD. Hospital mortality was 4.8% (4/84) and 30-day mortality was 6.0% (5/84). Overall morbidity was 48.8% (41/84) and liver related morbidity (hepatic bleeding, sepsis, and bile leak) was 34.5% (29/84); of the 29 patients, 5 required re-operation. Liver complications occurred in 12/27 (44.4%) in the FFT group as opposed to 17/57 (29.8%) in the USD group. The incidence of postoperative hepatic bleeding was significantly less by USD than by FFT(p=O.03). As well, intraoperative blood loss (p=O.01)number of intraoperative blood units used (p=0.002), and postoperative length of stay (p=O.O09) have been significantly reduced by USD. IOUS was used on 64 patients. Not only has it improved the sensitivity (99%) and specificity (98%) for detection of hepatic neoplasms, it has also helped increase the precision and accuracy of anatomical tumour localization. As a result, 11/64 patients (17.2%) had their preoperative plans changed: 8 were abandoned and 3 were revised. In summary, USD has significantly reduced intraoperative blood loss and hence reduced the number of intraoperative transfusions, incidence of postoperative complications and postoperative length of stay. IOUS should be routinely employed in patients undergoing liver resection since it provides critical information that could obviate oncologically useless resections.


ique for hepa
ic resection has focused mainly on maximizing hemostasis..Recently, ultrasonic dissection has been given considerable attention partly because of its wide- 234 spread surgi.calapplications".As well, intraoperative ultrasonography has been able to give higher resolution images for better tumour localization and anatomical definition over any other preoperative radiological de- vice5.By directly examining the liver with an ultrasound probe, it has allowed the surgeon to clearly map out each individuals' anatomical and pathological vari- ances for precise and accurate resections.

The purpose ofthis study is to retrospectively assess how the ultrasonic dissector compares to the standard finger fra ture technique with respect to hepatic re- lated morbidity both intra-and postoperatively in ma- jor and minor liver resections.The influence of intraoperative ultrasonography on the decisions made during laparotomy is also examined.


MATERIALS AND METHOD

Between 1980 and 1993, 109 patients were assessed for liver surgery by the primary author (S.

H) at the Division of
eneral Surgery, Sunnybrook Health Sci- ence Centre-University of Toronto (Toronto, Canada).Twenty and 89 patients presented with be- nign and malignant disease, respectively (Table 1).

Their ages ranged from 21 to 78 years with a mean age of 56.2 fifty-one were female and 58 were male.Preoperative work-up included ultrasonography (US), computed tomography (CT) scan, and a serum 121   biochemical liver profile in all patients.Where war- ranted, angiography, radionuclide scanning and percutaneous biopsies were also done to aid in the diagnosis.CT portography (CTP) was performed on 30 patients since its advent in 1990 and magnetic resonance imaging (MRI) was incorporated as a diag- nostic test since 1991

All 109 patients were surgically explored.The final decision as to whether or not to proceed with resection was made during laparotomy.All explorations con- sisted of general inspection and palpation of the abdo- men including careful examination of the liver and, since October 19896, intraoperative ultrasound (IOUS) of the liver.Patients with malignant disease were considered unresec able if there was diffuse he- patic disease, if there was extensive extrahepatic dis- ease, or if the patient had advanced cirrhosis.Of the 109 patients explored, 84 were amenable to resection.

The finger fracture technique (FFT) was used on 27 patients and the ultrasonic dissector (Cavitron Ultra- sonic Surgical Aspirator, CUSAR, Valleylab, Pfizer Co) was used on 57 patients; since 1988, the CUSA was routinely used on almost every patient thereafter.Fifty-one patients underwent major liver resections (lobectomies and triseg entectomies) and 33 under- went minor resections (removal of 3 or fewer separate segments as defined by Couinaud) (Table 2).All resec- tions performed by the primary author were intended for cure.

Since its introduction to the author's practice in October 1989, IOUS was used on 64 ofthe 109 patients analyzed.From the 27 patients in the FFT group, 3 had IOUS; from the 57 patients in the CUSA group, 31 had IOUS.Twenty of the 25 patients that were not resected also had IOUS.IOUS of the liver was done with an Aloka SSD-630 (Omnium Medical Devices, Ontario) sonographic unit with both linear 7.5 MHz (556T-75) and 5 MHz (587T-5) transducers.Liver surfaces were exposed for optimal imaging and were systematically examined in perpendicular planes.Neoplasms were carefully localized in accordance with Couinaud's segmental anatomy and in relation to key vascular structures such as the vena cava, the portal vein and the hepatic vein.

All patient records were maintained on a computer database (Smart and Smart II Database Systems, Informix Software, Inc.) for retrospective analysis.Statistical comparisons were made between the FFT group and the CUSA group.To make the compari- sons more statistically uniform, separate comparisons were made with patients only undergoing major resec- tions (lobectomies and trisegmentectomies) in the two groups.Estimated blood loss, the number of peri- operative blood units used, and the postoperative length of stay were compared using Wilcoxon Rank Sum Tests.The rates of postoperative hepatic compli- cations in the two groups were compared using Fish- er's Exact Test.Postoperative hepatic complications included bile leak, sepsis and hepatic bleeding.The influence of IOUS was measured by how many operative plans it altered and by how many neoplastic lesions it detected.This number was compared to the actual number determined by pathology in those pa- tients who were resected.All patients would proceed to surgery from pre-operative assessment and then would have IOUS.Its result would influence outcome in one of three ways: proceed with resection plan; alter resec- tion plan; or abandon the resection.If surgery was abandoned because of unforeseen tumour burden, it would be confirmed either by tissue biopsy examined by frozen section or by obvious gross and palpable intraopative assessment.A true positive was gauged by histological examination of each specimen.Five to 10 milimetre slices (depending on the size of the speci- men) would be histologically examinedfor malignant cells after gross inspection.A true negative per se could not be measured since each specimen contained tumour.

The presence of false positives, however, which could be found from histological examination of all preoperative wo

ed-up an
IOUS detected lesions, re- quires some definition of a true negative to measure specificity.Thus, a true negative will be arbitrarily defined as the number of specimens in which the number of lesions found from either preoperative work-up or IOUS was equal to the number of lesions found by pathology.


RESULTS

Overall morbidity was 48.8% (41/84) and hepatic related morbidity (hepatic bleeding, sepsis and bile leak) was 34.5% (29/84); 5/29 required re-operation (Table 4, Figure 1).All patients (7) who experienced hepatic bl

ding alone underwent ma
or resection; no patient experienced postoperative hepatic bleeding from minor resections.The patients that underwent major resections showed a hepatic related complica- tion rate of 31.2% (10/32) in the USD group and 50.0%(9/18, 19t patient was intraoperative death) in the FFT group.


Mortality andMorbidity

Overall hospital mortality (i.e.death within the same hospital admission) was 4.8% (4/84) and 30-day mortality was 6.0% (5/84).These included patients with advanced cirrhosis, severe jaundice and other premorbid factors (Table 3).


Operative Technique

Table 5 and Figures 2-4 summarize the comparison between the USD and FFT group with respect to estimated blood loss (EBL), number of blood units (NBU) required and postoperative length of stay (PLOS).Of the 64 patients that underwent IOUS, 11 plans were altered (17.2%).Eight plans for resections were abandoned and 3 plans were amended.IOUS revealed 2 patients with diffuse unresectable hepatic metastases from colorectal cancer (Figure 5); 3 patients with le- sions adherent to vascular structures (vena cava, the portal and hepatic veins) in which their nature made it technically unfeasible for resection (Figure 6); with a tumour thrombus in the right portal vein from a hepatoma; with a benign lesion from focal nodular hyperplasia that was left alone; and the eighth patient with a hepatoma and extensive cirrhosis whose resec- tion was abandoned because IOUS detected more no- dules which would have required an extended right lobectomy.Of the three remaining patients who had altered resections, IOUS revealed tha

one patient planned for a right l
bectomy from metastases from a Klatskin tumour had only segments 4B and 5 involved and therefore only those segments were removed.In Table 5 Comparison of USD and FFT in estimated blood loss (EBL), number of blood units used (NBU), and postoperative length of stay (PLOS) Intraoperatire Assessment Forty patients within the database had IOUS and records available revealing number of hepatic neo- plasms found by pathology; a total of 68 liver nodules were found.Sensitivity and specificity rates are com- pared in Table 6.The single false positive from IOUS was a hemangioma lesion detected in the caudate lobe that could not be confirmed by pathology.

another patient planned for a right lobectomy from colorectal metastases, IOUS showed that the tumour extended into segment 4A and thus a step resection of 4A with the right lobectomy was performed.In the third patient planned for left lateral sectorectomy (seg- ments 2 and 3) from colorectal metastases, IOUS re- vealed that the tumour involved portions of the left medial lobe and therefore an anatomic left lobectomy was performed.The principle behind ultrasonic dissection (USD) is based on the premise that different tissues have differ- ent water content in proportion to collagen and elastin density.Hepatic tissue with high water content can includes pre-op ultrasonography, CT scan +portography, +MRI scan, +radionuclide scanning.sensitivity TP/(TP+FN) specificity TN/(TN+FP) therefore be selectively fragmented at a given fre- quency leaving numerous collagen rich blood vessels and bile ducts intact for safe ligation.Thus, blood loss is minimized and a cleaner resection line is obtained.Hemorrhage is the most important intraoperative and postoperative complication in liver surgery.TM Little et al. found that USD significantly reduced perioperative blood loss and the number of transfu- sions required. 12Reports have shown that the higher the blood loss, the greater the incidence of postopera- tive complications. 8' Although the determinants for postoperative length of stay (PLOS) is multifactorial, the degree of intraoperative and postoperative blood loss plays a major role.Indeed, PLOS was, on average, approximately 10 days shorter

the USD group which reduced he
lth care costs by 30% in our series.


Liver

Figure 5 IOUS detection of extra lesion (MET) that was not detected by preoperative computed tomography.Similar lesions were found elsewhere in the liver and resection was subsequently abandoned.


H

Figure 6 IOUS detection of extra tumour involvement with the middle hopatic vein (MHV), left hepatic vein (LHV) and inferior vena cava (IVC) which were not discernable by preoperative computed tomography.Resection was subsequently abandoned.It has been further suggested that high perioperative blood loss requiting numerous blood transfusions may have an unfavourable prognostic influence on long-term survival in colorectal metastatic disease since blood loss may adversely influence immunocompetence. 315Al- though the exact mechanism is unclear, experiments performed on tumour beating mice and rats have clearly shown increased tumour growth as a result of blood transfusions possibly by affecting lymphocyte and macrophage activity. 1619SD also decreases the amount ofischemic tissue at the resection margin in controlled animal experiments.[24] It should be mentioned that subjectively, USD in- cr ased the time required for the liver transection phase compared to finger fracture, but once the liver transection was complete hemostasis was superior to finger fracture and this mo t likely resulted in no difference in overall operative time.


Intraoperative Ultrasonography

The advantage of scanning with an ultrasound probe directly across the liver surface has allowed for ore accurate anatomic localization of neoplasms.Because there is no need to penetrate the abdominal wall, high frequency transducers (7.5 MHz) can be used to generate high resolution images that cannot be 25 obtained preoperatively.As a result, IOUS saved 8 patients from unnecessary resections that would have otherwise resulted in considerable morbidity and possible mortality.It alte