Robotic caudo-peripheral approach for liver parenchymal transection in anatomical liver resections for hepatocellular carcinoma

Patient selection and preoperative planning

Diagnosis is based on typical imaging findings from dynamic-magnetic resonance imaging (MRI) and dynamic-computed tomography (CT), as widely described in clinical practice guidelines^[3]. The selection of adequate candidates for surgical resection is based on the Child-Pugh classification^[4,5]. Preoperative imaging with 3D routine reconstruction is essential to know the presence of anatomical variations [Figure 1]. The IWATE scoring criteria are used to assess the difficulty level preoperatively by considering tumor location and size, the proximity to major vessels, and the fibrosis severity^[6].

Figure 1. Example of preoperative 3D imaging with reconstruction of arterial, portal, and venous system and their relationships with the tumor. 3D: Three-dimensional.

Major exclusion criteria for minimally invasive approach are: the presence of thrombosis of the portal vein/hepatic vein requiring thrombectomy, and multiple previous laparotomies. Specific exclusion criteria for the down-to-up approach are HCC lesions close to the hepatic-caval confluence. For those cases, our preferred technique is the cranial-to-caudal one following the hepatic veins.

Patient positioning and port placement

The Da Vinci Xi system (Intuitive Surgical, Inc, Sunnyvale, CA, USA) is used. The French supine position with 15 degrees reverse Trendelenburg is adopted for all robotic hepatic resection. In the case of posterior segment lesions (segments VII and VIII), a cylindrical pillow is positioned along the right side of the patient under the right shoulder and the operating table is tilted toward the left side of the patient, 5 to 10 degrees according to the size of right liver, to facilitate liver mobilization. The arms are fixed at the sides of the trunk. The surgical assistant is positioned in between the open legs of the patient. Pneumoperitoneum is induced by a Veress needle via percutaneous puncture 2 to 4 cm from the edge of left hypochondrium ribs at the level of hemi-clavear point. Once a pneumoperitoneum of 14 mmHg is achieved, the robotic camera is inserted using a 10-12 mm visiport-type laparoscopic trocar at the site of the assisting ports in order to avoid the risk of blind port introduction. Standard port placement is employed with the Xi system docked from the side of the patient. Two assisting ports are used [Figure 2]. The camera is inserted through arm 2 but liberally shifted to arm 3 during the parenchymal liver transection in case the line of transection is not clearly visible from arm 2. Targeting is performed from arm 2.

Figure 2. Standard port positioning in robotic liver resections with the da Vinci Robotic Xi Surgical System (Intuitive Surgical, Sunnyvale, CA, USA).

Robotic instrumentation

Standard robotic instruments are used: two bipolar instruments (a long fenestrated bipolar forceps and the Maryland bipolar forceps), two Cadiere forceps, monopolar curved scissors, a robotic Hem-o-lok applier (size ML and L), and a needle driver.

Laparoscopic instrumentation

The following instruments are used: a long suction device, an advanced bipolar and ultrasonic dissection (Thunderbeat^TM Type S, Olympus Medical Systems Corp., Tokyo, Japan), laparoscopic Hem-o-lok applier (size ML, L, XL), and for complex cases, a monopolar electrosurgical electrode (Coolingbis^TM, Vecmedical, Montcada i Reixac, Barcelona, Spain). The assistant ports are used for suction, insertion of sutures, sponges, laparoscopic grasper, complementary coagulation, and staplers.

Caudo-peripheral liver parenchymal transection

Firstly, once the pneumoperitoneum is established and the trocars are placed, the hepatoduodenal ligament is exposed and the pars flaccida of the gastrohepatic ligament is opened. The Pringle is achieved by a Foley catheter surrounding the hepatic pedicle for the majority of the cases [Supplementary video 1], with the exception of HCC sited on segments VII and VIII. In these cases, due to the semilateral position of the patient, we prefer an extracorporeal Pringle with a tourniquet passing through a 5mm incision on the right side to facilitate the clamping maneuver. A transient hepatic vascular inflow occlusion with 15 min of clamping followed by 5 min of de-clamping is our reference technique.

Intraoperative ultrasound (IOUS) is realized with the ultrasound probe entered through the 12 mm assistant port by the bedside assistant. The TilePro^TM function of the daVinciSurgical System® enables the integration of the ultrasound images and the 3D reconstruction into the display during surgery for both the console surgeon and the assistant. After identifying the pedicles by ultrasound and marking the edges of tumoral margins on the hepatic surface, the right or left liver is mobilized according to the localization of the HCC. The caudo-peripheral liver parenchymal transection begins with the isolation of the pedicles to be divided by an extra Glissonean approach. During this step, avoiding forced maneuvers is essential in order to preserve Glissonean capsule, protecting biliary and vascular elements.

Subsequently to the pedicle ligation, the intraoperative administration of 2.5 mg/body of indocyanine green (ICG) is performed according to the negative staining technique: the near-infrared (NIR) light camera enables clear identification of the portal territory of interest, while the rest of the liver fluoresces green [Figure 3]. The Da Vinci’s Firefly modeTM represents another strength of the robotic system in identification of key landmarks. This integrated NIR fluorescence imaging function is activated directly by the surgeon at the console.

Figure 3. Indocyanine green dye negative staining technique during anatomical hepatectomy to clearly identify the portal territory of clamped pedicles using the Da Vinci’s Firefly modeTM.

Once the demarcation line of the portal territory has been clearly identified and marked by robotic monopolar scissors, the liver parenchymal transection can be started. This step begins with the incision of the Glisson capsule performed by laparoscopic ultrasonic dissection, penetrating the parenchyma by 1-2 cm. Then, transection is achieved using a Kelly clamp crush technique: the Maryland bipolar forceps on the 3rd arm dissects the liver from the pedicles with small bites, rapidly opening and closing the jaws. Complementary hemostasis is performed by diathermy with the bipolar forceps on the 1st arm for vessels up to 5 mm. In this phase, the role of the assistant is to follow the robotic instruments with quick irrigation and aspiration in order to always guarantee optimal functioning of the bipolar forceps.

Larger caliber vessels are dissected and surrounded at 360° so as to be able to apply the medium-large Hem-o-lok by the bedside surgeon when the maneuver is laparoscopically easy, otherwise using the robotic applier. A trick we use in the most challenging cases is to load the vessel on a Vicryl suture in order to suspend and expose it as best as possible, ensuring it is completely surrounded before cutting. The suspension of the vascular elements on a tape is a useful measure in robotic surgery because it allows indirect traction of delicate structures without tearing them due to the lack of tactile feedback.

When feasible and easy for the assistant, the sectioning of coagulated and clipped vessels is performed laparoscopically to save time otherwise spent on changing robotic instruments. The liver parenchyma is dissected from the caudal to the cranial aspects of the liver following the pedicles.

Supplementary videos 2-4 extensively show the caudal-to-cranial technique combined with the robotic Maryland Kelly clamp crushing parenchymal transection applied to anterior [Supplementary video 2], posterior [Supplementary video 3] segments resections, and major hepatectomies [Supplementary video 4], respectively.

Intraoperative complications management

With the advent of minimally invasive approaches, bleeding control has become one of the most important criteria for establishing not only the safety and feasibility of the technique but also the surgeon’s proficiency, and constitutes the most frequent reason for conversion to laparotomy. Moreover, intraoperative blood loss is an independent predictor of recurrence and survival after HCC surgery. Tumor size, location, and the extent of resection represent important risk factors for vascular injuries. Intraoperative bleeding control in hepatobiliary surgery is multimodal. The first step is communication with nurses and anesthesia team. In this phase, it is essential to decrease tidal volumes and positive end-expiratory pressure (PEEP) in order to reduce central venous pressure and, consequently, hepatic venous back-bleeding. Close monitoring of the patient is also necessary to identify the possibility of a CO₂ embolism as early as possible; in this case, conversion to laparotomy is mandatory. From a purely technical surgical point of view, we essentially follow three steps: control, exposure, and repair. The first thing to ensure is the control of bleeding, which is achieved by using the Pringle maneuver, clamps or temporary clips, and hemostatic gauze (SURGICEL^TM Absorbable Hemostat). Once the hemorrhage has been controlled, the structure of origin is clearly identified, and the surgeon can dissect it above and below, paying attention to not worsening the injury. In this phase, the bedside assistant plays a crucial role in ensuring correct vision through aspiration and irrigation of the operating field. The optimal exposure is essential to choose the correct repair strategy beyond suturing or stapling, clipping, or using advanced laparoscopic energy devices such as saline-coupled bipolar sealing device (Aquamantys®) or monopolar electrosurgical sealing dissector (Coolingbis®).

Another important precaution is to detect an intraoperative bile leak (IOBL) and properly managing it is essential to prevent postoperative biliary fistula. Bile leak check is performed with a white gauze and any bile leakage is closed with a 5/0 absorbable monofilament. According to our experience, the key to achieving good results in intraoperative complications management is to be ultra-selective in hemostasis and biliostasis, using sutures or bipolar diathermy and avoiding monopolar coagulation which, due to energy transmission, could itself cause iatrogenic ischemic lesions in the surrounding parenchyma.

Specimen extraction, hemostasis, and drains policy

The specimen is placed in an endobag and extracted through an 8- to 10-cm Pfannenstiel incision. Hemostasis is checked carefully once the Pringle maneuver is released, and whenever necessary, one or more TachoSil(®), a medicated sponge coated with human fibrinogen and human thrombin, are applied on the liver surface. One multi-lamellar drain is positioned in the right hypochondrium in case of major hepatectomy or doubtful bile leak test.

3D printing of livers with cancers: patient education and surgical planning

One of the emerging applications of AI is patient education. This is a critical component of healthcare, as it enables people to understand their own medical diagnosis, treatment options, and preventive measures. It is well known that a patient who is aware of his illness and well informed about therapeutic prospects is also much more compliant with the treatments offered to him, resulting in better health outcomes. In this context, several AI-based chat boxes are emerging with the purpose of educating and providing information to patients and their caregivers on dietary recommendations and lifestyle modifications such as smoking cessation and behavioral therapies^[7]. From a purely surgical point of view, a huge advantage that we found in the use of 3D reconstructions was that of being able to inform patients in detail about the size and location of the neoplasm and perform a simulation of the resection together with them, showing which part and how much of liver will remain after surgery. This illustration is helpful in further understanding possible complications and their management.

The most recent deep learning models can accurately predict surgical risk (i.e., intraoperative blood loss and complications stratified according to Clavien-Dindo classification) based on arterial preoperative CT imaging phases and other clinical data, including age, sex, body mass index, preoperative American Society of Anesthesiologists physical status classification, the occurrence of diabetes mellitus, Child-Pugh classification status, platelet count, and surgical approach^[8].

This tool therefore represents a real communication platform, and in our hospital, it is shared not only with patients but, by adding more technical details, also with students and residents who can benefit from their training and comprehension of the surgical procedures they will witness.

AI through ICG, simulation, navigation, and radiomics for HCC

AI technology is also used for augmented reality (AR) navigation, enabling surgeons to overcome the limits of already known systems. The application of AR provides a semi-transparent overlay of the preoperative images on the working area. Liver neoplasm, vascular and biliary structures reconstructed using preoperative CT are projected on the hepatic surface during parenchymal transection, giving the possibility of a simulation before reaching the point of no return and resulting in more accurate and safer resections^[8].

Research in the field of medical imaging has focused heavily on AI-based radiomics.

Radiomics is a new discipline that works by extrapolating many features from radiological images, which are then processed using appropriate data analysis methods. This not only enables more accurate diagnoses but also provides valuable information on specific tumor characteristics, such as predicting the response to a treatment, or highlighting the presence of particular genetic and epigenetic alterations, which could be obtained only by using other biomedical technologies. The application of AI in radiomics has significantly improved the precision and efficiency of ultrasound diagnosis, including intraoperative ultrasonography^[9].

Another routine technique is ICG fluorescence imaging, which has high sensitivity (> 90%) in tumor and resection margin detection. However, a relatively high number of false positives may represent a limitation of this technique. The integration of preoperative images into the intraoperative field provided by AI increases the fluorescence imaging precision^[10]. The most sophisticated models release even a virtual biliary tree depiction.

In our daily practice, the application of 3D reconstructions and AR navigation has proven to be very useful to improve the precision in liver resections and to facilitate communication between the main operator, the bedside assistant, and young surgeons who attend the operation from the outside.