Hpb最新文献

Aims: We conducted this systematic review to answer the questions: a. what are the timing categories for revision surgery (RS) in incidental gallbladder cancer (iGBC)? b. which RS timing achieves better oncological outcomes?

Methods: We performed literature search in 4 databases (PubMed, Scopus, Google Scholar and Cochrane Reviews) till 10th October 2025 and included studies which reported patient outcomes based on RS timing. Study characteristics, timing category definitions and RS outcomes were collected. (Study protocol PROSPERO ID CRD42023453990).

Results: Twelve retrospective studies were included, with 2067 iGBC patients (566 males and 1346 females). On the 'Joanna Briggs Institute' (JBI) tool, most studies scored a 'Yes' to 7-8 out of 10 questions. There was no consensus on the definitions of 'early', 'intermediate' and 'delayed' timings for RS. Successful RS, perioperative morbidity, R0 resection were similar. On individual patient data meta-analysis, there was no difference in overall survival between RS at '4 weeks' [hazard ratio: 1.29, 95 % CI: 0.79-2.10].

Conclusion: There was no consensus on the definitions of timing categories and optimum timing for RS in iGBC. Definitions of timing categories need to be standardised and future studies based on these categories may identify the ideal timing of RS in iGBC.

Background: Although biliary brush cytology has a high specificity (95-100 %), the sensitivity is poor (41-67 %). This study aimed to evaluate whether the use of an optimized protocol for brush cytology improves the results in patients with suspected perihilar or intrahepatic cholangiocarcinoma (pCCA/iCCA).

Methods: Patients were prospectively included after changing the protocol (June 2021-June 2023) and compared with a historical cohort (January 2017-May 2021). Changes included different brush processing, addition of next-generation sequencing (NGS), and additional sampling (two brush samples and intraductal biopsies). Primary outcome was the sensitivity and the specificity of the procedure.

Results: A total of 175 patients were evaluated (62 prospective, 113 historical) of which 165 patients had malignant disease (94 %). After implementation of the protocol, the sensitivity was 88.3 % (95%CI, 76.8-94.8 %) versus 50.5 % (95%CI, 40.6-60.3 %) prior to implementation. Sensitivity of only the first brush sample with the optimized processing did also significantly increase (78 %; 95%CI, 65.5-87.5 %). Specificity was 100 % in both groups (2/2 vs 8/8).

Conclusions: A modification in the processing of cytopathology led to a significant improvement in the sensitivity of the first bile duct brush to 78 %. Furthermore, adding NGS increased sensitivity to 83 %, an extra brush sample to 85 %, and intraductal biopsies to 88 %.

Background: The advantages of robotic hepatectomy (Rob-H) over laparoscopic hepatectomy (Lap-H) remain unclear. This study compares the outcomes between Rob-H and Lap-H in a single-center setting.

Methods: A retrospective analysis was conducted on patients who underwent minimally invasive liver resection between 2014 and 2023. Patient demographics, perioperative parameters, and postoperative outcomes were reviewed. Propensity score matching (PSM) was employed to reduce selection bias.

Results: A total of 2999 patients were included in this study. 2375 patients underwent Lap-H and 624 patients underwent Rob-H. After PSM, 42 patients who underwent right hemihepatectomy. The results showed that, compared to the Lap-H group, the Rob-H group had lower intraoperative blood loss (P = 0.016). A total of 108 patients who underwent left hemihepatectomy were included. The Rob-H group had shorter operative time (P = 0.005), lower intraoperative blood loss (P = 0.049).For 108 patients who underwent right posterior segmentectomy, the Rob-H group showed shorter operative time (P < 0.001), less intraoperative blood loss (P = 0.012), shorter Pringle duration (P = 0.008).

Conclusion: Compared with the Lap-H group, intraoperative blood loss and operative time were lower in the Rob-H group, and the results were consistent with previous studies, suggesting that the robotic platform overcome the limitations of laparoscopic liver resection.

Background: Iatrogenic bile duct injuries (IBDIs) remain complex and diverse, presenting significant challenges for preoperative evaluation and surgical repair.

Methods: Patients who underwent hepaticojejunostomy (HJ) for IBDIs from May 2019 to June 2024 were enrolled. Preoperatively, all patients underwent preoperative individualized 3D modelling of bile duct injury (3DM-BDI) for assessment. During surgery, augmented reality navigation (ARN) combined with indocyanine green fluorescence imaging (ICG-FI) was used for guidance. Perioperative indicators and short-term postoperative outcomes were evaluated to verify the safety and feasibility of this novel approach.

Results: In all patients, the 3DM-BDI accurately predicted the type and extent of bile duct and vascular injuries. The mean operation time was 380.7 ± 83.9 min, and the mean intraoperative blood loss was 135.0 ± 169.7 mL, with no patients requiring intraoperative blood transfusion. The mean postoperative hospital stay was 9.3 ± 2.1 days. The navigation efficiency was 75 % for ICG-FI and 87.5 % for ARN. The success rate of reconstruction was 85.7 % based on follow-up within 90 days.

Conclusion: The combination of ARN and ICG-FI as an auxiliary method in the reconstruction of IBDIs may be feasible and safe. These modalities may provide technical advantages in preoperative evaluation and precise dissection of hilar vessels and bile ducts during surgery.

Background: There is a scarcity of data on the feasibility of laparoscopic approach to associating liver partition and portal vein ligation for staged hepatectomy (Lap-ALPPS) regarding its impact on functional liver remnant (FLR) hypertrophy when compared with the open approach.

Methods: A retrospective study of patients who underwent open or lap-ALPPS for hepatocellular carcinoma (HCC) in a tertiary referral centre in Hong Kong during the period from December 2013 to April 2023.

Results: Fifty-seven (42 open and 15 laparoscopic) ALPPS were performed for HCC. The open group had more blood loss during stage I (500 ml vs. 300 ml, p = 0.001). The morbidity (Clavien-Dindo grade ≥ 3a) (14.3 % vs. 26.7 %, p = 0.43) and Grade B/C post-hepatectomy liver failure rates (20.0 % vs 35.7 %, p = 0.34) were similar. The open group had a higher percentage increment in remnant volume (50.6 % vs. 34.8 %, p = 0.02). Linear regression revealed that a small pre-operative FLR/ESLV (B = -1.75, 95 % CI -2.82-0.678, p < 0.001) and an open approach at stage I B = -20.2, 95 % CI -37.7-2.68, p < 0.001) predicted a higher percentage increment in remnant volume.

Conclusion: Lap-ALPPS had less blood loss but was associated with slower hypertrophy. Hence, a longer waiting time to stage II ALPPS may be required in selected patients.

Background: Laparoscopic cholecystectomy (LC), a common abdominal operation, is associated with significant morbidity, particularly bile duct injury. Artificial intelligence (AI) can enable real-time monitoring, assist decision-making, increase safety, and improve patient outcomes. This study systematically reviews AI applications in LC, evaluating different models and their performance.

Methods: A systematic review was conducted in accordance with the PRISMA guidelines. A comprehensive literature search was conducted using PubMed and ScienceDirect databases for studies published between 2014 and 2024. All studies assessing AI applications in LC were included. Data extraction focused on the study aims, types of AI tools, datasets, anatomical recognition capabilities, and accuracy metrics.

Results: The search yielded 413 citations; a final list of 43 citations was compiled after applying the inclusion and exclusion criteria. Different datasets and developed AI tools were used in LC. AI tools were utilised in risk-scoring models for complication identification and outcome prediction, as well as for recognising anatomical landmarks during LC and subdividing the procedure into subtasks.

Conclusion: AI integration in LC is promising for improving intraoperative guidance, enhancing surgical education, and supporting decision-making processes. Future large-scale studies are warranted to validate the role of AI in improving patient safety and outcomes in LC.

Background: Prognostic models for patients with hepatocellular carcinoma (HCC) undergoing surgery often fail to account for perioperative changes in liver function. This study evaluated a novel dynamic index that integrates changes in the albumin-bilirubin (ALBI) grade and platelet count to predict postoperative morbidity.

Methods: A multi-institutional database was queried for patients undergoing surgery for HCC (2000-2023). "Changes in ALBI and platelets" (CAP) were calculated as CAP = (ΔALBI²+ΔPlatelets²), comparing preoperative values with those from postoperative day 3. Associations between CAP and the Comprehensive Complication Index (CCI) were examined using restricted cubic spline and Rand Forest analyses.

Results: A total of 1155 patients were included. The median CAP was 1.1 (IQR 0.8-1.5). Postoperative complications occurred in ∼40 % of patients, with a mean CCI of 15. Recursive partitioning determined CAP = 1.0 and CAP = 1.6 as the primary and secondary optimal cut-offs. In adjusted analysis, each unit increase in CAP corresponded to a 4.90 (95%CI 0.98-8.82) increase in CCI. CAP>1.6 was associated with higher likelihood of any (aOR 2.35, 95%CI 1.66-3.34) and severe complications (aOR 2.27, 95%CI 1.61-3.20).

Conclusion: CAP independently predicted morbidity following HCC surgery, highlighting the prognostic utility of dynamic indices. An online calculator was made available at: https://jk-osu.shinyapps.io/CAP_Chatzipanagiotou/.

Background: Predicting surgical difficulty in robotic liver resection (RLR) is crucial for optimizing resource allocation, training programs, and patient outcomes. While several difficulty scoring systems (DSSs) have been validated for laparoscopic liver resection (LLR), their applicability to RLR remains uncertain. This study evaluates the predictive performance of five DSSs: Halls Southampton Score (HSS), Ban Iwate Score (BIS), Hasegawa Score (HGS), Institut Mutualiste Montsouris Score (IMM), and Tampa Difficulty Score (TAS), in the robotic setting.

Methods: A retrospective study was conducted on 124 patients who underwent RLR between January 2011 and June 2024 at two high-volume centers. Each DSS was retrospectively applied. Predictive accuracy for operative duration, intraoperative blood loss (>400 mL), transfusion need, postoperative complications, surgical reintervention, and 90-day readmission was assessed using R² (continuous variables) and AUC (categorical outcomes).

Results: HSS demonstrated the highest overall predictive power, particularly for transfusion need (AUC = 0,85), postoperative complications (AUC = 0,74), and 90-day readmission (AUC = 0,86). BIS was the most accurate for intraoperative blood loss (R² = 0,32). TAS showed the lowest predictive performance across most outcomes.

Conclusion: Laparoscopic DSSs are applicable to RLR, with HSS emerging as the most reliable. TAS requires further validation. A combined DSS approach could improve surgical planning and patient management.

Background: As cancer care increasingly prioritizes patient-centered outcomes, understanding predictors of postoperative quality of life (QOL) is essential. This study aimed to identify preoperative factors associated with early QOL outcomes following oncologic liver resection, using prospectively collected data from a multicenter randomized controlled trial.

Methods: This was a secondary analysis of the Hemorrhage During Liver Resection: Tranexamic Acid (HeLiX) trial (NCT02261415), conducted across 11 tertiary centers in Canada and the USA. Patients undergoing liver resection for malignancy completed the EORTC QLQ-C30 questionnaire preoperatively and at postoperative day (POD) 30. Multivariable regression was used to identify preoperative clinical and demographic predictors of clinically meaningful QOL changes across five domains. Inverse probability weighting addressed potential bias from missing data.

Results: Of 863 eligible patients, 796 completed baseline and 588 completed POD30 QOL assessments. Lower preoperative QOL was consistently associated with worse postoperative scores. Major and multivisceral resections predicted clinically meaningful declines in physical, role, and global functioning. Female sex, cardiovascular comorbidities, and smoking were also associated with decline. Prior liver resection and biliary or vascular reconstruction were linked to better physical function at POD30.

Conclusion: Preoperative factors significantly influence early postoperative QOL. These findings support personalized risk counseling and proactive recovery planning in liver cancer surgery.