Consequences of the COVID-19 pandemic on the diagnosis and treatment of gastric cancer in referral centers in Italy.

BACKGROUND: The coronavirus pandemic had a major impact in Italy. The Italian health system's re-organization to face the emergency may have led to significant consequences especially in the diagnosis and treatment of malignancies. This study aimed to assess the impact of the pandemic in the diagnosis and treatment of gastric cancer in nine Gruppo Italiano RIcerca Cancro Gastrico (GIRCG) centers. METHODS: All patients assessed for gastric adenocarcinoma at nine GIRCG centers between January 2019 and November 2020 were included. Patients were grouped according to the date of "patient 1's" diagnosis in Italy: preCOVID versus COVID. Clinico-pathological and outcome differences between the two groups were analyzed. RESULTS: A total of 632 patients were included in the analysis (205 in the COVID group). The cT4 weighted ratios were higher in 2020 from April to September, with the greatest differences in May, August and September. The cM+ weighted ratio was significantly higher in July 2020. The mean number of gastrectomies had the greatest reduction in March and May 2020 compared with 2019. The median times from diagnosis to chemotherapy, to complete diagnostic work-up or to operation were longer in 2019. The median time from the end of chemotherapy to surgery was 17 days longer in the preCOVID group. CONCLUSIONS: A greater number of advanced or metastatic cases were diagnosed after the spread of SARS-CoV-2 infection, especially after the "full lockdown" periods. During the pandemic, once gastric cancer patients were referred to one of the centers, a shorter time to complete the diagnostic work-up or to address them to the best treatment option was required.

Which gastric cancer patients could benefit from staging laparoscopy? A GIRCG multicenter cohort study.

BACKGROUND: This study aimed to investigate which gastric cancer patients could benefit the most from staging laparoscopy. METHODS: A retrospective cohort study was carried out, including 316 (216 cM- and 100 cM+) gastric cancer patients who had undergone staging laparoscopy between 2010 and 2020 in seven GIRCG centers. A model including easily-accessible clinical, biochemical and pathological markers was constructed to predict the risk of carcinomatosis. ROC curve and decision curve analyses were used to verify its accuracy and net benefit. RESULTS: In the cM-population staging laparoscopy could detect 67 cases who had peritoneal carcinomatosis or positive cytology, for a yield of 30.5%. In cM-patients, intestinal type tumors (0.25, 0.12-0.51; p = 0.002), cT4 tumors (2.18, 1.11-4.28; p = 0.023) and cancers of the lower third (0.31, 0.14-0.70; p = 0.004) were associated with the presence of peritoneal carcinomatosis and/or positive cytology. The ROC curve analysis of the model including the three variables showed an AUC of 0.75 (0.68-0.81, p < 0.001). The decision curve analyses showed that the model had a higher net benefit than the treating all strategy between threshold probabilities of 15 and 50%. CONCLUSIONS: Staging laparoscopy is a useful tool to address the patient with gastric cancer to the most adequate treatment. In cM-patients the assessment of the location of the tumor, the Lauren's histotype and the cT status may help in providing additional elements in indicating or not the use of staging laparoscopy.

Signet ring cell percentage in poorly cohesive gastric cancer patients: A potential novel predictor of survival.

BACKGROUND AND OBJECTIVES: Signet ring cells (SRC) are widely acknowledged as a prognostically unfavorable histotype amongst poorly cohesive gastric cancer. In this study we evaluated the impact of SRC percentage on the clinical, pathological and prognostic features of these tumors according to the classification by the European Chapter of the IGCA. METHODS: We retrospectively reviewed records of patients with poorly cohesive gastric cancer that underwent surgery between 1995 and 2016, whose tissue specimens were available in a biological bank. All slides were put under revision, patients were reclassified into three groups according to the proportion of signet ring cells: "pure" SRC (containing ≥90% of SRCs), Poorly Cohesive-Not Otherwise Specified (PC-NOS) (containing ≤10% of SRCs), and PC-NOS/SRC (containing <90% but >10% of SRCs). The clinicopathological factors between different types were analyzed and prognostic differences were compared. RESULTS: Among 143 enrolled patients, 51% were male and 49% were female. The mean (±SD) age at diagnosis was 61 ± 13.9 years. Eighty-seven patients (60.8%) were reclassified as PC-NOS, 56 (39.2%) as PC-NOS/SRC and none as "pure" SRC. Five-years overall survival was significantly higher in PC-NOS/SRC group (63.3%) compared with PC-NOS group (12.7%). The increase in mortality risk was more than four-fold in patients with PC-NOS pattern compared to those with PC-NOS/SRC (HR 4.32 [95% CI 2.5-7.4]. After adjustment for potential confounding factors, SRC pattern was still an independent predictor of survival. CONCLUSIONS: The percentage of SRCs is inversely related to tumor aggressiveness, confirming the role of SRC pattern as an independent predictor of survival.

Looking for a strategy in treating peritoneal gastric cancer carcinomatosis: an Italian multicenter Gastric Cancer Research group's analysis.

BACKGROUND: The present study provides a snapshot of Italian patients with peritoneal metastasis from gastric cancer treated by surgery in Italian centers belonging to the Italian Research Group on Gastric Cancer. Prognostic factors affecting survival in such cohort of patients were evaluated with the final aim to identify patients who may benefit from radical intent surgery. METHODS: It is a multicentric retrospective study based on a prospectively collected database including demographics, clinical, surgical, pathological, and follow-up data of patients with gastric cancer and synchronous macroscopic peritoneal metastases. Patients were surgically treated from January 2005 to January 2017. We focused on patients with macroscopic peritoneal carcinomatosis (PC) treated with upfront surgery in order to provide homogeneous evidences. RESULTS: Our results show that patients with peritoneal carcinomatosis cannot be considered all lost. Strictly selected cases (R0/R1 and P1 patients) could benefit from an aggressive surgical approach performing an extended lymphadenectomy and HIPEC treatment. CONCLUSION: The main result of the study is that GC patients with limited peritoneal involvement can have a survival benefit from a surgery with "radical oncological intent", that means extended lymphadenectomy and R0 resection. The retrospective nature of this study is an important bias, and for this reason, we have started a prospective multicentric study including Italian stage IV patients that hopefully will give us more answers.

Development and validation of deep learning classifiers to detect Epstein-Barr virus and microsatellite instability status in gastric cancer: a retrospective multicentre cohort study.

BACKGROUND: Response to immunotherapy in gastric cancer is associated with microsatellite instability (or mismatch repair deficiency) and Epstein-Barr virus (EBV) positivity. We therefore aimed to develop and validate deep learning-based classifiers to detect microsatellite instability and EBV status from routine histology slides. METHODS: In this retrospective, multicentre study, we collected tissue samples from ten cohorts of patients with gastric cancer from seven countries (South Korea, Switzerland, Japan, Italy, Germany, the UK and the USA). We trained a deep learning-based classifier to detect microsatellite instability and EBV positivity from digitised, haematoxylin and eosin stained resection slides without annotating tumour containing regions. The performance of the classifier was assessed by within-cohort cross-validation in all ten cohorts and by external validation, for which we split the cohorts into a five-cohort training dataset and a five-cohort test dataset. We measured the area under the receiver operating curve (AUROC) for detection of microsatellite instability and EBV status. Microsatellite instability and EBV status were determined to be detectable if the lower bound of the 95% CI for the AUROC was above 0·5. FINDINGS: Across the ten cohorts, our analysis included 2823 patients with known microsatellite instability status and 2685 patients with known EBV status. In the within-cohort cross-validation, the deep learning-based classifier could detect microsatellite instability status in nine of ten cohorts, with AUROCs ranging from 0·597 (95% CI 0·522-0·737) to 0·836 (0·795-0·880) and EBV status in five of eight cohorts, with AUROCs ranging from 0·819 (0·752-0·841) to 0·897 (0·513-0·966). Training a classifier on the pooled training dataset and testing it on the five remaining cohorts resulted in high classification performance with AUROCs ranging from 0·723 (95% CI 0·676-0·794) to 0·863 (0·747-0·969) for detection of microsatellite instability and from 0·672 (0·403-0·989) to 0·859 (0·823-0·919) for detection of EBV status. INTERPRETATION: Classifiers became increasingly robust when trained on pooled cohorts. After prospective validation, this deep learning-based tissue classification system could be used as an inexpensive predictive biomarker for immunotherapy in gastric cancer. FUNDING: German Cancer Aid and German Federal Ministry of Health.

A novel treatment protocol with 6 cycles of neoadjuvant chemotherapy followed by cytoreductive surgery and hyperthermic intraperitoneal chemotherapy (HIPEC) in stage III primary ovarian cancer.

BACKGROUND: Few prospective studies investigated neoadjuvant chemotherapy (NAC), interval cytoreductive surgery (CRS) and hyperthermic intraperitoneal chemotherapy (HIPEC) in advanced ovarian cancer. We report the results of a phase II study where 6 rather than 3 cycles of NAC, followed by CRS and HIPEC, were adopted (HIPEC_ovaio, EudraCT number 2007-005674-31). MATERIALS AND METHODS: Between 2007 and 2014, 56 patients with stage III primary ovarian cancer and peritoneal carcinomatosis were assigned to 6 cycles of platinum and taxane-based NAC. Of these, two had progression, 8 underwent palliative surgery, and 46 had CRS and HIPEC. RESULTS: A complete pathological response was observed in 9 patients. Of 46 patients who completed the treatment protocol, 29 had no macroscopic residual tumor. Postoperative grade III morbidity rate was 28.2%; no grade IV complications or mortality events were observed. Five-year overall survival (OS) of the entire series was 36 ± 7% (median: 36, 95% CI: 26-45 months). In 46 patients treated by CRS and HIPEC, 5-year OS was 42 ± 8% (median: 53, 95% CI: 29-76 months), and 5-year progression-free survival was 26 ± 7% (median: 23, 95% CI: 19-27 months). Completeness of cytoreduction, peritoneal cancer index and FIGO stage resulted as significant prognostic factors. CONCLUSIONS: A novel protocol consisting of 6 cycles of NAC, followed by CRS and HIPEC, is associated with notable improvement in peritoneal carcinomatosis, limited postoperative morbidity risk and high survival rates in responders, and could deserve further investigations in randomized clinical trials.

Robotic-assisted gastrectomy for gastric cancer: single Western center results.

A robotic approach to abdominal surgery procedures may improve postoperative outcomes compared to either open or laparoscopic approaches. The role of robotics for gastric surgery, however, is still being evaluated. A retrospective review of the prospectively maintained database for robotic gastric surgery at University of Siena between 2011 and 2020 was conducted. Data regarding surgical procedures, early postoperative outcomes, and long-term follow-up were analyzed. 38 patients underwent robotic partial or total gastrectomy. Conversion to open occurred in two patients (5.2%) due to locally advanced disease as well as difficult identification of primary lesion. Postoperative morbidity was 13.1% while no postoperative mortality was registered. The mean length of operation was 358.6 (220-650) minutes and the mean number of retrieved lymph nodes was 35.8 (range: 5-73). The median OS of all population was 70.9 months. The median 5-year OS for the patients with positive nodes was worse than that of patients without metastatic lymph nodes [51.4 months (95% CI 35.5-67.4) vs. 79.5 months (95% CI 67.1-91.8); p = 0.079]. The interesting results including postoperative morbidity as well as mortality rate, the surgical outcomes, and the 5-year OS, were to be acceptable considering the data recorded by previous studies on robotic gastrectomy. This study demonstrated that robotic gastrectomy is feasible and can be safely performed. However, further follow-up and randomized clinical trials are required to confirm the role of a robotic approach in gastric cancer surgery.

The effect of learning curve on perioperative outcomes of robotic gastrectomy in two western high-volume centers.

INTRODUCTION: To compare outcomes of robotic gastrectomy (RG) performed during the learning curve (P1) with those after its completion (P2). METHODS: In this retrospective study, all consecutive RG patients (n = 92) performed between 2008 and 2018 were included. Primary outcome was conversion rate. RESULTS: D2 lymphadenectomies were more common in P2 (41, 97.6%) than P1 (41, 82.0%) (p = 0.019). Conversions were 11 (22%) in P1 versus 2 (4.8%) in P2 (p = 0.006). Postoperative morbidity was comparable between the groups. Median hospital stay was significantly shorter in P2. The only factor significantly associated with conversion was P2 (odds ratio = 0.18; 95% confidence interval, 0.04-0.85; p = 0.039). The 5-year overall survival in P1 was 79.6% versus 79.7% in P2 (p = 0.373). CONCLUSIONS: The learning curve affected operative and postoperative outcomes: during the learning curve, conversion to open surgery was significantly more frequent, the number of D2 was higher and patients were discharged earlier.

Early Gastric Cancer: identification of molecular markers able to distinguish submucosa-penetrating lesions with different prognosis.

BACKGROUND: Early Gastric Cancer (EGC) reaches 25% of the gastric cancers surgically treated in some areas of Northeastern Italy and is usually characterized by a good prognosis. However, among EGCs classified according to Kodama's criteria, Pen A subgroup is characterized by extensive submucosal invasion, lymph node metastases and worse prognosis, whereas Pen B subgroup by better prognosis. The aim of the study was to characterize the differences between Pen A, Pen B and locally advanced gastric cancer (T3N0) in order to identify biomarkers involved in aggressiveness and clinical outcome. METHODS: We selected 33 Pen A, 34 Pen B and 20 T3N0 tumors and performed immunohistochemistry of mucins, copy number variation analysis of a gene panel, microsatellite instability (MSI), TP53 mutation and loss of heterozygosity (LOH) analyses. RESULTS: Pen A subgroup was characterized by MUC6 overexpression (p = 0.021). Otherwise, the Pen B subgroup was significantly associated with the amplification of GATA6 gene (p = 0.002). The higher percentage of MSI tumors was observed in T3N0 group (p = 0.002), but no significant differences between EGC types were found. Finally, TP53 gene analysis showed that 32.8% of Pen tumors have a mutation in exons 5-8 and 50.0% presented LOH. Co-occurrence of TP53 mutation and LOH mainly characterized Pen A tumors (p = 0.022). CONCLUSIONS: Our analyses revealed that clinico-pathological parameters, microsatellite status and frequency of TP53 mutations do not seem to distinguish Pen subgroups. Conversely, the amplification of GATA6 was associated with Pen B, as well as the overexpression of MUC6 and the TP53mut/LOH significantly characterized Pen A.

New technologies in breast cancer sentinel lymph node biopsy; from the current gold standard to artificial intelligence.

Sentinel lymph node biopsy is an important diagnostic procedure performed in early breast cancer patients with clinically negative axillary lymph nodes. Detection and examination of sentinel lymph nodes determine further therapy decisions, therefore a choice of optimal technique minimising the risk of false-negative results is of great importance. Currently, the gold standard is the dual technique comprising the blue dye and radiotracer, however, this method creates a logistical problem for many medical units. The intrinsic constraints of the existing methods led to the development of a very wide range of alternatives with varying clinical efficiency and feasibility. While each method presents with its own advantages and disadvantages, many techniques have improved enough to become a non-inferior alternative in the sentinel lymph node biopsy. Along with the improvement of the existing technologies, there are evolving trends such as multimodality of the techniques maximising the diagnostic outcome or an emerging use of artificial intelligence (AI) improving the workflow of the procedure. This literature review aims to give an overview of the current status of the standard techniques and emerging cutting-edge technologies in the sentinel lymph node biopsy.

Preoperative oral immunonutrition in gastrointestinal surgical patients: How the tumour microenvironment can be modified.

BACKGROUND & AIMS: This study is focused on the impact of enteral immunonutrition on the cell-mediated immune response in the microenvironment of gastric and colorectal cancers. METHODS: This is a prospective pilot study approved by the local Ethics Committee. The immunophenotypic structure of the immune cells before (on the biopsy) and after (on the surgical sample) the administration of the immunonutrition in 16 patients is compared with 8 patients receiving regular diet. The samples of non-tumour tissue from sleeve-gastrectomy are used as non-neoplastic control. Antibodies were tested: CD4, CD8, PD-1, FOX-P3, CD68, CD163, CD80, CD21, CD56, PD-L1. We applied already well-known scoring systems for the evaluation of the immunohistochemistry and compared our data in the different groups by statistical analysis. RESULTS: In treated patients, we detected a modulation of the immune response with higher number of cytotoxic and helper T-lymphocytes in the tumour microenvironment of the surgical specimens compared to the pre-operative biopsy, and a lower number of lymphocytes presenting an exhausted (i.e. double positive CD8 and PD-1 lymphocytes) and regulatory (i.e. double positive CD4 and FOX-P3 lymphocytes) phenotype. Moreover we observed the M1 polarization with a lower number of CD163 positive macrophages and the inhibition of the PD-1/PD-L1 pathway in treated patients. CONCLUSIONS: The immunonutrition impacts on the tumoral microenvironment of gastric and colorectal cancer activating the inflammatory pathway, in terms of humoral and cellular response.

Stage IV Gastric Cancer: The Surgical Perspective of the Italian Research Group on Gastric Cancer.

Background/Aim: This work explored the prognostic role of curative versus non-curative surgery, the prognostic value of the various localizations of metastatic disease, and the possibility of identifying patients to be submitted to aggressive therapies. Patients and Methods: Retrospective chart review of stage IV patients operated on in our institutions. Results: Two hundred and eighty-two patients were considered; 73.4% had a single metastatic presentation. In 117 cases, a curative (R0) resection of primary and metastases was possible; 75 received a R1 resection and 90 a palliative R2 gastrectomy. Surgery was integrated with chemotherapy in multiple forms: conversion therapy, HIPEC, neo-adjuvant and adjuvant treatment. Median overall survival (OS) of the entire cohort was 10.9 months, with 14 months for the R0 subgroup. There was no correlation between metastasis site and survival. At multivariate analysis, several variables associated with the lymphatic sphere showed prognostic value, as well as tumor histology and the curativity of the surgical procedure, with a worse prognosis associated with a low number of resected nodes, D1 lymphectomy, pN3, non-intestinal histology, and R+ surgery. Considering the subgroup of R0 patients, the variables pT, pN and D displayed an independent prognostic role with a cumulative effect, showing that patients with no more than 1 risk factor can reach a median survival of 33 months. Conclusions: Our data show that the possibility of effective care also exists for Western patients with stage IV gastric cancer.

Pathological response and outcome after neoadjuvant chemotherapy with DOC (docetaxel, oxaliplatin, capecitabine) or EOF (epirubicin, oxaliplatin, 5-fluorouracil) for clinical T3-T4 non-metastatic gastric cancer.

PURPOSE: In this prospective observational study, we sought to compare the efficacy and safety of docetaxel + oxaliplatin + capecitabine (DOC) with epirubicin + oxaliplatin + 5-fluouracil (EOF) as neoadjuvant chemotherapy (NAC) for clinical T3 or T4 non-metastatic gastric cancer (GC) patients. METHODS: The DOC NAC consisted of docetaxel 35 mg/m2 (days 1-8), oxaliplatin 85 mg/m2 (day 1), and capecitabine 750 mg/m2 twice daily (days 1-14), every 3 weeks. The EOF NAC consisted of intravenous (IV) epirubicin 50 mg/m2 combined with IV oxaliplatin 130 mg/m2 on day 1 and continuous infusion 5-fluouracil 750 mg/m2 on days 1-5, every 3 weeks. After 4 cycles of NAC or upon progression during chemotherapy, patients underwent gastrectomy with standard D2 or D3 lymphadenectomy. Pathological complete response rate per Becker tumor regression grading system was the primary endpoint and the secondary endpoints included progression-free survival (2-yr PFS) and 2-year overall survival (2-yr OS) and tolerability. RESULTS: Overall, we identified 63 patients with T3-4 non-metastatic GC starting either NAC regimen between January 2010 and December 2017 at our Institution: 34 in the DOC group and 29 in EOF group. Thirty patients (88%) in the DOC group and 22 (76%) in the EOF group completed the 4 planned cycles of NAC. Fifty-seven patients received surgery. Results indicated no statistical significant differences between the two groups, and only a trend for some better data in favour of the DOC group. The R0 resection rate was 90.6% and 88.0% for the DOC and EOF cohorts, respectively. The pathological complete response rate was 6.2% in the DOC group and 4.0% in the EOF group. Becker 1-2 pathological response was found in 46.8% of the DOC cohort and 28.0% of the EOF cohort (p = .14). The 2-yr PFS rate was 54.1% for DOC vs. 41.4% for EOF (p = .14) and the 2-yr OS rate was 80.8% for DOC vs. 58.6% for EOF (p = .05). Neutropenia was the most common grade ≥3 toxicity and occurred in 8 (23.5%) patients of the DOC group and 10 (34.4%) patients of the EOF group (p = .33). CONCLUSIONS: These findings seem to confirm the feasibility of NAC for clinically T3 and T4 non-metastatic GC and, despite no statistical significant difference was documented, suggest a trend for better activity and tolerability for the docetaxel-based regimen (DOC) compared to the epirubicin-based combination (EOF).

Conversion gastrectomy for stage IV unresectable gastric cancer: a GIRCG retrospective cohort study.

BACKGROUND: The aim of this study is to report the experience with conversion surgery from six Gruppo Italiano Ricerca Cancro Gastrico (GIRCG) centers, focusing our analysis on factors affecting survival and the risk of recurrence. METHODS: A retrospective, multicenter cohort study was performed in patients who had undergone conversion gastrectomy between 2005 and 2017. Data were extracted from a GIRCG database including all metastatic gastric cancer patients submitted to surgery. Only stage IV unresectable tumors/metastases which became resectable after chemotherapy were included in this analysis. RESULTS: Forty-five resected M1 patients were included in the analysis. Reasons for being deemed unresectable at diagnosis were peritoneal involvement (PCI > 6) (n = 38, 84.4%), distant metastatic nodes (n = 3, 6.6%) and extensive liver involvement (n = 4, 8.8%). Median follow-up was 25 months (IQR 9-50). Median overall survival from surgery was 15 months and 1-, 3- and 5-year survivals were 57.2, 36.1 and 24%, respectively. Median progression-free survival was 12 months with 1- and 3-year survival of 46.4 and 33.9%, respectively. At cox regression analysis the only independent prognostic factor for OS was the presence of more than one type of metastasis (HR 4.41, 95% CI 1.72-11.3, p = 0.002). A positive microscopic resection margin was the only risk factor for recurrence (HR 5.72, 95% CI 1.04-31.4, p = 0.045). CONCLUSIONS: Unresectable stage IV GC patients could benefit from radical surgery after chemotherapy and achieve long survivals. The main prognostic factor for these patients was the presence of more than one type of extra-gastric metastatic involvement.

Hereditary diffuse gastric cancer: how to look for and how to manage it.

With a current molecular revolution, hereditary gastric cancer represents a small group of patients that require a special multidisciplinary treatment. Surgeons being a member of the multidisciplinary teams are an important part of the diagnosis, treatment and follow-up of these patients. The prophylactic nature of the gastrectomy with all different problems associated with this procedure need to be widely discussed with patients. We present a review of how to look for and how to manage a hereditary diffuse-type gastric cancer.

Siewert III esophagogastric junction adenocarcinoma: does TNM 8th save us?

Siewert III cancers were classified as esophageal cancers by the TNM 7th edition (TNM7), while being defined as gastric cancers by the new TNM 8th edition (TNM8). Aim of this study was to compare previous and present TNM classifications of Siewert III. From 2000 to 2015, 309 patients with Siewert III adenocarcinoma were treated at ten high-volume centers, belonging to the GIRCG (Italian Research Group for Gastric Cancer). We retrospectively analyzed overall survival according to TNM classifications: gastric TNM8 was compared with either gastric TNM7 or esophageal TNM7. Median number of lymph nodes harvested was 31 (interquartile range 22-44). Agreement between gastric TNM7 and TNM8 was very good (weighted kappa 92.3%, IC 95% 90.3-94.1%). Accordingly, stage migration was observed in 54 of 309 patients (17.5%), with 12 patients upstaged (3.9%) and 42 downstaged (13.6%). Cox models including either gastric TNM7 or TNM8 achieved similar goodness-of-fit and c-index. Differences were much larger, when shifting from esophageal TNM7 to gastric TNM8: the agreement was much lower (weighted kappa 69.1%, 65.2-73.2%), with 196 of 309 patients (63.4%) downstaging. The corresponding Cox model presented the lowest goodness-of-fit and discrimination ability. Gastric TNM7 and TNM8 were largely superimposable, so that stage migration was minor and prognostic significance was similar. At variance, stage migration was substantial when shifting from esophageal TNM7 to TNM8. Moreover, survival models with esophageal TNM7 presented the worst goodness-of-fit and the lowest discrimination ability. This further supports placing Siewert III among gastric cancers, as done in TNM8.

Surgical treatment of hepatic metastases from gastric cancer.

The purpose of the study was to investigate the clinical factors influencing the prognosis of patients submitted to hepatectomy for metastases from gastric cancer and their clinical role. We conducted a retrospective multicentre review. We evaluated how survival from surgery was influenced by patient-related, tumour-related and treatment-related prognostic factors. We analysed data on 144 patients submitted to hepatectomy for metastases from gastric cancer, in the synchronous and metachronous setting. In 117 cases, an R0 resection was achieved, while in 27 an R + hepatic resection was performed. Chemotherapy was administered to 55 patients. Surgical mortality was 2.1% and morbidity 21.5%. One-, 3-, and 5-year OS rates after surgery were 49.9, 19.4 and 11.6%, respectively, with a median OS of 12.0 months. T4 gastric cancer, H3 hepatic involvement, non-curative resection, recurrence after surgery, and abstention from chemotherapy were associated with a worse prognosis. Factor T and H displayed a clear (p < 0.001) cumulative effect. Our data show that R0 resection must be pursued whenever possible. The treatment of T4 gastric cancer with hepatic bilateral and diffuse metastasis (H3) should be considered carefully or it should be probably avoided. Finally, a multimodal treatment associating surgery and chemotherapy offers the best survival results.