HB tumor rupture at diagnosis is rare, and its incidence greatly varies among different cooperative study groups (COG, GPLT, JPLT, and SIOPEL), ranging from 5% to 15%(2, 6). This may be explained by the fact that no common definition of tumor rupture is available at present. According to the latest 2017 PRETEXT staging system that provides a new and common set of definitions available in clinical trials, tumor rupture is clearly defined as the presence of free fluid in the abdomen or pelvis with at least one imaging criteria of hemorrhage after excluding biopsy-related hemorrhage, intraoperative rupture, non-hemorrhagic ascites and subcapsular fluid(10)(Fig3). Today, no consensus is reached on the predictive role of HB tumor rupture. In the CHIC database, tumor rupture is shown to be significantly associated with the poor outcome in univariate analysis. However, it is not taken as a high-risk factor in the JPLT and the COG groups(12, 13). Considering the potential risk of peritoneal seeding after tumor rupture, patients with HB tumor rupture at diagnosis receive the high-risk protocol in the SIOPEL group(2).
In our retrospective study, promising results were achieved in tumor rupture group, in which no patient died and only 2 events occurred. The 3-year EFS and OS in tumor rupture group were higher than those in high-risk group, and the difference in OS was statistically significant. These results suggested that HB tumor rupture at diagnosis was not a poor prognostic factor in our cohort. However, there was a noticeable clue in tumor rupture group, in which the strong poor prognostic factor of lung metastasis was not found in any patient. This might partly explain the superior outcome in tumor rupture group to high-risk group where some patients suffered from lung metastasis. Additionally, in the absence of tumor rupture, some patients in tumor rupture group had standard-risk HB (3 PRETEXTⅡ). In tumor rupture group, 7 (77.8%) patients achieved CR at the end of the treatment, and only 2 showed mildly increased AFP level without residual disease on imaging. Of these 2 patients, one elder patient developed intrahepatic relapse, while the other infant patient gradually had normal AFP level. Obviously, such findings further confirmed that tumor rupture at diagnosis was not a risk factor of peritoneal seeding and relapse.
The management of hepatic trauma has changed significantly from the early compulsory operative approach to the non-operative management approach(14). In our cohort, we adopted the far more conservative management. After tumor rupture, in addition to the increased risk of peritoneal seeding, the injury to the biliary tree and the hepatic vasculature will result in bile and blood leakage, which present with ascites and peritonitis. It has been confirmed that intraperitoneal instillation of IL-2 inhibits the formation of malignant ascites(14). Therefore, in our management protocol of HB tumor rupture, peritoneal perfusion with normal saline and IL-2 was an obligatory procedure to protect patients from peritonitis and peritoneal seeding. In our cohort, only one patient (1/9, 11.1%) suffered from peritoneal seeding, which was partly because that his postoperative chemotherapy was delayed for 4 months. After salvage chemotherapy and the second operation, the patient remained disease-free until the follow-up of 7.5 years. None of our patients experienced peritonitis or hematoma. Based on these results, conservative management with peritoneal perfusion of IL-2 and normal saline was an effective approach to prevent peritoneal seeding and peritonitis.
In the case of tumor rupture, blood loss to varying degrees was the most common presentation. 3 (33.3%) of our patients suffered from severe anemia. Additionally, the drop of hemoglobin level ≥2 g/L/day was closely associated with the potential active bleeding. The hemoglobin levels of all patients were dynamically monitored, and 2 (22.2%) of them received operative intervention to control bleeding. In some severe cases, the patients concomitantly presented with mild to moderate thrombocytopenia and/or hypofibrinogenemia. In our patients, the presentation of imbalanced serum electrolytes was rare, and no patient developed tumor lysis syndrome (TLS) based on the laboratory criterion. TLS results from the rapid destruction of malignant cells and the abrupt release of intracellular components into the extracellular space, thus causing metabolic disorders(15). Therefore, in HB tumor rupture, the tumor cell destruction rate and the intracellular component release rate should be mild and they are under the normal metabolic homeostasis of the body. The major challenge in HB tumor rupture should be the management of active bleeding.
In conclusion, the current goals of pediatric solid tumor management focus on minimizing toxicity while maintaining the excellent outcomes in low-risk disease, and improving outcomes in patients with high-risk disease. In this regard, it is necessary to progressively refine the risk-stratification systems and evaluate every potential annotation factor in HB. Frankly, our results are obtained from a small sample size. Hopefully, our findings can be tested in a randomized, multi-center clinical trial in the future.