PDF(Pubmed)
Abstract：
OBJECTIVE: Clinically significant portal hypertension (CSPH) seriously affects the feasibility and safety of surgical treatment for hepatocellular carcinoma (HCC) patients. The aim of this study was to establish a new surgical scheme defining risk classification of post-hepatectomy liver failure (PHLF) to facilitate the surgical decision-making and identify suitable candidates for individual hepatectomy among HCC patients with CSPH.
BACKGROUND: Hepatectomy is the preferred treatment for HCC. Surgeons must maintain a balance between the expected oncological outcomes of HCC removal and short-term risks of severe PHLF and morbidity. CSPH aggravates liver decompensation and increases the risk of severe PHLF thus complicating hepatectomy for HCC.
METHODS: Multivariate logistic regression and stochastic forest algorithm were performed, then the independent risk factors of severe PHLF were included in a nomogram to determine the risk of severe PHLF. Further, a conditional inference tree (CTREE) through recursive partitioning analysis validated supplement the misdiagnostic threshold of the nomogram.
RESULTS: This study included 924 patients, of whom 137 patients (14.8%) suffered from mild-CSPH and 66 patients suffered from (7.1%) with severe-CSPH confirmed preoperatively. Our data showed that preoperative prolonged prothrombin time, total bilirubin, indocyanine green retention rate at 15 min, CSPH grade, and standard future liver remnant volume were independent predictors of severe PHLF. By incorporating these factors, the nomogram achieved good prediction performance in assessing severe PHLF risk, and its concordance statistic was 0.891, 0.850 and 0.872 in the training cohort, internal validation cohort and external validation cohort, respectively, and good calibration curves were obtained. Moreover, the calculations of total points of diagnostic errors with 95% CI were concentrated in 110.5 (range 76.9-178.5). It showed a low risk of severe PHLF (2.3%), indicating hepatectomy is feasible when the points fall below 76.9, while the risk of severe PHLF is extremely high (93.8%) and hepatectomy should be rigorously restricted at scores over 178.5. Patients with points within the misdiagnosis threshold were further examined using CTREE according to a hierarchic order of factors represented by the presence of CSPH grade, ICG-R15, and sFLR.
CONCLUSIONS: This new surgical scheme established in our study is practical to stratify risk classification in assessing severe PHLF, thereby facilitating surgical decision-making and identifying suitable candidates for individual hepatectomy.
BACKGROUND: Hepatectomy is the preferred treatment for HCC. Surgeons must maintain a balance between the expected oncological outcomes of HCC removal and short-term risks of severe PHLF and morbidity. CSPH aggravates liver decompensation and increases the risk of severe PHLF thus complicating hepatectomy for HCC.
METHODS: Multivariate logistic regression and stochastic forest algorithm were performed, then the independent risk factors of severe PHLF were included in a nomogram to determine the risk of severe PHLF. Further, a conditional inference tree (CTREE) through recursive partitioning analysis validated supplement the misdiagnostic threshold of the nomogram.
RESULTS: This study included 924 patients, of whom 137 patients (14.8%) suffered from mild-CSPH and 66 patients suffered from (7.1%) with severe-CSPH confirmed preoperatively. Our data showed that preoperative prolonged prothrombin time, total bilirubin, indocyanine green retention rate at 15 min, CSPH grade, and standard future liver remnant volume were independent predictors of severe PHLF. By incorporating these factors, the nomogram achieved good prediction performance in assessing severe PHLF risk, and its concordance statistic was 0.891, 0.850 and 0.872 in the training cohort, internal validation cohort and external validation cohort, respectively, and good calibration curves were obtained. Moreover, the calculations of total points of diagnostic errors with 95% CI were concentrated in 110.5 (range 76.9-178.5). It showed a low risk of severe PHLF (2.3%), indicating hepatectomy is feasible when the points fall below 76.9, while the risk of severe PHLF is extremely high (93.8%) and hepatectomy should be rigorously restricted at scores over 178.5. Patients with points within the misdiagnosis threshold were further examined using CTREE according to a hierarchic order of factors represented by the presence of CSPH grade, ICG-R15, and sFLR.
CONCLUSIONS: This new surgical scheme established in our study is practical to stratify risk classification in assessing severe PHLF, thereby facilitating surgical decision-making and identifying suitable candidates for individual hepatectomy.
摘要：
目的：临床意义重大的门静脉高压症（CSPH）严重影响肝细胞癌（HCC）患者手术治疗的可行性和安全性。这项研究的目的是建立一种新的手术方案，定义肝切除术后肝功能衰竭（PHLF）的风险分类，以促进手术决策，并确定肝癌患者CSPH个体肝切除术的合适人选。
背景：肝切除术是肝癌的首选治疗方法。外科医生必须在HCC切除的预期肿瘤结果与严重PHLF和发病率的短期风险之间保持平衡。CSPH会加重肝脏失代偿，并增加严重PHLF的风险，从而使HCC的肝切除术复杂化。
方法：进行多元逻辑回归和随机森林算法，然后将严重PHLF的独立危险因素纳入列线图，以确定严重PHLF的风险。Further,通过递归分区分析验证的条件推理树（CTREE）补充了列线图的误诊阈值。
结果：本研究包括924名患者，其中137例（14.8％）患有轻度CSPH，66例（7.1％）患有重度CSPH。我们的数据显示术前凝血酶原时间延长，总胆红素,吲哚菁绿在15分钟时的保留率，CSPH等级,和标准的未来肝脏残余体积是严重PHLF的独立预测因子。通过结合这些因素，列线图在评估严重PHLF风险方面取得了良好的预测性能，在训练队列中，其一致性统计量为0.891、0.850和0.872，内部验证队列和外部验证队列，分别,并获得了良好的校准曲线。此外，95%CI的诊断错误总点数的计算集中在110.5(范围76.9~178.5).它显示出严重PHLF的低风险（2.3％），提示当评分低于76.9分时,肝切除术是可行的,而严重PHLF的风险极高(93.8%),肝切除术在评分超过178.5分时应严格限制.根据由CSPH分级表示的因素的分层顺序，使用CTREE进一步检查了在误诊阈值内的患者。ICG-R15和sFLR。
结论：在我们的研究中建立的这个新的手术方案对于评估严重PHLF的风险分类是实用的，从而促进手术决策和确定适合个体肝切除术的候选人。
背景：肝切除术是肝癌的首选治疗方法。外科医生必须在HCC切除的预期肿瘤结果与严重PHLF和发病率的短期风险之间保持平衡。CSPH会加重肝脏失代偿，并增加严重PHLF的风险，从而使HCC的肝切除术复杂化。
方法：进行多元逻辑回归和随机森林算法，然后将严重PHLF的独立危险因素纳入列线图，以确定严重PHLF的风险。Further,通过递归分区分析验证的条件推理树（CTREE）补充了列线图的误诊阈值。
结果：本研究包括924名患者，其中137例（14.8％）患有轻度CSPH，66例（7.1％）患有重度CSPH。我们的数据显示术前凝血酶原时间延长，总胆红素,吲哚菁绿在15分钟时的保留率，CSPH等级,和标准的未来肝脏残余体积是严重PHLF的独立预测因子。通过结合这些因素，列线图在评估严重PHLF风险方面取得了良好的预测性能，在训练队列中，其一致性统计量为0.891、0.850和0.872，内部验证队列和外部验证队列，分别,并获得了良好的校准曲线。此外，95%CI的诊断错误总点数的计算集中在110.5(范围76.9~178.5).它显示出严重PHLF的低风险（2.3％），提示当评分低于76.9分时,肝切除术是可行的,而严重PHLF的风险极高(93.8%),肝切除术在评分超过178.5分时应严格限制.根据由CSPH分级表示的因素的分层顺序，使用CTREE进一步检查了在误诊阈值内的患者。ICG-R15和sFLR。
结论：在我们的研究中建立的这个新的手术方案对于评估严重PHLF的风险分类是实用的，从而促进手术决策和确定适合个体肝切除术的候选人。
