Background: Failure to rescue (FTR), defined as death after major postoperative complications, is a critical quality indicator in pancreatic cancer surgery. Despite advances in surgical techniques and perioperative care, FTR rates remain high and vary across institutions.Methods: This systematic review uses a narrative synthesis followed by PRISMA 2020. A PubMed search (1992–2025) identified 83 studies; after screening, 52 studies (2010–2025) were included. Eligible designs were registry-based, multicenter, single-center, or prospective audits. Given substantial heterogeneity in study designs, FTR definitions, and outcome measures, a narrative synthesis was performed; no formal risk-of-bias assessment or meta-analysis was conducted.Results: Definitions of FTR varied (in-hospital, 30-day, 90-day, severity-based, and complication-specific cases). Reported rates differed by definition: average reported rates were 13.2% for 90-day CD ≥ III (G1); 10.3% for in-hospital/30-day CD ≥ III (G3); and 7.4% for 30-day “serious/major” morbidity (G8). Absolute differences were +3.0 and +2.9 percentage points (exploratory, descriptive comparisons). Five domains were consistently associated with lower FTR: (i) centralization to high-volume centers; (ii) safe adoption/refinement of surgical techniques; (iii) optimized perioperative management including early imaging and structured escalation pathways; (iv) patient-level risk stratification and prehabilitation; and (v) non-technical skills (NTSs) such as decision-making, situational awareness, communication, teamwork, and leadership. Among NTS domains, stress and fatigue management were not addressed in any included study.Limitations: Evidence is predominantly observational with substantial heterogeneity in study designs and FTR definitions; the search was limited to PubMed; and no formal risk-of-bias, publication-bias assessment, or meta-analysis was performed. Consequently, estimates and associations are descriptive/associative with limited certainty and generalizability.Conclusions: NTSs were rarely used or measured across the included studies, with validated instruments; quantitative assessment was uncommon, and no study evaluated stress or fatigue management. Reducing the FTR after pancreatic surgery will require standardized, pancreas-specific definitions of FTR, process-level rescue metrics, and deliberate strengthening of NTS. We recommend a pancreas-specific operational definition with an explicit numerator/denominator: numerator = all-cause mortality within 90 days of surgery; denominator = patients who experience major complications (Clavien–Dindo grade III–V, often labeled “CD ≥ 3”). Addressing the gaps in stress and fatigue management and embedding behavioral metrics into quality improvement programs are critical next steps to reduce preventable mortality after complex pancreatic cancer procedures.
**背景：** 挽救失败，定义为发生重大术后并发症后的死亡，是胰腺癌手术中一项关键的质量指标。尽管手术技术和围手术期护理有所进步，但挽救失败率仍然很高，且在不同机构间存在差异。 **方法：** 本系统综述采用叙述性综合法，并遵循PRISMA 2020指南。通过PubMed检索（1992–2025年）确定了83项研究；筛选后，纳入52项研究（2010–2025年）。符合条件的研究设计包括基于登记库的研究、多中心研究、单中心研究或前瞻性审计。鉴于研究设计、挽救失败定义和结局指标存在显著异质性，进行了叙述性综合；未进行正式的偏倚风险评估或荟萃分析。 **结果：** 挽救失败的定义各异（住院期间、30天、90天、基于严重程度以及特定并发症病例）。报告的发生率因定义不同而有所差异：90天Clavien-Dindo分级≥III级的平均报告发生率为13.2%；住院/30天Clavien-Dindo分级≥III级的平均报告发生率为10.3%；30天“严重/重大”并发症的平均报告发生率为7.4%。绝对差异分别为+3.0和+2.9个百分点（探索性、描述性比较）。有五个领域与较低的挽救失败率持续相关：（i）向高手术量中心集中；（ii）安全采用/改进手术技术；（iii）优化围手术期管理，包括早期影像学检查和结构化升级处理路径；（iv）患者层面的风险分层和预康复；（v）非技术技能，如决策、情境意识、沟通、团队合作和领导力。在非技术技能领域中，压力和疲劳管理在所有纳入研究中均未涉及。 **局限性：** 证据主要为观察性研究，研究设计和挽救失败定义存在显著异质性；检索仅限于PubMed；未进行正式的偏倚风险、发表偏倚评估或荟萃分析。因此，估计值和关联性是描述性/关联性的，其确定性和普遍性有限。 **结论：** 在纳入的研究中，非技术技能很少被使用或测量，使用经过验证工具的研究不多，定量评估不常见，且没有研究评估压力或疲劳管理。降低胰腺手术后挽救失败率需要标准化的、胰腺特异性的挽救失败定义、过程层面的挽救指标，并有意识地加强非技术技能。我们建议采用一个胰腺特异性的操作定义，明确分子/分母：分子 = 术后90天内全因死亡率；分母 = 发生重大并发症的患者。解决压力和疲劳管理方面的空白，并将行为指标纳入质量改进计划，是降低复杂胰腺癌手术后可预防死亡率的关键后续步骤。
肿瘤(癌症)患者之家