Elfanagely Omar, Mellia Joseph A, Othman Sammy, Basta Marten N, Mauch Jaclyn T, Fischer John P
Division of Plastic Surgery, Department of Surgery, University of Pennsylvania, Philadelphia, Pa.
Department of Plastic and Reconstructive Surgery, Brown University. Providence, R.I.
Plast Reconstr Surg Glob Open. 2020 Dec 16;8(12):e3307. doi: 10.1097/GOX.0000000000003307. eCollection 2020 Dec.
Ventral hernias are a complex and costly burden to the health care system. Although preoperative radiologic imaging is commonly performed, the plethora of anatomic features present and available in routine imaging are seldomly quantified and integrated into patient selection, preoperative risk stratification, and perioperative planning. We herein aimed to critically examine the current state of computed tomography feature application in predicting surgical outcomes.
A systematic review was conducted in accordance with the Preferred Reporting Items for a Systematic Review and Meta-Analysis (PRISMA) checklist. PubMed, MEDLINE, and Embase databases were reviewed under search syntax "computed tomography imaging" and "abdominal hernia" for papers published between 2000 and 2020.
Of the initial 1922 studies, 12 papers met inclusion and exclusion criteria. The most frequently used radiologic features were hernia volume (n = 9), subcutaneous fat volume (n = 5), and defect size (n = 8). Outcomes included both complications and need for surgical intervention. Median area under the curve (AUC) and odds ratio were 0.68 (±0.16) and 1.12 (±0.39), respectively. The best predictive feature was hernia neck ratio > 2.5 (AUC 0.903).
Computed tomography feature selection offers hernia surgeons an opportunity to identify, quantify, and integrate routinely available morphologic tissue features into preoperative decision-making. Despite being in its early stages, future surgeons and researchers will soon be able to integrate 3D volumetric analysis and complex machine learning and neural network models to improvement patient care.
腹疝给医疗保健系统带来了复杂且代价高昂的负担。尽管术前通常会进行放射学成像检查，但常规成像中存在的大量解剖特征很少被量化并纳入患者选择、术前风险分层和围手术期规划中。我们在此旨在严格审查计算机断层扫描特征在预测手术结果方面的应用现状。
按照系统评价和荟萃分析的首选报告项目（PRISMA）清单进行系统评价。在PubMed、MEDLINE和Embase数据库中，以搜索语法“计算机断层扫描成像”和“腹疝”对2000年至2020年发表的论文进行检索。
在最初的1922项研究中，有12篇论文符合纳入和排除标准。最常用的放射学特征是疝体积（n = 9）、皮下脂肪体积（n = 5）和缺损大小（n = 8）。结果包括并发症和手术干预需求。曲线下面积（AUC）中位数和比值比分别为0.68（±0.16）和1.12（±0.39）。最佳预测特征是疝颈比> 2.5（AUC 0.903）。
计算机断层扫描特征选择为疝外科医生提供了一个机会，可将常规可用的形态学组织特征识别、量化并整合到术前决策中。尽管尚处于早期阶段，但未来的外科医生和研究人员很快就能整合三维容积分析以及复杂的机器学习和神经网络模型，以改善患者护理。
