LifeGlimmer GmbH, Berlin, Germany; Laboratory of Systems and Synthetic Biology, Wageningen University and Research, Wageningen, the Netherlands.
Burn Centre, Red Cross Hospital, Beverwijk, the Netherlands; Department of Plastic, Reconstructive and Hand Surgery, Amsterdam Movement Sciences Amsterdam UMC, Amsterdam, the Netherlands; Association of Dutch Burn Centers, Beverwijk, the Netherlands.
Int J Med Inform. 2022 Nov;167:104878. doi: 10.1016/j.ijmedinf.2022.104878. Epub 2022 Sep 24.
Necrotizing Soft Tissue Infections (NSTI) are severe infections with high mortality affecting a heterogeneous patient population. There is a need for a clinical decision support system which predicts outcomes and provides treatment recommendations early in the disease course.
To identify relevant clinical needs, interviews with eight medical professionals (surgeons, intensivists, general practitioner, emergency department physician) were conducted. This resulted in 24 unique questions. Mortality was selected as first endpoint to develop a machine learning (Random Forest) based prediction model. For this purpose, data from the prospective, international INFECT cohort (N = 409) was used.
Applying a feature selection procedure based on an unsupervised algorithm (Boruta) to the > 1000 variables available in INFECT, including baseline, and both NSTI specific and NSTI non-specific clinical data yielded sixteen predictive parameters available on or prior to the first day on the intensive care unit (ICU). Using these sixteen variables 30-day mortality could be accurately predicted (AUC = 0.91, 95% CI 0.88-0.96). Except for age, all variables were related to sepsis (e.g. lactate, urine production, systole). No NSTI-specific variables were identified. Predictions significantly outperformed the SOFA score(p < 0.001, AUC = 0.77, 95% CI 0.69-0.84) and exceeded but did not significantly differ from the SAPS II score (p = 0.07, AUC = 0.88, 95% CI 0.83-0.92). The developed model proved to be stable with AUC > 0.8 in case of high rates of missing data (50% missing) or when only using very early (<1 h) available variables.
This study shows that mortality can be accurately predicted using a machine learning model. It lays the foundation for a more extensive, multi-endpoint clinical decision support system in which ultimately other outcomes and clinical questions (risk for septic shock, AKI, causative microbe) will be included.
坏死性软组织感染（NSTI）是一种严重的感染，死亡率高，影响异质患者群体。需要有一种临床决策支持系统，能够在疾病早期预测结果并提供治疗建议。
为了确定相关的临床需求，对 8 名医疗专业人员（外科医生、重症监护医生、全科医生、急诊医生）进行了访谈。这导致了 24 个独特的问题。选择死亡率作为第一个终点来开发一个基于机器学习（随机森林）的预测模型。为此，使用前瞻性国际感染队列（N=409）的数据。
应用基于无监督算法（Boruta）的特征选择程序对感染中可用的>1000 个变量（包括基线以及 NSTI 特异性和非 NSTI 特异性临床数据）进行处理，得到 16 个可在或之前 ICU 第 1 天获得的预测参数。使用这 16 个变量，可以准确预测 30 天死亡率（AUC=0.91，95%CI 0.88-0.96）。除年龄外，所有变量均与脓毒症有关（例如乳酸、尿量、收缩压）。未发现 NSTI 特异性变量。预测结果明显优于 SOFA 评分（p<0.001，AUC=0.77，95%CI 0.69-0.84），并且高于但不显著高于 SAPS II 评分（p=0.07，AUC=0.88，95%CI 0.83-0.92）。在数据缺失率较高（50%缺失）或仅使用早期（<1h）可用变量的情况下，该模型的 AUC>0.8，证明其具有稳定性。
本研究表明，使用机器学习模型可以准确预测死亡率。它为更广泛的、多终点的临床决策支持系统奠定了基础，最终将包括其他结果和临床问题（脓毒性休克、AKI、病原体）。
