Preoperative chemoradiotherapy improved survival among patients with potentially curable esophageal or esophagogastric-junction cancer. The regimen was associated with acceptable adverse-event rates. (Funded by the Dutch Cancer Foundation [KWF Kankerbestrijding]; Netherlands Trial Register number, NTR487.).

No abstract

BackgroundFor esophageal cancer, the number of retrieved lymph nodes (LNs) is often used as a quality indicator. The aim of this study is to analyze the number of retrieved LNs in The Netherlands, assess factors associated with LN yield, and explore the association with short-term outcomes. This is a population-based study on lymph node retrieval in patients with esophageal cancer, presenting results from the Dutch Upper Gastrointestinal Cancer Audit.Study DesignFor this retrospective national cohort study, patients with esophageal carcinoma who underwent esophagectomy between 2011 and 2016 were included. The primary outcome was the number of retrieved LNs. Univariable and multivariable regression analyses were used to test for association with ≥ 15 LNs.Patients and Results3970 patients were included. Between 2011 and 2016, the median number of LNs increased from 15 to 20. Factors independently associated with ≥ 15 LNs were: 0–10 kg preoperative weight loss (versus: unknown weight loss, odds ratio [95% confidence interval]: 0.71 [0.57–0.88]), Charlson score 0 (versus: Charlson score 2: 0.76 [0.63–0.92]), cN2 category (reference: cN0, 1.32 [1.05–1.65]), no neoadjuvant therapy and neoadjuvant chemotherapy (reference: neoadjuvant chemoradiotherapy, 1.73 [1.29–2.32] and 2.15 [1.54–3.01]), minimally invasive transthoracic (reference: open transthoracic, 1.46 [1.15–1.85]), open transthoracic (versus open and minimally invasive transhiatal, 0.29 [0.23–0.36] and 0.43 [0.32–0.59]), hospital volume of 26–50 or > 50 resections/year (reference: 0–25, 1.94 [1.55–2.42] and 3.01 [2.36–3.83]), and year of surgery [reference: 2011, odds ratios (ORs) 1.48, 1.53, 2.28, 2.44, 2.54]. There was no association of ≥ 15 LNs with short-term outcomes.ConclusionsThe number of LNs retrieved increased between 2011 and 2016. Weight loss, Charlson score, cN category, neoadjuvant therapy, surgical approach, year of resection, and hospital volume were all associated with increased LN yield. Retrieval of ≥ 15 LNs was not associated with increased postoperative morbidity/mortality.Electronic supplementary materialThe online version of this article (10.1245/s10434-018-6396-7) contains supplementary material, which is available to authorized users.

Aim We aim to compare machine learning with neural network performance in predicting R0 resection (R0), length of stay > 14 days (LOS), major complication rates at 30 days postoperatively (COMP) and survival greater than 1 year (SURV) for patients having pelvic exenteration for locally advanced and recurrent rectal cancer. Method A deep learning computer was built and the programming environment was established. The PelvEx Collaborative database was used which contains anonymized data on patients who underwent pelvic exenteration for locally advanced or locally recurrent colorectal cancer between 2004 and 2014. Logistic regression, a support vector machine and an artificial neural network (ANN) were trained. Twenty per cent of the data were used as a test set for calculating prediction accuracy for R0, LOS, COMP and SURV. Model performance was measured by plotting receiver operating characteristic (ROC) curves and calculating the area under the ROC curve (AUROC). Results Machine learning models and ANNs were trained on 1147 cases. The AUROC for all outcome predictions ranged from 0.608 to 0.793 indicating modest to moderate predictive ability. The models performed best at predicting LOS > 14 days with an AUROC of 0.793 using preoperative and operative data. Visualized logistic regression model weights indicate a varying impact of variables on the outcome in question. Conclusion This paper highlights the potential for predictive modelling of large international databases. Current data allow moderate predictive ability of both complex ANNs and more classic methods.

No abstract