Back

Prognostic significance of compound combined peri-operative biomarkers in gastric cancer

Powell, A. G.; Coxon, A. H.; Robinson, D.; James, O.; Christian, A.; Roberts, S. A.; Lewis, W. G.

2020-04-24 surgery
10.1101/2020.04.21.20066332 medRxiv
Show abstract

BackgroundSurvival after gastric cancer surgery is largely attributed to tumor biology, neoadjuvant chemotherapy (NAC), and surgical approach, yet other prognostic factors have been reported, including pre-operative systemic inflammatory response (SIR), and Morbidity Severity Score (MSS). The hypothesis tested was that a SIR, MMS, and pathological composite score, would be associated with disease-free (DFS) and overall survival (OS). MethodsConsecutive 358 patients undergoing potentially curative gastrectomy for adenocarcinoma were studied. Complications were defined as a MSS of Clavien-Dindo classification (CDSC) >1. Serum SIR measurements were performed on the day before surgery, and a composite score (CIMpN) (0-3) was developed based on CRP, morbidity, and pN-stage. Primary outcome measures were DFS and OS. ResultsPost-operative complications occurred in 138 (38.5%) patients, (8 (2.2%) deaths), and was associated with higher CRP (28.3% vs. 15.5%, p=0.003), vascular invasion (55.8% vs. 36.8%, p<0.001), and R1 status (26.1% vs. 9.5%, p=0.001). Five-year DFS and OS were 32.9% and 33.3% for patients with post-operative complications compared with 62.5% and 64.0% in controls (p<0.001). Five-year DFS and OS were 31.4% and 37.3% in patients with raised CRPs compared with 58.5% and 59.5% in controls (p=0.005, p=0.001, respectively). Five-year DFS for CIMpN scores of 0, 1, 2, and 3 were 85.9%, 50.0%, 26.2%, and 15.4% (p<0.001) respectively. On multivariable analysis CIMpN score was independently associated with DFS [HR 3.00, 95% Confidence Interval (CI) 1.90-4.73, p<0.001] and OS [1.93 (1.43-2.59), p<0.001]. ConclusionA novel composite score, CIMpN, based on SIR, MSS and pN-stage, offers important prognostic signals.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.