Neutrophil-to-lymphocyte, lymphocyte-to-monocyte and platelet-to-lymphocyte ratios as predictive markers of pathological response to FLOT neoadjuvant strategy in locally advanced gastric/gastroesophageal junction cancer

Locally advanced gastric and gastroesophageal junction (LAG/GEJ) cancers are responsible for the third leading cause of cancer-related mortality worldwide. FLOT perioperative chemotherapy showed improved prognosis in patients with LAG/GEF cancers, but predictive biomarkers have not been established. Neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR) and platelet-to-lymphocyte ratio (PLR) are inflammation biomarkers already reported as independent prognostic factors in several solid tumours. This study investigated the importance of NLR, PLR, and LMR in patients with resectable LAG/GEF and assesses the clinical potential of risk stratification for tumour behaviour prediction. We conducted a multi-institutional retrospective study in Portugal, including patients from 9 different oncological centres, diagnosed with LAG/GEJ cancer, that underwent surgery after at least one FLOT cycle, since its initial use in the institutions until December 31st, 2019. Analysis of receiver operating characteristics curve defined a cut-off value for group stratification. Univariate regression analysis was used to estimate the risk of disease progression/absence of response. A total of 107 patients (67,3% males) were included, with mean age of 65,2 years (40-84 years). All were adenocarcinomas (92,5% gastric; 6,5% GEJ; 0,9% oesophageal) and 41,1% poorly differentiated. Median number of preoperative FLOT cycles was 4 (2-8 cycles). Subtotal gastrectomy was performed in 53,2% of patients, while 38,3% had total gastrectomy, 7,4% Ivory-Lewis esophagectomy and 0,9% distal gastrectomy. Complete standard lymphadenectomies were 73.7% (56% D2, 17,7% D1+), with R0 resection achieved in 89,7% patients. Tumour downstaging (TD) was observed in 64,5% patients. Of the 86 surgical specimens evaluated for tumour regression grade (TRG), 45,3% showed partial response, 36% no response, and 18,6% pathological complete response. Optimal cutoff values for NLR, MLR and PLR were determined regarding TRG and TD, respectively. High NLR (>2.95 vs. >2.13) and PLR (>135.6 vs. >128.5), as well as low LMR (< 0.05, 95%CI): NLR OR 3.66, 1.59–8.41, P=.002 / MLR OR 2.88, 1.27–6.51, P=.011 / PLR OR 2.25, 1.0–5.0, P=.05). To better stratify risk, we combined ratios by ascending risk order, classifying two groups: 1) "Lower Risk” - [high LMR, low NLR & PLR] < [high LMR & PLR, low NLR] < [low LMR, NLR & PLR] < [high LMR & NLR, low PLR]; and 2) "Higher Risk” - [high PLR, low LMR & NLR] < [high NLR, LMR & PLR] < [high NLR, low LMR & PLR] < [high NLR & PLR, low LMR]); the second group had higher risk of progression or not responding (OR 4.1, 95%CI: 1.75–9.45, P=.001). NLR, LMR & PLR have the potential to be used as biomarkers to predict pathological response to preoperative FLOT. Through risk assessment, we suggest stratification of 2 groups to better evaluate patients and predict outcomes. Higher risk patients should be considered for more active surveillance during chemotherapy, to detect progression early and possibly reconsider therapeutic strategies.