Perioperative Treatment in Gastric Cancer—Developments in Patient Selection

Gastric cancer (GC) is a major public health issue. It is considered the 5th most common cancer diagnosed worldwide and it is one of the main causes of malignant disease-associated morbidity and mortality. The cornerstone of curative treatment is still surgery, and since the rate of relapse is high, a multidisciplinary approach is warranted in most developed countries. And while there have been recent developments in the perioperative scenario namely the FLOT regimen, little has advanced considering patient selection. We have reviewed the major trials in this setting and provide some insights from recently reported microsatellite instability (MSI) in a subgroup analysis in the MAGIC trial patients that seem to suggest an opportunity to patient selection. Furthermore, GC subtyping may prove helpful selecting candidates to immunotherapy or even multimodal therapy in the future. As the paradigm is moving towards a precision oncology model, GC patient selection remains one the biggest challenges in oncology but seems closer to clinical practice reality as new developments are being reported.


Introduction
Gastric cancer (GC) represents the 5 th most common cancer diagnosed worldwide, being one of the main causes of morbidity and mortality due to malignant disease.The incidence of GC varies geographically, being particularly higher in the Asiatic population compared to the Western countries.In Europe, almost 107,000 deaths occurred in 2012 [1] [2].Currently, the only curative treatment for GC is surgery, with survival rates at 5 years of 70% -75% in stage I and 35% in stage II and III for completely resected GC, with several patients experiencing recurrence during follow up [3].Those disappointing numbers lead to different strategies that should be discussed in a multidisciplinary scenario: neoadjuvant/perioperative and adjuvant therapy, all currently accepted as possible and valid strategies.
As the field is moving towards a precision oncology paradigm in clinical practice, patient selection becomes increasingly important-how to identify the patients that would benefit the most, while sparing them unnecessary toxicities.On the other hand, being Medical Oncologists who treat gastric patients in our routine clinical practice, we understand the need for better treatments, as relapse is predominantly a synonym of a dire prognosis.The authors will review some of the most relevant data and discuss its implications.

Materials and Methods
We identified the positive phase III trials in the perioperative setting in gastric or gastro-esophageal junction using only chemotherapy, not radiotherapy or chemoradiotherapy treatments.In parallel, we conducted a search for the latest and most credible evidence regarding molecular profiling for patient selection in perioperative gastric cancer patients.

Discussion
The evidence for perioperative chemotherapy in GC in Europe was based in two In June 2017, the results of the FLOT4-AIO trial were presented as abstract at the American Society of Clinic Oncology (ASCO) annual meeting and since then this regimen has been generally accepted as the promising new standard of care for perioperative chemotherapy in GC [6]  Additionally, one of the major pitfalls considering GC staging is accurate nodal disease assessment-that is generally considered to be under evaluated by standard imaging, such as computed tomography (CT) scan.The sensitivity for CT scan in diagnosing malignant lymph node disease is variable [10], and in our clinical practice experience it can be even lower than published in the literature (35% vs 60% -90%) [11].Still today, a global consensus on malignant lymph node disease diagnostic criteria is not available [10], so we believe that a better understanding of the role of the different imaging modalities is necessary for better patient selection.
In February 2017, an exploratory post hoc analysis from the MAGIC trial was published [12].The goal was to evaluate the proportion of patients with high MSI (MSI-H) or mismatch repair deficiency (MMRD) in the 2 cohorts, and if these biomarkers had prognostic effect on survival.Of the 456 patients who underwent surgery in the MAGIC trial, MSI results were available for 303 patients, and only 20 had MSI-H, all located in primary gastric tumors.Compared to MSS (stable) or MSI-L (low) tumors, patients with MSI-H were more frequent in female, with older median age and intestinal subtype, but none of these differences were statistically significant.Patients treated with surgery alone who had MSI-H or MMRD had a median OS that was not reached (95% CI, 11.5 months to not reached) versus 20.5 months (95% CI, 16.7 -27.8 months; HR 0.42; 95% CI, 0.15 -1.15; P = 0.09) among those who had neither MSI-H nor MMRD.In contrast, patients treated with chemotherapy plus surgery who had either high MSI or MMRD had a mOS of 9.6 months (95% CI, 0.1 -22.5 months) compared with the mOS among those who were neither MSI-H nor MMRD of 19.5 months (95% CI, 15.4 -35.2 months; hazard ratio, 2.18; 95% CI, 1.08 -4.42; P = 0.03).
Despite the limitations of this exploratory analysis, the authors concluded that MMRD and MSI-H were correlated with a positive prognostic effect in patients treated with surgery alone and a negative prognostic effect in the perioperative group, thus suggesting a potential development in patient selection.Although these results seem promising, we believe that due to the low patient numbers (around 20) the evidence may not be strong enough to change clinical practice yet.Furthermore, these results may not be extrapolated to predict "chemoresistance" in the general population, since that in the MAGIC trial the regimen included anthracycline but now the standard regimen FLOT does not.It would be useful to explore these principles in the FLOT4-AIO trial subpopulations in order to understand if the concept of "chemoresistance" is regimen-associated or not.
A major breakthrough in understanding GC heterogeneity was the Cancer Genome Atlas research network (TCGA) publication in 2014, that suggested 4 subtypes of gastric tumours, based on their molecular signature: Epstein Barr virus (EBV) positive, microsatellite unstable (MSI), genomically stable (GS) and chromosomal instability (CIN) [13].Of these, the EBV positive and MSI subtypes have a higher mutational burden and it may play a role in immune response, rendering them as potential candidates for "better responders" to immunotherapy drugs [14].For instance, the role of perioperative Pembrolizumab, an anti-PD-1 drug, is already ongoing (ClinicalTrials.govIdentifier: NCT02918162).

Conclusions and Future Directions
As tumor heterogeneity in GC is increasingly recognized and molecular subtyping is being further understood, the field is evolving at a fast pace and new standards of care may emerge in the near future.One of the latest practice changing novelties was the perioperative regimen FLOT in resectable GC that improved OS and PFS when compared to the MAGIC trial regimen.However, insights regarding MSI in subgroup analysis in the MAGIC trial patients seem to suggest an opportunity to patient selection.Furthermore, advancing GC subtyping may prove helpful selecting candidates to immunotherapy-still experimental in this setting, but demonstrated in the metastatic patients-or even multimodal therapy selection.Standardizing imaging criteria for lymph node disease assessment is a fundamental need for accurate staging.As the paradigm is moving towards a precision oncology model, GC patient selection remains one the biggest challenges in oncology today.