Multimodal treatment of gastric cancer : assessment of tumor regression and the role of minimally invasive surgery

Sammanfattning: The generally poor prognosis of advanced gastric cancer (AGC) has led to the development of multimodal therapeutic protocols in which surgery is combined with systemic treatment, perioperative chemotherapy being the preferred approach in Europe. In this context, there still are several issues that need to be addressed. Information on the in vivo efficacy of the administered cytotoxic therapy, as reflected by the histopathologic Tumor Regression Grade (TRG), can have numerous applications in research and clinical practice. At present, however, this is hampered by the diversity of the TRG systems in use. Another drawback is the fact that the current chemotherapy regimens are characterized by considerable toxicity and consequently, failure to resume the systemic treatment after surgery is a frequent clinical scenario. Apparently, factors related to the surgical procedure, such as complications and nutritional problems that are common after gastrectomy, play an important role. Accordingly, a potential positive effect of minimally invasive surgery should be further investigated. The overall aim of study I was to, through an international Delphi survey, reach consensus on a TRG system that could be internationally adopted. A review of the literature and subsequent evaluation of the existing systems resulted in the formation of a questionnaire consisting of 72 statements, categorized into nine topics: (1) specimen processing, (2) gross examination, (3) cross-sectioning/method of sampling (4) staining, (5) immunohistochemistry, (6) assessment of tumor regression in response to neoadjuvant therapy, (7) tumor regression grading, (8) assessment of regression of nodal metastases, and (9) role of histologic tumor type. The survey was undertaken between January and October 2017 and was terminated after four rounds, with consensus reached in 97% of the statements. The main objective was to form a proposal on how to assess and stratify TRG not only in the primary tumor, but also in lymph node (LN) metastases. For the primary tumor, a 4-tiered classification was advocated: grade 1 corresponds to complete response, grade 2 indicates <10% residual tumor (subtotal regression), grade 3 10-50% residual tumor (partial regression), and grade 4 >50% residual tumor (minimal/no regression). The novelty of the proposed system is that this is complemented with a 3- tiered grading for LN metastases, with grades a, b and c corresponding to complete, partial, or no regression, respectively. In the ensuing study II, the interobserver agreement of the proposed TRG system from study I was evaluated. Twenty observers assessed 60 histopathologic slides (30 primary tumors, 30 LNs), and the level of agreement was assessed by estimating the Kendall’s coefficient of concordance. With respect to primary tumors, the level of agreement was good and independent of previous experience. When assessing the regression of LN metastases, the agreement was good among subspecialized observers, whereas the estimated coefficient among non-subspecialized observers corresponded to a moderate agreement. Even though the interobserver agreement did not reach the highest level, the results are encouraging when taking into account the study conditions (relatively large number of assessors with varying previous experience, a single hematoxylin-eosin stained slide provided for each case, no ability to complement with immuno-histochemistry) and suggest that the proposed system is reproducible. Future studies should investigate whether this information on nodal response is of clinical significance, in a way that it enhances the prognostic value of conventional TRG systems that focus exclusively on the alterations that take place in the primary tumor. Study III addressed the question of whether laparoscopic gastrectomy (LG) for AGC yields oncologic results equivalent to those obtained by conventional open gastrectomy (OG). This was a population-based study with data from the Swedish national register between 2015-2020 and included 622 patients receiving curative intended treatment for cT2-4aN0-3M0 adenocarcinoma of the stomach or cardia type III. Comparison of shortterm postoperative and survival outcomes, as well as assessment of associations with the surgical approach, was done using multivariable logistic regression, the Kaplan-Meier method and multivariable Cox regression. The results confirmed that LG is safe, both in terms of morbidity and mortality and in terms of oncologic surgical quality (R0 resection rate, number of retrieved LNs). The main finding in study III was that LG was associated with better overall survival as compared to OG. However, stratified analyses revealed that this survival advantage was restricted to patients undergoing distal gastrectomy. The importance of these results lies in that: (i) they are based on a Western population with 53% of patients receiving neoadjuvant chemotherapy and 48% undergoing total gastrectomy, and (ii) the early period of implementation of LG was not excluded, meaning that they incorporate the learning curve phase. The mechanism that could be driving the observed survival benefit of LG in study III was not clear. In study IV, we examined whether LG has a positive impact on the compliance with adjuvant treatment and thus, could possibly partially explain this finding. Using the same cohort from the national register, we selected all patients who were treated with neoadjuvant chemotherapy and surgery and were planned to resume chemotherapy postoperatively. A total of 247 patients (126 LG and 121 OG) were included and data on the administered chemotherapy was extracted from their medical records. Omission of adjuvant chemotherapy or need for reduction of the chemotherapy regimen occurred in 44% of the patients in total. Multivariable analysis showed no evidence of a positive effect of LG in terms of the probability to start adjuvant treatment or the need for chemotherapy reduction. In contrast, severe complications had a significant, negative impact on these outcomes irrespective of surgical approach, emphasizing that further exploration of perioperative interventions with focus on minimizing postoperative morbidity is needed.