Requirement for a standardised definition of advanced gastric cancer

Each year, ~988,000 new cases of stomach cancer are reported worldwide. Uniformity for the definition of advanced gastric cancer (AGC) is required to ensure the improved management of patients. Various classifications do actually exist for gastric cancer, but the classification determined by lesion depth is extremely important, as it has been shown to correlate with patient prognosis; for example, early gastric cancer (EGC) has a favourable prognosis when compared with AGC. In the literature, the definition of EGC is clear, however, there is heterogeneity in the definition of AGC. In the current study, all parameters of the TNM classification for AGC reported in each previous study were individually analysed. It was necessary to perform a comprehensive systematic literature search of all previous studies that have reported a definition of ACG to guarantee homogeneity in the assessment of surgical outcome. It must be understood that the term ‘advanced gastric cancer’ may implicate a number of stages of disease, and studies must highlight the exact clinical TNM stages used for evaluation of the study.

The average time between EGC diagnosis and progression is ~37 months (19) and it commonly takes 8 months for EGC to progress to an advanced stage of disease. The percentage of EGC cases reported in Japan is higher (30-50%) compared with Western countries (5-15%), where screening programmes have not been implemented (20). For example, in Korea the proportion of T1 and N0 tumours was demonstrated to increase gradually with time, from 24.8 to 41.3% in period 1 (1986)(1987)(1988)(1989)(1990)) and from 48.9 to 58.6% in period 4 (2000-2006; both P<0.001) (21). The importance of correctly identifying EGC is highlighted by its correlation with the excellent results achieved by surgical treatment and the favourable prognosis of EGC patients following surgery.
The survival rate following surgical resection is 70-95% depending on LN involvement. Therefore, it is extremely useful to understand the involvement of LNs prior to surgery. However, there is no simple, high sensitivity test available to evaluate LN status pre-operatively. The present diagnostic imaging techniques remain unsatisfactory, however, double contrast-enhanced ultrasonography shows promise as a new, non-invasive, convenient and reproducible method for the evaluation of LN metastasis in EGC and the prognosis of EGC pre-operatively. The identification of EGC by marked hyperenhancement allows for closer post-operative follow-up and possibly the use of drugs for the targeted growth of new vessels to prevent recurrence (22).
Macroscopic classification is of extreme prognostic relevance, therefore, it is fundamental to standardise the definitions of EGC and AGC. The definition of EGC is well-codified in the literature, however, a clear definition of AGC is required.
There are certain examples of heterogeneity in the definition of AGC in the literature, and the majority highlight an unclear or heterogeneous definition of AGC. A non-systematic search for studies that highlight a clear definition of AGC confirmed the lack of consistency for a definition of AGC in various research fields. The definitions of AGC generated in single studies are reported in Table II. The aim of the current study was to individually analyse all parameters of the TNM classification for AGC reported in each previous study. The definitions of AGC provided in the previous singular studies were compared with the current classifications of the Japanese Society of Gastric Cancer (15) using only the parameters for TNM classification. The results showed that only 5 studies resulted in a AGC T3 or T4 classification of the gastric neoplasms (23)(24)(25)(26)(27), whilst the remaining studies were hypothesised to be AGC neoplasms with parietal invasion initiating from stage T2 (Table III) (26), LN involvement at a minimum of the N1 stage was required for the definition of AGC. Studies by Luinetti et al (28) and Theuer et al (29), however, defined AGC neoplasms as between T2 and T4, irrespective of the N parameter (Table IV).
Macroscopic classification of EGC and AGC is a simple and reliable method to prediction LN metastasis and determine the extent of LN dissection, unless endoscopic ultrasound (EUS) or CT scans show highly suspected LN metastasis.
All studies analysed in the current study, with the exception of those by Scaringi  In the Japanese Classification of Gastric Cancer (JCGC; 3rd English edition), stomach tumours are grouped by macroscopic classification into 2 types, superficial tumours (T1) and advanced tumours (T2-4) (15).
Superficial tumours, defined as type 0 are divided into 5 subtypes according to the Japanese Endoscopy Society Classification (1962 ; Table V), whereas AGC is divided into 5 types according to macroscopic appearance (Table VI). According to the Japanese Classification, stages Ia, Ib, IIa and IIb (T1a, T1b and N) are associated with EGC (Table VII).
In previous Japanese studies, AGC classification has been clear and univocal. By contrast, studies of other populations often exhibit heterogeneous definitions of AGC. Non-uniformity in the definition of AGC with regard to parameter T emerged from the studies analysed previously (Table III), with 11/16 studies (68.7%) considering that AGC started from T2, whilst 5/16 studies (31.2%) considered the start from T3.
Non-homogeneity of the definition for AGC is more evident when considering the parameters N and M. The definition prescinds from LN involvement or the presence of metastases. In the studies included in the current review, 9/16 were consistent with this definition of AGC. However, 5/16 (31.2%) considered N+ tumours as AGCs and 2/16 studies (12.5%) did not evaluate LN involvement.
For parameter M, 9/16 studies (56.2%) were consistent with the definition of AGC highlighted in Western studies, and such heterogeneity of the definition was reflected again in the assess-  The hypothesis that AGC starts from T3 leads to exclusion of T2 neoplasms from studies and is likely to modify the results. In addition, with regard to LN involvement and metastases, to consider only M+ or N+ as associated with AGC excludes N0 and M0 and leads to incomparable results.
In the context of such heterogeneity, it is impossible to use macroscopic classification since it is not certain that all studies identify neoplasms as starting from T2. Therefore, a number of studies incorrectly determine certain advanced neoplasms as early. In addition, other studies exclude neoplasms defined as advanced, therefore, making the interpretation of the results obtained difficult. surrounded by a thickened gastric wall without clear margins. 4 (diffuse infiltrative) Tumours without marked ulceration or raised margins, the gastric wall is thickened and indurated and the margin is unclear.
The TNM classification was standardised in 2008. The AJCC Cancer Staging manual (7 th edition) (31) and the JCGC (3rd English Edition) (15) are comparable, with the only differences being with regard to the M parameter. In the AJCC classification, MX is excluded and M1 includes positive peritoneal fluid cytology. However, there are no differences between the two classifications with regard to staging. When the stage is used for analysis, a c-prefix must be added to specify that the stage is a clinically predicted stage, for example, analysis by TNM stage, including cT1N0, cT2-4N+ and cM1. The definition of EGC is extremely important for prognostic purposes and for the choice of therapeutic strategy, however, the principal classification used has been the TNM classification, with univocal results. Macroscopic classification is less used and often inconsistent with the definition highlighted by the Japanese Gastric Cancer Association (JGCA).
Therefore, we recommend standardisation of the definition for AGC. To correctly evaluate the results of the previous studies, the classification supplied by the JGCA must be used to guarantee homogeneity in the assessment of surgical outcomes. It is evident that a comprehensive systematic literature search must be performed for all studies that report a definition of ACG, in order to confirm the hypothesis and to subsequently identify a uniform definition of AGC to be used worldwide for the management of patients. The cTNM stages must be identified in the systemic review for a detailed description, as EUS or CT may now be utilised to pre-operatively evaluate T and N stage, although not always with 100% accuracy; therefore, the KLASS-01 study used 'stage 1' cancer as an indication (32).

Conclusion
The definition of EGC and AGC by macroscopic classification regardless of LN metastasis, is made when the tools to determine LN metastasis (EUS and high quality CT) are unavailable and gastroscopy and sonography become the most reliable methods for the evaluation of clinical stages. However, this has been an extremely useful definition for the determination of the treatment strategy and the prediction of LN metastasis and the prognosis. Although EUS and CT are capable of predicting LN metastasis (although not perfectly), the definition of EGC remains useful in a number of cases, since the majority of EGC do not exhibit LN metastasis.
Use of the term 'advanced' is more complicated compared with EGC, since it may be defined as: i) T2 or deeper cancer according to the Japanese classification; ii) a bulky tumour with LN enlargement compared with an 'early stage tumour'; iii) T4b (adjacent organ invading) tumour or unresectable tumour due to adjacent organ invasion or bulky LNs without distant metastasis, often referred to as 'locally advanced gastric cancer'; iv) M1 cancer +/-unresectable M0 cancer referred to as 'far advanced gastric cancer' in the majority of studies on palliative chemotherapy; or v) 'locally advanced gastric cancer', which may implicate 'resectable' cancer when compared with M1 'advanced' cancer (33).
Therefore, it may be difficult to differentiate between resectable advanced stage, unresectable M0 and M1 cancers using the term 'advanced'. It must be understood that 'advanced gastric cancer' may implicate a number of stages of disease and studies must highlight the exact clinical TNM stages for evaluation of the study.