Gastric Cancer (2020) 23:765–779 https://doi.org/10.1007/s10120-020-01086-0 REVIEW ARTICLE Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from a comprehensive literature review and a large cohort study of Caucasian and Asian gastric cancer K. G. P. Kerckhoffs1 · D. H. W. Liu1 · L. Saragoni2 · R. S. van der Post3 · R. Langer4 · M. Bencivenga5 · M. Iglesias6 · G. Gallo7 · L. C. Hewitt1 · G. E. Fazzi1 · A. M. Vos3 · F. Renaud8 · T. Yoshikawa9 · T. Oshima10 · A. Tomezzoli11 · G. de Manzoni5 · T. Arai12 · R. Kushima13 · F. Carneiro14,15,16 · H. I. Grabsch1,17 Received: 23 March 2020 / Accepted: 14 May 2020 / Published online: 2 June 2020 © The Author(s) 2020 Abstract Background The literature on the prognostic relevance of signet-ring cell (SRC) histology in gastric cancer (GC) is contro- versial which is most likely related to inconsistent SRC classification based on haematoxylin–eosin staining. We hypothesised that mucin stains can consistently identify SRC-GC and predict GC patient outcome. Methods We performed a comprehensive literature review on mucin stains in SRC-GC and characterised the mucin expres- sion in 851 Caucasian GC and 410 Asian GC using Alcian Blue (AB)-Periodic Acid-Schiff (PAS), MUC2 (intestinal-type mucin), and MUC5AC (gastric-type mucin). The relationship between mucin expression and histological phenotype [poorly cohesive (PC) including proportion of SRCs, non-poorly cohesive (non-PC), or mucinous (MC)], clinicopathological vari- ables, and patient outcome was analysed. Results Depending on mucin expression and cut-offs, the positivity rates of SRC-GC reported in the literature varied from 6 to 100%. Patients with MUC2 positive SRC-GC or SRC-GC with (gastro)intestinal phenotype had poorest outcome. In our cohort study, PC with ≥ 10% SRCs expressed more frequently MUC2, MUC5AC, and ABPAS (p < 0.001, p = 0.004 and p < 0.001, respectively). Caucasians with AB positive GC or combined ABPAS-MUC2 positive and MUC5AC negative had poorest outcome (all p = 0.002). This association was not seen in Asian patients. Conclusions This is the first study to suggest that mucin stains do not help to differentiate between SRC-GC and non-SRC- GC. However, mucin stains appear to be able to identify GC patients with different outcome. To our surprise, the relationship between outcome and mucin expression seems to differ between Caucasian and Asian GC patients which warrants further investigations. Keywords Gastric cancer · Signet-ring cells · Mucin · Histological phenotype · Survival Introduction Gastric cancer (GC) is a heterogenous disease with respect to epidemiology, morphology, and clinical behaviour. Despite declining incidence, GC remains one of the major K. G. P. Kerckhoffs and D. H. W. Liu have contributed equally to causes of cancer-related death worldwide [1]. The incidence this work. of poorly cohesive GC, including signet-ring cell (SRC) can- cer, appears to be rising in the Western World [2, 3]. Several Electronic supplementary material The online version of this studies investigating the prognostic relevance of SRC histol- article (https​://doi.org/10.1007/s1012​0-020-01086​-0) contains ogy reported conflicting results [4–8]. In some studies, SRC supplementary material, which is available to authorized users. histology was associated with poor outcome, which was not * H. I. Grabsch confirmed in other studies [4–10]. Furthermore, it has been h.grabsch@maastrichtuniversity.nl suggested that the relationship between SRC histology and Extended author information available on the last page of the article outcome may depend on the disease stage in GC patients 13 Vol.:(0123456789) 766 K. G. P. Kerckhoffs et al. [11–17]. The clinical utility of the proportion of SRCs to and ‘signet-ring cell cancer’ (see Online Resource 1). The predict response to preoperative chemo(radio)therapy in GC title and abstract of resulting articles were screened based patients remains a matter of debate [18–24]. on the following inclusion criteria: the abstract or title men- A recent expert panel hypothesised that these inconclu- tioned SRC-GC, and results from histochemical or immu- sive results could be related to inconsistencies in the his- nohistochemical mucin stains were reported separately for tological classification of SRC-GC [25]. Whilst SRC-GC SRC-GC (see Fig. 1). If a study provided both, data from an have always been typed as diffuse-type cancers in the Lauren SRC-GC and non-SRC-GC cohorts, only the SRC-GC data classification [26], the WHO definition of SRC-GC changed were extracted and analysed. We excluded studies reporting several times between the 1st edition in 1977 [27] and the results from less than 10 SRC-GC, referring to hereditary 4th edition in 2010 [28]. Up to the 4th edition [27, 29, 30], diffuse GC, metastases with SRCs or unknown primary when SRC-GC became a subcategory of poorly cohesive tumour, cell culture-based studies, animal studies, diagno- GC, SRC-GC was classified as a separate specific subtype ses based on cytology, and letters to editors containing no of GC. Furthermore, the definition of the extent of SRCs additional information. We also excluded studies where we to qualify as SRC-GC changed over the years from ‘pre- were unable to retrieve the full-text version of the article, dominant’, to more than 50% SRCs in the 2nd edition WHO articles written in languages other than English or Japanese, [29] and back to “predominantly” or “exclusively” in the 4th and articles reporting only on gastro-oesophageal junction and 5th editions WHO [28, 31]. In an attempt to achieve a adenocarcinomas. more consistent classification of SRC-GC, a multidiscipli- From the included studies, we extracted information nary expert panel recently proposed specific cut-off values about the definition of SRC-GC, classification system used, for the percentage of SRCs to distinguish bona-fide SRC- frequencies of positivity for histochemical and immunohis- GC (more than 90% SRCs) from poorly cohesive GC with tochemical mucin stains, differences in mucin expression SRC component (between 10 and 90% SRCs) and poorly between SRC-GC and relationship between mucin expres- cohesive GC not otherwise specified (less than 10% SRCs) sion in SRC-GC, patient outcome, and other clinicopatho- [25]. However, what remains particularly challenging is the logical factors. unequivocal definition of what constitutes a SRC based on routine histology as exemplified by the 5 different types of Gastric cancer cohort study SRCs described in the 3rd edition WHO [30]. Therefore, there remains an urgent clinical need to identify a specific Patients biomarker for SRCs to standardise SRC-GC classification and establish the clinical importance of SRC-GC. We included 851 Caucasian patients from Leeds Teach- We hypothesised that (1) SRC containing gastric cancers ing Hospitals NHS Trust (LTHT), Leeds, UK, and 410 have a different mucin expression compared to non-SRC Asian patients from the Kanagawa Cancer Center Hospital gastric cancers, and (2) there is an association between SRC (KCCH), Yokohama, Japan, diagnosed with adenocarci- mucin expression, clinicopathological variables, and patient noma of the stomach or gastro-oesophageal junction. All outcome. patients underwent potentially curative gastrectomy or gas- The present study consists of two parts: (1) a comprehen- tro-oesophagectomy with lymph node dissection. 76 patients sive literature search to establish the frequency and clinical had neoadjuvant chemotherapy. Clinical variables (age, sex, importance of mucin stains in SRC containing GC, and (2) treatment, overall and 5-year survival, and mortality status) a large cohort study in Asian and Caucasian GC where the and histopathological variables [(y)pT, (y)pN, grade of dif- histological phenotype was classified according to recently ferentiation, and tumour location] were retrieved from hos- published consensus guidelines and the expression of sev- pital records and pathology reports. The study protocol was eral different mucin stains and its relationship to clinico- approved by the relevant local ethics committees. pathological variables, patient outcome, and ethnicity was Tissue microarrays (TMAs) had been previously con- investigated. structed from the resection specimen [32]. Three 0.7 mm cores where sampled from each LTHT GC and two 1.2 mm cores from each KCCH GC. The cores were taken from areas Materials and methods with the highest tumour density in both cohorts. Four micron thick sections were cut and stained with haematoxylin–eosin Literature review (H&E) using a standard protocol and subjected to histo- chemical stains and immunohistochemistry as described A comprehensive literature search was conducted in the below. Slides were scanned at × 40 magnification using an PubMed database including all publications up to October Aperio XT Scanner (Aperio Technologies, Vista, CA, USA) 1st, 2018 using synonyms and MESH terms for ‘gastric’ at the University of Leeds slide scanning facility and viewed 13 Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from… 767 Fig. 1  Flowchart of included papers in literature study. SRC Search Pubmed: signet-ring cell; GC gastric cancer ‘gastric’ AND ‘signet-ring cell carcinoma’ 1696 papers 1419 papers Not fullfilling in- and exclusion criteria 277 papers 193 papers - <10 SRC-GC (181) - No separate data on SRC- GC (11) - Hereditary carcinoma (1) 84 papers 49 papers - Not repor„ng on mucin stains 35 papers Mucin stains in SRC-GC using Aperio ImageScope (version 12.3.2.8013). Scoring characteristics of SRCs in poorly cohesive GCs with looka- of the immunohistochemical stains was performed indepen- likes SRCs in non-poorly cohesive and mucinous GCs, we dently by two observers (D.L. and K.K.), and discrepant decided to quantify the proportion of SRCs in all histo- scores were reviewed and discussed jointly until agreement logical subtypes based on the H&E appearance in 4 cat- was reached. egories: < 10% SRCs, ≥ 10–50% SRCs, ≥ 50–90% SRCs, and ≥ 90% SRCs. This classification was performed blinded Histological classification to (immuno)histochemical expression results or clinico- pathological information. The H&E stain was used to classify GC based on 5th edition WHO Classification of Tumours of the Digestive System [31] in combination with the recently published consen- Mucin histochemistry and immunohistochemistry sus [25]. A poorly cohesive GC was defined as a cancer composed of isolated neoplastic cells or small aggregates All GCs were stained for Alcian Blue (AB)-Periodic Acid- of neoplastic cells [31]. To be classified as a mucinous GC, Schiff (PAS) combined with pan-cytokeratin (CK-ABPAS), more than 50% of tumour area had to be occupied by pools MUC2, and MUC5AC. We chose ABPAS to differentiate of extracellular mucin [31]. GC not fulfilling the criteria for between acidic and neutral mucin. ABPAS stain was com- either poorly cohesive or mucinous GC were classified as bined with immunohistochemistry for pan-cytokeratin to non-poorly cohesive GC. distinguish between mucin within epithelial cells, mucin A classical SRC was defined as a cell with ample opti- within non-epithelial cells such as macrophages and extra- cally clear cytoplasmic mucin on H&E stain and eccentri- cellular mucin. We chose MUC2 as intestinal-type mucin cally placed nucleus. To compare the mucin expression stain and MUC5AC as gastric-type mucin stain, since these 13 768 K. G. P. Kerckhoffs et al. stains were most frequently used in studies in the literature Pearson Chi-square test for categorical variables. For the [33–40]. analysis of the association between histological phenotype After deparaffinisation in xylene and rehydration fol- and mucin expression, results from both cohorts (LTHT lowing a standard protocol, the pan-cytokeratin stain was and KCCH) were combined. Associations between mucin performed first (see below) followed by the ABPAS stain. expression and clinicopathological variables were assessed We used 1% AB solution (pH 2.5), 1% periodic acid solu- using the Kruskal–Wallis test. For survival analyses, indi- tion, and Schiff’s solution following our routine laboratory vidual mucin expression and combinations of mucin expres- protocol. Slides were dehydrated, coverslipped, and scanned. sions (see Online Resource 2) were used. Kaplan–Meier For all antibodies, antigen retrieval was performed in a survival analysis and log-rank test were used to compare microwavable pressure cooker using 10 mM citrate buffer, 5-year and overall survival between patients with different pH 6.0, and full pressure for 5 min. Hydrogen peroxide and histological tumour types or different mucin expression. egg white solution were used to block endogenous per- 16 LTHT patients and 60 KCCH patients underwent neo- oxidase activity and endogenous biotin, respectively. The adjuvant chemotherapy, respectively. To determine whether sections were incubated with antibodies against MUC2 neoadjuvant treatment influenced the results, all analyses (dilution 1:100, clone CCP58; Agilent/Dako), MUC5AC were repeated excluding these 76 patients. As all associa- (dilution 1:100, clone CLH2; Agilent/Dako), and pan- tions remained the same, we report here results based on the cytokeratin (dilution 1:200, clone AE1/AE3; Agilent/Dako) whole cohort. p values < 0.05 were considered statistically for 1 h at 37 °C. Dako Real streptavidin–biotin detection kit significant. was used as detection system and 3,3′-Diaminobenzidine (DAB) as chromogen (Dako) according to the instructions of the manufacturer. Slides were counterstained with Mayer’s haematoxylin, dehydrated, coverslipped, and scanned. Results Scoring of mucin stains Literature review of mucin stains in gastric signet‑ring cell cancers For all three stains (MUC2, MUC5AC, and ABPAS), a core with ≥ 10% stained tumour cells was classified as positive, The literature search in PubMed resulted in only 35 studies irrespective of expression intensity. For this scoring, all published over a period of 40 years (between 1977 and 2017) tumour cells were considered irrespective of their morphol- (see Fig. 1). The median number of SRC-GC patients per ogy (signet-ring cell or not). If one of the cores of a case was study was 37 (range 11–317), and the percentage of SRC-GC classified as being positive, the whole case was classified within individual studies ranged from 7 to 100%. The major- as positive. ity of studies (n = 25, 71%) originated from Asia. ABPAS-positive cores were initially subdivided into 5 categories: (1) AB positive; (2) PAS positive; (3) mixed pos- itivity—same cells AB and PAS positive; (4) mixed positiv- Definition of gastric signet‑ring cell cancers ity—different cells AB and PAS positive; (5) combination and classification system used of categories 3 and 4. A case was classified as positive for both ABPAS, if AB and PAS positivity was seen in the same The definition of SRCs and the classification system used core (category 3–5) or if one core was AB positive (category varied between studies (see Tables 2 and 3). Some authors 1) and another core was PAS positive (category 2). As the provided a very detailed description mentioning intracel- number of cases in certain ABPAS subcategories was very lular accumulation of PAS, AB, and/or mucicarmine posi- small, we subsequently combined groups (see Table 1). tive mucin [41–44] together with an eccentric nucleus as Furthermore, we created additional variables by combin- defining factors [44] or described ‘classical SRCs’ next to ing the results of the different mucin stains (ABPAS, MUC2, small SRCs with deep eosinophilic cytoplasm and a round and MUC5AC) (see Online Resource 2). larger hyperchromatic nucleus [45]. Other authors neither described the definition of SRC nor mentioned the classifica- Statistical analyses tion system used. Some studies used cut-off values to define SRC-GC, either in concordance with the 2nd and 3rd edi- Data were analysed using SPSS Statistics for Windows ver- tions WHO classification (50% rule) [35, 36, 38, 46–51] or sion 25.0 (IBM, Armonk, NY, USA). Patient characteris- using the 4th edition WHO classification [28] (predominant tics, histological tumour types, and mucin expression were or exclusive rule without specific percentage) [41, 42]. One compared between the LTHT and the KCCH cohorts using of the studies used a cut-off value of 90% to define a GC as the Mann–Whitney U test for continuous variables and the SRC-GC [53]. 13 Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from… 769 Table 1  Descriptive statistics of Overall (n = 1261) Caucasian (n = 851) Asian (n = 410) p value the two cohorts n % n % n % Gender (male) 852 68 557 66 295 72 0.026* Mean age (years) 66.6 (SD 11.0) 68.2 (SD 10.8) 63.3 (SD 10.7) < 0.001* T stage < 0.001* T1 131 11 96 12 35 9 T2 146 12 84 11 62 15 T3 307 26 252 33 55 13 T4 600 51 343 44 257 63 N stage < 0.001* N0 359 29 275 32 84 21 N1 264 21 177 21 87 22 N2 296 24 194 23 102 25 N3 340 27 204 24 136 33 Classification 0.641 PC < 10% SRCs 192 15 129 15 63 15 PC > 10% SRCs 67 5 46 5 21 5 Non-PC < 10% SRCs 905 72 608 71 297 72 Non-PC > 10% SRCs 38 3 23 3 15 4 MC < 10% SRCs 14 1 9 1 5 1 MC > 10% SRCs 27 2 22 3 5 1 Non-informative 18 1 14 2 4 1 Mucins MUC2 0.043* Negative 1050 83 698 82 352 86 Positive 173 14 129 15 44 11 MUC5AC < 0.001* Negative 844 67 544 64 300 73 Positive 387 31 289 34 98 24 ABPAS 0.037* Negative 974 77 674 79 300 73 Positive 255 20 159 19 96 23 *AB positive 62 5 47 6 15 4 *PAS positive 37 3 23 3 14 3 *Mixed, same cells 7 0.6 2 0.2 5 1 *Mixed, different cells 4 0.3 4 0.5 0 0 *Mixed, both expressions 145 12 83 10 62 15 n number of cases, SRC signet-ring cell, PC poorly cohesive cancer, non-PC non-poorly cohesive cancer, MC mucinous cancer, SD standard deviation Gastric signet‑ring cell cancer and histochemical mucin subtypes. Kubota et al. [53] used AB, AB-PAS, and stains LNAase stains to classify 64 SRC-GC as type A (imma- ture: PAS weak positive, AB negative, LNAase positive, We identified 12 studies reporting on histochemical mucin small cell size, and high nuclear/cytoplasmic ratio), type stains in SRC-GC published between 1977 and 2013. The B (intermediate: stronger PAS positivity, AB negative or median number of included SRC-GC cases was 28 (range weak positive, LNAase positive, and smaller nuclear/cyto- 11–102). plasmic ratio) or type C (mature: PAS strong positive, AB Different studies used single stain or combination of positive, and eccentric nucleus). The C-type was further stains to classify SRC-GC (see Table 4). divided in C1 and C2 subtypes with the C1 subtype show- Three studies combined morphological features and his- ing weaker AB positivity combined with LNAase positiv- tochemical mucin stains to classify SRC-GC into several ity and the C2 subtype showing stronger AB positivity 13 770 Table 2  Results literature review reporting on individual immunohistochemical mucin stains 13 Author Year Number SRC- Description Classification tumour type used Mucin stains (% ­positivea) Relationship with Comparison with GC SRC-GC survival other tumour types n % MUC2 MUC5AC MUC6 EMA/MUC1 Yes/No Sakamoto H 1997 38 17 No No 66 NA NA 7 Yes Yes Baldus SE 1998 62 48 No WHOb, Lauren 27 NA NA 60/45 Yes Yes Gürbüz Y 2002 31 19 Yes No 81 100 35 52 NR Yes Akyürek N 2002 37 26 No Borrmann, ­WHOc, Lauren 41 NA NA 43 NS Yes Aihara R 2004 69 100 Yes WHOb, JGCA​ 41 100 38 NA Yes No Chu PG 2004 30 100 Yes No 50 50 NA 17 NR No Zhang HK 2004 31 33 No no 81 55 71 NA NS Yes Nguyen MD 2006 21 100 Yes No 48 38 29 24 NR No Sentani K 2008 21 100 Yes No 71 48 NA NA NR No Chiaravalli AM 2009 24 20 Yes Own 60 60 NA NA NR Yes Seki T 2009 35 100 Yes Lauren, Nakamura 60 97 NA NA NR No Ilhan O 2010 19 7 No WHOc, Lauren 95 11 NA NA NS Yes Bartley AN 2011 22 100 Yes WHOb 57 100 NA NA NR No Yonezawa S 2012 69 35 No JGCA​ NA NA NA 6 to ­97e No Yes Konno-Shimizu M 2013 51 43 Yes No NA High NA NA NR No Terada T (II) 2013 30 100 Yes WHOd 13 67 70 10 (MUC1) NR No Terada T (III) 2013 30 100 Yes WHOd NA NA NA 57 (EMA) NR No Nakajima T 2016 35 100 No No 54 94 77 NA No No Fujimoto A 2017 12 12 No JGCA​ 50 75 NA NA NR Yes Kim YH 2017 317f 100 Yes WHOc 36 50 44 NA NR No Xiong ZF 2017 163 100 Yes No 47 74 29 21 Yes No n number of cases, % percentage of all cases, NA not applicable, NR not reported, NS not specific for SRC, SRC signet-ring cell, GC gastric cancer, WHO World Health Organisation, JGCA​ Japanese Gastric Cancer Association a Percentage rounded up b 2nd edition WHO c 3rd edition WHO d 4th edition WHO e Used different antibodies f 92 instead of 317 cases were used for immunohistochemistry K. G. P. Kerckhoffs et al. Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from… 771 Table 3  Results from studies in the literature reporting on mucin phenotypes based on immunohistochemical mucin stains Author Year Number Descrip- Classification Mucin phenotypes (% of total SRC-GCa) Relationship Comparison SRC-GC tion SRC- tumour type with survival with other GC used tumour types n % G I GI/mixed U Yes/No Bamba M 2001 54 100 Yes JGCA​ 28 2 69 NA Yes No Tsukashita S 2003 17 17 No NR 77 0 0 23 NR Yes Aihara R 2004 69 100 Yes WHOb, JGCA​ 60 0 41 0 Yes No Aihara R 2005 69 54 Yes WHOb 47 2 46 5 Yes Yes Ohkura Y 2005 79 28 No No SRC: 81; SRC: 2; SRC: 16; 0 NR No mixed: 35 mixed: 15 mixed: 50 Tian MM 2007 66 100 Yes WHOc 26 15 47 12 Yes No Nakajima T 2016 35 100 No No 0 0 35 0 NR No Xiong ZF 2017 163 100 Yes No 39 18 42 NA Yes No n number of cases, % percentage of all cases, G gastric phenotype, I intestinal phenotype, GI gastrointestinal phenotype, U unclassified, NA not applicable, NR not reported, SRC signet-ring cell, GC gastric cancer, WHO World Health Organisation, JGCA​Japanese Gastric Cancer Associa- tion a Percentages rounded up b 2nd edition WHO c 3rd edition WHO Table 4  Histochemical mucin stains and detection purpose Stains Purpose/specific cell-type References Alcian blue (AB) Identification of acid and neutral mucin; goblet cells [39, 47–49, 51, 52, 55–57] Periodic acid-Schiff (PAS) with or without diastase Identification of acid and neutral mucin [39, 47, 48, 50, 51, 56] High iron diamine (HID) with AB Differentiating sulphomucin from sialomucin [51, 56] Galactose oxidase-Schiff (GOS) Identification of terminal β-galactose and β-N- [56, 58] acetylgalactosamine; gastric surface mucous cells Periodic acid-sodium borohydride-potassium Identification of sialic acid O-acylated side chain; goblet cells [56] hydroxide (PA-SB-PH) in large intestine Mucicarmine Acid mucin; goblet cells [47, 48] Paradoxical Concanavlin A stain (PCS) Detection of stable class III mucin; cardiac glands, mucous [56, 58] neck cells, pyloric glands, and Brunner’s glands LNAase Marker enzyme for small intestine and intestinal metaplasia [55] combined with negative LNAase expression. Akamatsu and lower layer) based on morphology and histochemi- et al. [54] used five histochemical stains (AB-PAS, HID- cal mucin expression. Tatematsu [55, 56] combined his- AB, GOS, PA-SB-PH-PAS, and PCS) to classify 31 SRC- tochemical stains for PCS, GOS, and sialidase GOS with GC into 6 subtypes (1) surface mucous cell-type; (2) immunohistochemical stains for pepsinogen I and II to mucous neck cell-pyloric cell-type; (3) goblet cell (small classify 127 SRC-GC as gastric phenotype, intestinal phe- intestine)-type; (4) goblet cell (large intestine)-type; (5) notype, or mixed gastrointestinal phenotype. The gastric microcyst-type and (6) unclassified. Although different phenotype was the most prevalent subtype (in both studies SRC sub types could be present in the same tumour, types approximately 74%). 1, 2, 3, and 4 were observed most frequently as the domi- However, the above mentioned detailed histomorpho- nant cell type. In addition, the authors described the pres- logical subclassifications of SRC-GC have not been vali- ence of the so-called intramucosal laminated structures dated in any subsequent studies and the relationship with (ILS) in SRC cancer distinguishing three types: complete clinical variables has only been investigated by Akamatsu (upper, middle, and lower layers of the mucosa), incom- [54] who did not find any associations. plete (upper and middle layers), and inverted (middle 13 772 K. G. P. Kerckhoffs et al. Frequency of histochemical mucin positivity in gastric number of SRC-GC was 31 (12–317). The results are sum- signet‑ring cell cancers marised in Tables 2 and 3. We noted a wide variation of cut-off values used to Fujiyoshi et al. [43] categorised SRCs into ‘classical’ classify a cancer as being positive for a particular marker (33.3%) and ‘non-classical’ based on AB positivity, whereas which most likely explains the wide range in reported Bakkelund et al. [44] used PAS positivity as defining feature positivity frequencies. MUC4, STn, and trefoil factor fam- of SRCs, although the amount of PAS positive material var- ily peptide (TFF) 1 and TFF3 were only investigated in a ied between SRCs in the same GC and between GCs. Work single study and reported to be positive in 57% [58], 57% by Terada et al. [41, 42] suggested that SRCs were positive to 71% [59], 33.3% and 100% [60] SRC-GC, respectively. for PAS-D, AB and mucicarmine, whereas Santini et al. [45] The frequency of single mucin stain positive SRC-GC showed that 51% of SRC-GCs were positive for PAS, AB, ranged from 13% [41] to 95% [51] for MUC2, 11% [51] to and high-iron diamine (HID), 12% positive for AB or PAS 100% [35, 49, 61] for MUC5AC, 29% [58] to 71% [62] for and HID, and 37% showed PAS positivity only. Takenoshita MUC6 and 6% [63] to 97% [63] for EMA/MUC1. A single et al. [57] compared PAS and AB positivity between ‘pure’ study by Kim et al. [50] compared the frequency of mucin SRC-GC and tubular or poorly differentiated GC contain- expression between pure SRC-GC and SRC-GC combined ing SRCs. All pure SRC-GC were strongly PAS positive with < 50% tubular or papillary component and did not and 71% were also strongly AB positive. Interestingly, the find any differences for MUC2, MUC5AC, or MUC6. AB and PAS positivity rate was similar in SRC containing Some studies suggested a relationship between mucin tubular or poorly differentiated GCs, and only the expres- expression and depth of invasion. MUC5AC expression sion intensity was found to be lower. Furthermore, SRC-GC was seen in tumour cells in the superficial [33, 35, 49] seemed to have stronger AB expression compared to PAS part of the gastric wall, whereas tumour cells positive for expression (p < 0.01) [57]. MUC6 [35] or PCS III [33] were more frequently found in In summary, the frequency of AB positive SRC-GC var- deeper parts of the wall. The location of MUC2 positive ied from 56% [45] to 64% [58], and PAS positivity varied tumour cells appeared to be more variable and could be from 61% [57] to 95% [45]. One study described HID posi- superficially [49], central/marginal [35], or in a mosaic tivity in 63% of GCs [45]. pattern [33]. Several studies [33–40] investigated combinations Relationship of histochemical mucin positivity of mucin stains in SRC-GC (see Table 3) and described with clinicopathological variables four SRC-GC phenotypes: gastric (G), intestinal (I), gastrointestinal/mixed (GI), and unclassified (UC). The We did not find any study investigating the relationship G-type showed expression of one or more gastric mucin between AB and/or PAS positivity and clinicopathological stains (MUC5AC and/or MUC6 and/or PCSIII and/or variables or patient outcome. M-GGMC-1) and absence of intestinal-type mucin stain A single study combined histochemical mucin stains MUC2; the I-type showed expression of intestinal mucin (PCS, GOS, and sialidase GOS) and immunohistochemical stains and absence of gastric mucin stains; the GI-type stains [pepsinogen II, SH-9 (surface mucous cell stain), and showed a combination of stains and the unclassified type TKH-2 (goblet cells and parietal cell stain)] to define the is negative for any mucin stains (for details, see Online phenotype of 203 SRC-GC as gastric [> 90% surface mucous Resource 3). However, the reported frequency of these cell type or pyloric gland cell type, n = 130 (64%)], intesti- phenotypes in SRC-GC varied: G-type ranged from 0% nal [> 90% goblet cell type or microcyst type, n = 4 (2%)], [39] to 81% [37], I-type from 0% [34, 35, 39] to 18% [40], or mixed gastrointestinal [10–90% gastric and/or intestinal and GI-type 0% [34] to 69% [33]. Aihara et al. [36] found cell types, n = 69 (34%)] [46]. The proportion of gastric phe- a difference in mucin phenotype when comparing SRC-GC notype SRC-GC decreased with increasing depth of inva- with non-SRC-GC (SRC-GC: 41 G-type, 28 GI-type; non- sion, whereas the proportion of mixed phenotype SRC-GC SRC-GC: 20 G-type, 31 GI-type, p = 0.029). increased with depth of invasion. This could suggest that the Seki et al. [64] divided 35 intramucosal SRC-GC progression of SRC-GC may be associated with a pheno- in three groups based on the type of ILS (as described typic shift from gastric to intestinal-type mucin expression. by Akamatsu et al. [54]): complete type only (group A—40%), both complete and incomplete type (group Immunohistochemical stains for mucin expression B—49%), and incomplete type only (group C—11%). of gastric signet‑ring cell cancers MUC2 expression was different (29%, 76%, and 100% of GC in groups A, B, and C, respectively. p = 0.0006). Twenty-six studies using immunohistochemistry were pub- lished between 1997 and 2017, and the median (range) 13 Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from… 773 Relationship between immunohistochemical mucin GC patients; thus, it is not clear whether results in Caucasian stains and clinicopathological variables including patient patients would be similar. outcome All the above motivated us to complement the com- prehensive literature review with a large cohort study on Some studies reported a poorer outcome for patients with more than 1000 patients with gastric or gastro-oesophageal MUC1 positive SRC-GC [48, 60]. GI and I-type SRC-GC cancers. were associated with poorer overall outcome compared to G and UC-type (31.82% vs. 68.75%, p = 0.0146) [38]. No Results from our own gastric association was found between patient outcome and MUC2 and gastro‑oesophageal cancer cohort study or STn positivity [60]. MUC2 positive or GI-type SRC-GC was associated with This study included material from 958 (76%) patients with larger tumour diameter [35, 36, 38], increased depth of GC and 303 (24%) patients with junctional/lower oesopha- invasion [35, 36, 38, 40], presence of lymph-node metasta- geal adenocarcinoma. For clinicopathological and demo- ses [38], or lymphovascular invasion [36, 38, 40]. Patients graphic cohort characteristics, see Table 1. This cohort with G-type SRC-GC had smaller tumours, lower rates of included 851 Caucasian patients (LTHT cohort: 709 gastric lymph-node metastasis, or vascular invasion compared to and 142 junctional/lower oesophageal cancer) and 410 Asian other phenotypes (both p < 0.01) [38]. Furthermore, Xiong patients (KCCH cohort: 249 gastric and 161 junctional/ et al. [40] found that MUC5AC expression was inversely lower oesophageal cancer). associated with depth of invasion. Bamba et al. [33] showed that mucosal tumour size was related to abundance of I-type Frequency of mucin positivity tumour cells, whereas no such relationship was seen for tumour cells in deeper parts of the wall. Using the presence of expression in more than 10% tumour To the best of our knowledge, the relationship between cells as cut-off for all 3 stains (MUC2, MUC5AC, AB/PAS), mucin positivity and response to therapy has not been 670 (53%) cancers were classified as ‘triple negative’. 172 investigated. (14%) cancers were classified as MUC2 positive, 383 (31%) MUC5AC positive, and 253 (20%) AB and/or PAS positive Comparison of immunohistochemical mucin stains (see Table 1). 145 AB and/or PAS positive cancers showed between gastric signet‑ring cell cancer and other a mixed expression with both, the same and different cells histological tumour types being positive for AB and/or PAS. 330 (26%) cancers were positive for only one of the three mucin stains (48 (4%) only The relationship between immunohistochemical mucin MUC2 positive, 217 (17%) only MUC5AC positive, and 65 expression and histological tumour type in GC is still con- (5%) only ABPAS positive). 176 (14%) cancers were posi- troversial. Some studies reported that tubular/papillary/ tive for two mucin stains whereby combined positivity of glandular adenocarcinomas were more frequently MUC1 MUC5AC and ABPAS was most frequently seen (94 (8%) [47, 48, 59, 61, 62], MUC2 [47, 48] and/or MUC5AC cancers). 42 (3%) cancers were triple positive (see also [51] positive compared to SRC-GC. Whereas other studies Online Resource 2). found that SRC-GC were more often MUC2 [51, 59, 60, 65], TFF3 [60], or MUC5AC [61, 65] positive compared to Histological phenotypes and associations with mucin stains other types of GC, Ilhan et al. [51] found no relationship of MUC1 expression and histological tumour type. Zhang [62] 259 (21%) cancers were classified as poorly cohesive, 943 did not find an association between MUC2, MUC5AC, and (75%) as non-poorly cohesive, and 41 (3%) as mucinous. MUC6 expression and histological tumour type, whereas An overview of results from the non-poorly cohesive and other studies showed that MUC2 expression was associated mucinous cancers can be found in Table 1 and Fig. 2. In the with mucinous cancers [51, 61, 62]. group of poorly cohesive cancers, 192 (74%), 36 (14%), 21 In conclusion, the existing number of studies on (8%), and 10 (4%) cancers contained < 10% SRCs, ≥ 10–50% (immuno)histochemical mucin expression in SRC-GC is SRCs, ≥ 50–90% SRCs, and ≥ 90% SRCs, respectively. Sig- limited, and results are controversial most likely related to net-ring cells were seen in mucinous cancers: < 10% SRCs sample size, the use of different mucin stains, or combina- (n = 14, 34%), ≥ 10–50% SRCs (n = 12, 29%), ≥ 50–90% tions of stains and variable cut-offs. We noted differences SRCs (n = 9, 22%), and ≥ 90% SRCs (n = 6, 15%). As in frequencies of mucin positivity in SRC-GCs as well as expected, the number of non-poorly cohesive cancers with differences in the reported association with clinicopatho- cells looking like signet-ring cells was very low, only 38 logical variables including patient outcome. The majority of (3.8%) cancers contained more than 10% lookalike SRCs, published studies were performed using material from Asian 13 774 K. G. P. Kerckhoffs et al. (n = 192, 15%), poorly cohesive cancer with ≥ 10% SRCs (n = 67, 5%), non-poorly cohesive cancer with < 10% SRCs (n = 905, 72%), non-poorly cohesive cancer with ≥ 10% SRCs (n = 38, 3%), mucinous cancer with < 10% SRCs (n = 14, 1%), and mucinous cancer with ≥ 10% SRCs (n = 27, 2%) (see also Table 1). Both, poorly cohesive and non-poorly cohesive cancers with ≥ 10% SRCs, were more frequently MUC2, MUC5AC, or ABPAS positive compared to cancers with < 10% SRCs (see Fig. 2). MUC2 was highly expressed in all mucinous cancers irrespective of the percentage of SRCs. Mucinous cancers with < 10% SRCs were less often ABPAS positive compared to mucinous cancers with ≥ 10% SRCs. Comparison between the LTHT cohort and the KCCH cohort The proportion of cancers with a particular histomorpholog- ical phenotype was similar in the LTHT and KCCH cohorts (see Online Resource 4). In both cohorts, non-poorly cohe- sive cancers were the most common histological subtype [LTHT n = 608 (71%); KCCH n = 297 (72%)]. MUC2- and MUC5AC-positive cancers were more frequent in the LTHT cohort (MUC2 LTHT n = 129 (15%) vs. KCCH n = 44 (11%), p = 0.043; MUC5AC LTHT n = 289 (34%) vs. KCCH n = 98 (24%), p < 0.001), whereas ABPAS-positive cancers were more frequent in the KCCH cohort (KCCH n = 96 (23%) vs. LTHT n = 159 (19%); p = 0.037) (see also Table 1). AB positivity was associated with worse 5-year survival only in the LTHT cohort (p = 0.002) (see Fig. 3). PAS, MUC2, or MUC5AC positivity was not related to outcome in any of the cohorts. Presence of combined ABPAS positiv- ity, MUC2 positivity, and MUC5AC negativity (n = 35, 4%) was related to poorer 5-year survival in the LTHT cohort (p = 0.002) (see Fig. 3 and Online Resource 5). No other associations with patient outcome were found in either cohort. Fig. 2  Mucin expression in association with histological tumour type. a MUC2; b MUC5AC; c ABPAS. For PC and non-PC cases Discussion with ≥ 10%, RCs showed more MUC2 (p < 0.001 and p = 0.003), MUC5AC (p = 0.004 and p < 0.001), and ABPAS (p < 0.001 and There is an on-going debate whether patients with signet- p < 0.001) positivity compared to cases with < 10% SRCs. MC ring cell (SRC) type gastric or gastro-oesophageal cancer showed high MUC2 expression. MC with ≥ 10% SRCs were more have a different prognosis and response to chemotherapy often ABPAS positive compared to MC with < 10% SRCs (p = 0.003). PC poorly cohesive cancer; non-PC non-poorly cohesive cancer; MC [12–18] and, therefore, should be treated differently to mucinous cancer; SRC signet-ring cell patients with other gastric cancer (GC) subtypes. The incon- sistent results reported in the literature could be related to the fact that the histopathological classification of SRC-GC can be challenging. This is due to the variable morphological and none of the non-poorly cohesive cancers contained more appearances of individual SRCs [54–56] which may or may than 50% lookalike SRCs. not be recognised as SRCs by some investigators [54, 54], Because of relatively small number of cases in subgroups, as well as changing criteria in the WHO classification over we compared poorly cohesive cancer with < 10% SRCs the last decades [27–31]. Whilst an expert panel recently 13 Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from… 775 proposed cut-offs to enable separating pure SRC-GC (≥ 90% between mucin expression and patient outcome. Previous SRCs) from GC with an SRC component, a biomarker to studies had either investigated Caucasian cohorts or Asian unequivocal identify SRC-GC would be of potential great cohorts. value to clinicians and patients. Our study provides further evidence that there is no In search of a potential promising biomarker, we decided (immuno)histochemical mucin stain unique to SRC cancer, to focus on mucin-related stains as mucin appeared to be a since (1) a proportion of SRC containing GC was negative defining characteristic of SRCs. Our group was the first to for the (immuno)histochemical mucin stains and (2) a rela- conduct a comprehensive literature review into the reported tively large percentage of GC without SRCs were positive frequency and clinical importance of different mucin stains for one or more mucin stains, similarly to what had been in SRC-GC. To our surprise, the number of studies describ- reported in the literature [47, 48, 51, 59–62, 65]. However, ing (immuno)histochemical mucin expression in SRC-GC this is the first study to suggest that the mucin expression was very limited and results varied between studies making might be related to the quantity of SRCs within a given firm conclusions difficult. However, the existing literature tumour as we saw more frequently mucin expression in seemed to suggest that mucin is not only present in SRC-GC, poorly cohesive GC containing ≥ 10% SRCs. but can also be seen in other types of GC [47, 51, 59, 61, 62]. When directly comparing Caucasian (LTHT) and Asian Furthermore, some SRC-GC appeared to contain no mucin (KCCH) GC cohorts, MUC2 and MUC5AC positivity was [34, 35, 39]. Similarly to the well described morphological more frequent in the LTHT cases, whereas ABPAS positivity heterogeneity of GC, we found evidence that combinations was more frequent in the KCCH cases. This is the first study of different kinds of mucins can be seen in the same GC [33, to report a relationship between AB positivity and outcome 35–40, 45]. Our literature review supports the previous sug- in GC patients. Most interestingly, AB positivity alone or in gestion of an expert panel that inconsistent clinicopathologi- combination with other mucin expressions was only related cal findings can at least partly be explained by differences to poor outcome in Caucasian GC providing further support in the histological haematoxylin–eosin-based classification for the hypothesis that different outcome in Caucasian and of SRC-GC together with different cut-off values used for Asian GC patients may be related to an underlying biological considering a stain positive. Furthermore, even if the same difference [67]. As AB stains acidic mucins which are con- cut-off value was used, results remained contradictory [39, sidered to be present in an intestinal phenotype, our finding 40, 48, 49, 52, 62, 63] which could be related to the use of would support the described association between (gastro) different primary antibodies [64]. When comparing reported intestinal mucin phenotype and unfavourable outcome [38]. results from Asian and Caucasian SRC-GC [34, 39, 41, 49, Due to relatively low patient numbers in mucin-defined sub- 58, 62], we saw similar wide ranges of mucin positivity, sug- groups, we were unable to explore whether the AB expres- gesting that differences might not be simply related to eth- sion related outcome difference between cohorts was related nic origin of the cancers. Sampling of the tumours (luminal to a different disease stage mix. versus centre versus invasive front) could potentially explain Based on the literature review and the results of our different results in mucin expression as it has been reported cohort study, we had to reject our first working hypothesis by several investigators that intramucosal SRC-GC showed and conclude that SRC containing GC do not have a differ- a ‘layered structure’ both morphologically and (immuno) ent mucin expression compared to non-SRC-GC. Further histochemically [33, 35, 49]. Mucin characteristics of SRC- studies are needed to address this clinical need. On the GC varied depending on tumour size and disease stage [46, other hand, results reported in the literature and from our 66], thus results might vary depending on the case selection own cohort study confirmed a relationship between SRC for the study [35, 36, 38, 46]. mucin expression and patient outcome in the surgery alone In order to close the gap in the current literature setting. Furthermore, our cohort study suggests that irre- revealed by our comprehensive review, we decided to spective of histological phenotype, the mucin expression is conduct our own gastric and gastro-oesophageal cohort different in Caucasian and Asian GC patient and is associ- study into mucin stains. Our study is the largest study to ated differently with outcome in different ethnic groups. date where all cancers were re-classified in a standard- We recommend further studies comparing Cauca- ised manner according to WHO classification and recent sian with Asian GC to validate our findings and explore consensus [25]. Furthermore, our study is the first study underlying molecular mechanisms for difference in mucin to include both, Asian and Caucasian patients, enabling expression and outcome. Also, we recommend investigat- us to directly compare patient characteristics, histologi- ing whether the different mucin expression in SRC con- cal tumour types, mucin expression, and relationship taining GC is related to variable treatment response. 13 776 K. G. P. Kerckhoffs et al. Fig. 3  Kaplan–Meier plots showing probability of 5-year survival Example of CK-ABPAS stain showing positivity for AB (blue stain- stratified by AB expression and cohort a and b Kaplan–Meier survival ing). d Example of CK-ABPAS stain showing positivity for PAS and analysis showed in the LTHT cohort significantly worse outcome in combined AB-PAS positivity (pink and purple staining respectively) AB positive GC (p = 0.002), this was not seen in the KCCH cohort. c 13 Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from… 777 Acknowledgements KGP Kerckhoffs and DHW Liu would like to 10. Piessen G, Messager M, Leteurtre E, Jean-Pierre T, Mariette C. thank the British Division of the International Academy of Pathology Signet ring cell histology is an independent predictor of poor for the travel grant and the third poster price of the Leeds Pathology prognosis in gastric adenocarcinoma regardless of tumoral clini- 2019 meeting in Harrogate. cal presentation. Ann Surg. 2009;250(6):878–87. https​://doi. org/10.1097/SLA.0b013​e3181​b21c7​b. Funding No funding. 11. Bamboat ZM, Tang LH, Vinuela E, Kuk D, Gonen M, Shah MA, et al. Stage-stratified prognosis of signet ring cell histology in patients undergoing curative resection for gastric adenocarcinoma. Compliance with ethical standards Ann Surg Oncol. 2014;21(5):1678–85. https​://doi.org/10.1245/ s1043​4-013-3466-8. Conflict of interest The authors declare that they have no conflict of 12. Chon HJ, Hyung WJ, Kim C, Park S, Kim JH, Park CH, et al. interest. Differential prognostic implications of gastric signet ring cell car- cinoma: stage adjusted analysis from a single high-volume center Open Access This article is licensed under a Creative Commons Attri- in asia. Ann Surg. 2017;265(5):946–53. https​://doi.org/10.1097/ bution 4.0 International License, which permits use, sharing, adapta- sla.00000​00000​00179​3. tion, distribution and reproduction in any medium or format, as long 13. Ha TK, An JY, Youn HK, Noh JH, Sohn TS, Kim S. Indication for as you give appropriate credit to the original author(s) and the source, endoscopic mucosal resection in early signet ring cell gastric can- provide a link to the Creative Commons licence, and indicate if changes cer. Ann Surg Oncol. 2008;15(2):508–13. https:​ //doi.org/10.1245/ were made. The images or other third party material in this article are s1043​4-007-9660-9. included in the article’s Creative Commons licence, unless indicated 14. Hyung WJ, Noh SH, Lee JH, Huh JJ, Lah KH, Choi SH, et al. otherwise in a credit line to the material. If material is not included in Early gastric carcinoma with signet ring cell histology. Cancer. the article’s Creative Commons licence and your intended use is not 2002;94(1):78–83. permitted by statutory regulation or exceeds the permitted use, you will 15. Kim DY, Park YK, Joo JK, Ryu SY, Kim YJ, Kim SK, et al. Clin- need to obtain permission directly from the copyright holder. To view a icopathological characteristics of signet ring cell carcinoma of the copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/. stomach. ANZ J Surg. 2004;74(12):1060–4. https:​ //doi.org/10.11 11/j.1445-1433.2004.03268​.x. 16. Otsuji E, Yamaguchi T, Sawai K, Takahashi T. Characteriza- tion of signet ring cell carcinoma of the stomach. J Surg Oncol. 1998;67(4):216–20. References 17. Yokota T, Kunii Y, Teshima S, Yamada Y, Saito T, Kikuchi S, et al. Signet ring cell carcinoma of the stomach: a clinicopatho- 1. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo logical comparison with the other histological types. Tohoku J M, et al. Cancer incidence and mortality worldwide: sources, Exp Med. 1998;186(2):121–30. methods and major patterns in GLOBOCAN 2012. Int J Cancer. 18. Patel VR, Hofstetter WL, Correa AM, Agarwal A, Rashid A, 2015;136(5):E359–E386. https​://doi.org/10.1002/ijc.29210​. Bhutani MS, et al. Signet ring cells in esophageal adenocarci- 2. Wu H, Rusiecki JA, Zhu K, Potter J, Devesa SS. Stomach car- noma predict poor response to preoperative chemoradiation. Ann cinoma incidence patterns in the United States by histologic Thorac Surg. 2014;98(3):1064–71. https:​ //doi.org/10.1016/j.athor​ type and anatomic site. Cancer Epidemiol Biomarkers Prev. acsur​.2014.04.099. 2009;18(7):1945–52. https​://doi.org/10.1158/1055-9965. 19. Charalampakis N, Nogueras Gonzalez GM, Elimova E, Wadhwa 3. Henson DE, Dittus C, Younes M, Nguyen H, Albores-Saavedra R, Shiozaki H, Shimodaira Y, et al. The proportion of signet J. Differential trends in the intestinal and diffuse types of gastric ring cell component in patients with localized gastric adenocar- carcinoma in the United States, 1973–2000: increase in the signet cinoma correlates with the degree of response to pre-operative ring cell type. Arch Pathol Lab Med. 2004;128(7):765–70. chemoradiation. Oncology. 2016;90(5):239–47. https​: //doi. 4. Kim JP, Kim SC, Yang HK. Prognostic significance of signet ring org/10.1159/00044​3506. cell carcinoma of the stomach. Surg Oncol. 1994;3(4):221–7. 20. Chirieac LR, Swisher SG, Correa AM, Ajani JA, Komaki RR, 5. Kunisaki C, Takahashi M, Nagahori Y, Fukushima T, Makino H, Rashid A, et al. Signet-ring cell or mucinous histology after pre- Takagawa R, et al. Risk factors for lymph node metastasis in his- operative chemoradiation and survival in patients with esopha- tologically poorly differentiated type early gastric cancer. Endos- geal or esophagogastric junction adenocarcinoma. Clin Can copy. 2009;41(6):498–503. https​://doi.org/10.1055/s-0029-12147​ Res. 2005;11(6):2229–36. https​://doi.org/10.1158/1078-0432. 58. Ccr-04-1840. 6. Maehara Y, Sakaguchi Y, Moriguchi S, Orita H, Korenaga D, 21. Heger U, Sisic L, Nienhuser H, Blank S, Hinz U, Haag GM, et al. Kohnoe S, et al. Signet ring cell carcinoma of the stomach. Can- Neoadjuvant therapy improves outcomes in locally advanced sig- cer. 1992;69(7):1645–50. net-ring-cell containing esophagogastric adenocarcinomas. Ann 7. Taghavi S, Jayarajan SN, Davey A, Willis AI. Prognos- Surg Oncol. 2018;25(8):2418–27. https​://doi.org/10.1245/s1043​ tic significance of signet ring gastric cancer. J Clin Oncol. 4-018-6541-3. 2012;30(28):3493–8. https​://doi.org/10.1200/jco.2012.42.6635. 22. Heger U, Blank S, Wiecha C, Langer R, Weichert W, Lordick F, 8. Zhang M, Zhu G, Zhang H, Gao H, Xue Y. Clinicopathologic et al. Is preoperative chemotherapy followed by surgery the appro- features of gastric carcinoma with signet ring cell histology. J priate treatment for signet ring cell containing adenocarcinomas Gastrointest Surg. 2010;14(4):601–6. https​://doi.org/10.1007/ of the esophagogastric junction and stomach? Ann Surg Oncol. s1160​5-009-1127-9. 2014;21(5):1739–48. https:​ //doi.org/10.1245/s10434​ -013-3462-z. 9. Li C, Kim S, Lai JF, Hyung WJ, Choi WH, Choi SH, et al. 23. van Hootegem SJM, Smithers BM, Gotley DC, Brosda S, Advanced gastric carcinoma with signet ring cell histology. Thomson IG, Thomas JM, et al. The impact of signet ring cell Oncology. 2007;72(1–2):64–8. https​://doi.org/10.1159/00011​ differentiation on outcome in patients with esophageal and 1096. gastroesophageal junction adenocarcinoma. Ann Surg Oncol. 2019;26(8):2375–84. https​://doi.org/10.1245/s1043​4-019-07322​ -x. 13 778 K. G. P. Kerckhoffs et al. 24. Messager M, Lefevre JH, Pichot-Delahaye V, Souadka A, Piessen K-ras mutation. Genet Mol Res. 2017. https​://doi.org/10.4238/ G, Mariette C, et al. The impact of perioperative chemotherapy on gmr16​02918​1. survival in patients with gastric signet ring cell adenocarcinoma: 41. Terada T. An immunohistochemical study of primary signet-ring a multicenter comparative study. Ann Surg. 2011;254(5):684–93. cell carcinoma of the stomach and colorectum: II. Expression of https​://doi.org/10.1097/SLA.0b013​e3182​35264​7. MUC1, MUC2, MUC5AC, and MUC6 in normal mucosa and in 25. Mariette C, Carneiro F, Grabsch HI, van der Post RS, Allum W, 42 cases. Int J Clin Exp Pathol. 2013;6(4):613–21. de Manzoni G, et al. Consensus on the pathological definition and 42. Terada T. An immunohistochemical study of primary signet-ring classification of poorly cohesive gastric carcinoma. Gastric Can- cell carcinoma of the stomach and colorectum: III. Expressions of cer. 2019;22(1):1–9. https​://doi.org/10.1007/s1012​0-018-0868-0. EMA, CEA, CA19-9, CDX-2, p53, Ki-67 antigen, TTF-1, vimen- 26. Lauren P. The two histological main types of gastric carcinoma: tin, and p63 in normal mucosa and in 42 cases. Int J Clin Exper diffuse and so-called intestinal-type carcinoma an attempt at Pathol. 2013;6(4):630–8. a histo-clinical classification. Acta Pathol Microbiol Scand. 43. Fujiyoshi Y, Eimoto T. Chromogranin A expression correlates 1965;64:31–49. https​://doi.org/10.1111/apm.1965.64.1.31. with tumour cell type and prognosis in signet ring cell carcinoma 27. Oota K, Sobin LH. Histological typing of gastric and oesophageal of the stomach. Histopathology. 2008;52(3):305–13. https​://doi. tumours. 1st ed. WHO International histological classification of org/10.1111/j.1365-2559.2007.02950​.x. tumours, No 18. Geneva. 1977. 44. Bakkelund K, Fossmark R, Nordrum I, Waldum H. Signet ring 28. Bosman FT, Carneiro F, Hruban RH, Theise ND (eds). WHO cells in gastric carcinomas are derived from neuroendocrine cells. classification of tumours of the digestive system. 4th ed. 2010. J Histochem Cytochem. 2006;54(6):615–21. 29. Watanabe H, Jass JR, Sobin LH (eds). Histological typing of 45. Santini D, Bazzocchi F, Mazzoleni G, Ricci M, Viti G, Marrano D oesophageal and gastric tumours. 2nd ed. WHO: International et al (1987) Signet-ring cells in advanced gastric cancer. A clini- histological classification of tumours. Springer-Verlag, Berlin cal, pathological and histochemical study. Acta Pathol Microbiol Heidelberg; 1990. Immunol Scand A 95(5):225–231 30. Hamilton SR, Aaltonen LA, (eds). Pathology and Genetics of 46. Yamachika T, Inada K, Fujimitsu Y, Nakamura S, Yamamura Y, tumours of the digestive system. 3rd ed. WHO classification of Kitou T, et al. Intestinalization of gastric signet ring cell carcino- tumours: IARC press Lyon; 2000. mas with progression. Virchows Arch. 1997;431(2):103–10. 31. WHO Classification of Tumours Editorial Board (eds). Diges- 47. Akyurek N, Akyol G, Dursun A, Yamac D, Gunel N. Expression tive system tumours. 5th ed. Lyon: IARC Press; 2019. of MUC1 and MUC2 mucins in gastric carcinomas: their relation- 32. Hewitt LC, Inam IZ, Saito Y, Yoshikawa T, Quaas A, Hoels- ship with clinicopathologic parameters and prognosis. Pathol Res cher A, et al. Epstein-Barr virus and mismatch repair deficiency Pract. 2002;198(10):665–74. https​://doi.org/10.1078/0344-0338- status differ between oesophageal and gastric cancer: A large 00318​. multi-centre study. Eur J Cancer 2018;94:104–14. https​://doi. 48. Baldus SE, Zirbes TK, Engel S, Hanisch FG, Monig SP, Loren- org/10.1016/j.ejca.2018.02.014. zen J, et al. Correlation of the immunohistochemical reactivity 33. Bamba M, Sugihara H, Kushima R, Okada K, Tsukashita S, of mucin peptide cores MUC1 and MUC2 with the histopatho- Horinouchi M, et al. Time-dependent expression of intestinal logical subtype and prognosis of gastric carcinomas. Int J Cancer. phenotype in signet ring cell carcinomas of the human stomach. 1998;79(2):133–8. Virchows Arch. 2001;438(1):49–56. 49. Bartley AN, Rashid A, Fournier KF, Abraham SC. Neuroendo- 34. Tsukashita S, Kushima R, Bamba M, Nakamura E, Mukai- crine and mucinous differentiation in signet ring cell carcinoma sho K, Sugihara H, et al. Beta-catenin expression in intramu- of the stomach: evidence for a common cell of origin in com- cosal neoplastic lesions of the stomach. Comparative analy- posite tumors. Hum Pathol. 2011;42(10):1420–9. https​://doi. sis of adenoma/dysplasia, adenocarcinoma and signet-ring org/10.1016/j.humpa​th.2010.12.008. cell carcinoma. Oncology. 2003;64(3):251–8. https ​: //doi. 50. Kim YH, Park JH, Park CK, Kim JH, Lee SK, Lee YC, et al. org/10.1159/00006​9310. Histologic purity of signet ring cell carcinoma is a favorable risk 35. Aihara R, Mochiki E, Kamiyama Y, Kamimura H, Asao T, factor for lymph node metastasis in poorly cohesive, submucosa- Kuwano H. Mucin phenotypic expression in early signet ring cell invasive early gastric carcinoma. Gastric Cancer. 2017;20(4):583– carcinoma of the stomach: its relationship with the clinicopatho- 90. https​://doi.org/10.1007/s1012​0-016-0645-x. logic factors. Dig Dis Sci. 2004;49(3):417–24. 51. Ilhan O, Han U, Onal B, Celik SY. Prognostic significance of 36. Aihara R, Mochiki E, Nakabayashi T, Akazawa K, Asao T, MUC1, MUC2 and MUC5AC expressions in gastric carcinoma. Kuwano H. Clinical significance of mucin phenotype, beta-catenin Turk J Gastroenterol. 2010;21(4):345–52. and matrix metalloproteinase 7 in early undifferentiated gastric 52. Chu PG, Weiss LM. Immunohistochemical characterization of carcinoma. Br J Surg. 2005;92(4):454–62. signet-ring cell carcinomas of the stomach, breast, and colon. 37. Ohkura Y, Kawamata H, Chibana Y, Ichikawa K, Fujii S, Tomita Am J Clin Pathol. 2004;121(6):884–92. https​://doi.org/10.1309/ S, et al. Histopathological study of small (%3c2 cm) gastric carci- a09e-rymf-r64n-erdw. noma with signet-ring cell component: significance of the admixed 53. Kubota K, Yamada S, Ito M, Nakamura W, Nagayo T. Cytoplasmic glandular components. Int J Surg Pathol. 2005;13(2):197–203. leucine naphthylamidase activity expressed in signet-ring cell car- https​://doi.org/10.1177/10668​96905​01300​211. cinoma of the stomach. J Natl Cancer Inst. 1977;59(6):1599–604. 38. Tian MM, Zhao AL, Li ZW, Li JY. Phenotypic classification of 54. Akamatsu T, Katsuyama T. Histochemical demonstration of gastric signet ring cell carcinoma and its relationship with clin- mucins in the intramucosal laminated structure of human gastric icopathologic parameters and prognosis. World J Gastroenterol. signet ring cell carcinoma and its relation to submucosal invasion. 2007;13(23):3189–98. Histochem J. 1990;22(8):416–25. 39. Nakajima T, Uehara T, Maruyama Y, Iwaya M, Kobayashi Y, 55. Tatematsu M, Ichinose M, Miki K, Hasegawa R, Kato T, Ito N. Ota H. Distribution of Lgr5-positive cancer cells in intramucosal Gastric and intestinal phenotypic expression of human stomach gastric signet-ring cell carcinoma. Pathol Int. 2016;66(9):518–23. cancers as revealed by pepsinogen immunohistochemistry and https​://doi.org/10.1111/pin.12451​. mucin histochemistry. Acta Pathol Jpn. 1990;40(7):494–504. 40. Xiong ZF, Shi J, Fu ZH, Wan HP, Tu LX. Phenotypic classifica- 56. Tatematsu M, Hasegawa R, Ogawa K, Kato T, Ichinose M, tion of gastric signet ring cell carcinoma and its relationship with Miki K, et al. Histogenesis of human stomach cancers based on 13 Mucin expression in gastric‑ and gastro‑oesophageal signet‑ring cell cancer: results from… 779 assessment of differentiation. J Clin Gastroenterol. 1992;14(Suppl MUC1 cytoplasmic tail domain: use in sensitive identification 1):S1–7. of poorly differentiated cells in adenocarcinoma of the stomach. 57. Takenoshita S, Hashizume T, Katoh R, Koitabashi H, Nagamachi Gastric Cancer. 2012;15(4):370–81. https:​ //doi.org/10.1007/s1012​ Y. Mucin production and subsequent interstitial fibrosis in gastric- 0-011-0125-2. cancer with signet-ring cells. Oncol Rep. 1994;1(3):537–41. 64. Seki T, Ito T, Kawachi H, Sekine M, Funata N, Takizawa T. Rela- 58. Nguyen MD, Plasil B, Wen P, Frankel WL. Mucin profiles in sig- tionship between mucin expression of gastric intramucosal signet net-ring cell carcinoma. Arch Pathol Lab Med. 2006;130(6):799– ring cell carcinoma and its background mucosa. J Med Dent Sci. 804. https​://doi.org/10.1043/1543-2165(2006)130[799:Mpisc​ 2009;56(1):25–35. c]2.0.Co;2. 65. Chiaravalli AM, Klersy C, Tava F, Manca R, Fiocca R, Capella 59. Sakamoto H, Yonezawa S, Utsunomiya T, Tanaka S, Kim YS, Sato C, et al. Lower- and higher-grade subtypes of diffuse gastric can- E. Mucin antigen expression in gastric carcinomas of young and cer. Hum Pathol. 2009;40(11):1591–9. https​://doi.org/10.1016/j. old adults. Hum Pathol. 1997;28(9):1056–65. humpa​th.2009.04.004. 60. Fujimoto A, Ishikawa Y, Ishii T, Yamada A, Igarashi Y, Ohmoto 66. Yoshikawa A, Inada Ki KI, Yamachika T, Shimizu N, Kaminishi Y, et al. Differences between gastric signet-ring cell carcinoma M, Tatematsu M. Phenotypic shift in human differentiated gastric and poorly differentiated adenocarcinoma: a comparison of cancers from gastric to intestinal epithelial cell type during dis- histopathologic features determined by mucin core protein and ease progression. Gastric Cancer. 1998;1(2):134–41. https​://doi. trefoil factor family peptide immunohistochemistry. Pathol Int. org/10.1007/s1012​00050​007. 2017;67(8):398–403. https​://doi.org/10.1111/pin.12559​. 67. Yamada T, Yoshikawa T, Taguri M, Hayashi T, Aoyama T, Sue- 61. Gurbuz Y, Kahlke V, Kloppel G. How do gastric carcinoma classi- Ling HM, et al. The survival difference between gastric cancer fication systems relate to mucin expression patterns? An immuno- patients from the UK and Japan remains after weighted propensity histochemical analysis in a series of advanced gastric carcinomas. score analysis considering all background factors. Gastric Cancer. Virchows Arch. 2002;440(5):505–11. https​://doi.org/10.1007/ 2016;19(2):479–89. https​://doi.org/10.1007/s1012​0-015-0480-5. s0042​8-002-0620-2. 62. Zhang HK, Zhang QM, Zhao TH, Li YY, Yi YF. Expression of Publisher’s Note Springer Nature remains neutral with regard to mucins and E-cadherin in gastric carcinoma and their clinical sig- jurisdictional claims in published maps and institutional affiliations. nificance. World J Gastroenterol. 2004;10(20):3044–7. 63. Yonezawa S, Kitajima S, Higashi M, Osako M, Horinouchi M, Yokoyama S, et al. A novel anti-MUC1 antibody against the Affiliations K. G. P. Kerckhoffs1 · D. H. W. Liu1 · L. Saragoni2 · R. S. van der Post3 · R. Langer4 · M. Bencivenga5 · M. Iglesias6 · G. Gallo7 · L. C. Hewitt1 · G. E. Fazzi1 · A. M. Vos3 · F. Renaud8 · T. Yoshikawa9 · T. Oshima10 · A. Tomezzoli11 · G. de Manzoni5 · T. Arai12 · R. Kushima13 · F. Carneiro14,15,16 · H. I. Grabsch1,17 1 10 Department of Pathology, GROW School for Oncology Department of Gastrointestinal Surgery, Kanagawa Cancer and Developmental Biology, Maastricht University Center, Yokohama, Japan Medical Center+, P. Debyelaan 25, 6229HX Maastricht, 11 Department of Pathology, Verona University Hospital, The Netherlands Verona, Italy 2 Pathology Unit, Morgagni-Pierantoni Hospital, Forlì, Italy 12 Department of Pathology, Tokyo Metropolitan Geriatric 3 Department of Pathology, Radboudumc, Nijmegen, Hospital and Institute of Gerontology, Tokyo, Japan The Netherlands 13 Department of Pathology, Shiga University of Medical 4 Institute of Pathology, University of Bern, Bern, Switzerland Science, Shiga, Japan 5 14 Unit of General and Upper GI Surgery, University of Verona, Institute of Molecular Pathology and Immunology Verona, Italy at the University of Porto (Ipatimup), Porto, Portugal 6 15 Pathology Department, Hospital del Mar, Universitat Instituto de Investigação e Inovação em Saúde (i3S), Autonoma de Barcelona, Barcelona, Spain University of Porto, Porto, Portugal 7 16 Department of Anatomic Pathology, Azienda Pathology Department, Centro Hospitalar de São João Ospedaliero-Universitaria Policlinico di Modena, Modena, and Faculty of Medicine, Porto, Portugal Italy 17 Division of Pathology and Data Analytics, Leeds Institute 8 Department of Pathology, Univ. Lille, CNRS, Inserm, of Medical Research at St. James’s, University of Leeds, CHU Lille, UMR9020 – UMR-S 1277 - Canther – Cancer Leeds, UK Heterogeneity, Plasticity and Resistance to Therapies, Lille, France 9 Department of Gastric Surgery, National Cancer Center Hospital, Tokyo, Japan 13