www.nature.com/modpathol ARTICLE OPEN Results of a worldwide survey on the currently used histopathological diagnostic criteria for invasive lobular breast cancer Maxim De Schepper1,27, Anne Vincent-Salomon 2,27, Matthias Christgen 3,27, Karen Van Baelen1, François Richard 1, Hitoshi Tsuda4, Sasagu Kurozumi5, Maria Jose Brito6, Gabor Cserni 7, Stuart Schnitt 8, Denis Larsimont9, Janina Kulka10, Pedro Luis Fernandez11, Paula Rodríguez-Martínez11, Ana Aula Olivar12, Cristina Melendez12, Mieke Van Bockstal 13, Aniko Kovacs14, Zsuzsanna Varga 15, Jelle Wesseling 16, Rohit Bhargava 17, Pia Boström18, Camille Franchet19, Blessing Zambuko 20, Gustavo Matute21, Sophie Mueller3, Anca Berghian22, Emad Rakha23, Paul J. van Diest 24, Steffi Oesterreich 25, ✉ ✉ Patrick W. B. Derksen 24,28, Giuseppe Floris 26,28 and Christine Desmedt 1,28 © The Author(s) 2022 Invasive lobular carcinoma (ILC) represents the second most common subtype of breast cancer (BC), accounting for up to 15% of all invasive BC. Loss of cell adhesion due to functional inactivation of E-cadherin is the hallmark of ILC. Although the current world 1234567890();,: health organization (WHO) classification for diagnosing ILC requires the recognition of the dispersed or linear non-cohesive growth pattern, it is not mandatory to demonstrate E-cadherin loss by immunohistochemistry (IHC). Recent results of central pathology review of two large randomized clinical trials have demonstrated relative overdiagnosis of ILC, as only ~60% of the locally diagnosed ILCs were confirmed by central pathology. To understand the possible underlying reasons of this discrepancy, we undertook a worldwide survey on the current practice of diagnosing BC as ILC. A survey was drafted by a panel of pathologists and researchers from the European lobular breast cancer consortium (ELBCC) using the online tool SurveyMonkey®. Various parameters such as indications for IHC staining, IHC clones, and IHC staining procedures were questioned. Finally, systematic reporting of non- classical ILC variants were also interrogated. This survey was sent out to pathologists worldwide and circulated from December 14, 2020 until July, 1 2021. The results demonstrate that approximately half of the institutions use E-cadherin expression loss by IHC as an ancillary test to diagnose ILC and that there is a great variability in immunostaining protocols. This might cause different staining results and discordant interpretations. As ILC-specific therapeutic and diagnostic avenues are currently explored in the context of clinical trials, it is of importance to improve standardization of histopathologic diagnosis of ILC diagnosis. Modern Pathology (2022) 35:1812–1820; https://doi.org/10.1038/s41379-022-01135-2 INTRODUCTION types exist, accompanied by their specific clinical characteristics. Breast cancer is one of the most frequently diagnosed cancers Invasive breast carcinoma of no special type (IBC-NST), formerly known worldwide, as one in eight women will get the diagnosis of breast as invasive ductal carcinoma, remains the most frequent subtype cancer during her lifetime in Western countries1. Different histological (~75%), followed by invasive lobular breast cancer (ILC, ~10–15%)2. 1 Laboratory for Translational Breast Cancer Research, Department of Oncology, KU Leuven, Leuven, Belgium. 2Diagnostic and Theranostic Medicine Division, Institut Curie, PSL Research University, Paris, France. 3Institute of Pathology, Hannover Medical School, Hannover, Germany. 4Department of Basic Pathology, National Defense Medical College Hospital, Tokorozawa, Saitama, Japan. 5Department of Breast Surgery, International University of Health and Welfare, Narita, Chiba, Japan. 6Breast Unit, Champalimaud Clinical Center, Champalimaud Foundation, Lisbon, Portugal. 7Department of Pathology, Bács-Kiskun County Teaching Hospital, Kecskemét, Hungary & Department of Pathology, Albert Szent-Györgyi Medical Center, University of Szeged, Szeged, Hungary. 8Brigham and Women’s Hospital, Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA, USA. 9Department of Pathology, Institut Jules Bordet, Université Libre de Bruxelles (ULB), Brussels, Belgium. 10Department of Pathology, Forensic and Insurance Medicine, Semmelweis University, Budapest, Pest, Hungary. 11Hospital German Trias i Pujol, Universidad Autonoma de Barcelona, Barcelona, Spain. 12University Hospital Doctor Josep Trueta, Girona, Spain. 13Department of Pathology, Cliniques universitaires Saint-Luc Bruxelles, Woluwé-Saint-Lambert, Brussels, Belgium. 14Department of Clinical Pathology, Sahlgrenska University Hospital, Gothenburg, Sweden. 15Department of Pathology and Molecular Pathology, University Hospital Zurich, Zurich, Switzerland. 16Divisions of Molecular Pathology and Diagnostic Oncology, Netherlands Cancer Institute, Amsterdam, Netherlands. 17Department of Pathology, UPMC Magee-Womens Hospital, Pittsburgh, PA, USA. 18Department of Pathology, Turku University Hospital and University of Turku, Turku, Finland. 19Institut Claudius Regaud, Institut Universitaire du Cancer Toulouse - Oncopole, Toulouse, France. 20Department of Pathology, Sir Ketumile Masire Teaching Hospital, University of Botswana, Gaborone, Botswana. 21Clínica Universitaria Bolivariana, Universidad Pontificia Bolivariana, Medellín, Colombia. 22Department of Biopathology, Centre Henri Becquerel, Rouen, France. 23Department of Histopathology, Nottingham University Hospital NHS Trust, City Hospital Campus Hucknall Road, Nottingham, UK. 24Department of Pathology, University Medical Center Utrecht, Utrecht, The Netherlands. 25 Women’s Cancer Research Center, UPMC Hillman Cancer Center, Magee-Womens Research Institute, Pittsburgh, PA, USA. 26Department of Pathology, University Hospitals Leuven, UZ Leuven, Leuven, Belgium. 27These authors contributed equally: Maxim De Schepper, Anne Vincent-Salomon, Matthias Christgen. 28These authors jointly supervised this work: Patrick W. B. Derksen, Giuseppe Floris, Christine Desmedt. ✉email: giuseppe.floris@uzleuven.be; christine.desmedt@kuleuven.be Received: 8 April 2022 Revised: 22 June 2022 Accepted: 23 June 2022 Published online: 3 August 2022 M. De Schepper et al. 1813 Most characteristic of ILC is the infiltrative growth of resistance might differ between ILC and IBC-NST5,26,27. An dyscohesive cells in a single file or, the concentric growth of increased frequency of HER2 mutations has for instance been tumor cells around pre-existing breast structures also referred to reported in patients with ILC, especially in the metastatic as “targetoid pattern” (WHO classification of Tumours 5th edition, setting26and preliminary results of prospective trials targeting breast tumors 2019)3. Several other growth patterns and nuclear these mutations have shown a higher benefit for patients with ILC grades of ILC are also recognized, all sharing cell dyscohesion. The than patients with IBC-NST28. underlying molecular feature responsible for this specific appear- ILC is primarily a histopathological diagnosis based on standard ance are the loss of function mutations and a subsequent loss of hematoxylin and eosin staining (H&E). According to the WHO- heterozygosity (LOH) of CDH1, the gene coding for E-cadherin, classification of breast tumors (5th edition)3, immunohistochem- and occasionally epigenetic silencing4,5. E-cadherin is expressed at istry (IHC) is not required for the diagnosis of ILC, as a proportion the cell membrane where it safeguards epithelial cell-cell of these tumors express E-cadherin despite the typical morphol- adhesion6. Loss of E-cadherin protein function thus results in loss ogy of ILC. Nevertheless, many pathologists rely on E-cadherin IHC of cell-cell adhesion with loss of β-catenin membrane expression in routine practice to classify cases as ILC. Although classical ILC, and cytoplasmic translocation of p120-catenin leading to the especially when associated with lobular carcinoma in situ, is a typical morphological appearance of ILC, rendering it a true straightforward diagnosis on H&E slide, different subtypes and morpho-molecular entity in breast cancer7–9. growth patterns exist, making the differential diagnosis with IBC- From a clinical point of view, ILC has dual features, described as NST sometimes challenging7. Moreover, recent central pathology ‘initially indolent but slowly progressive’ by Rakha et al.10. Indeed, reviews of randomized clinical trials (MINDACT and West German ILC can relapse even > 10 years after diagnosis of the primary Plan B) have demonstrated relative overdiagnosis of ILC in local tumor11, largely arising in the post-menopausal setting and pathological laboratories, as only ~60% of the locally diagnosed presenting often as multifocal or bilateral tumors2. In addition to ILCs were confirmed by central pathology review29,30. However, frequent bone and liver metastasis, ILC is associated with metastatic the interobserver agreement analysis was not the primary aim of dissemination in unusual sites such as peritoneum, gastrointestinal- these studies, nor were the diagnostic procedures of the local and urinary tract, leptomeninges, skin, orbit and ovaries12–15. centers captured. Regarding management, ILC is less responsive than IBC-NST tumors When IHC is used, antibodies recognizing E-cadherin are to chemotherapy, with low pathological complete response (pCR) preferred because loss of E-cadherin supports the diagnosis of rates in the neoadjuvant setting16. This may be related to the low ILC, acting as a surrogate marker for the underlying genomic proliferation rate in ILC compared to other, grade matched, types of alteration. However, interpretation of E-cadherin IHC can be breast cancers. Classic lobular carcinoma in situ (LCIS), a non- problematic due to aberrant expression of E-cadherin in some obligate precursor of ILC, is considered as a risk factor for cases of ILC (2–26%, depending on the series and the used developing BC and important differential diagnosis needs to be antibody clones)31–34. This is challenging in diagnostic practice, as made with ductal carcinoma in situ (DCIS). Namely, when finding aberrant expression patterns may be misinterpreted as positive classical LCIS on diagnostic biopsy or in the resection margin of a and, as a result, a diagnosis of IBC-NST could be rendered35, surgical specimen, a more conservative approach can be main- eventually precluding patients from ILC-specific care. Additional tained in contrast to finding DCIS17. Also, correct identification of IHC, using β-catenin or p120-catenin36, can be used as additional non-lobular ILC mimickers like polymorphous carcinoma of the adjuncts to help assess E-cadherin functionality35. Absent or breast is clinically relevant because of the different hormone perinuclear staining of β-catenin and absent or cytoplasmic receptor status and the aggressive clinical behavior18. translocation of p120-catenin, would then support the diagnosis ILC is more difficult to recognize on screening mammograms, of ILC in cases with aberrant E-cadherin staining. ILC cases leading to underestimation of tumor extent, requiring sometimes however will be missed using these conventional tools, in which more extensive surgery or additional surgery2. The detection of rare somatic CDH1 mutations are present that do not affect ILC has dramatically improved due to the introduction of more protein expression but cause attenuation or a lack of trans sensitive modalities such as magnetic resonance imaging (MRI), homotypic interactions. These are the interactions of the contrast enhanced digital mammography and digital breast extracellular E-cadherin domains between adjacent cells, which tomosynthesis in the clinical workup, resulting in more adept is involved in the adherens junctions. Further, the adherens patient management and care19–22. Recognizing ILC as a type of junction may also be compromised by loss of function of other breast cancer is thus relevant not only for the clinical manage- proteins in the complex. Here, detailed IHC supported by ment of the patients, but also for treatment and follow-up sequencing of CDH1 and other adherens junction genes sequen- protocols23. This is especially applicable to the pre-operative cing will be required for the correct diagnosis. Finally, several diagnostic setting, since a diagnosis of ILC should prompt MRI clones exist for the aforementioned antibodies, some E-cadherin which in a high percentage reveals bigger, multifocal and bilateral antibodies also recognize P-cadherin (P.W. Dersken and M. cancers22. Christgen, personal communication), and there are no clear Increasing evidence is available for ILC specific therapeutic recommendations or guidelines for the use in diagnostic practice. avenues. In pre-clinical models, synthetic lethality between ROS1 In summary, there is a discordance between different pathology inhibition in E-cadherin deficient cells has been demonstrated, labs in terms of diagnostic preferences and techniques for ILC, characterized by antitumor effects on E-cadherin deficient breast leading to a clinically relevant problem: under- or overdiagnosis of cancer cells in vivo after the administration of ROS1 inhibitors ILC. The diagnosis of ILC (and LCIS) is of importance as it is (such as crizotinib and foretinib)24. This was the rationale for associated with specific care modalities and future specific recently launched ILC-specific trials in the primary and advanced therapies. In order to understand the variable practices in ILC setting (e.g., Rosaline, ClinicalTrials.gov Identifier: NCT04551495, diagnostics at the histopathological level, we undertook a and ROLO, ClinicalTrials.gov Identifier: NCT03620643). In the BIG1- worldwide survey on the currently favored histopathological 98 Trial, the effect of an adjuvant aromatase inhibitor (letrozole) diagnostic preferences and criteria for the diagnosis of ILC. versus the selective estrogen receptor modulator tamoxifen was investigated in post-menopausal patients with early, hormone receptor positive, HER2-negative breast cancer and compared MATERIALS AND METHODS between ILC and IBC-NST. The authors concluded that the benefit Survey of adjuvant letrozole over tamoxifen was greater in patients with An online survey was drafted by a panel of pathologists and researchers ILC25. Additionally, treatment targets and markers of treatment from the European lobular breast cancer consortium (ELBCC) using Modern Pathology (2022) 35:1812 – 1820 M. De Schepper et al. 1814 SurveyMonkey®. This survey was circulated by email to pathology laboratories all around the world and was open for entries from December Table 1. Overview of continents and countries of the participants. 14, 2020 until July, 1 2021. Pathologists were invited through the pathology network of ELBCC/Lobsterpot and their international connec- Europe 92 tions, through coordinators of cancer hospitals/oncology departments and Belgium 15 through larger cancer organizations. The main goals were to register the reporting of ILC diagnosis, with its variants, the use of ancillary IHC use of France 13 E-cadherin as a diagnostic marker for ILC and the systematic reporting of The Netherlands 12 the ILC subtypes. A total of 42 questions were asked, with emphasis on use Spain 10 of IHC. First, 7 questions were asked regarding the setting and institution the respondent works in. Next, 4 questions were asked about the Germany 8 indication for performing IHC in clinical practice. Finally, we sought to Hungary 8 delineate E-cadherin, β-catenin and p-120 catenin IHC use in more detail. For each antibody, participants were asked about their respective antibody Italy 6 clone, concentration, modality of antigen retrieval and tissue of validation. Albania 3 The list of questions from the survey can be found in the Supplementary Poland 3 Table 1. Portugal 2 Statistical methods England 2 Associations between categorical variables were assessed using Fisher Sweden 2 exact tests. P values were 2-sided and considered as statistically significant Denmark 1 at the conventional level of < 0.05. Correction for multiple testing was not applied given the descriptive nature of the work. The statistical analysis Austria 1 was performed using R 4.0.2. Croatia 1 Latvia 1 RESULTS Malta 1 Demographics and characteristics of the survey participants Finland 1 A total of 153 entries were recorded. 147 unique entries were Switzerland 1 recorded from 34 different countries spread over 6 continents. Ireland 1 Most entries were recorded for Europe (63%, 92/147), followed by Asia (18%, 27/147). The countries with the largest contribution Asia 27 were Japan (12%, 18/147), Belgium (10%, 15/147), France (9%, 13/ Japan 18 147) and The Netherlands (8%, 12/147, Table 1). Vietnam 3 The majority of the pathologists declared working in a Mongolia 3 university hospital (56%, 82/147) or large tertiary hospital (30%, 44/147). The vast majority of the pathologists (110/147, 75%) China 2 worked in a place with an average yearly volume of breast cancers Lebanon 1 above 300, with 80/147 (54%,) even above 500. Most pathologists South America 11 declared being specialized in breast pathology (78%, 115/147). Argentina 5 Only a minority reported working in a small community hospital (5%, 7/147) or a private laboratory (10%, 14/147). However, most Colombia 4 centers reported an average volume of >300 BC samples per year Brazil 2 in respectively 57% (4/7) and 93% (13/14) of the centers. North America 11 There was no significant difference between the continents USA 9 concerning the type of institution (academic, large tertiary, private, p = 0.14, Supplementary Table 2). There was however a Canada 2 significant difference in the reported yearly average breast Africa 4 cancer samples (p = 0.0005), and the number of pathologists Botswana 2 that handle breast cancer specimens per center between the Uganda 1 continents (p = 0.0005). In general, there is a relatively higher proportion of participants from North America handling more DRCongo 1 breast cancer specimens (>500 per year), or working in an Oceania 2 institution with more breast pathologists, in comparison to Australia 2 Europe and Asia (Supplementary Table 2). These results should Grand Total 147 however be interpreted with caution since some continents only had few participants. (increased density, architectural distortion), and often coincides Practice of ILC diagnostics with the classic histological variant. Not all lobular variants are systematically reported by all the Consensus diagnosis, i.e., consulting a colleague/colleagues surveyed pathologists: classic (145/147, 99%), pleomorphic (138/ within the institution for difficult cases, is performed by 97% (143/ 147, 94%), solid (107/147, 73%), histiocytoid (89/147, 61%), 147) of the participants: 58% reported that consensus diagnosis is alveolar (89/147, 61%), trabecular (54/147, 37%), mixed non- regularly done, while 42% reported that consensus diagnosis is classic (53/147, 36%), non-classic (52/147, 35%) and ILC with only sometimes made. (Fig. 1B). extracellular mucin (51/147, 35%), as illustrated in Fig. 1A. Some additional variants that were mentioned are ILC with signet ring Use of IHC for ILC diagnosis cell morphology, ILC with tubular elements and the diffuse type. Indications. The majority of the surveyed pathologists perform The latter one is actually a more a radio-pathological term based IHC for the diagnosis of in situ and invasive lobular neoplasia on the lack of a detectable tumor mass on mammograms (97%, 143/147). About half of them systematically perform IHC for Modern Pathology (2022) 35:1812 – 1820 M. De Schepper et al. 1815 Fig. 1 Current practice of ILC diagnostics. A Reported patterns of ILC. B Use of consensus diagnostics. Fig. 2 Use of IHC for ILC diagnostics. A Indication for using IHC for ILC. B Estimated proportion of ILC cases diagnosed by ancillary IHC. C Proportion of commonly used antibodies for ILC. D Proportion of commonly co-used antibodies for ILC. ILC diagnosis (52%, 75/143), while others only perform staining in Only a minority of the surveyed pathologists uses upfront case of doubt (45%, 64/143) or for differentiating DCIS from LCIS E-cadherin staining (12%, 16/138), i.e., perform IHC before (3%, 4/143). Only 3% (4/147) of the pathologists declared to never examination of the H&E slide. We did not observe any significant use IHC for the diagnosis of ILC (Fig. 2A). association between the use or indication of E-cadherin IHC and When looking at the proportion of cases diagnosed with the type of institution, the volume of breast cancers examined on ancillary IHC, 41% (57/139) reported that > 75% of ILC cases are a yearly basis and the number of pathologists handling breast diagnosed with ancillary IHC while 31% (43/139) reported that this cancer specimens (Tables 2 and 3). is the case in only 0–25% of ILC cases (Fig. 2B). When a lobular growth pattern is observed, but E-cadherin is Concerning the use of IHC, 139/143 (97%) participants use positive, 51% (70/138) opt for additional ancillary IHC (β-catenin antibodies recognizing E-cadherin, 35/143 (24%) β-catenin and and/or p120-catenin). 49/143 (35%) use p120-catenin (Fig. 2C). The majority (50%, 71/ 143) uses only E-cadherin, 13% (19/143) use E-cadherin in Antibodies combination with β-catenin, and 23% (33/143) use E-cadherin E-cadherin: For E-cadherin, 11 different antibody clones were with p120-catenin, while 11% (16/143) use all 3 antibodies reported, of which the NCH-38 was the most frequently used (Fig. 2D). (42%, 38/91), followed by Clone 36 (16%, 15/91) and EP700Y (15%, Modern Pathology (2022) 35:1812 – 1820 M. De Schepper et al. 1816 Table 2. Upfront E-cadherin per type of center. upfront e-cadherin no yes p-value Center of activities Large tertiary hospital 36 (29.5) 5 (31.2) 0.4533 Private laboratory 11 (9.0) 3 (18.8) Small community hospital 6 (4.9) 1 (6.2) University Hospital 69 (56.6) 7 (43.8) Average volume breast cancer samples/year 0–150 samples 5 (4.1) 1 (6.2) 0.7676 151–300 samples 23 (18.9) 4 (25.0) 301–500 samples 26 (21.3) 3 (18.8) >500 samples 68 (55.7) 8 (50.0) Number of pathologist handling breast cancer specimen 1 9 (7.4) 0 (0.0) 0.5802 1–4 61 (50.0) 10 (62.5) 5 or more 52 (42.6) 6 (37.5) Table 3. IHC use per type of center. IHC only in IHC only to Only by Systematically by p-value case of differentiate LCIS morphology morphology doubt from DCIS and IHC Center of activities Large tertiary 17 (26.6) 1 (25.0) 1 (25.0) 25 (33.3) 0.5802 hospital Private laboratory 6 (9.4) 1 (25.0) 0 (0.0) 7 (9.3) Small community 3 (4.7) 1 (25.0) 0 (0.0) 3 (4.0) hospital University Hospital 38 (59.4) 1 (25.0) 3 (75.0) 40 (53.3) Average volume breast 0–150 samples 6 (9.4) 0 (0.0) 1 (25.0) 1 (1.3) 0.1259 cancer samples/year 151–300 samples 9 (14.1) 2 (50.0) 1 (25.0) 17 (22.7) 301–500 samples 16 (25.0) 0 (0.0) 0 (0.0) 14 (18.7) >500 samples 33 (51.6) 2 (50.0) 2 (50.0) 43 (57.3) Number of pathologist 1 5 (7.8) 0 (0.0) 1 (25.0) 4 (5.3) 0.2804 handling breast cancer 1 to 4 30 (46.9) 3 (75.0) 0 (0.0) 41 (54.7) specimen 5 or more 29 (45.3) 1 (25.0) 3 (75.0) 30 (40.0) 14/91), as illustrated in Table 4 (For all reported clones in basic pH (79%, 60/76), most of which with predominantly pH > 8,0 Supplementary table 3). Heterogeneity is reported regarding the (63%, 38/60). There is however a minority using citrate buffer with used concentration per clone, with up to seven different acidic pH (4%, 3/76, Supplementary Table 5). concentrations for NCH-38, ranging from 1/50 to 1/300 and the For the evaluation of the validity of E-cadherin staining, 65% Ready to use (RTU) form. Looking at the distribution of the (91/139) of pathologists use an external control on the slide, different antibody clones per continent, we see that NCH-38 is whereas 36% (48/139) only use the internal control of the frequently reported in all continents, except for Oceania. Clone 36 examined tissue. These external control tissues are highly variable, was not reported in South America and Africa. Clone EP700Y was but almost always contain tissues with an epithelial component. not reported in Asia, North America and Oceania (Supplementary 50% are non-tumorous tissues (46/91) containing normal breast, Table 4). Caution should be used however for these regional liver, colon, appendix, tonsil, kidney, pancreas and skin tissue. 20% differences regarding continents with lower number of respon- (18/91) involve tumor tissue, of which 94% (17/18) contain a dents. When looking deeper into the European continent, the breast carcinoma (IBC-NST and/or ILC). In 27/91 (30%) of the NCH-38 was represented by 12/20 countries, except for France, entries, the exact tissue could not be attributed to one of these which has a representation of 13 respondents. The dominant categories, due to missing or unspecific answers) clone used here was EP700Y. As for the interpretation of the staining, the most acknowl- β-catenin: When considering β-catenin, 4 different antibody edged pattern is the complete absence of staining (52%, 150/290), clones were reported for use in clinical practice, with clone 14 followed by a fragmented membranous pattern (38%, 111/290), and β-catenin-1 being the most frequently used (in respectively perinuclear dot-like pattern (33%, 95/290), and complete but very 53% (20/38) and 32% (12/38)). Similar to E-cadherin, there were weak membranous pattern (27%, 78/290). Only 17% (49/290) of up to 8 different concentrations reported for clone 14, ranging the participants recognize all of these patterns in ILC diagnostics. from 1/75 to 1/300, including 1.25 µg/ml and the RTU form The most frequently used form of antigen retrieval is the heat (Supplementary Table 6A). HIER is again the preferred means of induced epitope retrieval (HIER) (84%, 76/91). The means however antigen retrieval (86%, 31/36), with different modalities of which for HIER are varying, with 6 different modalities of which the automated stainer and hot plate remain the most frequent automated stainer and hot plate are the most frequently used, in modalities (respectively 35% (11/31) and 29% (9/31)). The buffer respectively 37% (28/76) and 26% (20/76) of institutions. For HIER, used again is Tris and or EDTA based with basic pH (77% (24/31), the most frequently used buffers are Tris and/or EDTA based, with Supplementary Table 5). Modern Pathology (2022) 35:1812 – 1820 M. De Schepper et al. 1817 possible explanations for suboptimal consistency in BC classifica- Table 4. Overview of used concentration per most frequently used tion as ILC. As ILC-specific therapeutic avenues are currently being E-cadherin clone. explored, some of which are already being evaluated in clinical Concentration per clone Number of participants trials, it is of importance to further improve the standardization of NCH-38 38 (42%) the histopathologic diagnosis of ILC. Moreover, because of its peculiar biology, patients with ILC require adapted therapeutic RTU 12 (13%) decision and disease monitoring strategies2,23. 1/100 10 (11%) One of the biggest challenges is establishing a clear definition 1/50 4 (4%) or gold standard of ILC. The goal of the definition should be the 1/200 3 (3%) identification of patients that have a tumor with a distinct biological behavior. In current clinical practice this is proxied by 1/300 1 (1%) histopathological findings, in essence based on morphology, 1/50 to 1/100 1 (1%) aided by IHC and sometimes molecular findings. We do not know 1/170 1 (1%) however how cases with ambiguous features behave biologically 1/25 1 (1%) (e.g., lobular morphology with preserved E-cadherin-catenin complex, or IBC-NST morphology with canonical CDH1 mutations). Missing data 5 (5%) Two participants reported that, despite preserved E-cadherin Clone 36 15 (16%) staining, a case would still be considered ILC if it has the RTU 8 (9%) morphological appearance of an ILC. This is an important question 0,314 µg/ml 2 (2%) that was however not systematically asked to the pathologists in this survey. It is thus also of utmost importance to investigate 1/200 1 (1%) which features are indicative and/or decisive for recognizing a Missing data 4 (4%) tumor as ILC. EP700Y 14 (15%) In the WHO classification of breast cancer (3), several growth 1/200 4 (4%) patterns or cellular variants are described with an emphasis on classic ILC and pleomorphic ILC. The most commonly mentioned RTU 5 (5%) in a descriptive manner are the solid, the alveolar, the apocrine or 0,314 µg/ml 1 (1%) histiocytoid variant, and the ILC with signet ring cell features. A 1/700 1 (1%) tubulolobular variant is mentioned, which is most likely not a unknown 1 (1%) variant of ILC, not to be confused with an ILC with tubular elements37. We know however that there are more patterns Missing data 1 (1%) described in literature and also recognized in this survey. The Clone 36B5 7 (8%) importance of this subtyping has been demonstrated in several RTU 3 (3%) studies. First, a prognostic significance has been linked to these 1/100 1 (1%) variants with a favorable prognosis of the pure classic subtype in comparison to non-classical subtypes11. Second, it has been 1/40 1 (1%) demonstrated that these subtypes or variants display a different Missing data 2 (2%) molecular landscape. The mixed non-classic subtypes are enriched for ERBB2 and TP53 mutations, and frequently present with 1p36.22 (ARID1A) deletions. The solid subtype is characterized by P120-catenin: Twenty-nine entries were recorded for specifica- more ARID1A mutations, 1p36.22 (ARID1A) deletions and by 11p tions regarding p120-catenin. 4 different clones were reported, of and 6q25.1 (ESR1) gains. Last, the alveolar subtype is characterized which clone 98 and EP66 are the most frequently used products by gains in 11q13.3 (CCND1) and 11q14 (PAK1)27. Third, it has (respectively 48%, 14/29, and 24%, 7/29). For the most used Clone become clear that some of these subtypes that are not included in 98, again 7 different concentrations were reported ranging from the WHO classification, are a source of misdiagnosis of ILC, such as 1/150 up to 1/500 with inclusion of 0.076 µg/ml and the RTU form the trabecular subtype or ILC with tubular elements37,38. (Supplementary Table 6B). Antigen retrieval was again performed Remarkably, the majority of pathologists take note of the different preferably by HIER (79%, 23/29), by means of automated stainer variants and growth types of ILC, despite most of them are not (48%, 11/23) and hot plate (26%, 6/23), with basic Tris and/or considered for clinical decision making. EDTA-based buffer (83%, 19/23, Supplementary Table 5). Our results demonstrate that half of the institutions system- atically perform ancillary IHC, mainly E-cadherin based, in order to Other used IHC antibodies: Pathologists were given the option support the morphological diagnosis of ILC, a practice which is to enter antibodies other than the ones offered in the however not mandatory based on the current WHO essential questionnaire, and seven responded. Three different antibodies diagnostic criteria3. Interestingly, the diagnostic accuracy of ILC were given: Cytokeratin 34βE12 (5/7), Catenin delta 1 (D7S2M) between diagnosis based on H&E only and H&E supported with (1/7), and P-cadherin (clone 56) (1/7, Supplementary Table 6C). E-cadherin IHC has been recently investigated in a large international study. Here, the investigators demonstrated a Sequencing for CDH1. In case of normal or discordant features significantly increased inter-observer agreement when patholo- between histological findings and immunohistochemistry, we gists were provided with additional E-cadherin IHC, of note, when requested whether sequencing for the E-cadherin encoding gene the antibody clone and staining procedure was equal for all CDH1 is performed for the definitive diagnosis. Only 4/120 reviewers38. pathologists replied that they rarely perform CDH1 DNA sequen- There is however a great variability in E-cadherin antibody cing (in less than 5% of ILC cases). clones used as well as their staining procedures (concentration, antigen retrieval modalities, validation), which might result in differences in staining results and their interpretation. There DISCUSSION seems to be little heterogeneity between the continents, with We report the results of the first worldwide survey concerning the most antibody clones for E-cadherin being represented in most diagnosis of ILC in pathology practice, in order to explore the continents, with subtle differences between continents Modern Pathology (2022) 35:1812 – 1820 M. De Schepper et al. 1818 concerning presence of a single antibody clone or the reported diagnostic medical devices regulation (IVDR). The IVDR is already frequency (Supplementary Table 4). going into effect by 26 May 2022, and foresees a conditional The most straightforward pattern of E-cadherin staining, is the transition phase until 26 May 2024, when all medical devices are absence of staining. However, different aberrant patterns are required to be compliant to the IVDR. Not all manufacturers have described and used for E-cadherin in clinical practice: cytoplasmic; obtained a CE-label yet for the reported antibody clones, and it is dot-like, perinuclear Golgi-type pattern; fragmented, focal, or unsure if all of them will pursue the acquisition of it. Laboratories beaded membranous expression; a complete, but weak membra- would be still be able to use a product without CE-IVD label or a nous E-cadherin staining33. These patterns however could be CE-IVD product with deviation of the provided protocol. In these dependent on the antibody clone and concentration that is used. cases however, laboratories would need to provide an extensive Adding the extra variable of different antigen retrieval modalities, validation report of this product and/or protocol (laboratory results in a complex situation allowing high variability of staining developed test (LDT)). Given the heterogeneity of staining results and subsequent interpretation. Also, it has become clear procedures reported in this survey, it is also unsure whether that false positive E-cadherin staining could be observed, as there laboratories will pursue the validation of an LDT or comply to the exist different somatic mutations in CDH1 (frameshift mutation, CE-IVD protocol. The antibody market is thus highly likely to extracellular truncating mutation, extracellular missense mutation, change in the near future. It is however very important to keep the truncating mutation)38, and as each antibody may have its own issues mentioned in this manuscript in mind for the acquisition of specificity for detecting these mutations. In these cases, although CE-label. partial membranous E-cadherin staining may be observed, the To conclude, this is the first study to our knowledge to monitor protein is most probably non-functional. The exact staining the diagnostic practices for ILC across the world. We have pattern, sensitivities and specificities have not been systematically demonstrated possible explanations for the reported discordances investigated, but are probably of importance. in ILC diagnosis such as guidelines open for variable interpreta- Concerning the validation of E-cadherin, healthy tissues are tions for diagnosis of ILC and heterogeneity in antibody staining mostly used for the presence of a membranous staining in procedures, particularly for E-cadherin. Standardization of IHC epithelial structures, however the diagnostic performance of the staining procedures as well as their interpretation is thus antibody (i.e. being able to reliably distinguish a case between warranted. Pathologists active in breast pathology need to be IBC-NST and ILC) is only validated by a minority of the participants aware of these existing pitfalls, and discordances between (11%, 10/91) and not systematically reported for all antibodies (E- morphology (H&E) and IHC are a strong indication for false cadherin, p120-catenin and β-catenin). When interrogating for positive E-cadherin staining or the presence of a less known ILC instance the Nordic immunohistochemical Quality Control (Nor- variant. Initiatives are ongoing within the ELBCC to further diQC, 9000 Aalborg, Denmark) reports of E-cadherin assessment, investigate the staining patterns of the most frequently reported Clone EP700Y (the third most frequent clone used in this survey) E-cadherin clones in this survey according to different types of underperformed because of strong background staining in normal mutations in CDH1, as a further step in streamlining diagnostics, tissues and non-specific staining in plasma cells in the lamina clinical research and patient specific care for ILC. propria of the colon39. However, it is mentioned that the staining still allowed pathologists to distinguish ILC reliably from IBC-NST on the given Tissue Microarray. NordiQC reported that one of the DATA AVAILABILITY causes of insufficient staining for E-cadherin is the use of an acidic Data are available as Supplementary Tables 1-6: Supplementary Table 1: survey questionnaire. Supplementary Table 2: characteristics of participants per continent. buffer, which is reported by a minority of our participants (4%, 3/ Supplementary Table 3: Overview of used concentration per used E-cadherin clone. 76)39. It is however not reported if the diagnostic performance of Supplementary Table 4: E-cadherin antibody clone per continent. Supplementary differentiating ILC from IBC-NST was still preserved under these Table 5: staining procedures per antibody. Supplementary Table 6: concentration per conditions. reported clone. The added value of systematic concomitant catenin (p120- or β- catenin) staining or catenin staining alone without E-cadherin is not really recommended, as loss of membranous staining of REFERENCES catenins can be observed in IBC-NST40 and in non-breast 1. Ghoncheh, M., Pournamdar, Z. & Salehiniya, H. Incidence and mortality and malignancies41. The latter may reflect a state of dedifferentiation epidemiology of breast cancer in the world. Asian Pac J Cancer Prev 17, 43–46 or epithelial to mesenchymal transition, rather than a pathogno- (2016). monic mutation in CDH1. 2. Christgen, M., Steinemann, D., Kühnle, E., Länger, F., Gluz, O., Harbeck, N. et al. Given the large number of participating countries and different Lobular breast cancer: Clinical, molecular and morphological characteristics. Pathol Res Pract 212, 583–597 (2016). continents, we believe that the present survey provides a rather 3. WHO classification of Tumours Editorial Board. Breast tumours. fair representation of ILC diagnostic practices worldwide. Of note, 4. Desmedt, C., Zoppoli, G., Gundem, G., Pruneri, G., Larsimont, D., Fornili, M. et al. there might be a bias given the relative overrepresentation of Genomic Characterization of Primary Invasive Lobular Breast Cancer. J Clin Oncol European countries and overrepresentation of Japan within the 34, 1872–1880 (2016). Asian continent, probably resulting in significant differences 5. Ciriello, G., Gatza, M. L., Beck, A. H., Wilkerson, M. D., Rhie, S. K., Pastore, A. et al. observed in the characteristics of the participating centers Comprehensive Molecular Portraits of Invasive Lobular Breast Cancer. Cell 163, (Supplementary Table 2). North American countries are relatively 506–519 (2015). less well represented in terms of numbers and the respondents 6. Meng, W. & Takeichi, M. Adherens junction: molecular architecture and regula- mainly belong to large centers with high breast cancer sample tion. Cold Spring Harbor Perspect Biol vol. 1 (2009). 7. Christgen, M., Cserni, G., Floris, G., Marchio, C., Djerroudi, L., Kreipe, H. et al. volumes (Supplementary Table 2). Nevertheless, indications for IHC Lobular Breast Cancer: Histomorphology and Different Concepts of a Special staining do not seem to differ in the different types of centers Spectrum of Tumors. Cancers vol. 13 (2021). (Tables 2 and 3). Moreover, the gradual increase of missing data 8. Derksen, P. W. B., Liu, X., Saridin, F., van der Gulden, H., Zevenhoven, J., Evers, B. along the course of the survey, as exemplified by the smaller et al. Somatic inactivation of E-cadherin and p53 in mice leads to metastatic number of pathologists answering the questions related to the use lobular mammary carcinoma through induction of anoikis resistance and of IHC for ILC (i.e. the last questions in the survey; figure2), may angiogenesis. Cancer Cell 10, 437–449 (2006). render these entries less reliable in terms of data interpretation. 9. Schackmann, R. C. J., van Amersfoort, M., Haarhuis, J. H. I., Vlug, E. J., Halim, V. A., Within the European Union, all commercially available anti- Roodhart, J. M. L. et al. Cytosolic p120-catenin regulates growth of metastatic lobular bodies for clinical use will need to be compliant with the in vitro carcinoma through Rock1-mediated anoikis resistance. J Clin Invest 121, (2011). Modern Pathology (2022) 35:1812 – 1820 M. De Schepper et al. 1819 10. Rakha, E. A., El-Sayed, M. E., Powe, D. G., Green, A. R., Habashy, H., Grainge, M. J. 34. Lobo, J., Petronilho, S., Newell, A. H., Coach, J., Harlow, G., Cruz, A. et al. E-cadherin et al. Invasive lobular carcinoma of the breast: Response to hormonal therapy and clone 36 nuclear staining dictates adverse disease outcome in lobular breast outcomes. Eur J Cancer 44, 73–83 (2008). cancer patients. Mod Pathol 32, 1574–1586 (2019). 11. Pestalozzi, B. C., Zahrieh, D., Mallon, E., Gusterson, B. A., Price, K. N., Gelber, R. D. 35. Canas-Marques, R. & Schnitt, S. J. E-cadherin immunohistochemistry in breast et al. Distinct clinical and prognostic features of infiltrating lobular carcinoma of pathology: Uses and pitfalls. Histopathology vol. 68 57–69 (2016). the breast: combined results of 15 International Breast Cancer Study Group 36. Dabbs, D. J., Bhargava, R. & Chivukula, M. Lobular Versus Ductal Breast Neo- clinical trials. J Clin Oncol 26, 3006–3014 (2008). plasms: The Diagnostic Utility of P120 Catenin. Am J Surg Pathol 31, (2007). 12. Mathew, A., Rajagopal, P. S., Villgran, V., Sandhu, G. S., Jankowitz, R. C., Jacob, M. 37. Christgen, M., Bartels, S., van Luttikhuizen, J. L., Bublitz, J., Rieger, L. U., Christgen, et al. Distinct Pattern of Metastases in Patients with Invasive Lobular Carcinoma H. et al. E-cadherin to P-cadherin switching in lobular breast cancer with tubular of the Breast. Geburtshilfe Frauenheilkd 77, 660–666 (2017). elements. Mod Pathol 33, 2483–2498 (2020). 13. Arpino, G., Bardou, V. J., Clark, G. M. & Elledge, R. M. Infiltrating lobular carcinoma of the 38. Christgen, M., Kandt, L. D., Antonopoulos, W., Bartels, S., van Bockstal, M. R., Bredt, breast: tumor characteristics and clinical outcome. Breast Cancer Res 6, R149 (2004). M. et al. Inter-observer agreement for the histological diagnosis of invasive 14. Raap, M., Antonopoulos, W., Dämmrich, M., Christgen, H., Steinmann, D., Länger, lobular breast carcinoma. J Pathol Clin Res 8, 191–205 (2022). F. et al. High frequency of lobular breast cancer in distant metastases to the orbit. 39. NordiQC Assessment Run 53 (2018) E-Cadherin (ECAD). (2018). Cancer Med 4, 104–111 (2015). 40. Schonborn, I., Zschiesche, W., Behrens, J., Herrenknecht, K. & Birchmeier, W. 15. Blohmer, M., Zhu, L., Atkinson, J. M., Beriwal, S., Rodríguez-López, J. L., Rosenz- Expression of E-cadherin/catenin complexes in breast cancer. Int J Oncol 11, weig, M. et al. Patient treatment and outcome after breast cancer orbital and 1327–1334 (1997). periorbital metastases: a comprehensive case series including analysis of lobular 41. Yasui, W., Kuniyasu, H., Akama, Y., Kitahara, K., Nagafuchi, A., Ishihara, S. et al. versus ductal tumor histology. Breast Cancer Res 22, 70 (2020). Expression of e-cadherin, alpha-catenins and Beta-catenins in human gastric 16. O’Connor, D. J., Davey, M. G., Barkley, L. R. & Kerin, M. J. Differences in sensitivity carcinomas - correlation with histology and tumor progression. Oncol Rep 2, to neoadjuvant chemotherapy among invasive lobular and ductal carcinoma of 111–117 (1995). the breast and implications on surgery–A systematic review and meta-analysis. Breast 61, 1–10 (2022). 17. Wen, H. Y. & Brogi, E. Lobular Carcinoma In Situ. Surgical Pathology Clinics 11, 123–145 (2018). ACKNOWLEDGEMENTS 18. Asioli, S., Marucci, G., Ficarra, G., Stephens, M., Foschini, M. P., Ellis, I. O. et al. The authors wish to thank all the pathologists who consented of being acknowl- Polymorphous adenocarcinoma of the breast. Report of three cases. Virchows edged: Gabriela Acosta Haab, Georges Aftimos, Wiebke Antonopoulos, Laurent Arch 448, 29–34 (2006). Arnould, Leart Berdica, Laia Bernet Vegué, Rohit Bhargava, Peter Bult, Benjamin 19. Heidinger, O., Heidrich, J., Batzler, W. U., Krieg, V., Weigel, S., Heindel, W. et al. Calhoun, Marie-Pierre Chenard, Cecile Colpaert, Alicia Cordoba, Franceska Dedeur- Digital mammography screening in Germany: Impact of age and histological waerdere, James Degaetano, Raphaëlle Duprez, Gelareh Farshid, Pedro L. Fernandez, subtype on program sensitivity. Breast 24, 191–196 (2015). Bolorerdene Gantumur, Nina Helidon, Akira I. Hida, Naoko Honma, Rie Horii, Endre 20. Kirsh, V. A., Chiarelli, A. M., Edwards, S. A., O’Malley, F. P., Shumak, R. S., Yaffe, M. J. et al. Kalman, Naoki Kanomata, Ayaka Katayama, Yuka Katsurada, Tomonori Kawasaki, Tumor Characteristics Associated With Mammographic Detection of Breast Cancer in Anikó Kovács, Glen Kristiansen, Hajime Kuroda, Masafumi Kurosumi, Konrad Kurowski, the Ontario Breast Screening Program. J Natl Cancer Inst 103, 942–950 (2011). Inta Liepniece-Karele, Patricia Lopez Correa, Robert Lukande, Gaetan Macgrogan, 21. Nederend, J., Duijm, L. E. M., Voogd, A. C., Groenewoud, J. H., Jansen, F. H. & Aurélie Maran-Gonzalez, Takuya Moriya, Anupma Nayak, Yasuyo Ohi, Tomo Osako, Louwman, M. W. J. Trends in incidence and detection of advanced breast cancer Tetsunari Oyama, Alberto Ravarino, Peter Regitnig, Elisabeth Russ, Takashi Sakatani, at biennial screening mammography in The Netherlands: a population based Sandra Sarancone, Anne-Marie Schelfhout, Marcia Silveira- Graudenz, Luz F. Sua study. Breast Cancer Res 14, R10 (2012). Villegas, Miklos Torok, Joost Van Gorp, Zsuzsanna Varga, Andras Vörös, Willem Vreuks, 22. Amin, A. L., Sack, S., Larson, K. E., Winblad, O., Balanoff, C. R., Nazir, N. et al. Does Clive Wells, Pieter Westenend, Suzanne Wilhelmus, Rin Yamaguchi. the Addition of Breast MRI Add Value to the Diagnostic Workup of Invasive Lobular Carcinoma? J Surg Res 257, 144–152 (2021). 23. Mouabbi, J. A., Hassan, A., Lim, B., Hortobagyi, G. N., Tripathy, D. & Layman, R. M. Invasive lobular carcinoma: an understudied emergent subtype of breast cancer. Breast Cancer Res Treat 193, 253–264 (2022). AUTHOR CONTRIBUTIONS 24. Bajrami, I., Marlow, R., van de Ven, M., Brough, R., Pemberton, H. N., Frankum, J. M.D.S., M.C., A.S., P.D., G.F., and C.D. conceived the study and wrote the manuscript. et al. E-Cadherin/ROS1 Inhibitor Synthetic Lethality in Breast Cancer. Cancer All authors participated to the interpretation of the survey results and agreed with Discov 8, 498–515 (2018). the final version of the manuscript. 25. Metzger Filho, O., Giobbie-Hurder, A., Mallon, E., Gusterson, B., Viale, G., Winer, E. P. et al. Relative Effectiveness of Letrozole Compared With Tamoxifen for Patients With Lobular Carcinoma in the BIG 1-98 Trial. J Clin Oncol 33, 2772–2779 (2015). 26. Richard, F., Majjaj, S., Venet, D., Rothe, F., Pingitore, J., Boeckx, B. et al. Char- FUNDING acterization of stromal tumor-infiltrating lymphocytes and genomic alterations in This publication is partly based upon work from COST action LOBSTERPOT (CA19138), metastatic lobular breast cancer. Clin Cancer Res 26, 6254–6265 (2020). supported by COST (European Cooperation in Science and Technology). KVB and 27. Desmedt, C., Zoppoli, G., Gundem, G., Pruneri, G., Larsimont, D., Fornili, M. et al. MDS are funded by the KU Leuven Fund Nadine de Beauffort. Genomic Characterization of Primary Invasive Lobular Breast Cancer. J Clin Oncol (2016). 28. Ma, C. X., Luo, J., Freedman, R. A., Pluard, T., Nangia, J., Lu, J. et al. Abstract CT026: A phase II trial of neratinib (NER) or NER plus fulvestrant (FUL) (N+F) in HER2 mutant, non-amplified (HER2mut) metastatic breast cancer (MBC): Part II of COMPETING INTERESTS H.T.: recipient of scholarship donation from Chugai, Takeda, and Eli Lily and of MutHER. Cancer Res 81, CT026–CT026 (2021). research grant from Roche Diagnostics. S.K.: has received honoraria from Daiichi 29. Metzger, O., Cardoso, F., Poncet, C., Desmedt, C., Linn, S., Wesseling, J. et al. 6 Oral Sankyo co. ltd, Taiho Pharmaceutical co. ltd, Eli Lilly and Company, MSD K.K., - Clinical utility of MammaPrint testing in Invasive Lobular Carcinoma: Results AstraZeneca K.K., Chugai Pharmaceutical, Ltd., Dinow Inc., and Novartis Japan.A. from the MINDACT phase III trial. Eur J Cancer 138, S5–S6 (2020). Kovacs: received honoraria from Pfizer and R.J.W.: receives funding from KWF Dutch 30. Christgen, M., Gluz, O., Harbeck, N., Kates, R. E., Raap, M., Christgen, H. et al. Cancer Society, Cancer Research UK, Antoni van Leeuwenhoek Investment Fund, Differential impact of prognostic parameters in hormone receptor-positive lob- ZonMw Health Care Innovation funding from the Dutch government. G.F.: recipient ular breast cancer. Cancer 126, 4847–4858 (2020). of a post-doctoral fellowship sponsored by the KOOR from University Hospitals 31. Silva, L. da, Parry, S., Reid, L., Keith, P., Waddell, N., Kossai, M. et al. Aberrant Leuven. P.D.: received funding from Pfizer, ZonMw Health Care Innovation from the Expression of E-cadherin in Lobular Carcinomas of the Breast. Am J Surg Pathol Dutch government, University Utrecht. 32, (2008). 32. Rakha, E. A., Patel, A., Powe, D. G., Benhasouna, A., Green, A. R., Lambros, M. B. et al. Clinical and Biological Significance of E-cadherin Protein Expression in Invasive Lobular Carcinoma of the Breast. Am J Surg Pathol 34, (2010). ETHICS APPROVAL 33. Acs, G., Lawton, T. J., Rebbeck, T. R., LiVolsi, V. A. & Zhang, P. J. Differential No ethical approval was necessary for the set-up of this study. No patient data was expression of E-cadherin in lobular and ductal neoplasms of the breast and its recorded nor requested. Participants voluntarily consented to participate and to biologic and diagnostic implications. Am J Clin Pathol 115, 85–98 (2001). share contact information if they wanted to. Modern Pathology (2022) 35:1812 – 1820 M. De Schepper et al. 1820 ADDITIONAL INFORMATION Open Access This article is licensed under a Creative Commons Supplementary information The online version contains supplementary material Attribution 4.0 International License, which permits use, sharing, available at https://doi.org/10.1038/s41379-022-01135-2. adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Correspondence and requests for materials should be addressed to Giuseppe Floris Commons license, and indicate if changes were made. The images or other third party or Christine Desmedt. material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the Reprints and permission information is available at http://www.nature.com/ article’s Creative Commons license and your intended use is not permitted by statutory reprints regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http:// Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims creativecommons.org/licenses/by/4.0/. in published maps and institutional affiliations. © The Author(s) 2022 Modern Pathology (2022) 35:1812 – 1820