Recommendations on maximising the clinical value of tissue in the management of patients with intrahepatic cholangiocarcinoma

Background & Aims Patients with intrahepatic cholangiocarcinoma can now be managed with targeted therapies directed against specific molecular alterations. Consequently, tissue samples submitted to the pathology department must produce molecular information in addition to a diagnosis or, for resection specimens, staging information. The pathologist’s role when evaluating these specimens has therefore changed to accommodate such personalised approaches. Methods We developed recommendations and guidance for pathologists by conducting a systematic review of existing guidance to generate candidate statements followed by an international Delphi process. Fifty-nine pathologists from 28 countries in six continents rated statements mapped to all elements of the specimen pathway from receipt in the pathology department to authorisation of the final written report. A separate survey of ‘end-users’ of the report including surgeons, oncologists, and gastroenterologists was undertaken to evaluate what information should be included in the written report to enable appropriate patient management. Results Forty-eight statements reached consensus for inclusion in the guidance including 10 statements about the content of the written report that also reached consensus by end-user participants. A reporting proforma to allow easy inclusion of the recommended data points was developed. Conclusions These guiding principles and recommendations provide a framework to allow pathologists reporting on patients with intrahepatic cholangiocarcinoma to maximise the informational yield of specimens required for personalised patient management. Impact and Implications Biopsy or resection lesional tissue from intrahepatic cholangiocarcinoma must yield information about the molecular abnormalities within the tumour that define suitability for personalised therapies in addition to a diagnosis and staging information. Here, we have developed international consensus guidance for pathologists that report such cases using a Delphi process that sought the views of both pathologists and ‘end-users of pathology reports. The guide highlights the need to report cases in a way that preserves tissue for molecular testing and emphasises that reporting requires interpretation of histological characteristics within the broader clinical and radiological context. The guide will allow pathologists to report cases of intrahepatic cholangiocarcinoma in a uniform manner that maximises the value of the tissue received to facilitate optimal multidisciplinary patient management.


Data extraction and data items
Using best practice from the JBI Manual for Evidence Synthesis [4], a pre-specified digital data extraction form was customised in Excel, adapting the JBI 'Text and opinion data extraction tool' [5] to incorporate data fields specific to systematic reviews of textual opinion-based evidence, as previously used [3].The following data items were extracted from each resource: • Context within specimen reporting pathway -areas of specimen reporting pathway the guidance related to.• Phenomena of Interest -variables associated with histopathological iCCA reporting guidance (guidance development methodology, organisational and geographical representation in guidance development, clinical regulatory authority referenced, and presence of a reporting checklist).
• Verbatim extracts of guidance text.
• Interpretation of guidance statements; classifying guidance statements as 'explicit' or 'implicit' (explicit -'distinctly expressing all that is meant; leaving nothing merely implied or suggested': implicit -'suggested or understood but not directly expressed') [6].
Data from each resource were independently extracted by two reviewers (TK and GC) onto the data extraction form, followed by checks for consistency.Any discrepancies were discussed between the two reviewers and any disagreements resolved by a third reviewer (MG).

Data synthesis and analysis
The best fit framework synthesis method [7][8][9] that incorporates all elements of the JBI metaaggregation approach [10,11], was used for synthesis and analysis of descriptive qualitative data.The five stages of this method require 'familiarisation' with the literature to select the appropriate a priori framework; the distinct stages of the specimen pathway between specimen receipt in a pathology department and final authorisation of the histopathology report were used.The core group (TK, MG, GC, DO) examined the thematically linked extracted guidance from each source and generated a unifying statement or statements that were considered to accurately capture the meaning of the extracted text.

Ethical considerations
The UK Research Integrity Office's Code of Practice for Research and the Universities UK (UUK) Concordat to Support Research Integrity adopted as the University of Edinburgh's Research Integrity framework was followed and the project was assessed before study commencement using the UKRIO checklist for good practice in research in advance as a peer opinion study.Information was given to Delphi participants in the introductory page of the survey, and electronic informed consent, including for disclosure of participant name and affiliation at publication, was obtained following General Data Protection Regulation best practices and before survey completion (Supplementary Text 1).

Candidate item generation
The candidate items for the Delphi process were derived from the synthesised guidance generated by the systematic review.The final list was produced after discussion between the core group (TK, MG, GC, DO) allowing revision of, or addition to, the synthesised guidance to generate a complete offering to be assessed by the Delphi panellists.Statements were grouped into the stages of the reporting pathway between receipt of a specimen in the pathology department until the authorisation of the final pathology report.

Participants
Participants for the survey of histopathologists reporting iCCA were selected to include those practicing in multiple territories.The named secretary of every national division of the International Academy of Pathology were contacted to ask for a national expert.To democratise the process, and recent successes leveraging social networks of pathologists, an invitation to self-nominate was made on Twitter, amplified by liver pathologist experts on the network.
For the Delphi survey of 'end-users' (surgeons, oncologists, gastroenterologists) of histopathology reports, the membership list of working group 4 ('Epidemiology, clinical characterization and trials') of the EURO-CHOLANGIO-NET COST action was used.
Those responsible for developing the candidate items did not provide responses in the Delphi process to minimise potential bias.

Rating
Participants were asked to rate the importance of items on a nine-point Likert-type item scale, where '9' represented 'critical' and '1' represented 'not important'.Participants were able to use free-text boxes to comment about their scores, about the wording of the statement, or offer other items that were not included as candidates.
'End-user' participants were only asked to rate statements regarding what should be included in the written report.

Procedure
A minimum of two rounds was chosen to allow participants to reflect and change scores between rounds.A third round was not required because complete agreement was reached by the end of the second round.
Communication between the researchers and Delphi participants used private, individual email addresses so participants remained anonymous to one another.The Delphi was conducted via Google Forms with links to the Delphi provided in the emails.In Round One (22 nd September 2022), participants were given instructions, a consent form, and the items to rate.They were asked to complete the survey by 16 th October 2022 and reminder emails were sent to non-responders.
Round Two (17 th November 2022) invitations were sent only to those participants who completed Round One.Participants were advised which items did not reach consensus in Round One and how these had been amended based on feedback, and asked to re-rate these items.They were informed that their scores from Round One would be carried forward for any item for which a new score was not returned and asked to complete the survey by 2 nd December 2022.Reminder emails were sent to non-responders.

Analysis
In Rounds One and Two, response rate, median score and item consensus scores were calculated.Data were analysed in R. In addition, free text comments were reviewed to determine if items needed amendment or additional items required inclusion.
'So that the process is transparent, we will include your name and affiliation in a list of participants to be published with the guidelines.By providing us with these details you are giving your consent for us to do this.We are also collecting other data about your experience but this information and participant responses to Delphi items will only be published in aggregated form, not attributed to individuals.
You are able to withdraw your consent at any time prior to submission of the guidelines for publication by contacting Dr Tim Kendall, University of Edinburgh (Tim.Kendall@ed.ac.uk).'Although it has been proposed that the diagnosis of iCCA can be made based upon a combination of clinical presentation, laboratory analysis, and radiologic evaluation, pathological diagnosis is required for definitive diagnosis in most patients, particularly those with cirrhosis and small hepatic mass lesions as radiographic studies are nonspecific.
Pathological diagnosis is recommended for all patients who will be undergoing systemic chemotherapy or radiation therapy, or enrolling in a therapeutic clinical trial.The sensitivity of liver biopsy for pathological diagnosis will depend upon location, size, and operator expertise.Core biopsies are required for definitive diagnosis.Although a positive liver biopsy will establish a diagnosis, a negative biopsy does not exclude it because of the potential for sampling error.

Immunohistochemistry
iCCA needs to be distinguished from benign biliary lesions such as peribiliary glands, reactive ductular proliferation, biliary microhamartomas (von Meyenburg complexes) and bile duct adenomas (peribiliary tract hamartomas), particularly in the presence of inflammation which can result in reactive cellular atypia.The histological appearance of iCCA is similar to that of metastatic adenocarcinoma arising from extrahepatic primary tumors and especially those of foregut origin such as lung, pancreas, esophagus, and stomach.The differentiation of iCCA from metastatic adenocarcinoma cannot be readily ascertained on histological examination.Differentiation between iCCA and mixed HCC tumors may require evaluation of specific markers of hepatocellular or progenitor cell features such as Hep-Par-1, GPC3, HSP70, glutamine synthetase, EpCAM, and CK19.CK19 positivity is not specific for iCCA.The expression of cytokeratin 7 and cytokeratin 20 may be helpful to establish a biliary origin.

Written report
The 7th edition of the AJCC/UICCA staging schema is currently the preferred staging system for resected iCCA.

Burt et al., 2020
2. Macroscopy/cut up/trimming of the gross specimen When previous therapy has been administered microscopic examination of the entire tumour should be done when feasible.For selective sampling, sampling an entire cross section has been recommended if the tumour is <2 centimetres (cm) with an additional section for each 1 cm for larger tumours.
Additional sampling of areas that appear grossly viable is often necessary.
The following guidelines are provided for intrahepatic tumours: • Tumour with nearest hepatic resection margin (when this is close enough to the tumour to be included in the block).• Other blocks of tumour with adjacent liver tissue (for microscopic vascular invasion (MiVI)).
• Liver capsule if there is a possibility of capsular invasion, i.e., where there is subjacent tumour and overlying adherent tissue or macroscopic capsular invasion.Where the capsule appears intact over subcapsular tumour, with a smooth shiny surface, histology is not required to confirm capsular integrity.• Gallbladder bed and wall where there is adjacent intrahepatic tumour.
• Any site macroscopically suggestive of macrovascular or bile duct invasion.
• Background liver (taken as far away as possible from the tumour).
A block of representative background liver should be taken at a distance from the tumour, whether or not it looks abnormal macroscopically.
The number of tumours is also recognized as an important prognostic factor in intrahepatic cholangiocarcinoma.Location of all tumours (HCC and intrahepatic cholangiocarcinoma) should be reported since this is important for correlation with imaging when this is available.
Distinction from metastatic adenocarcinoma is based on the presence of a single or dominant intrahepatic mass and absence of a known extra-hepatic primary tumour.

Microscopic assessment
According to the current WHO Classification 5th edition, the two main histological subtypes of intrahepatic CC are the large duct type, arising in the intrahepatic large ducts and composed of mucin secreting tumour cells and the small duct type (non-mucin secreting and mainly occurring in the hepatic periphery).Cholangiolocarcinoma (CLC) and intrahepatic cholangiocarcinoma with ductal plate malformation pattern are subtypes of small duct intrahepatic CC.
Four tumour growth patterns of intrahepatic cholangiocarcinoma are described: the mass-forming type, the periductal infiltrating type, the intraductal growth type and the mixed type.Mass-forming intrahepatic cholangiocarcinoma (65% of cases) forms a well-demarcated nodule growing in a radial pattern and invading the adjacent liver parenchyma.The periductal-infiltrating type of cholangiocarcinoma (6% of cases) spreads in a diffuse longitudinal growth pattern along the bile duct, and the intra-ductal growth type (4% of cases) shows a polypoid or papillary tumour within the dilated bile duct lumen.
Definitive criteria for histological grading of cholangiocarcinomas have not been established; however, the following semiquantitative grading system based on the proportion of gland formation within the tumour is commonly used for intrahepatic cholangiocarcinomas: • Well differentiated (more than 95% of tumour composed of glands) • Moderately differentiated (50% to 95% of tumour composed of glands) • Poorly differentiated (up to 49% of tumour composed of glands)

Written report
Record extent of invasion.

Record other pathologies.
For dysplasia involving large bile duct radicles we recommend the use of the BilIN and Intraductal papillary neoplasm of the bile ducts (IPNB) classifications described in the WHO 5th edition guidelines, both of which distinguish low grade from high grade change.
Margin status -For intrahepatic cholangiocarcinoma there are a few publications citing margin status as a prognostic factor on multivariate analysis.A systematic review of intrahepatic CC did not include margin status among significant prognostic factors.
Lymph node metastases in intrahepatic and perihilar cholangiocarcinoma have been identified as an important predictor of prognosis.

Specimen fixation
The surgeons should fill an Application Form of Pathological Examination describing the clinical diagnosis, location and type of lesions and number of tissues.The surgical margin, suspected lesions, important vessels and bile duct margin should be marked with a dye or suture by surgeons.Small resected tissues, such as lymph nodes, should be placed into different containers and labeled with corresponding descriptions.
To maximally preserve the integrity of intracellular nucleic acids and proteins for avoiding autocytolysis, tumor specimens should be transferred to the Department of Pathology as soon as possible after resection, ideally within 30 min after surgical removal for sectioning and fixation.
The fresh specimens should be cut consecutively into 1cm thick multiple sections at the maximal diameter; a portion should remain unfixed fresh or cryopreserved for molecular examination.
At room temperature, tissues should be fixed in a neutral formalin solution (v:v, 1:4-5) for 12-24 h and embedded in paraffin.

Macroscopy/cut up/trimming of the gross specimen
Hepatic tumor samples should be collected using the 7 point baseline sampling protocol The location and number of liver tissues collected should be determined as appropriate according to the size, shape and number of the liver tumors as well as the adjacent liver tissues.
Because the detection rate of MVI and satellite nodules is related to the extent of adjacent liver tissues, it is necessary to describe the size of the adjacent liver tissues, and the suspected lesions should be sampled after reviewing several sections.
When the tumor tissue is close to the surgical margin, sampling should be done at the region vertical to the margin closest to the cancer.When the tumor tissue is far away from the surgical margin, sampling should be done parallel to the surgical margin.The status of the surgical margin should be determined using the section with maximal area.

Block processing and routine tinctorial staining
Sections of 5μm thickness should be cut from each block and stained with hematoxylin and eosin for histological examination.

Microscopic assessment
Adenocarcinoma is the most common histological type of ICC, although it may also present in other special histological and cell types and its differentiation degree can be classified as well, intermediate and poor.
Tumor growth patterns, including peritumoral invasion, capsule invasion, MVI and satellite nodules.
The presence of chronic liver disease, such as chronic hepatitis or hepatic cirrhosis.Although there are many systems for grading and staging chronic viral hepatitis, a simple histologic scoring system is recommended for routine pathological diagnosis, such as the Scheuer scoring system, etc.Furthermore, Masson's trichrome staining and reticular fiber staining can be routinely undertaken to assess the degrees of hepatic fibrosis and lobule reconstruction, respectively.

Immunohistochemistry
Currently used biomarkers for liver cancer are somewhat imperfect in their diagnostic specificity and sensitivity; thus, a biomarker panel in combination with other tissue-specific markers could represent a useful tool for diagnosis and differential diagnosis between benign and malignant hepatocellular tumors, HCC and ICC, other specific types of hepatic tumors, and primary and metastatic liver cancer.
Although immunohistochemical staining for CD34 does not directly label hepatic parenchymal cells, it is valuable for determining the extent of MVD and examining its unique distribution pattern in different liver tumors.For instance, a diffuse staining pattern is indicative for HCC, a scattered staining pattern for ICC, a patchy staining pattern for HCA, and a cord-like staining pattern for focal nodular hyperplasia, etc.

Fernandez Moro et al., 2016
5. Immunohistochemistry A combination of "generic" adenocarcinomamarkers (CK7, CK19, BerEP4, and polyclonal CEA) together with vimentin and a limited number of "pancreatobiliary" markers (MUC1 and CA19-9) defines its characteristic immunohistochemical profile.The present study identified cytoplasmic WT1 as a novel marker for intrahepatic cholangiocarcinoma.The characteristic immunohistochemical profile of intrahepatic cholangiocarcinoma positively supports its pathological diagnosis, which no longer needs to be regarded as a diagnosis of exclusion (of metastatic adenocarcinoma).

Microscopic assessment
Cholangiocarcinoma most often must be differentiated from metastatic carcinoma.Cholangiocarcinoma is almost always a desmoplastic tumor and often is mucin producing.

Immunohistochemistry
The immunophenotype of CCA is generally similar to that of extrahepatic pancreaticobiliary carcinomas.Almost all CCAs react with CK7, CK19, CEA (both monoclonal and polyclonal; noncanalicular pattern), and MOC-31.In contrast to other pancreaticobiliary carcinomas, CCAs are more likely to react with CK7 and less likely to react with CK17, CK20, and p53.Cytokeratins can also be useful in differentiating HCC from CCA, as can thyroid transcription factor 1 and claudins.Cholangiocarcinoma can also react with CA 125, and differentiation from mullerian carcinomas can be difficult, although CCA is almost never estrogen receptor positive.Newer studies with the mucins (eg, MUC4, MUC5AC, MUC5B, MUC6) suggest that they may be useful in the subclassification of CCAs and may be useful in evaluating prognosis.

Jiang et al., 2018
2. Macroscopy/cut up/trimming of the gross specimen Pathologic staging depends on pathologic documentation of the anatomic extent of disease, whether the primary tumor has been completely removed.If a biopsied tumor is not resected for any reason and if the highest T and N categories or the M1 category of the tumor can be confirmed microscopically, the criteria for pathologic classification and staging have been satisfied without complete removal of the primary cancer.The AJCC eight edition classifies the T category according to vascular invasion, extent of invasion, and the number of tumors.

Microscopic assessment
Compared to other types of cancer involving liver, cholangiocarcinomas are the ones that elicit a marked desmoplastic fibrotic reaction.
It has been well recognized that cholangiocarcinomas demonstrate a wide spectrum of growth patterns.Tumors can be composed of irregular cystic, branching/glandular tubular structures or as irregular aggregates of infiltrating glands.
Cholangiocarcinomas do not always show easily identifiable tubular contours; instead, tumor cells often grow in a nodular pattern or form sheets of solid nests that mimic neuroendocrine neoplasms and/or HCC.Other tumors may show copious mucin components, resembling mucinous adenocarcinoma or colloid tumors.
It should be kept in mind that for any cholangiocarcinoma, it is common to see a variety of the abovementioned histomorphological features in the same lesion.
Focal features of sarcomatoid or clear cell morphology are not uncommon.
Cholangiocarcinoma tumor cells frequently extend along the portal tracts by growing within the connective tissue, without directly invading into the bile ducts in the vicinity, even though this pattern of tumor invasion can be seen in not only cholangiocarcinoma but also other types of tumors, including metastatic carcinomas.Also, tumor cells can colonize and extend along the bile ducts and intermingle with native benign bile ducts and adjacent reactive ductules.
It should always be kept in mind that cholangiocarcinomas are neurotropic, and tumor cells frequently approach nerve tracts, with frequent perineural and intraneural involvement.In practice, this pattern can be very challenging to recognize, especially during frozen section evaluation; more so when the primary tumor is small and when there is no clinical evidence of metastatic disease.
Another malignant behavior of cholangiocarcinomas is lymphovascular invasion.Astute microscopic examination and extensive tumor sectioning is essential in reaching such a diagnosis.The usual intraluminal "dirty" necrosis, frequently associated with colorectal primary adenocarcinomas, are not typically seen in cholangiocarcinomas.
Histologically, cholangiocarcinoma follows a stepwise carcinogenesis process through a precursor lesion: BilIN, ranging from low to high grade (carcinoma in situ).The presence of these precursor lesions in the vicinity of an intrahepatic adenocarcinoma, combined with marked desmoplastic reaction and high-grade cytological atypia, is highly suggestive of cholangiocarcinoma, instead of metastatic colorectal adenocarcinoma; the latter usually demonstrates palisading, pencil-like nuclei and intraluminal "dirty" necrosis.However, a frequent obstacle encountered in practice is the small size of the biopsy material, which inevitably restricts the differential list and limits reaching a definitive diagnosis due to small areas of useful diagnostic material.

Immunohistochemistry
By immunostaining, cholangiocarcinomas are strongly positive for CK-7 and CAM5.2 and show cytoplasmic labeling by p-CEA.Mouse monoclonal antibody that recognizes human EPCAM on cell membrane (MOC31) is positive in around 90% of cases, whereas CK-19 is positive in 70% to 80% of cases.Interestingly, there have been findings to suggest that peripheral cholangiocarcinomas are more likely to be CK-7 positive but CK-20 negative, whereas central counterparts tend to be positive for both CK-7 and CK-20.CD56 also is reported to be positive in peripheral cholangiocarcinomas, which does not necessarily mean there is neuroendocrine differentiation, if other neuroendocrine markers are negative.
One particularly useful immunostain marker for confirming cholangiocarcinoma is S100P, which is negative in benign biliary epithelium but usually positive in cholangiocarcinomas, especially when combined with U3 small nucleolar ribonucleoprotein protein (IMP3) and a protein that in humans is encoded by the von Hippel -Lindau tumor suppressor gene (pVHL).In comparison, HCCs, including all the variants, are predominantly positive in the described pattern for the immunostain markers introduced earlier, including arginase, HepPar-1, glypican 3, CD34, CD10, and p-CEA; HCCs are also positive for AE1/3 and CAM5.2 in almost all cases but remain negative mostly for CK-7 and CK-20.In contrast, cholangiocarcinomas are mostly negative for the markers for HCC, except for p-CEA, which could be positive in various adenocarcinomas, irrespective of the organs of origin.
Rarely, HepPar-1 could be positive in cholangiocarcinoma cases, with foci of labeling, but given the negativity by other HCC markers and adenocarcinoma histomorphology and immunoprofiles, a diagnosis of cholangiocarcinoma should be beyond the question; this occasional cellular labeling is not specific.
In addition to HCC, a metastatic process from another organ or site is also frequently suspected when facing a potential cholangiocarcinoma, such as tumors originating in gastrointestinal luminal sites, especially those from colorectal regions.
It should be made clear to the entire clinical team that there are no specific markers that can be dependable for diagnosing cholangiocarcinoma.Cholangiocarcinomas are usually positive for CK-7, and sometimes for CK-20, similar to upper gastrointestinal tract and lung adenocarcinomas.A positive Thyroid Transcription Factor 1 (TTF-1) or napsin-A immunostain would strongly favor pulmonary origin; however, occasional lung tumors are negative for TTF-1 or napsin-A by immunostaining; these cases require detailed histomorphological evaluation and systemic radiological and clinical correlation to narrow down the most likely origin of the tumor.Likewise, the positive CK-20 labeling and/or focal CDX-2 immunoreactivity does not necessarily indicate gastrointestinal tract or colorectal origin.
Subsequently, it is extremely important to correlate radiological, clinical, and histopathological findings in order to draw a correct conclusion.

Written report
The pathological diagnosis of primary intrahepatic cholangiocarcinoma still mostly remains one of exclusion, because of a lack of specific markers.However, a novel RNA platform using in-situ hybridization for albumin RNA has shown specificity for primary liver cancers, including CC.80However, since this test is not widely available, diagnosis of primary intrahepatic cholangiocarcinoma still requires exclusion of all other other possibilities by analyzing histomorphology, immunoprofiles, imaging studies, as well as clinical evidence.Clinical history, endoscopic investigation, and ima-ging studies are all pivotal in reaching the correct conclusion.
Frequently, making a diagnosis of cholangiocarcinoma especially the intrahepatic ones, instead of a metastatic adenocarcinoma from another organ or site, can be difficult and challenging, more so when there are none or limited clinical and radiological information available for analysis.The clinical, histological, and molecular information are all essential for decision-making and management planning.
Clinically, cholangiocarcinomas are associated with chronic biliary inflammation and regeneration, as frequently seen in PSC, fluke infestation, chronic hepatitis C and hepatitis B viral infection, profession-related chronic exposure to chemicals, and pancreatobiliary malfunction.On the other hand, history of colorectal, gastrointestinal, or gynecological malignancies as well their stages are helpful in determining the possibility of dealing with a metastatic tumor.
It is very important to compare the histomorphology of the primaries and the adenocarcinoma in the liver to see whether there is any similarities; unfortunately in the setting of a comprehensive cancer center, it is not unusual to face a newly identified intrahepatic adenocarcinoma with no or minimal history and no previous cases for morphological comparison.
To summarize briefly, in order to decide whether an intra-hepatic adenocarcinoma is a primary cholangiocarcinoma or a metastatic process, morphological and ancillary test results need to be utilized in combination with the clinical and radiological findings to reach the most likely conclusion regarding the true nature of the tumor.

Khan et al., 2012
2. Macroscopy/cut up/trimming of the gross specimen Macroscopic features of intrahepatic CC Intrahepatic CCs are whiter and firmer than HCCs as they contain more desmoplastic stroma.They occur more commonly in non-cirrhotic livers than HCCs and are divided into four macroscopic types.

Microscopic assessment
Over 90% of CCs are adenocarcinomas and are classified according to the percentage of tumour composed of glandular tissue.Some types of adenocarcinoma are not graded (eg, carcinoma in situ, clear cell adenocarcinoma and papillary adenocarcinoma).Signet ring cell carcinoma is graded as 3 and small cell carcinoma as 4.
Tumours are usually adenocarcinomas and have prominent desmoplastic stroma.However, except in cases where there is co-existing biliary dysplasia, it may not be possible, even with immunohistochemistry, to differentiate between CC and metastatic tumour.Examples of this include intraductal papillary neoplasm with associated invasive neoplasia, and mucinous cystic neoplasm with associated invasive neoplasia.

Immunohistochemistry
Distinguishing intrahepatic CC from metastatic adenocarcinoma and other primary liver tumours can be difficult.Accurate differentiation, particularly from foregut metastases (lung, oesophagus, stomach, pancreas), often cannot be made histologically.Other modalities, especially imaging, are essential.
In distinguishing HCC from CC, lack of mucin production and expression of HepPar-1, CD10 and glypican-3 by HCC are useful.

Written report
CC staging is based on the tumour-node-metastasis (TNM) system.The 7th edition of the TNM classification introduced a specific staging system for intrahepatic CC, separate from HCC, providing better prognostic information.
Surgical resection specimens should be reported systematically, for example, according to Royal College of Pathologists' guide-lines.The final report should include the following information: Tumour -a.Histological type b.Histological grade c.Extent of invasion (according to the TNM system) d.Blood/lymphatic vessel invasion e. Perineural invasion: this is common and has prognostic significance.
Margins -These must be adequately sampled because local recurrence is related to involvement of the margins.This is particularly important because extrahepatic CC may be multifocal in up to 5% of cases.
Regional lymph nodes -To stage lymph nodes accurately, the node groups must be specifically identified.Peripancreatic nodes located along the body and tail of the pancreas are considered sites of distant metastasis.
Additional pathological findings -These must be noted if present (eg, carcinoma in situ, sclerosing cholangitis).
Metastases -Metastases to other organs or structures should be reported.

Microscopic assessment
Typically, pathologic analysis of biopsy specimens reveals adenocarcinoma, and the challenge is differentiating primary intrahepatic cholangiocarcinoma from metastasis from other gastrointestinal or pancreatic primary tumors, often requiring additional immunohistochemical evaluation.

Written report
Recognition of the biological and epidemiologic differences between hepatocellular carcinoma and intrahepatic cholangiocarcinoma led to the development of a distinct tumor, node, metastasis (TNM) staging system for intrahepatic cholangiocarcinoma that has been updated in the Eighth Edition of the American Joint Commission on Cancer (AJCC) Cancer Staging Manual.

Miyazaki et al., 2015
2. Macroscopy/cut up/trimming of the gross specimen Since the ductal margin status has an im-pact on patients' prognosis, the histological assessment of the ductal resection margin should be performed intraoperatively.

Microscopic assessment
Biliary intraepithelial neoplasia (BilIN) and intraductal papillary neoplasm of bile duct (IPNB) are premalignant lesions for bile duct cancer.

Macroscopy/cut up/trimming of the gross specimen
Sections should be prepared from each major tumor nodule, with representative sampling of smaller nodules if macroscopically different in appearance The evaluation of margins for total or partial hepatectomy specimens depends on the method and extent of resection.It is recommended that the surgeon be consulted to determine the critical foci within the margins that require microscopic evaluation.The transection margin of a partial hepatectomy may be large, rendering it impractical for complete examination.In this setting, grossly positive margins should be microscopically confirmed and documented.If the margins are grossly free of tumor, judicious sampling of the cut surface in the region closest to the nearest identified tumor nodule is indicated.In selected cases, adequate random sampling of the cut surface may be sufficient.outcome for intrahepatic cholangiocarcinoma.
Histologic examination of a regional lymphadenectomy specimen usually involves examination of 3 or more lymph nodes.

Microscopic assessment
Combined or mixed hepatocellular-cholangiocarcinoma accounts for less than 5% of primary liver carcinomas and should show histologic evidence of both hepatocellular differentiation and bile duct differentiation, such as production of mucin.
A system based on the proportion of gland formation within the tumor is suggested: • Grade X Grade cannot be assessed • Grade 1 Well differentiated (more than 95% of tumor composed of glands) • Grade 2 Moderately differentiated (50%-95% of tumor composed of glands) • Grade 3 Poorly differentiated (5%-49% of tumor composed of glands) • Grade 4 Undifferentiated (less than 5% of tumor composed of glands) Three tumor growth patterns of intrahepatic cholangiocarcinoma are described: the mass-forming type, the periductal-infiltrating type, and mixed mass-forming/periductal-infiltrating type.Massforming intrahepatic cholangiocarcinoma (60% of cases) forms a well-demarcated nodule growing in a radial pattern and invading the adjacent liver parenchyma (Figure 2).In contrast, the periductalinfiltrating type of cholangiocarcinoma (20% of cases) spreads in a diffuse longitudinal growth pattern along the bile duct.The remaining 20% of cases of intrahepatic cholangiocarcinoma grow in a mixed mass-forming/periductal-infiltrating pattern.

Written report
The histologic examination of the bile ducts at the cut margin is recommended to evaluate the lining epithelium for in situ carcinoma or dysplasia.If the neoplasm is found near the surgical margin, the distance from the margin should be reported.For multiple tumors, the distance from the nearest tumor should be reported.The TNM staging system of the American Joint Committee on Cancer and the International Union Against Cancer applies to all primary carcinomas of the intrahepatic bile ducts and mixed hepatocellular-cholangiocarcinomas.It does not apply to hepatic sarcomas or to metastatic tumors of the liver.

Wyatt et al., 2012
2. Macroscopy/cut up/trimming of the gross specimen Specimens can be dissected in the fresh or fixed state.Although formalin penetrates the liver poorly, intrahepatic tumours are usually clearly demarcated within the liver and examination after 24-48 hours does not significantly impair morphology.
Specimen hardening following fixation facilitates accurate slicing.
If fresh tumour is required, this can be obtained either by slicing the specimen fresh after painting the resection margin, or (if identifiable from the external appearance) by excising a portion of tumour through the capsule, so long as the capsule appears intact and is not covered by adherent fatty tissue which may result from underlying capsular breach by the tumour.The surfaces of the specimen other than the capsule (i.e.parenchymal resection plane, extrahepatic biliary tree, any tissue adherent to the liver capsule) should be painted with ink or silver nitrate to allow identification in histological sections.
Block-taking strategy for all liver specimens: • Tumour with nearest hepatic resection margin (when this is close enough to the tumour to be included in the block).• Other blocks of tumour with adjacent liver tissue (for microscopic vascular invasion).
• Liver capsule if there is a possibility of capsular invasion, i.e.where there is subjacent tumour and overlying adherent tissue or macroscopic capsular invasion.Where the capsule appears intact over subcapsular tumour, with a smooth shiny surface, histology is not required to confirm capsular integrity.• Gall bladder bed where there is adjacent intrahepatic tumour.
• Any site macroscopically suggestive of vascular or bile duct invasion.Background liver (taken as far away as possible from the tumour).
The number of tumour blocks will depend on the tumour type, but should include samples from areas of differing macroscopic appearance in heterogeneous tumours.
A block of representative background liver should be taken, whether or not it looks abnormal macroscopically.The appearance should be included in the text description of the specimen, but is a core item in the dataset proforma only in the histology section, since microscopy provides the more reliable assessment of fibrosis/cirrhosis.
The following additional blocks are required as appropriate: • where there is tumour tissue close to the hepatic hilum, the hilum should be sampled to include large vessels.Specifically label blocks of main left or right portal vein or bile duct, if present hepatic vein margin (if there is tumour nearby) extrahepatic biliary tree (when included) • gall bladder -optional when this is macroscopically normal The site of lymph nodes should be specified if known (hilar, hepatic artery, portal vein, cystic duct).More distant nodes (coeliac axis, periduodenal, periaortic) may be submitted separately by the surgeon.Large nodes that do not show macroscopic involvement should be serially sliced and embedded in their entirety since nodes at this site are often enlarged as a result of reactive changes and may harbour micrometastases, especially in cholangiocarcinoma.
Record the segments resected and the specimen weight after opening the gall bladder and rinsing out the bile.
The specimen dimensions (antero-posterior, medio-lateral and supero-inferior) should also be measured, particularly in cases where much of the specimen is occupied by tumour.
When included with the specimen, record the length of extrahepatic duct, number and site of lymph nodes, size and appearance of gall bladder.
The specimen should be sliced at right angles to the parenchymal resection plane, and preferably in the horizontal plane to facilitate correlation with preoperative cross-sectional imaging.Slices should be as thin as possible, and no more than 10 mm thick.The minimum size of tumours detectable by imaging is now less than 5 mm.
Record the number, site, maximum diameter and distance from hepatic margin of the tumour(s).For multiple tumours, the sites should be recorded in the text of the report in such a way that allows correlation with preoperative imaging.For example, this can conveniently be recorded by numbering the horizontal slices from the top, and specifying the slices and approximate segments of each tumour.'Multiple tumours' encompasses satellitosis, multifocal tumours and intrahepatic metastases.The presence of satellite nodules should be noted.
The appearance of the background liver (normal, bile-stained, fibrotic/cirrhotic) should be recorded.It is good practice to keep a photographic record of the macroscopic features of the specimen for use during MDT meetings.
The specimen should be inspected carefully for macroscopically apparent vascular invasion, and any suspected vascular invasion should be sampled for histological confirmation.Involvement of the main left or right branch of the portal vein or any of the three main hepatic veins should be specifically recorded, as this information is relevant to TNM staging.
The type of intrahepatic tumour may not be known until histology, however the same core items are required for all: • type of specimen • specimen weight • specimen dimensions (where orientation is known, provide antero-posterior, medio-lateral and superior-inferior dimensions) • tumour number and size • presence of satellite lesions (regarded as multiple tumours for TNM staging) • distance from nearest hepatic resection margin(s) • for HCC and intrahepatic CC: macroscopic involvement of vessels; specify if main left or right portal vein or a main hepatic vein, and record diameter of vessel involved • integrity of liver capsule (including bare area on postero-superior aspect) and presence of adherent tissues (e.g.diaphragm) or other organs • presence and number of lymph nodes received.

Block processing and routine tinctorial staining
Needle biopsy -Sectioning and staining -Initially one or two shallow levels stained with H&E should be examined.The pathologist can then determine whether tumour is present and what further investigations are required based on the morphology of the tumour in the biopsy and clinical circumstances.

Microscopic assessment
Provide comment on: • Tumour type • Tumour differentiation • Minimum distance to resection margin (hepatic, and where appropriate bile duct or vascular) measured microscopically when less than 5 mm.Microscopic involvement (R1) is generally defined as a clearance of <1 mm.• Invasion through liver capsule (Glisson's capsule) • Vascular invasion including confirmation of macroscopic vessel invasion.?Perineural invasion (cholangiocarcinoma). ?Effects of ablative or neoadjuvant therapy on tumour (if applicable).• Background liver -presence and stage of fibrosis, and other chronic liver.
• Lymph node involvement (where appropriate); number of nodes with metastasis.
In addition to the location of the CC, the Japanese classification according to the growth pattern into mass-forming, periductal infiltrating and intraductal papillary has gained international recognition.In general, mass-forming CC arises peripherally in the liver.The periductal infiltrating pattern is characteristic of perihilar CC, which includes CC arising in right, left and common hepatic ducts.
Intrahepatic CCs form an expansile tumour mass with obvious borders, and usually arise peripherally in the liver.The maximum diameter is readily determined.The size of the tumour is not a staging criterion since prognosis is independent of size however, it is important for correlation with preoperative imaging.
Staging depends on the number of tumours (single or multiple) and vascular invasion.Unlike HCC, vascular invasion (any size of vessel, pT2a) is distinguished from multiple tumours (pT2b, with or without vascular invasion).Tumour perforating the visceral peritoneum or direct infiltration of adjacent organs constitutes pT3 disease.The periductal infiltrating pattern, when associated with a tumour mass, indicates late-stage disease with a poorer prognosis and is classified as pT4 in TNM7.If the tumour is entirely of periductal infiltrating pattern, consideration should be given as to whether it arises in a main right or left duct; if so, it should be staged as perihilar CC.
Intrahepatic CC typically has a microacinar glandular pattern with central sclerosis, and distinction from metastatic adenocarcinoma particularly from stomach or pancreas is based on the single or dominant intrahepatic mass and absence of a known extra-hepatic primary tumour.Currently, available immunohistochemistry is not contributory.
Locally advanced intrahepatic mass-forming CCs invade through the liver capsule and directly into adjacent adherent organs.Perforation of the visceral peritoneum constitutes pT3 disease, and any roughened area of capsule over the tumour should be sampled for histology.
For right intrahepatic CC, the regional lymph nodes include the hilar (common bile duct, hepatic artery, portal vein and cystic duct), periduodenal and peripancreatic lymph nodes.For left intrahepatic CC, regional lymph nodes include hilar and gastrohepatic lymph nodes.For either type of intrahepatic CC, spread to the coeliac and/or periaortic and caval lymph nodes represent distant metastases (M1).

Written report
Needle biopsy -The report should include the following: • the clinical information received with the biopsy • a macroscopic description including biopsy size • the presence or absence of tissue from the focal lesion, and of liver tissue (hepatocytes, bile ducts) as histological confirmation that the specimen is indeed from the liver • a morphological description of the lesion • the results of any additional stains carried out, including immunohistochemistry a comment on the background liver, if sufficient is included • a definite diagnosis of the focal lesion where possible, or a discussion of the differential diagnosis.This would include a discussion of tumours compatible with or excluded by immunohistochemistry • an appropriate SNOMED code Supplementary Text 3 Delphi participants Pathologist survey Supplementary Figures

Fig
Fig. S1 PRISMA flow diagram of the selection process of guidance resources Fig. S2

Fig. S2
Fig. S2 Number of guidance resources showing low, moderate, and high quality across AGREE-GRS domains Fig. S3

Fig
Fig. S3 iCCA recommendation development diagram