Quo Vadis colorectal intramucosal adenocarcinoma?

The classification and clinical implications of colorectal intramucosal adenocarcinoma remain controversial. Given the increasing frequency of diagnosis through colorectal cancer screening programmes, a reassessment of terminology and its impacts is necessary. This paper critically examines the diagnostic criteria, biological behaviour and clinical consequences of labelling these lesions as colorectal intramucosal carcinoma. While intramucosal adenocarcinoma exhibits cytological and architectural atypia beyond high-grade dysplasia, it remains confined to the mucosa and has minimal metastatic potential, with rare documented exceptions. Conversely, the use of carcinoma terminology has been associated with potential overtreatment, including unnecessary surgical resection, increased patient anxiety and financial burdens such as insurance complications. We explore geographic variations in classification and analyse the impact of terminology shifts. We propose a standardised framework that restricts the term intramucosal adenocarcinoma to intramucosal lesions exhibiting tumour budding or poorly differentiated clusters, signet ring cells, desmoplasia, vascular invasion, mucinous differentiation or features of neuroendocrine carcinoma, while reclassifying adenomas with cribriform architecture and complex glands as high-grade dysplasia. This nomenclature shift aims to reduce overtreatment, align with current oncologic understanding and ensure optimal patient care and communication. - Colorectal Neoplasms - DIAGNOSIS - Histology Statistics from Altmetric.com

Several terms have been used to describe large intestinal tumours with morphological features overlapping with adenocarcinoma but confined to the mucosa; these include intramucosal adenocarcinoma and carcinoma in situ (CIS). 1 2 The American Joint Committee on Cancer (AJCC) classifies this as pTis.1 The WHO acknowledges the term intramucosal adenocarcinoma but does not endorse it.2 The term adenocarcinoma or carcinoma for intramucosal colorectal lesions has been controversial precisely because the risk of adenocarcinoma confined to the mucosa is low but not zero. For this discussion, colorectal intramucosal adenocarcinomas are defined by cytological and architectural atypia exceeding that of high-grade dysplasia, with the neoplastic process confined to the mucosa and not extending beyond the muscularis mucosae (except vascular invasion). This analysis focuses on endoscopically polypoidal or flat lesions (table 1) and excludes ulcerating or fungating tumours, where deeply invasive carcinoma is typically diagnosed endoscopically. Only lesions examined histologically in their entirety are included, and this discussion excludes patients with inflammatory bowel disease, Lynch syndrome or patients with high-risk factors for colorectal cancer (CRC). Considering the extensive screening programmes in the Global North and the significant number of patients who could be diagnosed with intramucosal adenocarcinoma, it is crucial to enhance case reporting. A 2006 study reported 40 cases of intramucosal adenocarcinoma (Vienna category 5.1) out of 1552 cases, accounting for 2.6% of the cohort.3 The term intramucosal adenocarcinoma for colorectal lesions raises concerns about overtreatment, such as unnecessary surgical resection. Recent data support this concern, showing that patients with intramucosal adenocarcinoma were significantly more likely to undergo surgery postendoscopic resection compared with those with high-grade dysplasia (10.5% vs 0%, p<0.001).4 However, one challenge in interpreting these data is the uncertainty about whether the diagnosis of intramucosal carcinoma drove the decision for resection or if other factors, such as endoscopic unresectability, made it difficult to understand these findings fully. In ‘Western’ countries, particularly in the USA, the term ‘intramucosal adenocarcinoma’ can lead to problematic insurance implications. This term may result in denial or increased insurance rates despite these carcinomas having little to no metastatic potential.5 In the USA, the Health Insurance Portability and Accountability Act grants patients the right to access their medical records directly. This means patients can receive their pathology reports, sometimes simultaneously with their treating physician. This immediate access can lead to significant anxiety, especially when reports contain terms like ‘carcinoma.’ Even with a detailed explanation from the treating physician, patients may struggle to understand that, despite the term ‘carcinoma,’ such lesions often have a low risk of behaving like invasive cancer. This misunderstanding can cause considerable emotional distress, even after successful endoscopic resection. This situation raises an ethical question: while the traditionally accepted criteria would support using this term, should we continue using the nomenclature, however accurate, if it negatively impacts patients? This dilemma underscores the need to balance precise medical terminology with potential socioeconomic consequences for patients. If the term is used, under what circumstances and with what accompanying explanation should it be employed to ensure accuracy and minimal harm to patients? Alternatively, it can be replaced with a term that conveys the necessary clinical information without causing unnecessary anxiety. These are ambitious goals, and we acknowledge several challenges in this process. First, all-encompassing literature on this topic is scarce. Case series, often confined to a handful of cases that describe adverse events in colorectal intramucosal adenocarcinoma, frequently lack detailed histology descriptions and illustrations. This absence makes it difficult for the reader to confidently exclude submucosal invasion. Additionally, defining submucosal invasion is often tricky due to various artefacts. Furthermore, the case series presented in the literature typically has relatively short follow-up time frames, limiting the weight that can be placed on these findings. It must be noted that only a minority of submucosal carcinomas with limited submucosal invasion will be associated with involved lymph nodes or distant metastases per se. Nevertheless, this document represents the collective experience of experienced pathologists, combined with existing literature, enabling us to offer some insightful comments. Challenging the use of the term ‘intramucosal colorectal adenocarcinoma’ Studies have shown that patients with cancer are 2.65 times more likely to declare bankruptcy, although the precise impact of the term ‘colorectal intramucosal adenocarcinoma’ on this statistic remains unclear.6 However, a relatively innocuous diagnosis such as Barrett’s dysplasia can significantly complicate life insurance coverage. Shaheen et al surveyed 20 national life insurance companies and found that a diagnosis of Barrett’s oesophagus led to a 118% increase in premiums.7 Patients were often denied coverage altogether, even when their physicians provided documentation indicating the minimal effect of Barrett’s oesophagus on life expectancy. Therefore, we expect that a diagnosis of intramucosal carcinoma is likely to have even more severe insurance consequences. Avoiding this term might reduce financial strain for those affected by this challenge.6 Although guidelines currently do not recommend colectomy for intramucosal adenocarcinoma, the effects of this terminology have not been thoroughly investigated. Recent evidence suggests that the diagnosis of ‘intramucosal adenocarcinoma’ has led to an increase in colectomy surgeries.4 One study found that 80.5% of intramucosal adenocarcinomas were managed with endoscopic resection, while 19.5% required surgical intervention. Factors influencing surgical management included polyp size (≥30 mm), sessile morphology and proximal location. Importantly, no patients with intramucosal adenocarcinomas, even those managed surgically, had lymph node metastases.8 However, the study relied on aggregated data without a centralised pathology review, making distinguishing between intramucosal adenocarcinoma and high-grade dysplasia challenging. A Canadian study also found that patients with intramucosal adenocarcinoma were more likely to undergo surgical intervention after complete endoscopic resection than those with high-grade dysplasia (10.5% vs 0%, p<0.001).4 The risks of synchronous advanced neoplasia were comparable between high-grade dysplasia (24.2%), intramucosal adenocarcinoma (26.7%), and T1 colorectal carcinoma (25.4%). Similarly, the adjusted risk of metachronous advanced neoplasia was comparable between intramucosal adenocarcinoma and high-grade dysplasia (18.3% vs 26.8%). While these findings provide some insight, the results remain inconclusive, as it is challenging to ascertain the precise reasons for colectomy in these patients. For example, it is plausible that endoscopically unresectable lesions were a driving factor, particularly in the first study. Nevertheless, this raises concern that the term ‘intramucosal adenocarcinoma’ may inadvertently lead to unnecessary colectomies. Reclassifying these cases could alleviate the emotional and psychological distress associated with the term ‘cancer,’ potentially avoiding unnecessary surgical procedures, their associated risks and the financial burden on patients. The AJCC approach to intramucosal adenocarcinoma The AJCC lacks consistent terminology across different anatomical sites (table 2). For instance, pTis is characterised as high-grade dysplasia in the oesophagus. At the same time, it is referred to as CIS in the stomach, though the term high-grade dysplasia is also noted in this context. Both designations describe intraepithelial tumours without invasion of the lamina propria. The AJCC specifies that pTis in the oesophagus is defined by a tumour confined to the basement membrane—a definition that, as we demonstrate later, is scientifically inaccurate. In the colon, pTis is also characterised as CIS, as in the stomach, and is described as intramucosal carcinoma involving the lamina propria. Meanwhile, pT1a refers to tumours invading the oesophagus and stomach’s lamina propria or muscularis mucosae. In contrast, in the colon, pT1 indicates invasion into the submucosa. These inconsistencies, variable terminology and scientifically inaccurate definitions highlight the need for greater precision and standardisation in AJCC staging criteria. The concept of cancer overdiagnosis Overdiagnosis refers to identifying a condition that would not have led to symptoms or death if left undetected and may lead to overtreatment as well. 9 ,10 This phenomenon in cancer can arise from two scenarios: (1) The cancer does not progress or even regresses or (2) The cancer progresses so slowly that the patient dies from other causes before any symptoms appear. A prime example of cancer overdiagnosis, recently addressed through renaming, involves papillary thyroid carcinoma.11 The renaming of the encapsulated follicular variant of papillary thyroid carcinoma—specifically, those without vascular or capsular invasion, known for their indolent behaviour—to non-invasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) has led to a notable decrease in diagnoses classified as papillary thyroid carcinoma. NIFTP, estimated to represent 20% of thyroid cancer cases, potentially impacts the clinical care and management of about 45 000 individuals annually. The significance of nomenclature in medical diagnosis is profound. Some experts suggest adopting the term ‘indolent lesion of epithelial origin’ for lesions presently termed as cancer and their precursors, which are unlikely to cause harm if untreated, to mitigate overdiagnosis and its implications. 9 Geographic variation in the use of the term intramucosal adenocarcinoma of the colon and the Vienna classification At the risk of oversimplification, pathologists in the ‘Western’ world and Japan identified a significant gap in their interpretation of preneoplastic gastrointestinal tract lesions, with a considerable proportion of high-grade dysplasia classified as adenocarcinoma by pathologists from Asia. In Japan, carcinoma and stromal invasion are assessed separately, and CIS is an accepted diagnostic term in many organs.12 13 That approach has not caused a significant issue in Japan, as the concept of CIS is widely shared with physicians and surgeons. The financial consequences of using CIS are negligible, as Japan’s national insurance scheme covers all residents. On the other hand, UK guidelines do not recognise intramucosal carcinoma or Tis (CIS) in the lower gastrointestinal tract, so they opt for the term ‘high-grade dysplasia’. 14 At the heart of this interpretative difference lies the fact that ‘Western’ pathologists, applying WHO criteria, require invasion into the lamina propria before calling intramucosal adenocarcinoma; in contrast, Japanese pathologists consider cytological changes (variably sized and enlarged nuclei, rounded nuclei, loss of polarity, prominent nucleoli) and architectural changes (complex budding or branching of glands, back to back glands) sufficient for a diagnosis of adenocarcinoma. The Vienna classification was introduced to bridge differences between Japanese and Western pathologists (tables 3 and 4).15 The study aligned the classification of neoplastic lesions into five categories; the categories most relevant for this discussion are category 4: non-invasive high-grade neoplasia and category 5: invasive neoplasia. The Vienna classification initially classified intramucosal adenocarcinomas as invasive neoplasia (category 5). Still, later updates aligned it with Western views as non-invasive, high-grade neoplasia (category 4) alongside high-grade dysplasia.15 16 Placing intramucosal adenocarcinoma in category 4 aligns with our current understanding of colorectal carcinoma biology: tumours confined to the mucosa seldom metastasise and can be managed conservatively. Nevertheless, the term intramucosal adenocarcinoma, while justified in another era, needs to be challenged by two developments over the last two decades: (1) screening programmes have dramatically increased the number of diagnoses of intramucosal adenocarcinoma and (2) it is incorrect to equate colorectal with gastric or oesophageal intramucosal adenocarcinomas, sites associated with a much higher incidence of metastatic disease when the lesion is intramucosal. Distinguishing high-grade dysplasia from intramucosal adenocarcinoma Distinguishing high-grade dysplasia from intramucosal adenocarcinoma is often challenging due to significant overlap in features. High-grade dysplasia typically shows simpler, elongated tubular glands, whereas intramucosal adenocarcinoma is usually diagnosed based on greater architectural complexity, more pronounced loss of polarity and luminal necrosis. While these distinctions may appear theoretically sound, the overlap between the two categories frequently complicates reproducibility in diagnosis. This group argues that the presence of complex or cribriform architecture without definitive evidence of single-cell infiltration, tumour budding or poorly differentiated clusters should not be considered sufficient for a diagnosis of intramucosal adenocarcinoma. Such cases are more appropriately classified as high-grade dysplasia. Interobserver variability further highlights these challenges. In one study, interobserver agreement for architectural features yielded a kappa value of 0.47, and for dysplasia grade, a kappa of 0.57, reflecting only moderate agreement.17 Similarly, a study reviewing 104 advanced adenomas from the Dutch cancer screening programme found that five cases were upgraded to adenocarcinoma or suspected adenocarcinoma, while 11 were downgraded to low-grade dysplasia.18 The challenges distinguishing between high-grade dysplasia and intramucosal adenocarcinoma are not confined to the lower gastrointestinal tract but are equally prominent in the upper tract. A study on Barrett’s oesophagus-related dysplasia found only moderate agreement for high-grade dysplasia (κ=0.47) and fair agreement for intramucosal adenocarcinoma (κ=0.30) despite using standardised histological criteria.19 One notable misconception distinguishing high-grade dysplasia from intramucosal carcinoma hinges on carcinomas’ purported absence of a basement membrane.20 Contrary to this belief, it has been observed that a neo-basement membrane develops around intramucosal carcinoma in the oesophagus, although this phenomenon has not been demonstrated in the colon. Notably, while there are discernible distinctions in the basement membranes between non-dysplastic glands, dysplastic glands and invasive adenocarcinomas, these variations are not detectable using standard H&E staining techniques.20 Risk of metastasis with colorectal intramucosal adenocarcinoma The largest body of literature addressing colorectal intramucosal adenocarcinoma comes from the SEER database, which analysed 109 953 carcinoma cases. 2383 cases (2.17%) were identified as intramucosal carcinomas or pTis.21 Notably, 95 of these cases reported N1 disease (1–3 positive lymph nodes) and 19 cases of N2 disease (≥4 positive lymph nodes), totalling 114 cases (4.8%) of pTis with lymph node metastases. Survival outcomes were reported as follows: 5-year survival for pTis N0 was 95%, for pTis N1 was 87%, and for pTis N2 was 83%. While these findings might seem alarming, the SEER database does not clarify whether the deaths in these groups were due to disease progression or unrelated causes. More importantly, these results highlight a critical challenge: whether these cases were indeed pTis. It is difficult to confirm their classification without a thorough and centralised histological review of each case. Among the cases with relatively detailed histologic descriptions or accompanying histological images, metastasis was reported in only five cases of intramucosal adenocarcinoma. Table 5 Virtually all of these patients exhibited alarming histological features, including desmoplasia, signet ring cells, solid, uninterrupted sheets of cells, lymphatic invasion and tumour budding. Risk of metastasis in oesophageal and gastric intramucosal adenocarcinoma While drawing parallels between the colorectal and upper gastrointestinal carcinomas might offer insights, it is essential to recognise the significant differences between these cancer types. There are notable discrepancies in their biological behaviour, risk factors and the surrounding microenvironment. Therefore, while studying intramucosal adenocarcinomas across various locations can be informative, it is critical to acknowledge and consider each type’s specific characteristics and caveats. Oesophageal intramucosal adenocarcinoma A recent American Gastroenterological Association practice update reported that the risk of lymph node metastasis for pT1a intramucosal adenocarcinoma is less than 2%, supporting the argument for conservative endoscopic therapy over oesophagectomy in these cases.22 However, a review of the literature suggests that the incidence of lymph node metastasis in patients with oesophageal intramucosal adenocarcinoma is more variable, with one series reporting 17% of high-risk pT1a patients with metastasis table 6 .23–27 Common patterns that may prompt the diagnosis of intramucosal adenocarcinoma of the colorectum Defining common patterns of intramucosal adenocarcinoma is challenging, as there is no consensus on what constitutes intramucosal carcinoma or on the boundaries between high-grade dysplasia and intramucosal cancer. However, several common histological patterns have emerged based on the literature and our practices. It is important to emphasise that the discussion here pertains exclusively to the colorectal, not other organs. Cribriform architecture: This pattern is characterised by the partial to complete extinction of the lamina propria and the fusion of glands (figure 1). The tumour cells exhibit vesicular nuclei with prominent nucleoli and loss of polarity. Notably, this presentation does not include isolated single cells or tumour buds. Poorly differentiated component: This variant shows tumour cell clusters lacking glandular formation. It encompasses tumours exhibiting budding and poorly differentiated clusters, without numerical restrictions on the number of poorly differentiated clusters, tumour budding or single tumour cells (figure 2). The presence of any of these features qualifies for this pattern. Mucinous component: Tumours in this category contain cells floating in mucin, singly or in clusters. Although reports of this entity are rare, it is critical to distinguish it from extracellular mucin accumulation, which is commonly seen in adenomatous colorectal lesions. Neuroendocrine carcinoma: This rare variant is characterised by distinct tumour clusters that differ from the epithelial component and stain positively for chromogranin and/or synaptophysin. Signet ring cells: Signet ring cells are infiltrating the lamina propria (figure 3). Lymphovascular invasion: This pattern acknowledges the propensity for the mucosal epithelium to develop lymphatic channels and signifies invasion of these newly formed channels by tumour cells (figure 4). This includes examples where the neoplasm, although confined to the mucosa, shows submucosal vascular invasion. Intramucosal adenocarcinomas can access lymphatic channels: D2-40-positive lymphatic vessels are identifiable within the mucosa, providing a route for tumour cells to reach submucosal lymphatics. In the largest series, Hashimoto et al documented intramucosal CRCs with submucosal lymphovascular invasion, no nodal metastasis, although these tumours showed enrichment for high-grade tumour budding and an ‘ERI (eosinophilic cytoplasm with round nuclei and neutrophilic inflammation)’ pattern .31 Complementing this, Rodrigo-Calvo et al showed that pTis lesions harboured mucosal lymphatics in 100% of cases and that, despite pN0 by H&E, 28% had molecular lymph node tumour burden as detected by low-level CK19 mRNA copies.32 Taken together, these data support managing intramucosal lesions with lymphovascular involvement as T1-equivalent disease (formal resection when fit) or, if surgery is deferred, with close multidisciplinary surveillance. Desmoplasia: An unequivocal desmoplastic stromal reaction is a strong indicator of submucosal invasion, warranting a thorough evaluation, including multiple levels and, if necessary, a request for additional material to rule out an invasive process (figure 5).33 34 However, it is essential to note that an altered stromal reaction—often secondary to prior biopsy—can show overlapping features with fibrosis associated with biopsy sites. Recommendations In experience, the most common reason for diagnosing colorectal intramucosal adenocarcinoma is the presence of cribriform architecture. The authors of this manuscript advocate for referring to such lesions as high-grade dysplasia. While this pattern may justify a diagnosis of intramucosal adenocarcinoma in the upper gastrointestinal tract, such as the oesophagus and stomach, this classification should not apply to colorectal neoplasia. The authors endorse that intramucosal neoplasms exhibiting tumour budding, poorly differentiated components and signet ring cells can be appropriately classified as intramucosal adenocarcinomas. Among these, there is strong support for designating tumours with lymphovascular invasion—whether in the mucosa or submucosa—as intramucosal adenocarcinoma. Although available data are limited, it would be imprudent, in the interest of avoiding conflation of tumour type with stage, to reclassify these lesions as invasive adenocarcinomas. We therefore advocate designating them as intramucosal adenocarcinomas. Nonetheless, because of their potential for lymph-node or systemic spread, such cases warrant a distinct management approach and should be treated clinically as T1-equivalent tumours. Intramucosal adenocarcinomas with unequivocal desmoplasia should be presumed to represent a submucosal invasive neoplasm. Intramucosal neoplasms showing associated neuroendocrine carcinoma are exceptionally rare. Nonetheless, the authors believe such tumours should also be included under the term intramucosal carcinoma and typed according to the proportions of the respective morphologies. Intramucosal neoplasms with mucinous differentiation are exceptionally rare. The authors believe such tumours should also be included under the term intramucosal adenocarcinoma and typed according to the proportions of the respective morphologies. One potential reason for using the term intramucosal adenocarcinoma is a biopsy with inadequate sampling from a patient with a clinically overt invasive carcinoma, such as a large ulcerated or annular lesion. The term ‘at least high-grade dysplasia’ may be used in such cases. However, we discourage using the term adenocarcinoma in this context. Instead, a note such as ‘The clinical/endoscopic impression of a mass is noted and is worrisome for a possibility of an unsampled malignancy, especially given the superficial biopsy fragments. If clinical suspicion of malignancy is high, then repeat deeper biopsies should be considered for a definitive diagnosis.’ may be added. The term adenocarcinoma is discouraged since it then becomes part of the medical record. It may be valuable to distinguish between microsatellite instability-associated intramucosal lesions, particularly those in Lynch syndrome. Early lesions in Lynch syndrome, often poorly differentiated and flat, may exhibit unique characteristics distinct from those observed in sporadic settings. For example, intramucosal lesions with signet ring cells or tumours resembling gastric adenocarcinomas—characterised by anastomosing glands without desmoplastic stroma—should raise suspicion for Lynch syndrome.35 Notably, such features have also been observed in other hereditary cancer syndromes, such as Li-Fraumeni syndrome, and in inflammatory bowel disease, emphasising the importance of considering correlating with the clinical context when these findings are encountered and considering a syndromic context when these findings are encountered.

The guidelines proposed in this paper require validation and reflect the authors’ perspectives. They aim to foster dialogue that inspires professional societies to develop standardised recommendations. We advocate caution in using the term intramucosal adenocarcinoma. This approach minimises overtreatment and mitigates the emotional and psychological distress associated with a cancer diagnosis. The term intramucosal adenocarcinoma should be reserved for poorly differentiated tumours, including those with signet ring cells, tumour budding, poorly differentiated clusters, lymphovascular invasion or specific subtypes such as mucinous adenocarcinoma (table 8). The term intramucosal adenocarcinoma would also be preferable for neoplasms with neuroendocrine carcinoma. Intramucosal adenocarcinoma should not apply to tumours with cribriform architecture, partial loss of the lamina propria or high-grade cytological atypia in moderately differentiated tumours. This is particularly relevant when patients access their pathology reports directly, which can lead to misinterpretation or distress due to information encountered online. These recommendations align with the International Collaboration on Cancer Reporting (ICCR) guidelines, which classify colorectal neoplasms without invasion as non-invasive neoplasia/high-grade dysplasia due to their negligible metastatic risk. We support their approach and stress the importance of reviewing such cases within a multidisciplinary team. The ICCR guidelines also include tumours with poorly differentiated or signet-ring cell morphologies in the non-invasive neoplasia category.36 Although intraepithelial neoplasia was considered a potential alternative to mitigate the consequences of labelling mucosal lesions as carcinoma, its use introduces significant conceptual and practical issues. The term is not widely adopted in routine pathology practice. Furthermore, it does little to resolve the core diagnostic ambiguity surrounding intramucosal adenocarcinoma and instead risks adding confusion for both pathologists and clinicians. Its historical use was primarily aimed at replacing inconsistent dysplasia terminology rather than clarifying the biological behaviour or management of early colorectal neoplasia. Given these limitations, and to promote clarity and consistency, we recommend avoiding the term intraepithelial neoplasia in this context and confining diagnostic terminology to intramucosal adenocarcinoma where justified by specific histologic features. The US Multi-Society Task Force on Colorectal Cancer advises against using terms like cancer or carcinoma for colorectal lesions confined to the mucosa.37 Our conclusions are consistent with the study by Vink-Börger et al, who advocate for retaining intramucosal adenocarcinoma but with a strict definition.38 Their definition requires features such as tumour budding or signet ring cells and limits the diagnosis to excisional specimens. By delineating specific histological patterns as compatible with a diagnosis of intramucosal adenocarcinoma, we aim to promote consistent diagnoses and effective management strategies. This approach balances the clinical need for accurate categorisation with the emotional, psychological and societal implications of a cancer diagnosis, ensuring that treatment decisions are both evidence-based and patient-centred. Ethics statements Patient consent for publication Ethics approval Not applicable.

Footnotes Handling editor Murali Varma. Contributors Conception: VD. Methodology and data curation: VD. Writing–original draft: VD. Writing–review, editing and paper editing: all authors. Supervision: all authors. AI was used to edit this manuscript. Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Competing interests None declared. Provenance and peer review Commissioned; externally peer reviewed.