Diagnostic Challenges in Sessile Serrated Lesions: Progression to Adenocarcinoma in a High-Risk Patient

Introduction

Serrated polyps are precursors of up to 30% of colorectal cancer (CRC) cases [1]. According to the World Health Organization, these lesions are categorized into 3 types: hyperplastic polyps, sessile serrated lesions (SSLs), and traditional serrated adenomas (TSAs). The serrated pathway to CRC is commonly associated with BRAF proto-oncogene (BRAF) mutations and the CpG island methylator phenotype (CIMP) [2].

On endoscopic examination, SSLs can be flat and/or sessile. Once a polyp is identified, resection can be completely or incompletely performed, using a variety of techniques [2]. Histopathologically, distinguishing SSLs with extensive dysplasia from conventional adenomas remains a diagnostic challenge, particularly if piecemeal resection is done, with uncertainty existing about completion of polypectomy [2]. Once dysplasia occurs in an SSL, the progression to cancer is rapid, which underscores the importance of timely diagnosis. SSLs with dysplasia are characterized by architectural distortion and cytological atypia [2]. Key histologic features include crypt dilation at the basal third, serrations extending into the crypt base, and asymmetrical proliferation. The diagnosis of an SSL can be made even if only a single crypt shows these characteristic features [2].

In this report, we present a case of an SSL initially diagnosed as an SSA with a tubular adenoma component. Six months later, this lesion was found to have progressed to an invasive, moderately-differentiated adenocarcinoma.

Case Report

A 78-year-old man with a diagnosis of acute myeloid leukemia (AML) with janus kinase 2 (JAK2) mutation underwent chemotherapy followed by allogeneic stem cell transplantation in 2017. He had been diagnosed with ulcerative colitis in 2016 and underwent complete remission after successful stem cell transplantation. He was taking pantoprazole 40 mg daily for short-segment Barrett’s esophagus and undergoing routine colon cancer surveillance for a prior history of colon polyps.

An 11-mm, sessile, non-granulated polyp in the transverse colon was resected piecemeal via a 13-mm Boston scientific captivator single-use cold snare, labeled as “transverse colon biopsy” and submitted to the pathology department (Figure 1A, 1B). Histology documented tubular adenoma (TA) and sessile serrated adenoma (SSA) (Figures 2, 3). Three additional ascending polyps (1–3 mm) in size were removed using single-use, radial-jaw, jumbo-capacity forceps and were read as TAs. Diverticulosis was present throughout the colon. Random colonic biopsies showed normal mucosa without evidence of changes suggestive of chronic colitis.

At follow-up 6 months later, colonoscopy findings were reviewed, and a repeat colonoscopy was planned to assess for residual polyp in the transverse colon. A residual lesion was seen at the prior resection site (Figure 4). This lesion was biopsied and labeled as “colon polyp 65.0 cm from the anal verge.” Pathology reported invasive, moderately-differentiated adenocarcinoma with loss of DNA mismatch repair gene (MMR) MLH1, PMS2 protein expression, and a BRAF V600E mutation. A subsequent colonoscopy 1 month later revealed the same residual 9-mm polyp in the mid-transverse colon, which was tattooed, resected (Figure 5), and was read by the pathologist as tubulovillous adenoma with high-grade dysplasia.

One month later, he underwent robot-assisted transverse colectomy, which was converted to open laparotomy to stage his colon cancer. The transverse colon specimen documented well-differentiated adenocarcinoma with 1.7-cm submucosal invasion (Figure 6). One of 12 lymph nodes was positive for metastasis. Molecular analysis confirmed the previous findings consistent with MMR-deficient CRC.

A diagnosis of stage III colorectal cancer was established and he was started on adjuvant FOLFOX (folinic acid, fluorouracil, and oxaliplatin) chemotherapy, scheduled for 12 cycles. A surveillance CT scan of the abdomen showed no recurrence or metastasis.

Discussion

This case shows the diagnostic complexity and clinical implications of SSLs with dysplasia, particularly in the context of fragmented polypectomy and evolving histopathologic classification criteria. Initially, the piecemeal-resected proximal colon polyp in our patient was classified as a combination of SSA and TA, as the pathologist believed the specimen from “transverse colon biopsy” consisted of 2 lesions instead of 1. Upon re-examination of this pathological specimen by a second pathologist, when the diagnosis of invasive CRC was established, and after input from the endoscopist, it was confirmed that the histological features of serrated crypts and cytological atypia were conclusive for SSL with dysplasia (SSL-D), not SSA and TA. If Figures 2 and 3 were interpreted as showing 2 separate lesions, the dysplastic features in Figure 3 could be suggestive of a tubular adenoma (TA). However, when considered together as originating from a single lesion, the combination of serrated architecture (Figure 2) and cytological atypia (Figure 3) were diagnostic of SSL with dysplasia.

SSLs with dysplasia are considered advanced precursors to colorectal cancer, with a higher malignant potential than non-dysplastic SSLs. As Utsumi et al (2023) emphasized, these lesions often exhibit subtle histologic changes that may be overlooked, especially when only fragments are evaluated or when the serrated architecture is not well preserved due to piecemeal resection [1]. Histologic differentiation from conventional adenomas can be particularly challenging, necessitating expert pathology review and, ideally, examination of the polyp in its entirety. Immunohistochemical markers such as annexin A10, MLH1, and β-catenin have also provided insights into the molecular pathways and progression of SSLs with dysplasia, supporting more precise diagnosis and prognostication [3]. In our case, immunohistochemical analysis of the residual transverse specimen demonstrated loss of MLH1 and PMS2 expression, along with a positive BRAF V600E mutation – molecular findings characteristic of the serrated neoplasia pathway, further reinforcing the diagnosis of SSL-D and its progression to invasive adenocarcinoma. Additionally, recent findings have emphasized that inflammatory bowel disease-associated serrated lesions with dysplasia are frequently associated with advanced neoplasia, further supporting a unified classification approach and reinforcing the need for heightened vigilance in these cases [4].

In this case, a key missed opportunity was the lack of early interdisciplinary communication between the gastroenterologist and the pathologist regarding the initial histologic findings. Payne et al (2014) have shown that significant variability exists in the pathologic interpretation of serrated lesions across centers, particularly in the identification of SSLs and their dysplastic features [5]. Enhanced communication could have clarified whether the lesion was a mixed histology or an ulcerative single SSL with evolving dysplasia, which would have impacted both surveillance intervals and follow-up planning.

The endoscopic assessment also warrants examination. During the initial piecemeal resection, the use of image-enhanced endoscopy and tools such as the Narrow-band Imaging International Colorectal Endoscopic (NICE) classification could have provided additional clues to the presence of dysplasia. Li et al (2022) demonstrated that advanced imaging techniques, when combined with classification systems, improve detection and characterization of serrated lesions, particularly when dysplastic changes are present [2]. The absence of such standardized assessment may have contributed to the initial misclassification in our case. Murakami et al reported that unlike hyperplastic polyps, sessile serrated adenomas/polyps (SSA/Ps), which typically measure over 5 mm in size, are often found in the proximal colon, and are frequently obscured by a characteristic mucous layer. Endoscopically, features such as open type II pit patterns seen with magnifying chromoendoscopy, along with dark crypt openings and dilated, varicose-like vessels visible under magnifying narrow-band imaging, can aid in distinguishing SSA/Ps from hyperplastic polyps [6]. In hindsight, use of the cold snare polypectomy (CSP) technique for initial resection of a lesion that exhibited depressed morphology was suboptimal. While CSP is considered safe for small, non-dysplastic lesions, it is not recommended for suspected neoplastic or invasive polyps, especially those >10 mm or with concerning surface pit patterns [7]. Under white-light endoscopy (WLE), SSL-Ds tend to demonstrate more conspicuous features compared to non-dysplastic SSLs. Specifically, SSL-Ds often have a depressed surface, disrupted or irregular vascular patterns, and more pronounced granularity or nodularity. These findings contrast with the smoother, pale, mucus-covered appearance typically seen in non-dysplastic SSLs. Lesions with dysplasia or carcinoma frequently exhibit elevated areas with reddish discoloration, erosion, or disrupted pit patterns – features that magnifying endoscopy can detect with greater clarity than standard WLE, thereby enhancing diagnostic confidence and allowing the use of the most appropriate resection techniques [8]. A large multicenter colonoscopy-based study by Burnett-Hartman et al found that older age, male sex, and smoking are independently associated with the presence of serrated polyps, including those with dysplastic changes, further supporting the importance of risk stratification in screening and surveillance [9].

Emerging technologies also play a role in improving detection. The combined use of artificial intelligence (AI) tools like GI Genius with adjunctive devices such as EndoCuff has been shown to significantly improve the detection of small and right-sided adenomas and sessile serrated lesions in real-world settings [10]. This suggests that the integration of such tools into routine colonoscopy practice could enhance detection rates for subtle and easily missed lesions like SSLs [10]. AI-based tools integrated into real-time colonoscopy platforms have demonstrated utility in identifying subtle mucosal patterns and neoplastic polyps. Facanali et al reported a rare case of a sessile serrated lesion overlying a submucosal lipoma, in which even with the aid of AI technology it was difficult to distinguish a hyperplastic polyp from a serrated polyp. This highlights how atypical presentations can remain challenging to diagnose despite advanced tools [11]. The incorporation of AI systems in colonoscopic evaluation – especially when used alongside image-enhanced and magnifying endoscopy – may help overcome human error in the visual assessment of serrated lesions, particularly those with dysplastic features that are easily overlooked under white-light endoscopy alone [12]. As AI becomes more refined and accessible, its integration into routine surveillance colonoscopy could play a pivotal role in improving early detection and preventing interval cancers associated with SSL-D.

Furthermore, the management of malignant or potentially malignant polyps requires a structured, evidence-based approach. Rex et al (2019) stated that the decision-making process from initial endoscopic resection to histologic review and potential surgical referral hinges on coordinated communication among gastroenterologists, pathologists, and surgeons [7]. In fragmented resections, such coordination becomes even more critical, as incomplete pathology due to piecemeal resection can obscure the true nature of the lesion.

In retrospect, labeling the biopsy specimens as components of a single polyp, along with early engagement of pathology to review and clarify the complex histologic findings, would have improved the diagnostic accuracy. In our case, the lesion’s non-granular, sessile appearance and transverse colon location – both known risk factors for malignancy [1] – should have prompted more thorough assessment using submucosal lift to evaluate for deeper invasion. Injection-lift and hot snare resection or en bloc endoscopic mucosal resection (EMR) would have been more appropriate. A randomized trial by Jacques et al demonstrated that en bloc resection led to significantly lower recurrence rates and improved complete resection rates for large nonpedunculated colonic adenomas, reinforcing the importance of selecting appropriate resection strategies for lesions with malignant potential [13]. Furthermore, a clinicopathological study comparing CSP, EMR, and underwater EMR (UEMR) in the treatment of SSLs found that CSP was associated with higher incomplete resection rates, especially in lesions ≥10 mm, while EMR and UEMR had superior efficacy for larger and potentially dysplastic lesions [14]. Surgical referral is still recommended if submucosal invasion or high-risk features are present [15]. This oversight highlights the need for a structured diagnostic algorithm when encountering proximal colon lesions with subtle atypia, particularly given the high malignant potential of sessile serrated lesions with dysplasia (SSL-D) [7].

We are not sure if any of these interventions would have led to the early diagnosis of cancer; however, it is entirely possible that the lesion was already cancerous at the time of the first colonoscopy. While this patient’s colorectal cancer was not directly related to inflammatory bowel disease (IBD)-associated dysplasia, his history of allogeneic bone marrow transplantation and remote ulcerative colitis placed him at an elevated baseline risk for gastrointestinal malignancies. Studies have shown that recipients of blood or marrow transplants are at increased risk for malignant neoplasms of the gastrointestinal tract, including colorectal cancer, potentially accelerated by chronic inflammation or immunosuppression [16]. These risk factors should have prompted heightened vigilance during the initial resection – particularly in the setting of an endoscopically misclassified lesion in the transverse colon. A more cautious approach, including en bloc resection and early multidisciplinary communication, may have improved diagnostic accuracy. Taken together, this case of SSL-to-cancer transformation underscores the critical importance of incorporating patient-specific risk factors into endoscopic and pathological decision-making to prevent delayed recognition of advanced precursor lesions.

Conclusions

This case emphasizes the diagnostic and management challenges associated with sessile serrated lesions (SSLs), particularly when dysplasia is present, and piecemeal resection limits histologic clarity. The initial misclassification of a single lesion as a combination of sessile serrated adenoma and tubular adenoma underscores the importance of complete resection, careful labeling, and collaborative communication between endoscopists and pathologists. Incorporating standardized endoscopic classification systems, such as the NICE criteria, during resection can aid in the real-time identification of dysplastic features. Early interdisciplinary discussion and unified interpretation of pathology are critical to guiding appropriate surveillance intervals and reducing the risk of misdiagnosis. This case demonstrates the need for a structured, team-based approach to optimize outcomes in patients with complex or high-risk colorectal lesions.