Bowel obstruction in mismatch repair–deficient colorectal cancer with excellent response to immunotherapy: a case series Authors: Sarah Johnston, Yu Jo Chua, Alisha Azmir Published: 2026 INTRODUCTION In Australia, colorectal cancer is the fourth most commonly diagnosed malignancy and the second leading cause of cancer-related death [1]. Approximately 15% of colorectal cancers show mismatch repair deficiency (dMMR), arising from either sporadic or inherited mutations [2]. These dMMR tumors respond poorly to conventional chemotherapy. Recent evidence indicates that dMMR tumors exhibit upregulated programmed cell death-ligand 1 (PD-L1) expression [3]. This protein facilitates tumor immune evasion by delivering inhibitory signals to T cells through binding to programmed cell death protein 1 (PD-1) [4]. Pembrolizumab is a monoclonal antibody that binds PD-1 and disrupts its interaction with PD-L1, thereby allowing activation and proliferation of T cells [5]. Numerous studies have demonstrated that immunotherapy can induce a complete pathological response and improve progression-free survival in patients with dMMR colorectal cancer [6, 7]. Accordingly, PD-1 monoclonal antibodies are now considered first-line therapy for unresectable or metastatic dMMR disease. Following the increased use of pembrolizumab for unresectable or metastatic dMMR colorectal adenocarcinoma at our institution, we observed that several patients developed bowel obstruction due to persistent luminal narrowing at the primary site despite an excellent oncological response. We present this case series to improve recognition and management of this rare but clinically significant complication that may require surgical intervention. CASE PRESENTATION Case 1 A 40-year-old woman was diagnosed with dMMR adenocarcinoma of the distal transverse colon with extensive metastatic disease involving the intra-abdominal lymph nodes and liver. Computed tomography (CT) demonstrated a thick-walled colonic lesion with shouldering and luminal narrowing (Fig. 1A). Colonoscopy could not traverse the primary lesion. She completed 7 cycles of pembrolizumab over 5 months before experiencing recurrent presentations with partial large bowel obstruction. Interval CT demonstrated an apple-core lesion in the distal transverse colon with proximal dilatation of the large and small bowel (Fig. 1C). She subsequently underwent laparoscopic extended right hemicolectomy with wedge resection of the adherent stomach. Macroscopically, there was an ill-defined, deeply ulcerated fibrotic region involving one-third of the circumference, with associated serosal puckering. Microscopically, a tiny focus of poorly preserved residual invasive adenocarcinoma was present, with a fibroinflammatory tissue reaction adjacent to the area of ulceration. The primary tumor showed a near-complete response to immunotherapy, with a tumor regression grade (TRG) of 1 and a final pathological stage of ypT3N1b. Pembrolizumab was recommenced postoperatively. Fluorodeoxyglucose positron emission tomography (FDG-PET) performed 1 month after surgery demonstrated no residual colonic disease or metastatic disease. Case 1 images. (A) Computed tomography (CT) image at diagnosis demonstrating an obstructing distal transverse colon mass (arrow) and numerous enlarged mesenteric lymph nodes (arrowhead). (B) Fluorodeoxyglucose positron emission tomography image at diagnosis demonstrating an avid primary colonic lesion (arrow), para-aortic lymphadenopathy (star), and liver lesions (circles). (C) CT image immediately before surgery demonstrating obstruction, with dilatation of the colon and small bowel (arrowhead) and an obstructing mass in the distal transverse colon (arrow). Case 2 A 78-year-old man was diagnosed with locally advanced dMMR adenocarcinoma of the distal transverse colon, with an adherent loop of small bowel and para-aortic nodal involvement. The primary tumor could not be traversed on colonoscopy and, because of obstructive symptoms, he underwent formation of a loop ileostomy before commencing pembrolizumab. Restaging FDG-PET after 8 cycles administered over 7 months demonstrated a marked metabolic and radiological response at the primary site, with persistent uptake in the left internal iliac nodes. Repeat colonoscopy demonstrated a nontraversable stricture in the distal transverse colon. Following multidisciplinary discussion, he underwent laparoscopic subtotal colectomy, ileosigmoid anastomosis, and bilateral pelvic lymph node dissection. Macroscopically, the specimen demonstrated complete luminal occlusion in an area of serosal puckering. Microscopically, there was no residual viable adenocarcinoma (TRG 0), and the tumor bed was characterized by fibrosis and scattered inflammatory cells (Fig. 2). The final pathological stage was ypT0N0. He made an excellent recovery, and surveillance CT demonstrated no evidence of recurrent or metastatic disease. Microscopic section of hematoxylin-eosin–stained bowel wall demonstrating replacement of the tumor bed by fibrosis (original magnification ×4). Case 3 An 82-year-old woman was diagnosed with poorly differentiated dMMR adenocarcinoma of the descending colon with multiple liver metastases. Colonoscopy could not traverse the primary lesion at diagnosis. She commenced pembrolizumab and tolerated 2 cycles over 2 months before presenting with symptoms of large bowel obstruction, which resolved with conservative management. One month later, she presented again with obstruction after the third cycle and underwent placement of a colonic stent for a severe stricture with normal overlying mucosa (Fig. 3). Following stent placement, her obstructive symptoms resolved, and pembrolizumab was recommenced. Her most recent CT demonstrated a significant reduction in the size of her liver metastases. Endoscopic images. (A) Initial colonoscopic image demonstrating a circumferential malignant stricture in the descending colon. Biopsies demonstrated mismatch repair–deficient adenocarcinoma. (B) Image from flexible sigmoidoscopy after 3 cycles of pembrolizumab demonstrating severe intrinsic stenosis at the site of known malignancy in the descending colon, with normal colonic mucosa. (C) Image demonstrating successful stent deployment across the stricture. Ethics statement The patients provided written informed consent for the publication of the research details and clinical images. DISCUSSION Immunotherapy is emerging as a transformative treatment for dMMR colorectal adenocarcinoma in both metastatic and localized settings. A recent systematic review demonstrated that neoadjuvant immunotherapy in nonmetastatic dMMR tumors resulted in a ypT0N0 response in 50% to 90% of cases [8], whereas an objective response rate of 30% to 40% has been reported in metastatic dMMR disease [9]. However, as the use of immunotherapy expands, a greater understanding of its risks and complications is needed. Immune-related toxicities are commonly reported in the literature and include hypothyroidism, colitis, and pneumonitis [10]. In contrast, bowel obstruction is a rarely reported but surgically significant complication. One other case series described 9 patients who developed bowel obstruction during immunotherapy for locally advanced or metastatic dMMR colonic adenocarcinoma [11]. Similar to our findings of near-complete or complete pathological response in the 2 patients who underwent surgical resection, that series reported a pathological response in all 7 patients who underwent resection, including 4 complete pathological responses. Risk factors for obstruction in that series included hepatic flexure tumors, radiological evidence of stricturing, and endoscopic obstruction. Although all of our patients had endoscopically obstructing primary tumors and radiological evidence of stricturing, the tumors in our series were located in the distal transverse colon and descending colon. This is consistent with the variable primary tumor locations reported by Xiao et al. [12], in which 3 of 73 patients developed bowel obstruction during neoadjuvant immunotherapy for localized dMMR colorectal cancer. These findings suggest that the advanced nature of the primary tumor, rather than its location, may contribute to obstruction risk through the mechanisms proposed below. Consistent with the published literature, our data suggest that an aggressive immune response to the primary tumor, correlating with an impressive pathological response, may be associated with the development of obstruction. Limited information is available regarding the mechanism of tumor response to immunotherapy in colorectal adenocarcinoma, with 1 study involving dual cytotoxic T-lymphocyte–associated antigen 4 (CTLA-4) and PD-1 inhibitors suggesting a serosal-to-mucosal regression pattern [13]. Our resection specimens support this observation, with marked serosal puckering in the region of tumor response. This regression pattern may contribute to bowel obstruction through fibrotic stricturing extending from the serosal surface to the lumen. Another contributing factor may be replacement of tumor cells by fibrotic or fibrofatty tissue, as observed in our specimens. Together, these mechanisms may explain why patients with a locally advanced and/or colonoscopically obstructing lesion at diagnosis subsequently develop clinically obstructing lesions during immunotherapy, as replacement of tumor by fibrotic tissue may lead to an unyielding stricture. These findings also help explain why, as seen in case 2, a complete pathological response does not necessarily restore luminal patency. This study demonstrates the importance of multidisciplinary care for patients with dMMR colonic adenocarcinoma. High-risk patients whose primary tumors cannot be traversed on colonoscopy, who have imaging evidence of stricture, or who have any clinical symptoms of obstruction should be reviewed by a colorectal surgeon before commencing immunotherapy. Early review enables individualized assessment, selection of the most appropriate intervention, and patient counseling. It may also help minimize interruption of immunotherapy and reduce the morbidity associated with emergency surgery. There is little evidence to guide assessment of primary disease response to immunotherapy, which is typically evaluated radiologically or endoscopically. Endoscopy permits more precise assessment of the structural luminal response at an earlier time point [14], but bowel preparation may precipitate obstruction [15]. Until stronger evidence is available to standardize care, the timing and modality of response assessment should be individualized within the multidisciplinary team. Finally, we recommend counseling high-risk patients that a complete response to immunotherapy does not necessarily correlate to avoidance of surgical intervention. As demonstrated in our case series, an excellent oncological response does not necessarily correspond to restoration of structural patency and may even worsen clinical symptoms because of replacement of tumor by rigid fibrotic tissue. Footnotes Conflict of interest No potential conflict of interest relevant to this article was reported. Funding None. Acknowledgments The authors thank Mitali Fadia (Department of Anatomical Pathology, Canberra Hospital) for assisting with interpretation of pathological findings. Author contributions Conceptualization: all authors; Investigation: all authors; Methodology: SJ, AA; Writing–original draft: SJ; Writing–review & editing: all authors. All authors read and approved the final manuscript. REFERENCES