Short-term surgical outcomes following neoadjuvant immunotherapy in mismatch repair-deficient colorectal cancer: initial experience from a tertiary referral center

Abstract

Purpose

Immunotherapy has demonstrated remarkable efficacy in mismatch repair-deficient (MMR-D) colorectal cancer (CRC). Due to their significant response rates, immune checkpoint inhibitors have emerged as a promising neoadjuvant therapy. However, data regarding short-term surgical outcomes following immunotherapy remain limited. The aim of this study is to evaluate the safety and feasibility of surgical resection after immunotherapy, as well as its short-term clinical outcomes.

Methods

A retrospective review of prospectively collected data was performed at a tertiary referral center from January 2020 to July 2024. Fifteen consecutive patients with MMR-D CRC treated with pembrolizumab were analyzed. The patients’ demographics, tumor characteristics, clinical outcomes, and histopathological responses were assessed.

Results

In total, 15 patients diagnosed with MMR-D locally advanced or metastatic colorectal cancers received neoadjuvant immunotherapy followed by surgery. Of the 15 patients, 11 (73.3%) were male, 12 (80.0%) presented with T3/T4 tumors, and 3 (20.0%) had metastatic disease at diagnosis. The median number of immunotherapy cycles was 5 (range, 3–13). Surgery was performed without any anastomotic leaks or 30-day mortality. The median length of hospital stay was 5 days (range, 3–14 days). All surgical specimens had negative resection margins. Major pathological response was observed in 11 patients (73.3%), including complete response in 8 (53.3%) and near-complete response in 3 (20.0%). The median follow-up was 14 months (range, 1–56 months). One patient developed liver metastasis, which was successfully resected.

Conclusion

Surgical resection of MMR-D CRC following neoadjuvant immunotherapy is safe and associated with low morbidity. Neoadjuvant immunotherapy in MMR-D CRC facilitates high rates of major pathological response.

INTRODUCTION

Colorectal cancer (CRC) is one of the most prevalent malignancies worldwide, accounting for significant cancer-related morbidity and mortality [1]. Mismatch repair deficiency (MMR-D), present in approximately 15% of CRC cases, is associated with high microsatellite instability (MSI-H) and increased responsiveness to immune checkpoint blockade therapy. Recent advances in immunotherapy have transformed treatment paradigms, especially for MMR-D tumors, which exhibit increased neoantigen loads and robust immune infiltration, making them ideal candidates for checkpoint inhibitor therapy [2, 3]. The pivotal KEYNOTE-177 trial demonstrated superior progression-free survival with pembrolizumab compared to standard chemotherapy in MMR-D metastatic CRC [3].

Similarly, the NICHE-2 trial revealed that neoadjuvant immunotherapy resulted in a 68% complete pathological response, raising the possibility of organ preservation strategies [4]. However, data on surgical outcomes post-immunotherapy, including perioperative complications and histopathological responses, remain limited. This study aims to evaluate the short-term surgical outcomes in patients who underwent surgical resection following neoadjuvant immunotherapy.

METHODS

Ethics statement

The study was approved by the Medical Research Center, Hamad Medical Corporation (No. MRC-01-24-496). The requirement for informed consent was waived as this was a retrospective review of prospectively collected data.

Study design and patient selection

This was a retrospective review of prospectively collected data conducted at a tertiary referral center. Patients diagnosed with locally advanced or metastatic MMR-D CRC who received neoadjuvant pembrolizumab between January 2020 and July 2024, followed by surgery, were included. All patients were discussed at a multidisciplinary team (MDT) meeting before commencement of therapy.

Treatment protocol

To determine MMR status, immunohistochemical stains with appropriate controls were performed on paraffin blocks in all cases to detect DNA mismatch repair proteins (MLH1, MSH2, MSH6, and PMS2) and classified as MMR-D by the absence of expression of these proteins. Tumors with MLH1 or PMS2 loss with BRAF^V600E mutation and/or MLH1 hypermethylation, in the absence of an MMR germline mutation, were considered sporadic. After pathological confirmation of MMR-D status, the MDT recommended the immunotherapy regimen, which was administered by the oncologist. All patients received intravenous pembrolizumab (200 mg) as a single agent every 3 weeks. The median number of immunotherapy cycles was 5 (range, 3–13), depending on patient tolerance and based on interval response assessments conducted with computed tomography (CT) scans every 3 months. Following completion of the immunotherapy regimen, patients underwent a final assessment with CT scans of the abdomen, pelvis, and chest, as well as colonoscopy, before being discussed again in the MDT prior to surgery.

Surgical and histopathological evaluation

Surgical intervention was performed after the MDT reached a consensus. Resection specimens were evaluated for pathological response using the modified Ryan tumor regression grading system, as recommended by the College of American Pathologists (CAP) [5]. Major pathological response was defined as no tumor cells or less than 10% tumor cells in the specimen. Immune-related adverse events were reported according to the Common Terminology Criteria for Adverse Events (CTCAE) ver. 5.0 (US National Cancer Institute) [6].

Outcome measures

Primary outcomes included perioperative morbidity (Clavien-Dindo classification), length of hospital stay, and blood loss. Secondary outcomes included the pathological response rate, recurrence, and overall survival during follow-up.

Follow-up

Patients received follow-up in surgical oncology clinics 2 weeks after surgery to detect any postoperative complications. Subsequently, follow-up consisted of clinical examination with tumor markers and CT scans every 3 months for 2 years, and then every 6 months for the next 3 years, to detect any recurrence.

Statistical analysis

Frequencies and percentages were calculated for categorical variables, while median and range were calculated for continuous variables.

RESULTS

Patient and tumor characteristics

Of the 15 patients, 11 (73.3%) were male, and the median age was 43 years (range, 33–68 years). Three patients (20.0%) had metastatic disease at diagnosis. Tumor staging revealed that 12 (80.0%) had T3/T4 tumors. Most tumors were located in the right colon (Table 1). Following immunotherapy, reassessments were carried out with CT scans and endoscopy (Figs. 1, 2).

Table 1.

Baseline characteristics (n=15)

Characteristic	Value
Age (yr)	43 (33–68)
Sex
Male	11 (73.3)
Female	4 (26.7)
Tumor site
Cecum	5 (33.3)
Ascending	1 (6.7)
Transverse	3 (20.0)
Sigmoid	1 (6.7)
Hepatic flexure	1 (6.7)
Splenic flexure	1 (6.7)
Synchronous tumor (hepatic and splenic flexure)	1 (6.7)
Recurrence at ileocolic anastomotic site	1 (6.7)
Liver metastases from colorectal primary	1 (6.7)
T category
T1 or T2	3 (20.0)
T3	2 (13.3)
T4	10 (66.7)
Nodal status
Positive	5 (33.3)
Negative	10 (66.7)
Metastasis
Yes	3 (20.0)
No	12 (80.0)
Mismatch repair deficiency
MLH1 or PMS2, or both	10 (66.7)
MSH2 or MSH6, or both	5 (33.3)
Lynch syndrome	4 (26.7)
Histological grade
Well differentiated	1 (6.7)
Moderately differentiated	7 (46.7)
Poorly differentiated	7 (46.7)

Values are presented as median (range) or number (%). Percentages may not total 100 due to rounding.

Fig. 1.

Axial view of contrast-enhanced computed tomography. (A) Pre-immunotherapy. The arrow indicates the tumor invading the abdominal wall. (B) Post-immunotherapy. The arrow indicates the residual mass in the ascending colon.

Fig. 2.

Endoscopy images. (A) A tumor in the ascending colon is visible pre-immunotherapy. (B) The residual tumor is visible post-immunotherapy.

Surgical outcomes

After completion of immunotherapy, the median time to surgery was 5 weeks (range, 2–13 weeks). Twelve patients underwent resections using a minimally invasive approach, with laparoscopy in 11 cases and robotics in 1. No conversions to open surgery were recorded. Three patients underwent open resections: 1 right hemicolectomy with Whipple operation, 1 right hemicolectomy with peritonectomy, and 1 partial hepatectomy. There were no intraoperative complications. One patient experienced postoperative bleeding, managed conservatively with transfusion. No anastomotic leaks were reported. The median length of hospital stay was 5 days (range, 3–14 days). No 30-day mortality occurred (Table 2).

Table 2.

Surgical outcomes following neoadjuvant immunotherapy (n=15)

Characteristic	Value
Surgical procedure
Laparoscopic right hemicolectomy	8 (53.3)
Laparoscopic subtotal colectomy	1 (6.7)
Laparoscopic left hemicolectomy	1 (6.7)
Liver resection for metastatic colorectal cancer	1 (6.7)
Right hemicolectomy with peritonectomy	1 (6.7)
Robotic excision of local recurrence with total hysterectomy and bilateral salpingo-oophorectomy	1 (6.7)
Right hemicolectomy and Whipple operation	1 (6.7)
Laparoscopic anterior resection	1 (6.7)
Intraoperative blood loss (mL)	50 (50–400)
Complication (Clavien-Dindo classification)
Grade I–II	1 (6.7)
Grade III–IV	0 (0)
Length of stay (day)	5 (3–14)
No. of lymph nodes	27 (15–66)
Pathological response
Complete	8 (53.3)
Near complete	3 (20.0)
Partial	4 (26.7)
None	0 (0)

Values are presented as number (%) or median (range). Percentages may not total 100 due to rounding.

Pathological and oncological outcomes

Histopathological examination revealed a major pathological re­sponse in 11 cases (73.3%), with complete pathological response in 8 (53.3%), near-complete response in 3 (20.0%), and partial response in 4 (26.7%). None of the patients had poor or no response. The median number of cycles required to achieve complete pathological response was 5 (range, 3–7) as shown in Table 3. All resected specimens had negative margins (R0 resection). At a median follow-up of 14 months (range, 1–56 months), 1 patient developed liver metastases, which were subsequently resected with no evidence of recurrence to date.

Table 3.

Details of the patients who underwent surgery following neoadjuvant immunotherapy (n=15)

Patient no.	Age (yr)	Sex	Tumor location	Initial staging	Mismatch repair status	No. of cycles received	Radiological response	Tumor regression score	Final staging
1	65	Male	Cecum	T4NxM0	MLH1, PMS2, BRAF-negative	5	Near-complete response	0	ypT0N0
2	37	Male	Cecum	T4NxM0	MSH2	7	Partial response	0	ypT0N0
3	35	Male	Hepatic flexure	T4NxM0	MSH2	4	Complete response	0	ypT0N0
4	44	Female	Splenic flexure	T4N1M0	MLH1, PMS2, BRAF-negative, Lynch-positive	5	Stable disease	0	ypT0N0
5	43	Male	Transverse	TxN1M0	MLH1, PMS2, BRAF-negative	5	Partial response	0	ypT0N0
6	37	Male	Liver metastasis	T3N1M1	MSH2, MSH6, BRAF-negative	5	Partial response	0	ypT0N0
7	40	Male	Hepatic flexure and splenic flexure	T3N1M0	MSH2, MSH6, BRAF-negative	4	Complete response	0	ypT0N0
8	33	Male	Cecum	TxNxM1	MLH1, PMS2, BRAF-negative, Lynch-positive	5	Partial response	0	ypT0N0
9	44	Female	Transverse	T4NxM0	MSH2, MSH6	3	Near-complete response	1	ypT0N1a
10	48	Female	Anastomotic site	T4NxM0	MLH1, PMS2, Lynch-positive	4	Partial response	1	Few metastatic cells with mucin in the mass
11	40	Female	Sigmoid	T4NxM0	MLH1, PMS2, Lynch-positive	5	Partial response	1	ypT3N1
12	34	Male	Transverse	T4NXM0	MLH1, PMS2, BRAF-negative	5	Disease progression	2	ypT3N1
13	68	Male	Cecum	T4NxM0	MLH1, PMS2	13	Partial response	2	ypT3N0
14	51	Male	Ascending colon	T2N1M1	MLH1, PMS2	4	Partial response	2	ypT2N0
15	50	Male	Cecum	T4NxM0	MLH1, PMS2, BRAF-negative	10	Stable disease	2	ypT1N1

Immunotherapy-related side effects

Six patients (40.0%) experienced immunotherapy-related adverse effects. Four patients (26.7%) developed large bowel obstruction during the treatment period: 2 after the second cycle and the other 2 after the third cycle of pembrolizumab. Two patients were managed conservatively and two underwent successful endoscopic stenting. One patient developed grade 2 thyroid disorder and one developed grade 1 hepatitis. None of the patients developed grade 4 immunotherapy-related side effects.

DISCUSSION

Immune checkpoint blockade by programmed cell death 1 (PD-1) receptor inhibitors and cytotoxic T-lymphocyte–associated protein 4 (CTLA-4) inhibitors has shown promising outcomes in both early-stage and advanced MMR-D CRC [2, 4]. This study highlights the efficacy of neoadjuvant immunotherapy in achieving high pathological response rates in MMR-D CRC and confirms its safety from a surgical perspective.

In our study cohort, we used single-agent immunotherapy. Various studies have employed different regimens of immunotherapy, either as monotherapy or in combination, with variable results [2, 3]. The US National Comprehensive Cancer Network (NCCN) guidelines (ver. 5.2024) recommend neoadjuvant immunotherapy, specifically single-agent pembrolizumab or nivolumab±ipilimumab, over chemotherapy for MMR-D locally advanced colon cancers as well as for inoperable tumors [7]. However, it remains unclear which drug or regimen is superior, as the evidence is still evolving [8].

In our study, the median number of cycles to achieve complete pathological response was 5, which is comparable to published literature [9]. The NICHE-3 study demonstrated a 68% complete response rate after 2 cycles of nivolumab/relatlimab, suggesting the possibility of new treatment strategies and early reassessments during treatment, potentially obviating the need for prolonged therapy and its side effects.

One adverse event associated with immunotherapy is the development of bowel obstruction, reported to range from 1.5% to 22% [10, 11]. In our series, 4 of 15 patients (27%) developed large bowel obstruction during the treatment period. In all of these cases, pretreatment colonoscopy either could not traverse the lesion or did so with difficulty. A possible explanation for this is that the majority of patients (87%) in our cohort had locally advanced tumors, leading to treatment-related fibrosis (Fig. 3). Following resection, 3 of 4 patients (75%) had a major pathological response. A recently published series demonstrated complete pathological response in 44% of patients who developed bowel obstruction, indicating that the development of obstruction is due to fibrosis and stricture formation related to tumor response, rather than disease progression [10].

Fig. 3.

Post-immunotherapy large bowel obstruction followed by resection with complete pathological response (score 0). Surface mucosal ulceration with granulation tissue formation (black arrow) and transmural fibrosis (white arrow) is shown (hematoxylin-eosin, original magnification ×10).

In locally advanced CRCs, achieving an R0 resection margin is particularly challenging, as it requires en bloc tumor excision, often with multivisceral resection. In such tumors, R0 resection rates range from 60% to 92%, which has a significant impact on survival [12]. Data from the FOXTROT trial showed that with neoadjuvant chemotherapy, R0 resection is achievable in up to 94% of patients [13]. However, MMR-D tumors respond poorly to chemotherapy. Therefore, neoadjuvant immunotherapy is advocated in such cases. A recent meta-analysis showed a 97% R0 resection rate in MMR-D locally advanced colon cancers following neoadjuvant immunotherapy [14]. In our study, complete resection of the primary tumor with negative margins was achieved in 100% of cases.

The majority of patients underwent resections utilizing a minimally invasive technique, with no conversions to open surgery. There were no intraoperative complications. The absence of anastomotic leaks and low postoperative complication rates suggest that prior immunotherapy does not compromise surgical safety.

In our cohort, the median time from completion of immunotherapy to surgery was 5 weeks (range, 2–13 weeks), which is comparable to the published literature [4]. Currently, there is no consensus regarding the optimal timing of surgery following immunotherapy, and whether this timing affects tumor response remains to be determined.

As evidence for the use of immunotherapy grows, various therapeutic strategies and drugs are being investigated to optimize outcomes, as summarized in Table 4 [4, 8, 9, 15–21]. The NICHE-3 study demonstrated a 68% complete pathological response in locally advanced MMR-D colon cancers treated with a combination of nivolumab and relatlimab [15]. Cercek et al. [22] reported a 100% complete clinical response in 41 patients with locally advanced rectal cancer treated with dostarlimab, with half maintaining a sustained clinical response at a median follow-up of 28.9 months. This brings new hope and represents a paradigm shift in the management of MMR-D colorectal cancers.

Table 4.

Summary of the published studies and outcomes following neoadjuvant immunotherapy

Study	No. of cases	Type of immunotherapy	Duration of treatment (wk)	No. of complete clinical response (%)	No. of patients who underwent surgery	No. of postoperative complications	No. of R0 resection(%)	No. of complete pathological response (%)	No. of immune-related adverse events (%)
									Grade 1–2	Grade 3–4
Pan et al. [16] (2024)	11	Nivolumab+ipilimumab	6–12	NA	11	0 (0)	NA	9 (81.8)	7 (63)	0 (0)
Deng et al. [17] (2024)	20	Pembrolizumab/sintilimab/tislelizumab, nivolumab+ ipilimumab	6–12	2 (10)	19	0 (0)	NA	15 (78.9)	9 (45)	0 (0)
Xiao et al. [18] (2023)	73	Pembrolizumab, nivolumab, sintilimab, toripalimab, camrelizumab, tislelizumab, ipilimumab	6–52	17 (23.2)	50	4 (8)	49 (98)	28 (56)	0 (0)	8 (10)
Pei et al. [8] (2023)	11	Sintilimab	18	3 (27.2)	11	0 (0)	NA	10 (90.9)	8 (72)	0
Li et al. [19] (2024)	24	Anti-PD-1 therapy	NA	3 (12.5)	21	2 (9)	20 (95.2)	10 (47.6)	NA	NA
Hu et al [20] (2022)	34	Toripalimab	12	NA	34	NA	34 (100)	26 (76.4)	20 (58)	1 (2)
Kothari et al. [21] (2022)	9	Pembrolizumab, nivolumab, nivolumab/ipilimumab	9.5–117	NA	9	NA	NA	8 (88.8)	NA	NA
Zhang et al. [9] (2022)	32	Pembrolizumab, sintilimab, tislelizumab	12–30	3 (9.3)	29	3 (10)	29 (100)	22 (75.8)	12 (37)	0 (0)
de Gooyer et al. [15] (2024)	59	Nivolumab+relatlimab	6–34	NA	59	22 (37)	59 (100)	40 (67.7)	41 (69)	6 (10)
Chalabi M et al. [4] (2024)	115	Nivolumab+ipilimumab	4–33	NA	115	22 (19)	115 (100)	77 (66.9)	73 (63)	5 (4)

NA, not available; PD-1, programmed cell death 1.

The role of a “watch-and-wait” approach remains debatable. While complete clinical response has been documented in rectal cancer following checkpoint inhibitor therapy, long-term data in colon cancer are lacking [22]. Prospective trials, such as the phase 2 trial PREMICES study (ClinicalTrials.gov identifier: NCT06646445) [23], aim to further assess this strategy.

Limitations

The main limitation of this study is the small number of patients from a single center, which reflects the rarity of MMR-D cancers. Due to its retrospective design, there is heterogeneity in chemotherapy cycles and tumor stages, which may affect external validity. Furthermore, the granularity of immunotherapy-related complications may have been underreported or missed. Prospective studies with longer follow-up are needed to determine if the pathological response translates into improved disease-free and overall survival.

Conclusions

Surgical resection following neoadjuvant immunotherapy in MMR-D CRC is safe and yields high rates of complete pathological response. These findings support further exploration of nonoperative management strategies and underscore the need for standardized guidelines on patient selection for organ preservation.