The New Horizon for Non-Metastatic dMMR Colorectal Cancer: A Systematic Review of the Adjuvant Chemoimmunotherapy and Neoadjuvant Immunotherapy Revolution

Abstract

Background: The management of resected stage III colorectal cancer (CRC) has long been reliant on fluoropyrimidine-based adjuvant chemotherapy. However, the 10–15% of patients with mismatch repair-deficient (dMMR) tumors derive limited benefit from this approach. While immunotherapy has revolutionized the treatment of metastatic dMMR CRC, its role in the early-stage setting is rapidly evolving, creating a paradigm shift. Methods: A systematic literature review was conducted to identify pivotal clinical trials evaluating therapeutic strategies for non-metastatic dMMR CRC. Databases including PubMed/MEDLINE and conference proceedings from the American Society of Clinical Oncology (ASCO) and European Society for Medical Oncology (ESMO) were searched up to June 2025. The review focused on phase II and III trials reporting on disease-free survival (DFS), pathological complete response (pCR), and safety. Study selection followed PRISMA guidelines. Results: The systematic review identified 14 key studies that were included for narrative synthesis. The evidence base encompassed three areas: (1) Foundational adjuvant chemotherapy trials (e.g., MOSAIC, IDEA); (2) Pivotal metastatic trials (e.g., KEYNOTE-177) validating immunotherapy efficacy in dMMR CRC; and (3) Modern trials in non-metastatic disease. The phase III ATOMIC trial demonstrated that adding atezolizumab to mFOLFOX significantly improved 3-year DFS versus chemotherapy alone (86.4% vs. 76.6%; Hazard Ratio [HR] 0.50, 95% Confidence Interval [CI] 0.34–0.72; p < 0.001). Concurrently, phase II neoadjuvant immunotherapy trials (e.g., NICHE-2) reported remarkable pCR rates of 68% and a 3-year DFS of 100%, with a more favorable safety profile compared to chemoimmunotherapy. Conclusions: The landscape for non-metastatic dMMR CRC is shifting from a chemotherapy-based model to an immunotherapy paradigm. The ATOMIC trial establishes adjuvant chemoimmunotherapy as a new standard, while robust neoadjuvant data suggest a potential future where short-course, chemotherapy-free immunotherapy could become a preferred strategy. Ongoing trials directly comparing these approaches are awaited.

1. Introduction

Colorectal cancer (CRC) is the third most common cancer and the second leading cause of cancer-related mortality worldwide. In 2022, it was estimated that there were 1,926,118 new diagnoses of CRC and 903,859 related deaths globally [1]. Since 1990, the management of resected stage III CRC has been transformed by adjuvant chemotherapy, beginning when Dr. Moertel demonstrated that 5-fluorouracil (5FU) led to a 33% reduction in the risk of death [2]. A subsequent major advance came with the MOSAIC trial, published in 2004, which confirmed that 6 months of FOLFOX chemotherapy resulted in a significant improvement in disease-free survival (DFS) and overall survival (OS) compared to LV5FU2 [3]. For low-risk stage III CRC, the International Duration Evaluation of Adjuvant Chemotherapy (IDEA) collaboration demonstrated in 2018 the non-inferiority of 3 months of chemotherapy compared to 6 months, with a lower frequency and severity of adverse events, particularly neuropathy [4]. However, despite adjuvant chemotherapy, 30% of stage III CRC patients will experience recurrence. Consequently, there is an unmet need for more effective strategies to improve patient outcomes.

Approximately 10 to 15% of patients diagnosed with CRC have mismatch repair deficiency (dMMR), representing 300,000 new cases each year. dMMR cancers are biologically distinct; given their DNA repair defect, they accumulate mutations that enhance the anti-tumor immune response [5]. These cancers do not derive any benefit from 5FU monotherapy, which should not be used in these patients. In the metastatic setting, immunotherapy has demonstrated meaningful improvements in progression-free survival (PFS), OS, and quality of life for patients with advanced dMMR CRC [6,7].

Recently, the second interim analysis of the ATOMIC trial, presented by Dr. Sinicrope on 1 June 2025, at the ASCO annual meeting (Plenary session, LBA 1001), showed that mFOLFOX plus atezolizumab significantly improved DFS compared to mFOLFOX alone in patients with R0-resected stage III dMMR CRC [8]. Nevertheless, the optimal therapeutic strategy for this population is not yet fully defined. Critical uncertainties persist, including the identification of patients who require treatment, the potential of adjuvant immunotherapy to improve cure rates, the role of adjuvant chemotherapy in the context of effective immunotherapy, and the comparative efficacy of neoadjuvant versus adjuvant immunotherapy.

Therefore, this systematic review aims to critically appraise and synthesize the current evidence regarding the efficacy and safety of immunotherapy in the management of non-metastatic dMMR CRC. We will systematically evaluate and compare the two emerging paradigms, adjuvant chemoimmunotherapy (as exemplified by the ATOMIC trial) and neoadjuvant immunotherapy (as investigated in the NICHE-2 and other trials), against the historical data of adjuvant chemotherapy alone (MOSAIC and IDEA). The objective is to provide a clear overview of this treatment landscape, discuss the implications for current clinical practice, and highlight key unanswered questions to guide future research.

2. Materials and Methods

A systematic literature review was conducted to identify pivotal clinical trials evaluating therapeutic strategies for mismatch repair-deficient (dMMR) or microsatellite instability-high (MSI-H) non-metastatic CRC. Databases including PubMed/MEDLINE were searched, along with proceedings from the annual meetings of the American Society of Clinical Oncology (ASCO) and the European Society for Medical Oncology (ESMO), up to June 2025. The search strategy combined keywords related to the following concepts: “colorectal cancer”, “stage II/III”, “non-metastatic”, “dMMR”, “MSI-H”, “immunotherapy”, “immune checkpoint inhibitors”, “neoadjuvant”, “adjuvant”, “atezolizumab”, “pembrolizumab”, “nivolumab”, and “ipilimumab”.

2.1. Eligibility Criteria

Studies were eligible if they (1) involved adult patients with resectable, non-metastatic dMMR/MSI-H CRC; (2) evaluated immunotherapy (in the neoadjuvant, adjuvant, or perioperative setting) or were historical adjuvant chemotherapy trials that defined the previous standard of care; (3) were Phase II or III clinical trials; and (4) reported relevant efficacy outcomes (DFS, pathological complete response [pCR], OS). Preclinical studies, Phase I trials, case reports, and publications in languages other than English were excluded.

2.2. Study Selection and Data Extraction

The study selection process followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (Figure 1). The completed PRISMA 2020 checklist is provided as Supplementary File S1. After duplicate removal, titles and abstracts were screened for relevance, followed by a full-text review of potentially eligible articles. The final selection was based on the predefined eligibility criteria. Data from the included studies were extracted into a standardized table, capturing key information such as study design, patient population, intervention, comparator, and primary efficacy and safety outcomes.

2.3. Protocol Registration

This systematic review was registered with the International Prospective Register of Systematic Reviews (PROSPERO) under registration number CRD420251174716. The protocol was developed prior to the data synthesis but after the initial literature search had commenced.

2.4. Data Synthesis

A quantitative meta-analysis was not performed due to the significant clinical heterogeneity of interventions (chemotherapy alone, chemoimmunotherapy, neoadjuvant immunotherapy alone) and settings (adjuvant vs. neoadjuvant) among the included studies. Therefore, a narrative synthesis was conducted. The results are presented to chronologically and thematically illustrate the evolution of the treatment paradigm, from the establishment of adjuvant chemotherapy to the current era of immunotherapy. While the present study adopted a narrative approach, the similarity of the endpoints in the studies included suggests that a quantitative meta-analysis could be conducted in the future to present robust pooled estimations.

3. Results

3.1. Study Selection

The initial database and manual search identified 420 records. After the removal of duplicates and the screening of titles and abstracts, 52 full-text articles were assessed for eligibility. Of these, 38 were excluded with reasons, the most common being incorrect disease focus (other cancer types), inappropriate patient population (pMMR/MSS tumors), or suboptimal interventions (chemotherapy-only trials not considered foundational). The final review included 14 pivotal studies for qualitative synthesis (

Table 1. Evolution of Adjuvant Therapy in Stage III Colon Cancer: Key Clinical Trials.

Trial (Phase)	Year of Publication	Patient Population	Investigation vs. Comparator	3-Year DFS	Key Grade 3/4 Adverse Events
Moertel et al. (III) [2]	1990	Resected stage III CRC (n = 1296)	Fluorouracil + levamisole vs. Observation	65% vs. 50%	Nausea (5%), vomiting (2%), diarrhea (7%), stomatitis (3%), dermatitis (1%), leukopenia (3%)
MOSAIC (III) [3]	2004	Resected stage II/III colon cancer (n = 2246); Included dMMR (n = 95)	FOLFOX vs. LV5FU2	78.2% vs. 72.9% (HR 0.76)	Sensory Neuropathy: 12.4% (FOLFOX4) vs. 0% (LV5FU2)
IDEA (Meta-Analysis) [4]	2018	Resected low-risk stage III CRC	3 months (CAPEOX/FOLFOX) vs. 6 months (CAPEOX/FOLFOX)	75.9% vs. 74.8% (HR 1.07)	Sensory Neuropathy: 24.2% (3-mo) vs. 36.9% (6-mo)
		Resected high-risk stage III CRC	3 months (CAPEOX/FOLFOX) vs. 6 months (CAPEOX/FOLFOX)	62.7% vs. 64.4% (HR 1)
ATOMIC (III) [8]	2025	R0-resected stage III dMMR CRC (n = 712)	Atezolizumab + mFOLFOX6, then Atezolizumab vs. mFOLFOX6	86.4% vs. 76.6% (HR 0.50)	Higher rates of PN (19% vs. 15%), hypothyroidism (20.5% vs. 0%), diarrhea (60.1% vs. 12.1%) neutropenia (43% vs. 36%) (Atezo arm)

Abbreviations: CAPEOX: Capecitabine and Oxaliplatin; dMMR: Mismatch Repair Deficient; DFS: Disease-Free Survival; FOLFOX: 5-Fluorouracil, Leucovorin, and Oxaliplatin; HR: Hazard Ratio; LV5FU2: Leucovorin and 5-Fluorouracil; PN: Peripheral Neuropathy; R0: Complete Resection; Note: The DFS benefit in MOSAIC was significant for stage III (HR 0.76; 95% CI 0.62–0.92) but not stage II. The ATOMIC trial showed a statistically significant improvement in DFS (HR 0.50; 95% CI 0.34–0.72).

and

Table 2. Pivotal Clinical Trials in the Evolution of Treatment for Non-Metastatic dMMR CRC.

#	Reference/Study Status (Trial Name)	Study Design	Study Population (n)	Intervention	Key Outcomes (DFS, pCR)
1	Sinicrope et al. ASCO 2025 (ATOMIC) [8]	Randomized Phase 3 Trial	Resected stage III dMMR CRC, R0 resected (n = 712)	mFOLFOX6 + Atezolizumab vs. mFOLFOX6 (adjuvant)	3-year DFS: 86.4% vs. 76.6% (HR 0.50). Potential new standard.
2	Chalabi et al. NEJM 2024 (NICHE-2) [9]	Phase 2 Trial	Locally advanced dMMR CRC (n = 115)	Nivolumab + Ipilimumab (neoadjuvant, 2 cycles)	pCR: 68%; Major Pathological Response (MPR): 95%; 3-year DFS: 100%.
3	Xu et al., 2024 [10]	Phase 1b Trial	Early MSI-H/dMMR CRC	IBI310 (anti-CTLA-4) + Sintilimab (anti-PD-1) (neoadjuvant)	Promising pCR rates.
4	de Gooyer et al. Nat Med 2024 (NICHE-3) [11]	Phase 2 Trial	Locally advanced dMMR CRC (n = 59)	Nivolumab + Relatlimab (neoadjuvant, 4 cycles)	pCR: 68%; MPR: 92%.
5	Hu et al. Lancet Gastro Hep 2022 (PICC) [12]	Randomized Phase 2 Trial	Locally advanced dMMR CRC (n = 28)	Toripalimab + Celecoxib (neoadjuvant) vs. Toripalimab	pCR: 88% (combo) vs. 65% (mono).
6	Shiu et al., 2024 (NEOPRISM-CRC) [13]	Phase 2 Trial	Stage II–III dMMR CRC (n = 32)	Pembrolizumab (neoadjuvant, 3 cycles)	pCR: 53%
7	de la Fouchardière et al. Ann Oncol 2024 (IMHOTEP) [14]	Phase 2 Trial	CRC Cohort: Localized dMMR CRC cancer (n = 87)	Pembrolizumab (neoadjuvant, 1–2 cycles)	pCR: 53.8%
8	Cercek et al. NEJM 2025 (Cohort-1) [15]	Phase 2 Trial	Locally Advanced dMMR Rectal Cancer (n = 50)	Dostarlimab (neoadjuvant, 6 months)	pCR: 74%; 2-year Event-Free Survival: 92%.
9	André et al. NEJM 2004 (MOSAIC) [3]	Randomized Phase 3 Trial	Resected stage II/III CRC (n = 2246); Included dMMR (n = 95)	FOLFOX4 vs. LV5FU2 (adjuvant)	Established FOLFOX as adjuvant standard; less benefit in dMMR.
10	Grothey et al. NEJM 2018 (IDEA) [4]	Meta-Analysis	Resected stage III CRC	3 months vs. 6 months of (CAPEOX/FOLFOX) (adjuvant)	Non-inferiority of 3 months for low-risk disease; less neuropathy.
11	Moertel et al. NEJM 1990 [2]	Randomized Phase 3 Trial	Resected stage III CRC (n = 1296)	5-FU + Levamisole vs. Observation (adjuvant)	Foundation of adjuvant chemotherapy; 33% reduction in death risk.
12	Sargent et al. JCO 2010 [16]	Pooled Analysis	Resected stage II/III CRC	5-FU-based adjuvant therapy (adjuvant)	dMMR is a predictive marker for lack of benefit from 5-FU.
13	André et al. Ann Oncol 2025 (KEYNOTE-177) [6]	Randomized Phase 3 Trial	Metastatic MSI-H/dMMR CRC	Pembrolizumab vs. Chemotherapy	Pivotal proof of efficacy of immunotherapy in metastatic dMMR CRC.
14	Andre et al. NEJM 2024 (CheckMate 8HW) [7]	Randomized Phase 3 Trial	Metastatic MSI-H/dMMR CRC	Nivolumab plus Ipilimumab vs. Chemotherapy	-

Abbreviations: CAPEOX: Capecitabine and Oxaliplatin; CRC: Colorectal Cancer; dMMR: Mismatch Repair Deficient; DFS: Disease-Free Survival; FOLFOX: 5-Fluorouracil, Leucovorin, and Oxaliplatin; HR: Hazard Ratio; LV5FU2: Leucovorin and 5-Fluorouracil; MPR: Major Pathological Response; MSI-H: Microsatellite Instability-High; pCR: Pathological Complete Response; R0: Complete Resection.

). The selection process is detailed in the PRISMA flow diagram (Figure 1).

3.2. Overview of Included Studies and Synthesized Findings

The 14 included studies were categorized into three groups that chart the evolution of treatment for non-metastatic dMMR/MSI-H CRC (Table 1): (1) Foundational Adjuvant Chemotherapy Trials: This group includes landmark studies such as Moertel et al. (establishing 5-FU), MOSAIC (establishing FOLFOX), and the IDEA collaboration (refining treatment duration). These trials set the historical standard against which new strategies are measured. The MOSAIC trial demonstrated a significant DFS benefit (78.2% vs. 72.9%) specifically in the stage III population, while the benefit in stage II was not statistically significant. Crucially, pooled analyses like Sargent et al. demonstrated the lack of benefit from 5-FU in dMMR/MSI-H tumors, highlighting a critical unmet need. (2) Immunotherapy in Metastatic dMMR CRC: Pivotal trials like KEYNOTE-177 (pembrolizumab) and CheckMate 8HW (nivolumab + ipilimumab) provided the foundational proof-of-concept for the high efficacy of immune checkpoint blockade in dMMR CRC, supporting their investigation in earlier disease stages. (3) Modern Trials in Non-Metastatic Disease: This group forms the core of the current paradigm shift. It is subdivided into two approaches: Adjuvant Chemoimmunotherapy: The Phase III ATOMIC trial demonstrated that adding one year of atezolizumab to adjuvant mFOLFOX6 resulted in a statistically significant and clinically meaningful improvement in 3-year DFS compared to chemotherapy alone (86.4% vs. 76.6%; HR 0.50; 95% CI 0.34–0.72). Neoadjuvant Immunotherapy: Multiple Phase II trials reported groundbreaking efficacy. The NICHE-2 trial, using nivolumab + ipilimumab, showed a pCR rate of 68% and a 3-year DFS of 100%. Similarly, the Cercek et al. trial in locally advanced dMMR rectal cancer demonstrated a remarkable pCR rate of 74%, with most patients achieving a clinical complete response and avoiding surgery and radiotherapy. Other studies like NICHE-3, IMHOTEP, and NEOPRISM-CRC have further confirmed high pCR rates (53–68%) with neoadjuvant immunotherapy. Collectively, these regimens demonstrate a consistently more favorable safety profile with notably lower rates of high-grade adverse events compared to chemotherapy-containing regimens (Table 1, Table 2 and

Table 3. Efficacy and Safety of Neoadjuvant Immunotherapy in Non-Metastatic dMMR Colorectal Cancer.

Study (Tumor Location)	Patient Population	Neoadjuvant Immunotherapy Regimen	pCR	Other Efficacy	Grade ≥ 3 Adverse Events
NICHE-2 [9] (Colon)	Stage cT3-4/cNany dMMR CRC (n = 115)	Nivolumab + ipilimumab (2 cycles)	68%	MPR: 95%; 3-year DFS: 100%	4%
NICHE-3 [11] (Colon)	Stage cT2-4/cNany dMMR CRC (n = 59)	Nivolumab + Relatlimab (4 cycles)	68%	MPR: 92%	Not Specified (Mostly G1-2)
Hu et al. [12] (Colon/Rectum)	Stage cT3 4/cN+ dMMR CRC	Toripalimab + Celecoxib (6 cycles)	88%	-	Not Specified (59% G1-2)
		Toripalimab (6 cycles)	65%	-	Not Specified
NEOPRISM-CRC (Colon/Rectum) [13]	Stage II–III dMMR CRC (n = 32)	Pembrolizumab (3 cycles)	53%	-	0%
IMOTHEP [14] (Colon/Rectum)	CRC cohort: Stage cT2-4/cNany dMMR CRC (n = 87)	Pembrolizumab (1–2 cycles)	53.8%	-	9%
Cercek et al. [15] (Rectum)	Locally Advanced dMMR Rectal Cancer (n = 50)	Dostarlimab (6 months)	74%	2-yr DFS: 92%	5% (Mostly G1-2)

Abbreviations: CRC: Colorectal Cancer; dMMR: Mismatch Repair Deficient; DFS: Disease-Free Survival; MPR: Major Pathologic Response; pCR: Pathologic Complete Response.

).

4. Discussion

4.1. The Landmark ATOMIC Trial: Establishing a New Standard with Adjuvant Chemoimmunotherapy

Alliance A021502 (ATOMIC) was a randomized, multicenter, open-label phase III study that enrolled patients (n = 712) with histologically confirmed, completely resected (R0) stage III dMMR CRC. Both sporadic cases and those associated with Lynch syndrome were included. Patients who had received prior systemic therapy or radiotherapy were excluded. Prespecified stratification factors were T stage, N stage, and tumor location. Patients were randomized 1:1 to receive either mFOLFOX6 plus atezolizumab (an anti–PD-L1 antibody) for 6 months, followed by atezolizumab monotherapy for an additional 6 months (investigational arm), or mFOLFOX6 for 6 months (comparator arm). The primary endpoint was DFS, and the key secondary endpoints were OS and safety [8].

Demographics and patient characteristics were well-balanced between the two arms. Notably, more than 45% of patients in the ATOMIC trial had low-risk disease; according to the IDEA collaboration findings, these patients would now typically receive only 3 months of CAPEOX chemotherapy. The study met its primary endpoint, demonstrating a clinically and statistically significant 10% absolute improvement in 3-year DFS for mFOLFOX plus atezolizumab compared to mFOLFOX alone (86.4% vs. 76.6%; HR 0.50, 95% CI 0.34–0.72; p < 0.001). The benefit of atezolizumab was observed across most subgroups. Notably, this therapeutic benefit appeared relevant regardless of whether the mismatch repair deficiency was of hereditary (Lynch syndrome) or somatic origin. However, 13% of patients experienced a recurrence despite receiving 6 months of chemotherapy plus one year of immunotherapy. Overall survival data are not yet mature. Regarding safety, atezolizumab and chemotherapy are generally manageable. However, grade 3/4 adverse events were more frequent in the atezolizumab arm, including peripheral neuropathy, hypothyroidism, diarrhea, and non-febrile neutropenia. Additionally, six (1.7%) treatment-related deaths were reported in the atezolizumab arm compared to three (0.6%) in the control arm [8].

4.2. The Paradigm Shift: Neoadjuvant Immunotherapy as a Potentially Transformative Strategy

The ATOMIC trial establishes adjuvant chemoimmunotherapy as a new, more effective standard. However, the treatment landscape for non-metastatic dMMR CRC is now evolving with two distinct pathways, as neoadjuvant immunotherapy emerges as a potentially transformative strategy. This approach finds strong support from successful applications in other cancer types, particularly non-small cell lung cancer (NSCLC) where the phase III CheckMate 816 trial demonstrated significantly improved outcomes with neoadjuvant nivolumab plus chemotherapy, leading to regulatory approval and establishing a new paradigm in thoracic oncology. Similarly, the NADINA trial in melanoma showed two cycles of neoadjuvant nivolumab–ipilimumab outperformed one year of adjuvant nivolumab, creating a new treatment standard in that disease [17,18]. In colorectal cancer itself, the NICHE-2 trial reported a 68% pathological complete response rate and 95% major pathological response with neoadjuvant nivolumab–ipilimumab, achieving 100% 3-year DFS despite 65% of patients having T4 tumors, while the Cercek trial in locally advanced dMMR rectal cancer demonstrated a 74% complete response rate with dostarlimab monotherapy, enabling organ preservation and showing 92% 2-year DFS [9,15].

The scientific basis strongly supports beginning immunotherapy before surgery, as giving checkpoint inhibitors while the primary tumor remains enables the immune system to identify a wide spectrum of tumor targets, generating more vigorous and varied activation of tumor-specific T-cells that can eliminate micrometastatic disease throughout the body [19]. Research from NSCLC and initial data from dMMR CRC trials verify that the neoadjuvant approach creates a stronger and broader T-cell response than usually possible after tumor removal [20,21]. This scientific advantage leads to important clinical benefits including the possibility of avoiding chemotherapy while providing better safety, as brief immunotherapy courses can achieve substantial results while avoiding neuropathy, hematological toxicity, and other treatment-related side effects that greatly affect quality of life (Table 3). This strategy also shows exceptional effectiveness against high-risk disease, controlling locally advanced tumors as effectively as adjuvant chemoimmunotherapy methods while additionally permitting treatment response evaluation where tissue examination at surgery provides an immediate indicator for adjusting therapy, allowing treatment de-escalation for complete responders or additional therapy for those with poor response or R1/2 disease, creating a more customized treatment strategy compared to fixed-duration regimens [8,9].

4.3. Immunotherapy Alone in dMMR Stage III CRC

Chemotherapy has limited efficacy in dMMR stage III CRC [16]. In contrast, neoadjuvant immunotherapy has demonstrated high efficacy and a favorable safety profile in phase II trials (Table 3). For instance, the NICHE-2 study reported a pCR rate of 68% and a 3-year DFS of 100%, while the Cercek Cohort-1 study in locally advanced rectal cancer reported a pCR rate of 74% [9,15]. A recent study including a large number of patients treated with neoadjuvant pembrolizumab (IMHOTEP) (n = 87) reported a lower pCR rate compared with NICHE-2 (53%), suggesting lower efficacy of pembrolizumab in the neoadjuvant setting [9,14].

Notably, despite NICHE-2 including 65% of patients with T4-stage disease, the 3-year DFS was 100%, underscoring the profound efficacy of this neoadjuvant immunotherapy protocol in high-risk, locally advanced dMMR tumors [9]. The safety profile of these regimens is encouraging, with NICHE-2 and IMHOTEP reporting only 4% and 9% high-grade adverse events, respectively [9,14].

4.4. Limitations of the Current Evidence

A key limitation of the ATOMIC trial was the absence of an immunotherapy-only arm, leaving unanswered the fundamental question of whether chemotherapy is a necessary component of treatment for all patients [8]. This critical question is being addressed by the ongoing AZURE-2 trial, a phase III study randomizing patients to a chemotherapy-free regimen of perioperative Dostarlimab or to the current standard of care from MOSAIC/IDEA, adjuvant FOLFOX/CAPEOX, which will definitively determine if a chemotherapy-free approach is non-inferior. Beyond this specific design limitation, other considerations warrant attention. The influence of factors such as the prevalence of Lynch syndrome and underlying tumor heterogeneity on the outcomes within the ATOMIC cohort requires further exploration. Furthermore, while the 100% 3-year DFS in the NICHE-2 trial is remarkable, it applies to the per-protocol population that underwent surgery, and longer-term follow-up across all these studies is essential to confirm the durability of responses and OS benefits [9]. Acknowledging these limitations and awaiting the results of definitive trials like AZURE-2 are crucial next steps in refining this rapidly evolving treatment paradigm.

4.5. Implications for Clinical Practice and Future Research

MMR deficiency status is a highly predictive biomarker for immunotherapy response and should be routinely analyzed in all non-metastatic colorectal cancer patients prior to surgery. The ATOMIC trial establishes that adding adjuvant atezolizumab to mFOLFOX cures more patients with stage III R0-resected dMMR/MSI-H CRC, demonstrating a 10% absolute improvement in 3-year DFS and represents a potential new standard of care.

Based on current evidence, we propose the following treatment algorithm (Figure 2): (1) For patients with clinically low-risk tumors (cT3/N0), we recommend upfront surgery. If pathological staging confirms stage III disease, treatment with 3 months of adjuvant chemotherapy (per IDEA collaboration data) plus one year of atezolizumab is advised. (2) For patients with clinically high-risk tumors (cT4/N+) at elevated risk of R1 resection, we recommend short-course neoadjuvant immunotherapy followed by surgery. If R1 resection is confirmed, post-operatively, adjuvant chemotherapy plus atezolizumab, should be administered, with a chemotherapy duration of 6 months (Figure 2).

There is an urgent need to develop reliable predictive biomarkers to prevent overtreatment and to better select patients who will benefit from neoadjuvant or adjuvant strategies.

5. Conclusions

The treatment paradigm for resected stage III dMMR CRC is undergoing a fundamental transformation, moving beyond the long-standing era of fluoropyrimidine-based chemotherapy. The recent results from the ATOMIC trial establish a new benchmark for the adjuvant setting, demonstrating that adding atezolizumab to mFOLFOX significantly improves DFS and offers a new, more effective standard of care. Concurrently, the remarkable efficacy and favorable safety profile of short-course neoadjuvant immunotherapy, as exemplified by the NICHE-2 trial, present it as a compelling and potentially chemotherapy-free alternative that can induce high rates of pathological complete response and possibly prevent recurrence altogether.

This evolution creates a critical dilemma for clinicians: choosing between the proven, enhanced efficacy of adjuvant chemoimmunotherapy and the highly effective, less toxic, and organ-preserving potential of a neoadjuvant immunotherapy-first approach. The answer to this question is not yet clear and underscores an urgent need for predictive biomarkers to guide personalized therapy selection. Future results from ongoing trials, such as AZURE-2, which directly compare perioperative immunotherapy to standard chemotherapy, are essential to definitively determine the optimal sequence, combination, and duration of treatment.

Ultimately, the management of non-metastatic dMMR CRC has entered a new era of personalization. The determination of MMR status is now a critical gateway, not just for prognosis but also for selecting between two powerful, and concurrently emerging, paradigm-shifting strategies: adjuvant chemoimmunotherapy and neoadjuvant immunotherapy.