Colorectal cancer (CRC) is the third most common malignancy diagnosed worldwide and the second in terms of cancer-related mortality. Approximately 20% of patients are metastatic at diagnosis, while half of the initially localized ones will develop metastases throughout the disease. Most of these patients will be treated with a palliative purpose, with a median overall survival of 30 months (1).

For several years, the standard of care (SOC) for unresectable locally advanced or oligometastatic (resectable or potentially resectable) CRC has been based on a combination of 5-Fluorouracil (5-FU)-based chemotherapy with or without an anti-vascular endothelial growth factor (VEGF) antibody when RAS or BRAF were mutated, or an anti-epidermal growth factor receptor (EGFR) agent, when they were found in the wild-type status. Therefore, nowadays, the detection of KRAS, NRAS and BRAF mutations as well as the deficit of mismatch repair proteins (dMMR) and subsequent high microsatellite instability (MSI-H) testing are compulsory in order to establish not only the best treatment sequence but also the prognosis of the disease. Depending on the response, surgery was performed after several cycles, and the same chemotherapy scheme was considered to complete 6 months-treatment between pre and post-surgery (2).

However, as it happened with metastatic CRC, this precept may change when the dMMR/MSI-H status is identified since this status is a trustworthy biomarker of long-term response to immune checkpoint inhibitors (ICIs) targeting programmed death-1/programmed death ligand-1 (PD-1/PD-L1) and cytotoxic T lymphocyte antigen- 4 (CTLA-4); therefore, the use of ICIs has become the SOC in this subgroup of patients in the metastatic first-line setting. The dMMR/MSI-H status is harbored by 5-15% of all CRC patients, with a higher incidence in early-stage CRC (3). Besides, the identification of dMMR/MSI-H tumors can assist in the necessity of genetic counselling for Lynch Syndrome (LS). However, 70-85% of these dMMR/MSI-H tumors are formed sporadically, mostly by the hypermethylation of the MLH1 promoter. It can take place concurrently with a BRAF V600E mutation and may or may not be associated with the loss of PMS2 (4, 5). Only 1 to 3% of CRC corresponds to LS, which is caused by a germline mutation in the MMR genes (MLH1, MSH2, MSH6 or PMS2) or in the epithelial cell adhesion molecule (EPCAM), which provokes an epigenetic silencing of MSH2 (5).

The family of MMR proteins coded by MMR genes oversee the repairing of insertions and deletions formed during DNA replication, especially in the microsatellite regions of DNA, which consist of repeated sequences of 1-6 nucleotides located throughout the DNA, especially near the coding regions. Therefore, the accumulation of errors in those regions leads to the mentioned MSI-H, which is associated with an increased mutational rate that can be up to 100-fold greater (>12 mutations per 10⁶ DNA bases) than proficient MMR, microsatellite-stable tumors or low microsatellite instability (pMMR/MSS-MSI-L; <8.24 mutations per 10⁶ DNA bases) (6, 7).

The identification of the dMMR and MSI-H status should be done by using an immunohistochemistry (IHC) assay and a multiplex polymerase chain reaction (PCR) test along with capillary electrophoresis, respectively; it is highly recommendable to use both methods, since using only one assay may be misjudging (8). Besides, both MSI and dMMR can also be assessed through next-generation sequencing (NGS) techniques but they are not available universally (6). Even though the terms of dMMR and MSI-H are usually associated, some exceptions may be found since the mutation of one of the MMR, such as MSH6 or PMS2, may reach an MSS or MSI-L status and not meet the criteria for MSI-H and, much more infrequently, the other way around can be seen (pMMR status with MSI-H) (9).

Therefore, the rationale for the use of immunotherapy in dMMR/MSI-H CRC is based on the vast amount of neoantigens that these tumors can produce and present as peptides through major histocompatibility complex (MHC) class I molecules on their surfaces. These peptides are recognized by T cell receptors (TCR), but this interaction is not enough to launch an immune response against tumoral cells since this TCR-MHC complex is modulated by co-stimulatory and co-inhibitory signals. These checkpoint inhibitors may be blocked using ICIs which can target either the co-inhibitory receptors PD-1 or CTLA-4, located on the surface of T-cells and other immune cells, or their ligands, such as PDL-1, which are presented on tumor cells and other immune cells. Therefore, by using ICIs the escape mechanism to the immune system of tumoral cells is revoked and T-cell response is enhanced against malignant cells (10). Besides, these tumors are known as immunogenic since their tumor microenvironment is heavily infiltrated by CD8+ tumor-infiltrating lymphocytes (TILs), CD4+ T helper 1 TILs and macrophages in their M1 status, as well as other antigen-presenting cells. The number of TILs is comparatively higher in MSS tumors and may also be even higher than the number of neoplastic cells but there is also an upregulation of the expression of the co-inhibitory signalization through CTLA-4, PD-1, PD-L1, LAG-3, and IDO that support the immune evasion. However, the PD-L1 expression in CRC is strikingly scarce in tumoral cells, being expressed on myeloid cells surrounding and, therefore, it is not a reliable predictive biomarker of response (11–14).

The role of immunotherapy in the neoadjuvant setting still lacks randomized trials to be completely defined, although some light has been shed on this field in the last few years. Up to now, the role of neoadjuvant chemotherapy in locally advanced tumors has not been proven successful in general, especially in MMR–deficient tumors, and surgery keeps being the standard at these stages, except for the unresectable cases which must be treated as metastatic with or without an anti-VEGF or anti-EGFR agent, depending on the mutational status and resectability. In the phase II/III FOxTROT trial, the use of 3 cycles of preoperative oxaliplatin-fluoropyrimidine chemotherapy showed a clear benefit in down-staging the locally advanced tumors, with fewer incomplete resections, and a better 2-year disease control. However dMMR tumors experimented a statistically lower response to chemotherapy than pMMR tumors and 70% of these patients experiencing no regression (15). The recent data from clinical trials using ICIs on this subgroup of patients (16–20) and case reports like the ones we are presenting point out the role of immunotherapy in the neoadjuvant setting. However, this is a small case series and, therefore, in order to validate our findings, larger patient cohorts and phase III clinical trials are required.

Here we present our four case reports of two unresectable locally advanced CRC and two oligometastatic CRC, including a patient with non-locoregional nodal disease and another one with a large unique liver relapse (oligometastatic disease). They all experienced great responses to pembrolizumab in monotherapy. Several cases of pathological complete response (pCR) or major PR (MPR) following immunotherapy have been described in the literature but, up to our knowledge, there have been no published reports of pCR or MPR in oligometastatic disease. Moreover, genetic profiling was performed at our centre with potentially targetable molecular alterations found which could be useful if a relapse occurs. The characteristics of the four case reports are summarized in Table 1 .

Nowadays, immunotherapy is the SOC for irresectable and metastatic dMMR/MSI-H CRC; on the one hand, pembrolizumab, an antiPD-1 agent, is approved in 1 ^st and further lines because of the remarkable results presented in the KEYNOTE 177 and KEYNOTE 016 vs. SOC chemotherapy (13, 25, 26). Besides, Nivolumab and the combination of Nivolumab plus Ipilimumab are also approved for the same settings as pembrolizumab due to the efficacy demonstrated in the CHECKMATE 142 phase III clinical trial (13, 27).

The role of immunotherapy in the neoadjuvant setting of dMMR/MSI-H CRC will be soon consolidated due to the remarkable results presented in recent studies: In the phase II study NICHE2, patients with locally advanced CRC were treated with one dose of ipilimumab and two doses of nivolumab and underwent surgery ≤6 weeks since starting therapy. An impressive mPR rate of 95%, including 67% pCR, was reported and the incorporation of immunotherapy as SOC at the neoadjuvant setting in these tumors is a matter of time (16). The phase II PICC trial randomized locally advanced dMMR/MSI-H CRC patients to received toripalimab, an antiPD-1, with or without celecoxib, for six cycles before surgical resection, with 88% of pCR in the celecoxib group (vs 65% in toripalimab monotherapy group) (19). The role of pembrolizumab in this setting has been evaluated in another phase II trial in which patients with localized unresectable or high-risk resectable dMMR/MSI-H solid tumors, including 27 CRC, were allocated to receive pembrolizumab for 6 months followed by surgical resection with an option to continue therapy for 1 year: a pCR was observed in 79% of CRC tumors operated (20). Recently, results from NICHE-3 were presented, which showed even better results in terms of pCR in the same setting of patients: after two cycles of nivolumab plus relatlimab, all the patients obtained a response, including 89% of mPR and 79% of pCR (18). The role of dostarlimab, another antiPD-1 checkpoint inhibitor, as neoadjuvant treatment in locally advanced rectal cancer also showed a remarkable 100% clinical complete response (cCR), allowing these patients to spare chemoradiotherapy and surgery, with no relapse evidence during follow up. These results have been confirmed in a 20-patient retrospective study: a 90% cCR was observed and pCR was confirmed in all the 11 patients that underwent surgery; the other 7 patients chose a watch-and-wait strategy, with no local or distant recurrence observed in neither of the groups (28). Interestingly, none of these trials are evaluating the role of immunotherapy as conversion therapy in oligometastatic disease and data from patients that could be elected for surgery in the metastatic trials published were censored. The principal clinical trials investigating the role of ICIs in the neoadjuvant setting and the results available are summarized in Table 2 .

However, several questions are raised because of these studies and the ones coming: The necessity of using double immunotherapy instead of monotherapy, even if it is only one cycle, may be one of the issues to be assessed. Grade 3 or 4 adverse effects were only observed in 3 (3%) patients in the NICHE2 and 1 (5%) of patients in the NICHE3 trial. Curiously in the latter, 4 out of the 19 patients included experienced an endocrinopathy, including a grade 3 hyperthyroidism and 3 hypophysitis with secondary adrenal and/or thyroid disfunction that require a supplementation for a lifetime. Besides, a late adverse effect may appear, and the toxicity of double therapy is notably higher than the one observed with monotherapy. However, the necessity of response may be the argument used for this neoadjuvant unresectable setting (27, 40). In the metastatic scenario, the phase III Checkmate 8HW (NCT04008030) trial compares monotherapy (nivolumab) vs. double therapy (ipilimumab/nivolumab) (41), whose results may bring some light on this issue. Moreover, different doses of nivolumab plus ipilimumab are also being tested in dMMR/MSI-H mCRC in a phase II trial to analyze the efficacy and safety (NIPISAFE) (42). On the other hand, several trials are testing the percentage of mPR achieved with antiPD-1 monotherapy in the neoadjuvant setting in patients with locally advanced irresectable dMMR/MSI-H CRC ( Table 2 ).

Another question to be addressed is the number of cycles necessary: in the NICHE2 trial, only two cycles (one combined and one monotherapy) were used before surgery, whereas other trials are proposing other numbers of cycles, an undetermined number of cycles for up to two years or until the tumour becomes resectable (16, 29, 30). The use of other dynamic biomarkers may be useful to determine the number of cycles or even the necessity of adjuvant therapy after surgery, such as ctDNA (43). A phase II trial is testing the utility of tumour mutation burden (TMB) as a biomarker of response to immunotherapy and if TMB is high (>20 mutations per Mb) or medium (>5 mutations per Mb) we can continue with pembrolizumab as neoadjuvant treatment for up to three cycles and if TMB is low (≤5 mutations per Mb) the patient can forgo pembrolizumab and proceed to surgery (39). Curiously, none of our cases had an elevated CEA despite the locoregional advanced setting or oligometastatic setting and, therefore, tumoral markers may not be trustworthy for making decisions or in the follow-up of these patients. Besides, there was no correlation either between the radiologic evolution and the pathological responses, even if remarkable responses were observed the mass was still visible; hence, if an organ-preservation strategy becomes an option in the future alternative modalities of re-staging may be superior and should be considered, such as PET-CT and/or a new colonoscopy to take biopsies (44).

Besides, the necessity of adjuvant therapy after applying neoadjuvant treatment also seems controversial; none of the patients of NICHE2 have had disease recurrence reported yet. However, this question may be more useful among the 5% that did not achieve a mPR but “only” partial response (4%) or no pathological response (1%) (16). In the KEYNOTE-177 trial, patients that could be elected for surgery were censored from the published data and, hence, we still have no data supporting or not supporting adjuvant therapy in these patients (25).

In previously untreated patients with stage II/III dMMR/MSI-H locally advanced rectal cancer, the use of dostarlimab, an antiPD-1 ICI, showed cCR and surgery could be avoided, with no relevant toxicity reported (no grade¾4 adverse effects) (45). This organ-preservative strategy is another question that should be tabled in dMMR/MSI-H locally advanced CRC, since some patients may benefit from disposing of surgery and some phase II trials are addressing this strategy either with monotherapy, such as dostarlimab (NCT05239546) or using a PD-1/CTLA-4 bispecific antibody (NCT05815290, NCT04556253) (30, 32, 37), with recent results that support this strategy following pembrolizumab (20).

The role of neoadjuvant treatment in oligometastatic disease, resectable or potentially resectable, has not been addressed in clinical trials either. Standard treatment includes 5-FU-based chemotherapy with or without monoclonal antibodies depending on the possibility of resection (3). However, as we have reported in our third and fourth clinical cases, these patients could benefit from using ICI due to the responses achieved and further investigation is required to establish its role either as monotherapy or the necessity of using double therapy if a response is needed. Although previous data on non-operable oligometastatic dMMR/MSI-H had shown deep and durable responses (46), no previous reports of oligometastatic dMMR/MSI-H CRC that underwent surgery after neoadjuvant ICI have been published and data from KEYNOTE-177 were censored from initial (25, 47).

Nowadays, a chemotherapy-sparing strategy seems reasonable for these patients and the use of combination strategies may have a place in selected cases (31) or if no response is achieved: a phase II study proposed a chemotherapy-salvage therapy when stable or progression disease is achieved after three cycles of Camrelizumab (antiPD-1) plus Apatinib (anti-VEGF) (35). The rates of pCR available from trials are remarkable, ranging from 60% (NICHE) to 88% (PICC), and almost reaching 100% when considering MPR (89-95%). Low rates of non-responder patients are reported (1% in NICHE2) (16, 17, 19), and patients that experienced “only” pathological response (50% or less residual viable tumour) are also scarce: 5% in NICHE, 4% in NICHE2 and 1% in PICC trial (corresponding to the combination arm) (16, 19). These data is striking when comparing to the data of progression disease as best response reported in metastatic trials (29,4% in KEYNOTE 177, 13% in CHECKMATE 142), although the substantially greater number of patients in the metastatic trials should be considered when analyzing these data (25, 48).

Therefore, biomarkers of response to ICIs are required in order to make an adequate selection of the patients that are candidate to neoadjuvant treatment, especially if a surgery option is available. Several retrospective studies have tried to elucidate clinical patterns of response to ICI: TMB^high (cut-point was estimated between 37 and 41 mutations/Mb) was predictive of response whereas the presence of peritoneal metastases and poor performance status were predictive of low benefit (13, 49). Focusing on the differences that these patients may have on their samples that may guide the response to ICIs, when using a multiplex immunofluorescence (MIF) staining and image analysis a higher pre-existing CD8+ T-cell density has been significantly associated with pCR response (767.47 per.mm2 vs. 326.64 per.mm2 [p= 0.013]) as well as a negative or low expression of PD-1 CD8+ cells (defined by an expression of PDCD1 of 0; s (700.24 per.mm2 vs. 288.29 per.mm2, [p=0.007]). This associates with the already known absence of correlation of response with PD-L1 expression in the metastatic setting. The signature genes that were highly expressed in these PD-1^low-CD8+ lymphocytes were TRGC2, CD160, and KLRB1, among others (50). On the other hand, the presence of Caudal-related homeobox transcription factor 2 (CDX-2) seem to be correlated with higher sensitivity to immunotherapy. This transcription factor is inversely associated with poor prognosis since it is downregulated by the oncogenic pathways activated in CRC. It promotes the expression of chemokine (C-X-C motif) ligand 14 (CXCL14) expression by activating its enhancer, promoting the migration and cytotoxicity of NK cells and, therefore, enhancing the immune response to ICIs. Although small (n=14) and performed in metastatic patients, this study showed a PFS rate at 12 months of 81% in CDX-2 positive patients (tested by IHC) vs 0% in CDX-2 negative patients (p = 0.0011) (51). Further studies are warranted to identify new biomarkers that could guide clinicians to choose the right therapy for these dMMR/MSI-H CRC patients.

Finally, the molecular characterization could also be useful in these patients: although BRAF mutation is classically connected with a worse prognosis, the possibility of using immunotherapy when dMMR/MSI-H status is associated has changed the scenario of these patients. Although one of our cases could not have an NGS test done due to lack of material, two of our cases had a BRAF V600E mutation using molecular testing, with no further alterations found in NGS and one of the wild-type cases had a RET::NCOA4 fusion that could be targetable if needed in the future. When reviewing the literature, a higher proportion of mutations can be found even in the presence of BRAF mutations (52), and we can hypothesize that additional mutations may be accumulated throughout the evolution of the disease.

On the other hand, immunotherapy has not demonstrated yet its efficacy in advanced CRC patients that are microsatellite stable with proficient mismatch-repair proteins (pMMR/non-MSI-H) that represent most of CCR. However, an interesting overall response rate of 30% has been observed in this subgroup of patients in the NICHE1 clinical trial ( Table 2 ), so these localized patients may be an interesting group on whom to investigate the reasons why they obtained these response rates and try combinational strategies (17). There is a need for trusty biomarkers of response to immunotherapy in this subgroup of patients, such as POLE mutation (53), as well as new therapeutic strategies that may allow these patients to benefit from the remarkable responses associated with ICIs (12).

This case series underlines the potential efficacy that pembrolizumab has in the neoadjuvant setting of the dMMR/MSI-H locoregional CRC, concurring with results from phase II neoadjuvant trials available. Nonetheless, our case reports also showed remarkable efficiency in oligometastatic disease, in which the use of ICIs as conversion therapy is still an unresolved topic. Several questions such as the best treatment scheme and duration, the necessity of adjuvant therapy or the establishment of organ-preservation strategies. are being questioned and, therefore, further investigation is warranted.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

MS-R-G: Conceptualization, Investigation, Visualization, Writing – original draft, Writing – review & editing, Methodology, Validation. IM-D: Investigation, Methodology, Resources, Supervision, Validation, Writing – review & editing. VA-F: Writing – review & editing. PG-S: Writing – review & editing. JP-P: Writing – review & editing. JC-P: Writing – review & editing. DR-R: Writing – review & editing. PS-R: Writing – review & editing. AB-C: Writing – review & editing. VA-N: Writing – review & editing. CG-M: Writing – review & editing. CG-D-Q-S: Writing – review & editing. VL: Resources, Writing – review & editing. IR-C: Resources, Writing – review & editing. CP-M: Resources, Writing – review & editing. RF-M: Investigation, Methodology, Resources, Supervision, Validation, Writing – review & editing.