In this single-center retrospective study, all patients received immunotherapy before surgery. After surgery, more than 40% of patients achieved pCR, with 66.7% and 29.8% of patients with dMMR and pMMR colorectal cancer achieving pCR, respectively. Additionally, this study suggests that the factors predicting pCR include younger age, lower CEA levels before treatment, clinical stage III, negative clinical lymph nodes, and molecular pathology including MSI-H, dMMR, pole gene mutations, and higher TMB levels.
The traditional neoadjuvant treatment methods for colon cancer include chemotherapy, radiotherapy, targeted therapy, and combination therapy. Currently, neoadjuvant treatment for rectal cancer is based on a combination of radiotherapy and chemotherapy [12–14]. The ORR of colorectal cancer with monotherapy is 40%, and the pCR rate is 5% [15]. The pCR rate of combined radiotherapy and chemotherapy is 10–15% [16]. Garcia-Aguilar et al. found that TNT therapy achieved pCR in 38% of rectal cancer patients [17]. Our research found that compared to traditional treatment methods, neoadjuvant combination immunotherapy can increase the rate of complete pathological remission after surgery, with more than 40% of patients achieving pCR, including 39.4 (13/43) of patients with rectal cancer who achieved pCR. This shows that neoadjuvant combination immunotherapy can improve the rate of pCR compared with traditional treatment methods. In addition, a Voltage study [6] explored a new neoadjuvant immunotherapy with nivolumab after long-distance synchronous radiotherapy and chemotherapy for locally advanced rectal cancer. The pCR rate in the dMMR group was 60%, which is consistent with our results. Among 24 patients with dMMR, 16 (66.7%) achieved pCR. Lin et al. [18] reported that the pCR rate in 26 patients with pMMR rectal cancer receiving short-term radiotherapy combined with camrelizumab and CapeOx was 46.2%. In our study, 29.8% (17/57) of patients achieved complete pathological remission. Because our study included both colon and rectal cancer cases and there were mixed factors, such as treatment plan, treatment cycle, small data volume, and tumor staging, the optimal treatment plan for combination therapy cannot be determined at present.
In patients with pMMR or MSS CRC, the optimization of new neoadjuvant immunotherapy regimens and the identification of effective biomarkers to predict therapeutic efficacy are important research topics. Currently, a few potential markers such as POLE or POLD1 gene mutations, TMB, immune score, PD-L1 expression, CD8+/Treg ratio, and intestinal flora [5, 6, 19, 20]. In a study of patients with MSI-H mCRC receiving PD-1 and PD-L1 treatment, all 13 patients with high TMB achieved objective remission, whereas six of nine patients with low TMB experienced tumor progression [21]. Therefore, TMB is an effective biomarker; however, there is controversy over its optimal cutoff value when applied as a marker for predicting the efficacy of immunotherapy. Because different studies are based on different detection platforms and panels, they have different cutoff values. Determining the optimal cutoff value for TMB is another direction that should be explored in future studies. Colorectal tumors with POLE or POLD1 gene mutations are of the MSS type and can benefit from PD-1 monoclonal antibody treatment; however, this type of mutation accounts for only 1% of colorectal cancer cases [22]. In a retrospective study using CEA levels to predict pCR, lower CEA levels were associated with pCR. Additionally, CEA levels before and after neoadjuvant therapy can be used to predict pCR [23, 24]. Clinical staging is related to patient risk stratification, which distinguishes patient risk by analyzing patient images and pathological results combined with the T stage, N stage, and other factors. However, the specific application of these indicators to predict the efficacy of neoadjuvant therapy and to define the weight of each indicator requires further exploration.
Among the 81 patients included in the study, AEs such as fever (4.9%), fatigue (4.9%), rash or pruritus (7.4%), and diarrhea (8.6%) were observed; however, there were no treatment-related deaths or severe adverse reactions. In the KEYNOTE-177 study [25], the pembrolizumab group had a 31% incidence of immune- and infusion-related adverse reactions, of which 9% were grade 3 or 4, with colonic inflammation and hepatitis being the most common. The adverse reaction outcomes of this study were similar to those of the KEYNOTE-177 study; however, no grade 3–4 AEs were observed, indicating that neoadjuvant immunotherapy for colorectal cancer is safe. Grade 1–2 AEs caused by neoCRT/NAC included hand-and-foot syndrome (19.8%), nausea (17.3%), and vomiting (12.3%). In the NICHE study (2), 20% of the 40 patients who underwent colon resection surgery had grade III surgical complications, and the incidence of anastomotic leakage was as high as 10%, including five cases of abdominal (or incision) infection, one case of intestinal obstruction, one case of pneumonia, and one case of small intestinal injury; the high incidence of anastomotic leakage may be related to the combined use of PD-1 and CTLA-4. In this study, there were only two cases of anastomotic leakage after CRC surgery, both of which improved with conservative treatment. Seven patients had intraoperative bleeding of > 100 mL, one had ileus, one had a chylous fistula, one had an incisional hernia, three had a perineal incision, and one had a stoma infection; however, there were no perioperative deaths.
This single-center, retrospective study has several limitations. First, the study only included patients with CRC who underwent surgical treatment and excluded patients who achieved complete clinical remission after neoadjuvant therapy, those whose disease progressed, and those who could not undergo surgery, which may have caused a selection bias. Second, the study included patients with colon and rectal cancers, including those with advanced CRC, and there were differences in the preoperative medication plans and cycles, which may have affected the statistical results. Finally, the sample size was relatively small. Further multicenter prospective studies with larger sample sizes are warranted.