Deep learning facilitates complex medical data analysis and is increasingly being explored in colorectal cancer diagnostics. However, the training cost of the deep learning model limits its real-world medical utility. In this study, we present a composite network that combines deep learning and unsupervised K-means clustering algorithm (RK-net) for automatic processing of medical images. RK-net was more efficient in image refinement compared with manual screening and annotation. The training of a deep learning model for colorectal cancer diagnosis was accelerated by two times with utilization of RK-net-processed images. Better performance was observed in training loss and accuracy achievement as well. RK-net could be useful to refine medical images of the ever-expanding quantity and assist in subsequent construction of the artificial intelligence model.

Importance: Patients with pathological complete response (pCR) of rectal cancer following neoadjuvant treatment had better oncological outcomes. However, reliable methods for accurately predicting pCR remain limited. Objective: To evaluate whether transrectal ultrasound-guided tru-cut biopsy (TRUS-TCB) adds diagnostic value to conventional modalities for predicting pathological complete response in patients with rectal cancer after neoadjuvant treatment. Design, setting, and participants: This study evaluated data of patients with rectal cancer who were treated with neoadjuvant treatment and reassessed using TRUS-TCB and conventional modalities before surgery. This study is registered with ClinicalTrials.gov. Main outcomes and measures: The primary outcome was accuracy, along with secondary outcomes including sensitivity, specificity, negative predictive value, and positive predictive value in predicting tumour residues. Final surgical pathology was used as reference standard. Results: Between June 2021 and June 2022, a total of 74 patients were enroled, with 63 patients ultimately evaluated. Among them, 17 patients (28%) exhibited a complete pathological response. TRUS-TCB demonstrated an accuracy of 0.71 (95% CI, 0.58–0.82) in predicting tumour residues. The combined use of TRUS-TCB and conventional modalities significantly improved diagnostic accuracy compared to conventional modalities alone (0.75 vs. 0.59, P=0.02). Furthermore, TRUS-TCB correctly reclassified 52% of patients erroneously classified as having a complete clinical response by conventional methods. The occurrence of only one mild adverse event was observed. Conclusions and relevance: TRUS-TCB proves to be a safe and accessible tool for reevaluation with minimal complications. The incorporation of TRUS-TCB alongside conventional methods leads to enhanced diagnostic performance.

Although programmed death 1 blockade has significantly improved the survival of advanced colorectal cancer patients with DNA mismatch repair-deficient (dMMR)/microsatellite instability-high (MSI-H), clinical data in neoadjuvant and adjuvant setting are limited. The role of circulating tumor DNA (ctDNA) in precision oncology is promising, but its clinical significance in immunotherapy needs to be validated. We report a case series of 3 colon patients who received neoadjuvant and adjuvant immunotherapy and serial ctDNA analysis. This report summarizes clinical and molecular details for 3 patients with locally advanced or recurrent dMMR/MSI-H/polymerase epsilon (POLE) mutation-positive tumors treated with neoadjuvant/adjuvant immunotherapy. One stage IV recurrent colon cancer patient diagnosed with Lynch syndrome received adjuvant sintilimab monotherapy and had a progression-free survival (PFS) over 16 months, one stage Ⅲc colon cancer patient with MSI-H/high tumor mutation burden received neoadjuvant toripalimab monotherapy, was assessed as clinical complete response before surgery, continued with adjuvant sintilimab monotherapy and had a PFS over 17 months, one stage Ⅱ colon cancer patient with POLE P286R also received adjuvant sintilimab monotherapy and had a PFS over 17 months. All patients had detectable ctDNA after radical surgery and clearance of ctDNA during adjuvant immunotherapy. All 3 patients are free of tumor disease at the time of this report. Further studies are warranted to evaluate the long-term efficacy of neoadjuvant and adjuvant programmed death 1 blockade in locally advanced and metastasis in dMMR/MSI-H/POLE mutated colorectal cancer and the role of ctDNA monitoring.

The synergistic effect of combining immune checkpoint inhibitors (ICIs) with neoadjuvant chemo(radio)therapy (nCRT) in colorectal cancer is still limited. We aimed to understand the impact of nCRT on the tumor microenvironment and to explore favorable immune markers of this combination. Herein, we investigated the expression of cytotoxic T lymphocyte-associated antigen 4 (CTLA-4), CD86, CD4, and CD8 after nCRT and its association with clinicopathological characteristics. Immunostaining of immune-related molecules was performed in 255 surgically resected specimens from rectal cancer patients treated with nCRT. CD4 and CD8 expression on the tumor (tCD4/CD8), stroma (sCD4/CD8), and invasive front (iCD4/CD8) was evaluated. The expression levels of immune-related molecules were significantly lower in the nCRT-treated group, except for CTLA-4 and sCD8. However, patients with higher sCD8⁺ cell density and CTLA-4 expression had better progression-free survival (PFS) and distant metastasis-free survival (DMFS). In addition, higher CD86 expression was associated with poorer overall survival (OS). Higher CTLA-4 expression was associated with higher tCD8⁺ cell density, whereas CD86 expression was correlated with the cell density of t/sCD8. Prognostic analysis confirmed that the relationships between CTLA-4 and DMFS as well as CD86 and OS were significantly correlated in low rather than high CD8⁺ cell density. Further the combination of CD8⁺ cell density and CD86 expression was shown to be an independent prognostic factor of OS, whereas the combination of CTLA-4 was not for DMFS. Together, these results demonstrate significant correlations between CD86 expression and t/sCD8⁺ cell density in rectal cancer after nCRT and could potentially have clinical implications for combining ICIs and nCRT.

CDC6 confers resistance to L-OHP-induced apoptosis of CRC cells.