Combination of CDX2 H-score quantitative analysis with CD3 AI-guided analysis identifies patients with a good prognosis only in stage III colon cancer

Machine learning evaluation of immune infiltrate through digital tumour score allows prediction of survival outcome in a pooled analysis of three international stage III colon cancer cohorts

BACKGROUND

T-cell immune infiltrates are robust prognostic variables in localised colon cancer. Evaluation of prognosis using artificial intelligence is an emerging field. We evaluated whether machine learning analysis improved prediction of patient outcome in comparison with analysis of T cell infiltrate only or in association with clinical variables.

METHODS

We used data from two phase III clinical trials (Prodige-13 and PETACC08) and one retrospective Italian cohort (HARMONY). Cohorts were split into training (N = 692), internal validation (N = 297) and external validation (N = 672) sets. Tumour slides were stained with CD3mAb. CD3 Machine Learning (CD3ML) score was computed using graphical parameters within the tumour tiles obtained from CD3 slides. CD3 infiltrates in tumour core and invasive margin were automatically detected. Associations of CD3 infiltrates and CD3ML with 5-year Disease-Free Survival (DFS) were examined using univariate and multivariable survival models by Cox regression.

FINDINGS

CD3 density both in the invasive margin and the tumour core were significantly associated with DFS in the different sets. Similarly, CD3ML score was significantly associated with DFS in all sets. CD3 assessment did not provide added value on top of CD3ML assessment (Likelihood Ratio Test (LRT), p = 0.13). In contrast, CD3ML improved prediction of DFS when combined with a clinical risk stage (LRT, p = 0.001). Stratified by clinical risk score (High or Low), patients with low CD3ML score had better DFS.

INTERPRETATION

In all tested sets, machine learning analysis of tumour cells improved prediction of prognosis compared to clinical parameters. Adding tumour-infiltrating lymphocytes assessment did not improve prognostic determination.

FUNDING

This research received no external funding.

Deep Learning-Based H-Score Quantification of Immunohistochemistry-Stained Images

Immunohistochemistry (IHC) is a well-established and commonly used staining method for clinical diagnosis and biomedical research. In most IHC images, the target protein is conjugated with a specific antibody and stained using diaminobenzidine (DAB), resulting in a brown coloration, whereas hematoxylin serves as a blue counterstain for cell nuclei. The protein expression level is quantified through the H-score, calculated from DAB staining intensity within the target cell region. Traditionally, this process requires evaluation by 2 expert pathologists, which is both time consuming and subjective. To enhance the efficiency and accuracy of this process, we have developed an automatic algorithm for quantifying the H-score of IHC images. To characterize protein expression in specific cell regions, a deep learning model for region recognition was trained based on hematoxylin staining only, achieving pixel accuracy for each class ranging from 0.92 to 0.99. Within the desired area, the algorithm categorizes DAB intensity of each pixel as negative, weak, moderate, or strong staining and calculates the final H-score based on the percentage of each intensity category. Overall, this algorithm takes an IHC image as input and directly outputs the H-score within a few seconds, significantly enhancing the speed of IHC image analysis. This automated tool provides H-score quantification with precision and consistency comparable to experienced pathologists but at a significantly reduced cost during IHC diagnostic workups. It holds significant potential to advance biomedical research reliant on IHC staining for protein expression quantification.

Machine-Learning Identifies a Strategy for Differentiation Therapy in Solid Tumors

BACKGROUND

Although differentiation therapy can cure some hematologic malignancies, its curative potential remains unrealized in solid tumors. This is because conventional computational approaches succumb to the thunderous noise of inter-/intratumoral heterogeneity. Using colorectal cancers (CRCs) as an example, here we outline a machine learning(ML)-based approach to track, differentiate, and selectively target cancer stem cells (CSCs).

METHODS

A transcriptomic network was built and validated using healthy colon and CRC tissues in diverse gene expression datasets (~5,000 human and >300 mouse samples). Therapeutic targets and perturbation strategies were prioritized using ML, with the goal of reinstating the expression of a transcriptional identifier of the differentiated colonocyte, CDX2, whose loss in poorly differentiated (CSC-enriched) CRCs doubles the risk of relapse/death. The top candidate target was then engaged with a clinical-grade drug and tested on 3 models: CRC lines in vitro, xenografts in mice, and in a prospective cohort of healthy (n = 3) and CRC (n = 23) patient-derived organoids (PDOs).

RESULTS

The drug shifts the network predictably, induces CDX2 and crypt differentiation, and shows cytotoxicity in all 3 models, with a high degree of selectivity towards all CDX2-negative cell lines, xenotransplants, and PDOs. The potential for effective pairing of therapeutic efficacy (IC50) and biomarker (CDX2-low state) is confirmed in PDOs using multivariate analyses. A 50-gene signature of therapeutic response is derived and tested on 9 independent cohorts (~1700 CRCs), revealing the impact of CDX2-reinstatement therapy could translate into a ~50% reduction in the risk of mortality/recurrence.

CONCLUSIONS

Findings not only validate the precision of the ML approach in targeting CSCs, and objectively assess its impact on clinical outcome, but also exemplify the use of ML in yielding clinical directive information for enhancing personalized medicine.

The expression and clinical significance of PLK1/p-PLK1 protein in NK/T cell Lymphoma

AIMS

To investigate the expression of polo-like kinase 1 protein (PLK1) and its phosphorylation level (p-PLK1) in extranodal NK/T cell lymphoma (NKTCL) and their correlation with clinical characteristics and prognosis.

METHODS

We collected 40 cases of NKTCL (referred to as the experimental group), which received diagnoses at the First Affiliated Hospital of Zhengzhou University between January 2018 and October 2022. Concurrently, we assembled a control group, including 20 cases afflicted with nasopharyngeal mucosal lymphoid hyperplasia diseases during the same timeframe. We utilized immunohistochemical techniques to evaluate the levels of PLK1 and p-PLK1 expression in both the experimental and control groups. Subsequently, we conducted an analysis to identify disparities in their expression and explore their relationships with clinical characteristics and patient prognosis.

RESULTS

Among the 40 NKTCL patients, there were 27 males and 11 females, with a median age of 51 years (range 12-80 years). Compared to the control group, the tissue samples of NKTCL patients exhibited significantly elevated expression levels and active phosphorylation levels of PLK1 (P < 0.05). Correlation analysis of the immunohistochemical H score and Ki-67 positive rate of PLK1 and p-PLK1, revealed a significant positive correlation for both (P < 0.0001, each). No statistically significant differences were observed in the distribution of PLK1 and p-PLK1 expression in NKTCL patients with respect to gender, age, Ann Arbor stage, PINK-E score, B-symptoms, lactate dehydrogenase, β2-microglobulin, blood EBV-DNA, bone marrow invasion, and lymph node metastasis (p > 0.05). Grouping based on PLK1 and p-PLK1 immunohistochemical H-scores revealed that the high expression of PLK1 and p-PLK1 was associated with poor prognosis.

CONCLUSIONS

The expression levels and active phosphorylation levels of PLK1 were significantly increased in NK/T cell lymphoma, and patients with overexpression of PLK1 and p-PLK1 had a poorer prognosis.

Association of CDX2 and mucin expression with chemotherapeutic benefits in patients with stage II/III gastric cancer

BACKGROUND

Better predictors of patients with stage II/III gastric cancer (GC) most likely to benefit from adjuvant chemotherapy are urgently needed. This study aimed to assess the ability of CDX2 and mucin markers to predict prognosis and fluorouracil-based adjuvant chemotherapy benefits.

METHODS

CDX2 and mucin protein expressions were examined by immunohistochemistry and compared with survival and adjuvant chemotherapy benefits in a prospective evaluation cohort of 782 stage II/III GC patients. Then, the main findings were validated in an independent validation cohort (n = 386) and an external mRNA sequencing dataset (ACRG cohort, n = 193).

RESULTS

In the evaluation cohort, CDX2, CD10, MUC2, MUC5AC, and MUC6 expressions were observed in 59.7%, 26.7%, 27.6%, 55.1%, and 57.7% of patients, respectively. However, only the expression of CDX2 was found to be associated with adjuvant chemotherapy benefits. Most importantly, CDX2-negative patients had a poorer prognosis when treated with surgery only, while the prognosis of CDX2-negative and CDX2-positive patients was similar when receiving postoperative adjuvant chemotherapy. Further analysis revealed that patients with CDX2 negative tumors benefited from chemotherapy (5-year overall survival rates: 60.0% with chemotherapy vs. 23.2% with surgery-only, p < 0.001), whereas patients with CDX2 positive tumors did not (pinteraction  = 0.004). Consistent results were obtained in the validation and ACRG cohorts.

CONCLUSIONS

Negative expression of CDX2 is an independent risk factor for survival in stage II/III GC, but subsequent adjuvant chemotherapy is able to compensate for this unfavorable effect. Therefore, active chemotherapy is more urgent for patients with negative CDX2 expression than for patients with positive CDX2 expression.

Identification of an Autophagy-Related Signature for Prognosis and Immunotherapy Response Prediction in Ovarian Cancer

BACKGROUND

Ovarian cancer (OC) is one of the most malignant tumors in the female reproductive system, with a poor prognosis. Various responses to treatments including chemotherapy and immunotherapy are observed among patients due to their individual characteristics. Applicable prognostic markers could make it easier to refine risk stratification for OC patients. Autophagy is closely implicated in the occurrence and development of tumors, including OC. Whether autophagy -related genes can be used as prognostic markers for OC patients remains unclear.

METHODS

The gene transcriptome data of 374 OC patients were downloaded from The Cancer Genome Atlas (TCGA) database. The correlation between the autophagy levels and outcomes of OC patients was identified through the single sample gene set enrichment analysis (ssGSEA). Recognized molecular markers of autophagy in different clinical specimens were detected by immunohistochemistry (IHC) assay. The gene set enrichment analysis (GSEA), ESTIMATE, and CIBERSORT analysis were applied to explore the correlation of autophagy with the tumor immune microenvironment (TIME). Single-cell RNA-sequencing (scRNA-seq) data from seven OC patients were included for characterizing cell-cell interaction patterns of autophagy-high or low tumor cells. Machine learning, Stepwise Cox regression and LASSO-Cox analysis were used to screen autophagy hub genes, which were used to establish an autophagy-related signature for prognosis evaluation. Four tumor immunotherapy cohorts were obtained from the GEO (Gene Expression Omnibus) database and the literature for autophagy risk score validation.

RESULTS

The autophagy levels were closely related to the prognosis of the OC patients. Additionally, the autophagy levels were correlated with TIME status including immune score, and immune-cell infiltration. The scRNA-seq analysis found that tumor cells with high or low autophagy levels had different interactions with immune cells, especially macrophages. Eight autophagy-hub genes (ZFYVE1, AMBRA1, LAMP2, TRAF6, PDPK1, ATG2B, DAPK1 and TP53INP2) were screened for an autophagy-related signature. According to this signature, higher risk score was correlated with poor prognosis and better immunotherapy response in the OC patients.

CONCLUSIONS

The autophagy-related signature is applicable to predict the prognosis and immune checkpoint inhibitors (ICIs) therapy efficiency in OC patients. It is possible to identify OC patients who will respond to ICIs therapy and have a favorable prognosis, although more verification is needed.