Identifying genetic mutations in cancer patients have been increasingly important because distinctive mutational patterns can be very informative to determine the optimal therapeutic strategy. Recent studies have shown that deep learning-based molecular cancer subtyping can be performed directly from the standard hematoxylin and eosin (H&E) sections in diverse tumors including colorectal cancers (CRCs). Since H&E-stained tissue slides are ubiquitously available, mutation prediction with the pathology images from cancers can be a time- and cost-effective complementary method for personalized treatment.

To predict the frequently occurring actionable mutations from the H&E-stained CRC whole-slide images (WSIs) with deep learning-based classifiers.

A total of 629 CRC patients from The Cancer Genome Atlas (TCGA-COAD and TCGA-READ) and 142 CRC patients from Seoul St. Mary Hospital (SMH) were included. Based on the mutation frequency in TCGA and SMH datasets, we chose APC, KRAS, PIK3CA, SMAD4, and TP53 genes for the study. The classifiers were trained with 360 × 360 pixel patches of tissue images. The receiver operating characteristic (ROC) curves and area under the curves (AUCs) for all the classifiers were presented.

The AUCs for ROC curves ranged from 0.693 to 0.809 for the TCGA frozen WSIs and from 0.645 to 0.783 for the TCGA formalin-fixed paraffin-embedded WSIs. The prediction performance can be enhanced with the expansion of datasets. When the classifiers were trained with both TCGA and SMH data, the prediction performance was improved.

APC, KRAS, PIK3CA, SMAD4, and TP53 mutations can be predicted from H&E pathology images using deep learning-based classifiers, demonstrating the potential for deep learning-based mutation prediction in the CRC tissue slides.

This article clarifies prognostic and predictive markers in the treatment of colorectal cancer. Multiple chemotherapeutic drugs are approved for metastatic colorectal cancer (mCRC), but available guidelines are often not helpful in directing drug selections. It would be desirable to define patient populations before chemotherapy by biomarkers that predict outcome and toxicities. RAS mutational evaluation remains the only established biomarker analysis in the treatment of mCRC. BRAF mutant tumors are associated with poor outcome. Chemotherapeutic combination therapies still remain the most active treatments in the armamentarium, and future trials should address the need to prospectively investigate and validate biomarkers.

Colorectal cancer (CRC) is one of the most common cancers worldwide, with ~700,000 mortalities occurring due to CRC in 2012. The treatment options are effective in a small percentage of patients, and it is important to identify specific biomarkers in order to determine patients for whom the available therapies will be beneficial. It has been hypothesised that the PIK3CA gene mutation may affect the response to therapy of patients with metastatic CRC. In the present study, primary tumour specimens were collected from 156 patients with CRC who were treated in the Military Institute of Medicine in Warsaw (Warsaw, Poland). Codons 12 and 13 of exon 1 of KRAS, exons 11 and 15 of BRAF and exons 9 and 20 of PIK3CA were analysed for mutation using direct sequencing. The prognostic value of each mutation and the clinical and pathological variables of these tumours were estimated. The results revealed that PIK3CA mutations were present in 15 patients (9.6%), of whom seven (46.7%) possessed mutations in codon 9 and eight (53.3%) possessed mutations in codon 20. Mutation in the PIK3CA gene was detected in six patients with KRAS gene mutations, which accounted for 40% of PIK3CA-mutated tumours, and in one patient with BRAF mutations, which accounted for 6.6% of PIK3CA-mutated tumours. No significant differences were identified between the overall survival (OS) rates of patients with PIK3CA mutations (median OS, 56.7 months) and those with wild-type PIK3CA genes (median OS, 47.6 months) (P=0.1270). Univariate analysis identified that the following prognostic factors affected the OS rate in the current patient cohort: Gender, female patients survived for 57.5 months compared with 39.3 months for male patients (P=0.0111); and lymph node involvement grade, as survival of patients without lymph node metastases was 61.4 months compared with 45.4 months in patients presenting with metastases (P=0.0122). The findings of the present analysis indicate that PIK3CA mutation status is not a prognostic factor in CRC patients. In addition, no statistically significant association exists between tumours with PIK3CA mutations and clinical or pathological factors.