Rapid multi-dynamic algorithm for gray image analysis of the stroma percentage on colorectal cancer

Digital image analysis provides robust tissue microenvironment-based prognosticators in stage I-IV colorectal cancer patients

AIMS

In colorectal cancer (CRC) patients, a promising marker is tumor-stroma ratio (TSR). Quantification issues highlight the importance of precise assessment that might be solved by artificial intelligence (AI)-based digital image analysis systems. Some alternatives have been offered so far, although these platforms are either proprietary developments or require additional programming skills. Our aim was to validate a user-friendly, commercially available software running in everyday computational environment to improve TSR assessment and also to compare the prognostic value of assessing TSR in three distinct regions of interests (ROIs), like hotspot, invasive front and whole tumor. Furthermore, we compared the prognostic power of TSR with newly suggested carcinoma percentage (CP) and carcinoma-stroma percentage (CSP).

METHODS AND RESULTS

Slides of 185 stage I-IV CRC patients with clinical follow up data were scanned and evaluated by a senior pathologist. A machine learning-based digital pathology software was trained to recognize tumoral and stromal compartments. The aforementioned parameters were evaluated in the hotspot, invasive front and whole tumor area, both visually and by machine learning. Patients were classified based on TSR, CP and CSP values. On multivariate analysis, TSR-hotspot was found to be an independent prognostic factor of overall survival (hazard ratio for TSR-hotspot_software: 2.005 (95% confidence interval (CI): 1.146-3.507), p=0.011, for TSR-hostpot_visual: 1.781 (CI: 1.060-2.992) p=0.029). Also, TSR was an independent predictor for distant metastasis and local relapse in most settings. Generally, software performance was comparable to visual evaluation and delivered reliable prognostication in more settings also with CP and CSP values.

CONCLUSIONS

This study presents that software assisted evaluation is a robust prognosticator. Our approach used a less sophisticated and thus easily accessible software without the aid of convolutional neural network; however, it was still effective enough to deliver reliable prognostic information.

Development and Validation of the Individualized Prognostic Nomograms in Patients With Right- and Left-Sided Colon Cancer

Background

The overall survival (OS) of patients diagnosed with colon cancer (CC) varied greatly, so did the patients with the same tumor stage. We aimed to design a nomogram that is capable of predicting OS in resected left-sided colon cancers (LSCC) and right-sided colon cancers (RSCC), and thus to stratify patients into different risk groups, respectively.

Methods

Records from a retrospective cohort of 577 patients with complete information were used to construct the nomogram. Univariate and multivariate analyses screened risk factors associated with overall survival. The performance of the nomogram was evaluated with concordance index (c-index), calibration plots, and decision curve analyses for discrimination, accuracy, calibration ability, and clinical net benefits, respectively, which was further compared with the American Joint Committee on Cancer (AJCC) 8th tumor-node-metastasis (TNM) classification. Risk stratification based on nomogram scores was performed with recursive partitioning analysis.

Results

The LSCC nomogram incorporated carbohydrate antigen 12-5 (CA12-5), age and log odds of positive lymph nodes (LODDS), and RSCC nomogram enrolled tumor stroma percentage (TSP), age and LODDS. Compared with the TNM classification, the LSCC and RSCC nomograms both had a statistically higher C-index (0.837, 95% CI: 0.827-0.846 and 0.780, 95% CI 0.773-0.787, respectively) and more clinical net benefits, respectively. Calibration plots revealed no deviations from reference lines. All results were reproducible in the validation cohort.

Conclusions

An original predictive nomogram was constructed and validated for OS in patients with CC after surgery, which had facilitated physicians to appraise the individual survival of postoperative patients accurately and to identify high-risk patients who were in need of more aggressive treatment and follow-up strategies.