Related Experiment Video
Updated: Aug 26, 2026
Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
Harmonization Framework for Translating Clinical Cancer Staging Systems for Public Health and Surveillance Use in
Gokul Sarveswaran1, Abhinav Ramraj1, Jayasankar Sankarapillai1
1Indian Council of Medical Research-National Institute of NCD Epidemiology (ICMR-NINE), Bengaluru, India.
Introduction:
Although accurate information on cancer stage at diagnosis is critical for surveillance, cancer registries in low- and middle-income countries (LMICs) report incomplete stage data. The use of multiple stage classification systems-UICC/AJCC-TNM-system, SEER-Summary-Stage (SSS), Essential-TNM (ETNM), and Condensed-TNM (CTNM) complicates their utility and interoperability. The lack of a unified framework for translating between their categories limits data harmonization for global epidemiological analyses.
Materials And Methods:
This study introduces a mapping framework for translating these four staging systems for common screening-amenable cancers (breast, cervical, colorectal, and oral). Using the TNM criteria as reference, all categories were systematically compared and mapped to corresponding categories in SSS, ETNM, and CTNM. The fields were color-coded to visualize prognostic gradients across classifications. This mapping was then validated using a registry-wide representative sample of 4,000 cancer cases with an assigned TNM stage. Matches were categorized as either being complete or as first- or second-degree mismatches to assess the extent of concordance and discordance.
Results:
Our analysis highlighted cancer-specific variations in prognostic gradation and concordance. Cervical and colorectal cancers had a steady gradient, representing the best-case scenarios, while breast and oral cancers showed greater discordance. Overall, SSS aligned with TNM categories, depicting maximal one-to-one associations. While ETNM is more concordant for colorectal cancer (100% match vs. 89%), SSS is most concordant for cervical cancer (80%), and to a lesser extent, breast cancer (77%). CTNM offers the least concordance with traditional TNM across cancer sites. Localized and metastatic stages aligned almost universally, except in oral cancer.
Conclusion:
This mapping exercise provided a comprehensive framework for translating cancer stage categories, offering LMIC registries a way to balance clinical relevance, epidemiological utility, and operational feasibility. Concordance among the systems varied by cancer type, underscoring the need for cancer-specific adaptations.
