@pipeworx/globocan

Connect: https://gateway.pipeworx.io/globocan/mcp · Install: one-click buttons

Tools: 4

Cancer incidence, mortality and 5-year prevalence for 185 countries and 41 cancer sites: new cases, deaths, age-standardised rates per 100,000, and cumulative risk to age 74.

Tools

Auth

None. Keyless, no registration, no quota.

Data source

https://gco.iarc.who.int/ — the Global Cancer Observatory. Estimates are GLOBOCAN 2024.

The API is on a different host than the site. The app is served from gco.iarc.who.int; its data comes from gco-api.iarc.fr. This matters when probing: every path under the who.int host returns the single-page-app shell with HTTP 200 and text/html, so a probe there looks like an API that exists and answers — it just never returns JSON. The real base was read out of the app bundle (VITE_APP_API), where the data version is a path segment.

Gotchas

1. The dimensions are bare integers and type is the dangerous one. Rows carry sex, type and cancer_code as numbers with no labels. type is 0 = incidence, 1 = mortality, 2 = prevalence. Handing a raw row to a caller lets 632,166 US cancer deaths be read as new cases, or an 8-million prevalence figure as an annual count. Every number this pack returns is labelled with which measure it is.

   These encodings were pinned against published figures rather than assumed: for the United States the three all-cancer totals come back as 2,462,729 / 632,166 / 8,069,773, which match known US incidence, mortality and 5-year prevalence respectively.

2. cancer_code 39 is “All cancers” (ICD C00-97), not a site. It is the all-sites total and is excluded from the leading-site rankings — leaving it in would put a row worth 100% at the top of every list.

3. Sex-specific sites have no both-sexes row. Asking for a male-only site with sex: "female" returns no_estimate with the reason, because that is the standard being coherent rather than a gap in the data.

4. These are modelled estimates, not case reports. IARC estimates national burden from registry data of varying completeness; the method_incidence / method_mortality fields upstream grade that per country. Responses say so rather than presenting the numbers as counts.

Tools

• globocan_country_profile — Cancer statistics for one country from the WHO/IARC Global Cancer Observatory (GLOBOCAN): total new cases, deaths and 5-year prevalence, plus the leading cancer sites ranked by new cases and by deaths
• globocan_cancer_in_country — Figures for ONE cancer type in one country — new cases, deaths, 5-year prevalence, age-standardised rate and cumulative risk to age 74. Use when the question names both a cancer and a place, e.g. “bre
• globocan_countries — List or search the 185 countries and territories GLOBOCAN publishes cancer estimates for, with ISO3 codes, world region and Human Development Index band. Call this to find the exact country spelling t
• globocan_cancer_types — List the 41 cancer sites GLOBOCAN reports, with their ICD-10 code ranges and whether a site is sex-specific. Call this to find the exact site name globocan_cancer_in_country expects.