Do those stats count all detections where the patient died, or all deaths where the cause of death was determined to be cancer? I can imagine in the USA someone

Do those stats count all detections where the patient died, or all deaths where the cause of death was determined to be cancer? I can imagine in the USA someone who can’t afford treatment might end up committing suicide, thereby artificially lowering the cancer rate and the deaths from cancer statistic. Remembering that different countries have different definitions for, say, infant mortality (eg: difference of opinion as to what counts as a live birth, meaning that some countries report lower infant mortality but higher perinatal mortality)