Why the Pacific total hides a Samoan rate of 9.5

From 23 to 27 October 2026, the ESMO Congress in Madrid gave a Presidential Symposium slot to SAFFRON, an international Phase III trial in lung cancer. Investigators tested savolitinib with osimertinib in 338 patients whose tumours showed an EGFR mutation alongside MET overexpression or amplification. The patients had progressed on osimertinib alone. Results released on 17 August showed significant gains in progression-free survival and overall survival against platinum doublet chemotherapy. China approved the combination first, on the earlier SACHI trial. Mainland coverage of the US new drug application announced on 29 September names the 2 drugs 赛沃替尼 (savolitinib) and 奥希替尼 (osimertinib).

In July, CLARITY-Gastric01 reported a statistically significant overall survival gain for sonesitatug vedotin, an antibody drug conjugate aimed at the Claudin-18.2 protein. The randomised trial set the drug against the investigator's choice of therapy in patients treated in the second line or later. Eligible tumours showed the protein on at least 25% of their cells. There were 594 patients enrolled at 175 centres in 19 countries, spread over North America, Europe, South America and Asia.

Phyu Sin Aye and colleagues used registry data for Northern New Zealand, the region of Auckland and Northland, to count 3,815 patients diagnosed with non-squamous non-small cell lung cancer between February 2010 and July 2017. Clinicians tested 45% of those patients for EGFR mutations. Of the tumours tested, 22.5% came back positive. Against New Zealand Europeans, the standardised incidence ratio for EGFR-positive disease reached 3.47 for Pacific patients, 3.35 for Asian patients and 2.02 for Māori. Adjusting for the untested patients left the ratios close to those figures, in the team's 2021 PLoS ONE paper.

For gastric cancer, Michael Walsh and colleagues put the Pacific age-standardised rate for 2020 to 2022 at 14.4 per 100,000, against 10.7 for Māori and 4.6 for European and Other New Zealanders. Their projection puts Pacific cases at 87 a year by 2045, a 93% rise on the 45 a year recorded for 2020 to 2022. The New Zealand Medical Journal published it on 8 May 2026. The authors name Helicobacter pylori as a key contributor. They also point to inherited CDH1 variants over-represented in some Māori whānau (extended families).

Satherley and Sporle split the Pacific population into its Level 2 ethnic groups, the finer code in national ethnicity data. For diagnoses between 1995 and 2022 they found gastric cancer incidence of 10.7 per 100,000 among Niuean people and 9.5 among Samoan people, against 6.1 for Cook Islands Māori and 6.4 for Tokelauans. The pooled Pacific rate came to 8.0. Among Tongans, the male rate reached 1.9 times the female rate. The authors concluded in April 2025 that aggregation hides real differences between the groups.

Walsh's team flags a parallel limit on its own numbers. Its model used prioritised ethnicity, a coding choice the authors say may understate the Pacific burden.

Follow-up raises a second problem for survival data. Satherley and Sporle found that 3.8% of Pacific people with cancer left Aotearoa within 5 years of diagnosis, against 1.0% of all cases. Fijians led at 4.2%, and Samoans, at 4.0%. The 2 trials report overall survival, an endpoint counted in deaths across years of follow-up.

Bowel screening pamphlets from Te Whatu Ora (Health New Zealand) come out language by language. HE2711, the Samoan version for people with a positive result, bears the title Maua iuga o le suesuega o le taufale ua aafia (positive bowel screening test result) and explains the follow-up colonoscopy. Reviewers last checked it in July 2024. Te Kāhui Matepukupuku o Aotearoa (the Cancer Society of New Zealand) titles its Samoan breast cancer booklet kanesa o le susu (breast cancer).

Trial consent in Aotearoa answers to the Code of Health and Disability Services Consumers' Rights. Right 7(6)(a) makes written consent compulsory for a consumer taking part in research. Right 9 applies the Code to research settings. Right 5(1) gives the consumer a right to information in a form, language and manner suited to that consumer. It adds a competent interpreter as far as reasonably practicable.

On a research consent short form Huri translated into 2 Pacific languages for a drug developer, the sponsor commissioned independent back-translations and compared them with the English. Back-translation returned the word for consent as plain agreement and the word for research as everyday labour. A participant signing that form agrees to something other than a study. The 2 terms were rebuilt by Huri linguists in their clinical sense. Technical terms such as whole genome sequencing stayed in English in brackets. A signed declaration of fidelity to the source closed the file.

Te Whatu Ora bars staff and contractors from entering personal or sensitive patient information into unapproved large language models and generative AI tools. Where staff use a tool outside that bar, the policy asks them to check the output and acknowledge its use on patient-facing material. The policy also names data sovereignty among the risks. Huri's Samoan panel sets a general-purpose engine's output beside a version signed off by a native linguist, from an English source.

Walsh's team recommends an H. pylori test-and-treat programme delivered through primary care, the setting for most Māori and Pacific patients. The authors also raise linking that testing to the National Bowel Screening Programme, the programme behind HE2711. They ask for care with the screening inequities already present. Satherley and Sporle ask for basic incidence and mortality rates by Level 2 group, produced from the New Zealand Cancer Registry outside the Integrated Data Infrastructure.