What a finding looks like when it reaches your desk.

The subject. Etcamah (camizestrant) is approved for HR-positive, HER2-negative advanced breast cancer, given with a CDK4/6 inhibitor when a blood test finds an ESR1 mutation in circulating tumour DNA during first-line hormonal therapy, before scans show the cancer progressing. Saudi Arabia’s regulator, the SFDA, approved it on 15 May 2026; the US FDA granted accelerated approval 112 days later, on 4 September 2026.

The earlier trigger. The approval moves the decision point: testing, and switching treatment, on a blood-test result during first-line therapy, rather than at recurrence or progression, the established point for ESR1 testing.

Four AI engines described the earlier trigger only when camizestrant was named.

Asked when ESR1 testing should be done, all four placed it at recurrence or progression. Asked when treatment should be switched to camizestrant, all four described the switch as triggered during first-line therapy. Every engine could describe the earlier trigger; none offered it in answer to the clinical question.

The FDA approval is accelerated: the agency states that whether acting at this earlier point improves outcomes is not yet confirmed, and confirmatory studies are required. The finding concerns what the answers said, not when testing should happen.

Three questions from the same reading, each put once to the same four engines in a Saudi Arabia market frame between 31 August 2026 and 2 September 2026. They differ in how much they disclose, from nothing about the treatment to the product’s name.

One answer shows the gap plainly. Asked the testing question, Gemini cited the Saudi approval and noted that it “applies to mutations that emerge during first-line hormonal therapy”, then recommended testing “at the point of disease recurrence or progression”.

The three questions differ in what they ask as well as what they disclose: when to test is a different decision from when to switch. This is a descriptive contrast, not a controlled comparison. Only the engines and the market frame were held constant.

The answers named who sets the standard and who orders the test, but not where care is delivered or where the sample goes.

An indication that depends on a blood test depends on a route to that test. Two other questions in the same reading asked who arranges it. What a customer learns here is where the answers stop guiding a patient.

This describes the answers only. It does not establish that ESR1 testing is unavailable in Saudi Arabia, or that any laboratory does or does not offer it: only that these answers did not name one.